TRM_PSoC5 Cypress PSo C5LP Technical Reference Manual
User Manual:
Open the PDF directly: View PDF .
Page Count: 452
Download | |
Open PDF In Browser | View PDF |
PSoC 5LP TRM PSoC® 5LP Architecture TRM (Technical Reference Manual) Document No. 001-78426 Rev. *F May 31, 2017 Cypress Semiconductor 198 Champion Court San Jose, CA 95134-1709 Phone (USA): 800.858.1810 Phone (Intnl): 408.943.2600 http://www.cypress.com Copyrights Copyrights © Cypress Semiconductor Corporation, 2012-2016. This document is the property of Cypress Semiconductor Corporation and its subsidiaries, including Spansion LLC ("Cypress"). This document, including any software or firmware included or referenced in this document ("Software"), is owned by Cypress under the intellectual property laws and treaties of the United States and other countries worldwide. Cypress reserves all rights under such laws and treaties and does not, except as specifically stated in this paragraph, grant any license under its patents, copyrights, trademarks, or other intellectual property rights. If the Software is not accompanied by a license agreement and you do not otherwise have a written agreement with Cypress governing the use of the Software, then Cypress hereby grants you a personal, non-exclusive, nontransferable license (without the right to sublicense) (1) under its copyright rights in the Software (a) for Software provided in source code form, to modify and reproduce the Software solely for use with Cypress hardware products, only internally within your organization, and (b) to distribute the Software in binary code form externally to end users (either directly or indirectly through resellers and distributors), solely for use on Cypress hardware product units, and (2) under those claims of Cypress's patents that are infringed by the Software (as provided by Cypress, unmodified) to make, use, distribute, and import the Software solely for use with Cypress hardware products. Any other use, reproduction, modification, translation, or compilation of the Software is prohibited. TO THE EXTENT PERMITTED BY APPLICABLE LAW, CYPRESS MAKES NO WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, WITH REGARD TO THIS DOCUMENT OR ANY SOFTWARE OR ACCOMPANYING HARDWARE, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. To the extent permitted by applicable law, Cypress reserves the right to make changes to this document without further notice. Cypress does not assume any liability arising out of the application or use of any product or circuit described in this document. Any information provided in this document, including any sample design information or programming code, is provided only for reference purposes. It is the responsibility of the user of this document to properly design, program, and test the functionality and safety of any application made of this information and any resulting product. Cypress products are not designed, intended, or authorized for use as critical components in systems designed or intended for the operation of weapons, weapons systems, nuclear installations, life-support devices or systems, other medical devices or systems (including resuscitation equipment and surgical implants), pollution control or hazardous substances management, or other uses where the failure of the device or system could cause personal injury, death, or property damage ("Unintended Uses"). A critical component is any component of a device or system whose failure to perform can be reasonably expected to cause the failure of the device or system, or to affect its safety or effectiveness. Cypress is not liable, in whole or in part, and you shall and hereby do release Cypress from any claim, damage, or other liability arising from or related to all Unintended Uses of Cypress products. You shall indemnify and hold Cypress harmless from and against all claims, costs, damages, and other liabilities, including claims for personal injury or death, arising from or related to any Unintended Uses of Cypress products. Cypress, the Cypress logo, Spansion, the Spansion logo, and combinations thereof, WICED, PSoC, CapSense, EZ-USB, FRAM, and Traveo are trademarks or registered trademarks of Cypress in the United States and other countries. For a more complete list of Cypress trademarks, visit cypress.com. Other names and brands may be claimed as property of their respective owners. 2 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Contents Overview Section A: Overview 21 1. Introduction ........................................................................................................... 23 2. Getting Started ...................................................................................................... 29 3. Document Construction ......................................................................................... 31 Section B: CPU System 35 4. Cortex™-M3 Microcontroller .................................................................................. 37 5. PSoC 5LP Cache Controller.................................................................................. 51 6. PHUB and DMAC .................................................................................................. 55 7. Interrupt Controller ................................................................................................ 73 Section C: Memory 83 8. Nonvolatile Latch ................................................................................................... 85 9. SRAM ................................................................................................................... 89 10. Flash Program Memory.......................................................................................... 93 11. EEPROM............................................................................................................... 95 12. EMIF ..................................................................................................................... 97 13. Memory Map ....................................................................................................... 105 Section D: System Wide Resources 107 14. Clocking System .................................................................................................. 109 15. Power Supply and Monitoring .............................................................................. 125 16. Low-Power Modes ............................................................................................... 135 17. Watchdog Timer .................................................................................................. 141 18. Reset .................................................................................................................. 143 19. I/O System .......................................................................................................... 151 20. Flash, Configuration Protection ............................................................................ 167 Section E: Digital System 173 21. Universal Digital Blocks (UDBs) ........................................................................... 175 22. UDB Array and Digital System Interconnect ......................................................... 217 23. Controller Area Network (CAN) ............................................................................ 225 24. USB .................................................................................................................... 241 25. Timer, Counter, and PWM .................................................................................... 257 26. I 2 C ...................................................................................................................... 273 27. Digital Filter Block (DFB) ..................................................................................... 287 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 3 Contents Overview Section F: Analog System 303 28. Switched Capacitor/Continuous Time................................................................... 305 29. Analog Routing ................................................................................................... 319 30. Comparators ....................................................................................................... 335 31. Opamp ................................................................................................................ 341 32. LCD Direct Drive ................................................................................................. 345 33. CapSense ........................................................................................................... 359 34. Temperature Sensor ............................................................................................ 365 35. Digital-to-Analog Converter ................................................................................. 369 36. Precision Reference ............................................................................................ 373 37. Delta Sigma Converter ........................................................................................ 377 38. Successive Approximation Register ADC ............................................................. 397 Section G: Program and Debug 401 39. Test Controller .................................................................................................... 403 40. Cortex-M3 Debug and Trace................................................................................ 415 41. Nonvolatile Memory Programming ....................................................................... 423 Glossary 429 Index 445 4 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Contents Section A: Overview 21 Document Revision History ..............................................................................................................21 1. Introduction 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 2. 3. Top Level Architecture............................................................................................................23 Features..................................................................................................................................25 CPU System ...........................................................................................................................25 1.3.1 Processor...............................................................................................................25 1.3.2 Interrupt Controller .................................................................................................25 1.3.3 DMA Controller ......................................................................................................25 1.3.4 Cache Controller ....................................................................................................25 Memory...................................................................................................................................25 System Wide Resources ........................................................................................................26 1.5.1 I/O Interfaces .........................................................................................................26 1.5.2 Internal Clock Generators ......................................................................................26 1.5.3 Power Supply.........................................................................................................26 1.5.3.1 Boost Converter .....................................................................................26 1.5.3.2 Sleep Modes ..........................................................................................26 Digital System.........................................................................................................................26 Analog System........................................................................................................................27 1.7.1 Delta Sigma ADC...................................................................................................27 1.7.2 Successive Approximation Register ADC..............................................................27 1.7.3 Digital Filter Block ..................................................................................................27 1.7.4 Digital-to-Analog Converters..................................................................................27 1.7.5 Additional Analog Subsystem Components...........................................................27 Program and Debug ...............................................................................................................27 Getting Started 2.1 2.2 2.3 29 Support ...................................................................................................................................29 Product Upgrades...................................................................................................................29 Development Kits ...................................................................................................................29 Document Construction 3.1 3.2 23 31 Major Sections ........................................................................................................................31 Documentation Conventions ..................................................................................................31 3.2.1 Register Conventions.............................................................................................31 3.2.2 Numeric Naming ....................................................................................................31 3.2.3 Units of Measure....................................................................................................32 3.2.4 Acronyms ...............................................................................................................32 Section B: CPU System 35 Top Level Architecture .....................................................................................................................35 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 5 4. Cortex™-M3 Microcontroller 4.1 4.2 4.3 4.4 4.5 5. PSoC 5LP Cache Controller 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 6. 6.2 6 51 Features ................................................................................................................................. 51 Block Diagram ........................................................................................................................ 51 Cache Enabling and Disabling ............................................................................................... 52 Code Protection and Security................................................................................................. 52 Invalidating the Cache Line .................................................................................................... 52 5.5.1 Measuring Cache Hits or Misses ........................................................................... 52 Cache Induced Flash Low-power Mode ................................................................................. 52 Sleep Mode Behavior ............................................................................................................. 52 Cache Limitations ................................................................................................................... 53 PHUB and DMAC 6.1 37 Features ................................................................................................................................. 37 How it Works .......................................................................................................................... 39 4.2.1 Registers ............................................................................................................... 39 4.2.1.1 Special Registers ................................................................................... 41 4.2.2 Operating Modes ................................................................................................... 42 4.2.3 Pipelining ............................................................................................................... 43 4.2.4 Thumb-2 Instruction Set ........................................................................................ 43 4.2.4.1 Data Processing Operations .................................................................. 43 4.2.4.2 Load Store Operations ........................................................................... 44 4.2.4.3 Branch Operations ................................................................................. 44 4.2.4.4 Instruction Barrier and Memory Barrier Instructions............................... 44 4.2.4.5 Saturation Operations ............................................................................ 44 4.2.5 SysTick Timer ........................................................................................................ 44 4.2.6 Debug and Trace: .................................................................................................. 45 Memory Map........................................................................................................................... 45 4.3.1 Bus Interface to SRAM Memory ............................................................................ 45 Exceptions .............................................................................................................................. 46 4.4.1 Priority Definitions.................................................................................................. 47 4.4.2 Fault Exceptions .................................................................................................... 47 4.4.3 System Call Exceptions ......................................................................................... 48 Nested Vector Interrupt Controller (NVIC).............................................................................. 48 4.5.1 Basic Interrupt Configuration ................................................................................. 49 4.5.1.1 Example Procedures in Setting Up an Interrupt..................................... 49 4.5.2 Nested Interrupts ................................................................................................... 49 4.5.3 Tail-Chaining Interrupts.......................................................................................... 49 4.5.4 Late Arrivals...........................................................................................................49 4.5.5 Interrupt Latency.................................................................................................... 50 4.5.6 Faults Related to Interrupts ................................................................................... 50 55 PHUB...................................................................................................................................... 55 6.1.1 Features ................................................................................................................ 55 6.1.2 Block Diagram ....................................................................................................... 55 6.1.3 How It Works ......................................................................................................... 56 6.1.4 Arbiter .................................................................................................................... 57 DMA Controller ....................................................................................................................... 57 6.2.1 Local Memory ........................................................................................................ 57 6.2.2 How the DMAC Works........................................................................................... 57 6.2.2.1 Interspoke Transfers .............................................................................. 58 6.2.2.2 Intraspoke Transfer ................................................................................ 59 6.2.2.3 Handling Multiple DMA Channels .......................................................... 60 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 6.3 6.4 7. 6.2.2.4 DMA Channel Priority.............................................................................60 6.2.2.5 DMA Latency in case of Nonideal Conditions ........................................63 6.2.2.6 Request per Burst Bit .............................................................................68 6.2.2.7 Work Sep Bit ..........................................................................................69 DMA Transaction Modes ........................................................................................................69 6.3.1 Simple DMA ...........................................................................................................69 6.3.2 Auto Repeat DMA ..................................................................................................69 6.3.3 Ping Pong DMA .....................................................................................................69 6.3.4 Circular DMA..........................................................................................................69 6.3.5 Indexed DMA .........................................................................................................69 6.3.6 Scatter Gather DMA...............................................................................................70 6.3.7 Packet Queuing DMA ............................................................................................70 6.3.8 Nested DMA...........................................................................................................70 Register List ...........................................................................................................................71 Interrupt Controller 7.1 7.2 7.3 7.4 7.5 73 Features..................................................................................................................................73 Block Diagram ........................................................................................................................73 How It Works ..........................................................................................................................74 7.3.1 Enabling Interrupts.................................................................................................74 7.3.2 Pending Interrupts..................................................................................................75 7.3.3 Interrupt Priority .....................................................................................................75 7.3.4 Level versus Pulse Interrupt ..................................................................................76 7.3.5 Interrupt Execution.................................................................................................76 PSoC 5LP Interrupt Controller Features.................................................................................77 7.4.1 Active Interrupts .....................................................................................................78 7.4.2 Interrupt Nesting ....................................................................................................78 7.4.3 Interrupt Vector Addresses ....................................................................................80 7.4.4 Tail Chaining ..........................................................................................................80 7.4.5 Late Arrival Interrupts.............................................................................................80 7.4.6 Exceptions .............................................................................................................81 7.4.7 Interrupt Masking ...................................................................................................81 Interrupt Controller and Power Modes....................................................................................81 Section C: Memory 83 Top Level Architecture .....................................................................................................................83 8. Nonvolatile Latch 8.1 8.2 8.3 8.4 8.5 9. Features..................................................................................................................................85 Device Configuration NV Latch...............................................................................................85 8.2.1 PRTxRDM[1:0].......................................................................................................85 8.2.2 XRESMEN .............................................................................................................86 8.2.3 DEBUG_EN ...........................................................................................................86 8.2.4 CFGSPEED ...........................................................................................................86 8.2.5 DPS[1:0].................................................................................................................86 8.2.6 ECCEN ..................................................................................................................86 8.2.7 DIG_PHS_DLY[3:0] ...............................................................................................86 Write Once NV Latch ..............................................................................................................86 Programming NV Latch ..........................................................................................................87 Sleep Mode Behavior .............................................................................................................87 SRAM 9.1 9.2 85 89 Features..................................................................................................................................89 Block Diagram ........................................................................................................................89 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 7 9.3 How It Works .......................................................................................................................... 91 10. Flash Program Memory 10.1 10.2 10.3 10.4 10.5 93 Features ................................................................................................................................. 93 Block Diagram ........................................................................................................................ 93 How It Works .......................................................................................................................... 94 Flash Memory Access Arbitration........................................................................................... 94 ECC Error Detection and Interrupts........................................................................................ 94 11. EEPROM 95 11.1 Features ................................................................................................................................. 95 11.2 Block Diagram ........................................................................................................................ 95 11.3 How It Works .......................................................................................................................... 96 12. EMIF 97 12.1 Features ................................................................................................................................. 97 12.2 Block Diagram ........................................................................................................................ 97 12.3 How It Works .......................................................................................................................... 98 12.3.1 List of EMIF Registers ........................................................................................... 98 12.3.2 External Memory Support ...................................................................................... 98 12.3.3 Sleep Mode Behavior ..........................................................................................101 12.4 EMIF Timing .........................................................................................................................102 12.5 Using EMIF with Memory-Mapped Peripherals ....................................................................104 12.6 Additional Configuration Guidelines .....................................................................................104 12.6.1 Address Bus Configuration ..................................................................................104 12.6.2 Data Bus Configuration........................................................................................104 12.6.3 16-bit Memory Transfers......................................................................................104 12.6.4 8-bit Memory Transfers........................................................................................104 13. Memory Map 105 13.1 Features ...............................................................................................................................105 13.2 Block Diagram ......................................................................................................................105 13.3 How It Works ........................................................................................................................105 13.3.1 PSoC 5LP Memory Map ......................................................................................106 Section D: System Wide Resources 107 Top Level Architecture ...................................................................................................................107 14. Clocking System 109 14.1 Features ...............................................................................................................................109 14.2 Block Diagram ......................................................................................................................110 14.3 Clock Sources ......................................................................................................................110 14.3.1 Internal Oscillators ............................................................................................... 110 14.3.1.1 Internal Main Oscillator ........................................................................110 14.3.1.2 Internal Low-Speed Oscillator ..............................................................111 14.3.2 External Oscillators.............................................................................................. 112 14.3.2.1 MHz Crystal Oscillator.........................................................................112 14.3.2.2 32.768 kHz Crystal Oscillator...............................................................113 14.3.3 Oscillator Summary ............................................................................................. 114 14.3.4 DSI Clocks ........................................................................................................... 114 14.3.5 Phase-Locked Loop............................................................................................. 114 14.4 Clock Distribution..................................................................................................................116 14.4.1 Master Clock Mux ................................................................................................ 117 14.4.2 USB Clock ........................................................................................................... 117 8 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 14.4.3 Clock Dividers ......................................................................................................118 14.4.3.1 Single Cycle Pulse Mode .....................................................................119 14.4.3.2 50% Duty Cycle Mode..........................................................................119 14.4.3.3 Early Phase Option ..............................................................................120 14.4.4 Clock Synchronization .........................................................................................120 14.4.5 Phase Selection and Control ...............................................................................120 14.4.6 Divider Update .....................................................................................................121 14.4.7 Power Gating of Clock Outputs............................................................................122 14.4.8 System Clock .......................................................................................................122 14.4.9 Asynchronous Clocks ..........................................................................................122 14.5 Low-Power Mode Operation .................................................................................................122 14.6 Clock Naming Summary .......................................................................................................123 15. Power Supply and Monitoring 125 15.1 Features................................................................................................................................125 15.2 Block Diagram ......................................................................................................................126 15.3 How It Works ........................................................................................................................127 15.3.1 Regulator Summary .............................................................................................127 15.3.1.1 Internal Regulators ...............................................................................127 15.3.1.2 Sleep Regulator....................................................................................127 15.3.1.3 Hibernate Regulator .............................................................................127 15.3.2 Boost Converter ...................................................................................................127 15.3.2.1 Operating Modes..................................................................................129 15.3.2.2 Status Monitoring .................................................................................130 15.3.2.3 Boost Firmware Requirements.............................................................130 15.3.2.4 Boost Design Process ..........................................................................130 15.3.3 Voltage Monitoring ...............................................................................................131 15.3.3.1 Low-Voltage Interrupt ...........................................................................131 15.3.3.2 High Voltage Interrupt ..........................................................................132 15.3.3.3 Processing a Low/High Voltage Detect Interrupt..................................132 15.3.3.4 Reset on a Voltage Monitoring Interrupt...............................................132 15.4 Register Summary ................................................................................................................133 16. Low-Power Modes 135 16.1 Features................................................................................................................................135 16.2 Active Mode ..........................................................................................................................136 16.2.1 Entering Active Mode...........................................................................................136 16.2.2 Exiting Active Mode .............................................................................................136 16.3 Alternative Active Mode ........................................................................................................137 16.3.1 Entering Alternative Active Mode.........................................................................137 16.3.2 Exiting Alternative Active Mode ...........................................................................137 16.4 Sleep Mode...........................................................................................................................137 16.4.1 Entering Sleep Mode ...........................................................................................137 16.4.2 Exiting Sleep Mode ..............................................................................................137 16.5 Hibernate Mode ....................................................................................................................137 16.5.1 Entering Hibernate Mode .....................................................................................138 16.5.2 Exiting Hibernate Mode........................................................................................138 16.6 Timers...................................................................................................................................138 16.6.1 Central Timewheel (CTW)....................................................................................138 16.6.2 Fast Timewheel (FTW).........................................................................................138 16.7 Register List .........................................................................................................................139 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 9 17. Watchdog Timer 141 17.1 Features ...............................................................................................................................141 17.2 Block Diagram ......................................................................................................................141 17.3 How It Works ........................................................................................................................142 17.3.1 Enabling and Disabling the WDT.........................................................................142 17.3.2 Setting the WDT Time Period and Clearing the WDT .........................................142 17.3.3 Operation in Low-Power Modes ..........................................................................142 17.3.4 Watchdog Protection Settings .............................................................................142 17.4 Register List .........................................................................................................................142 18. Reset 143 18.1 Reset Sources ......................................................................................................................143 18.1.1 Voltage Monitor Resets........................................................................................143 18.1.1.1 Initial Power-On Reset (IPOR) .............................................................143 18.1.1.2 Precise Low-Voltage Reset (PRES).....................................................143 18.1.1.3 Low-Voltage Interrupt/Reset (LVI)........................................................143 18.1.1.4 Hibernate Reset (HRES)......................................................................143 18.1.2 External Reset (XRES) ........................................................................................143 18.1.3 Software Initiated Reset (SRES) .........................................................................144 18.1.4 Watchdog Reset (WRES) ....................................................................................144 18.1.5 Identifying Reset Sources....................................................................................144 18.1.5.1 Preservation of Reset Status ...............................................................144 18.2 Reset Diagram......................................................................................................................145 18.3 Reset Summary....................................................................................................................146 18.4 Boot Process and Timing .....................................................................................................147 18.4.1 Manufacturing Configuration NV Latch................................................................148 18.4.1.1 Device Configuration NV Latch ............................................................148 18.4.2 Boot Phase ..........................................................................................................148 18.4.3 User Mode ...........................................................................................................149 18.5 Register List .........................................................................................................................149 19. I/O System 151 19.1 Features ...............................................................................................................................151 19.2 Block Diagrams ....................................................................................................................152 19.3 How It Works ........................................................................................................................154 19.3.1 Usage Modes and Configuration .........................................................................154 19.3.2 I/O Drive Modes...................................................................................................154 19.3.2.1 Drive Mode on Reset ...........................................................................155 19.3.2.2 High Impedance Analog.......................................................................155 19.3.2.3 High Impedance Digital ........................................................................155 19.3.2.4 Resistive Pull Up or Resistive Pull Down .............................................155 19.3.2.5 Open Drain, Drives High and Drives Low ............................................156 19.3.2.6 Strong Drive .........................................................................................156 19.3.2.7 Resistive Pull Up and Pull Down..........................................................156 19.3.3 Slew Rate Control................................................................................................156 19.3.4 Digital I/O Controlled by Port Register.................................................................156 19.3.4.1 Port Configuration Registers ................................................................156 19.3.4.2 Pin Wise Configuration Register Alias .................................................156 19.3.4.3 Port Wide Configuration Register Alias................................................157 19.3.5 Digital I/O Controlled Through DSI ......................................................................158 19.3.5.1 DSI Output ...........................................................................................158 19.3.5.2 DSI Input ..............................................................................................158 19.3.5.3 DSI for Output Enable Control .............................................................158 10 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 19.3.6 Analog I/O ............................................................................................................159 19.3.7 LCD Drive ............................................................................................................159 19.3.8 CapSense ............................................................................................................159 19.3.9 External Memory Interface (EMIF).......................................................................160 19.3.10 SIO Functions and Features ................................................................................160 19.3.10.1 Regulated Output Level........................................................................160 19.3.10.2 Adjustable Input Level ..........................................................................160 19.3.10.3 Hot Swap..............................................................................................161 19.3.11 Special Functionality ............................................................................................161 19.3.12 I/O Port Reconfiguration ......................................................................................162 19.3.13 Power Up I/O Configuration .................................................................................162 19.3.14 Overvoltage Tolerance .........................................................................................162 19.3.15 I/O Power Supply .................................................................................................163 19.3.16 Sleep Mode Behavior...........................................................................................163 19.3.17 Low-power Behavior ............................................................................................163 19.4 Port Interrupt Controller Unit.................................................................................................163 19.4.1 Features...............................................................................................................163 19.4.2 Interrupt Controller Block Diagram.......................................................................163 19.4.3 Function and Configuration..................................................................................164 19.5 Register Summary ................................................................................................................165 20. Flash, Configuration Protection 167 20.1 Flash Protection....................................................................................................................167 20.2 Device Security.....................................................................................................................168 20.3 Configuration Segment Protection........................................................................................168 20.3.1 Locking/Unlocking Segment Configuration Register............................................169 20.3.2 Locking and Protecting Segments .......................................................................169 20.3.3 Example ...............................................................................................................171 20.4 Frequently Asked Questions About Flash Protection and Device Security ..........................171 Section E: Digital System 173 Top Level Architecture ...................................................................................................................174 21. Universal Digital Blocks (UDBs) 175 21.1 Features................................................................................................................................175 21.2 Block Diagram ......................................................................................................................175 21.3 How It Works ........................................................................................................................176 21.3.1 PLDs ....................................................................................................................176 21.3.1.1 PLD Macrocells ....................................................................................177 21.3.1.2 PLD Carry Chain ..................................................................................179 21.3.1.3 PLD Configuration ................................................................................179 21.3.2 Datapath ..............................................................................................................179 21.3.2.1 Overview ..............................................................................................181 21.3.2.2 Datapath FIFOs....................................................................................182 21.3.2.3 FIFO Status ..........................................................................................189 21.3.2.4 Datapath ALU.......................................................................................189 21.3.2.5 Datapath Inputs and Multiplexing .........................................................192 21.3.2.6 CRC/PRS Support................................................................................192 21.3.2.7 Datapath Outputs and Multiplexing ......................................................195 21.3.2.8 Datapath Parallel Inputs and Outputs ..................................................197 21.3.2.9 Datapath Chaining................................................................................197 21.3.2.10 Dynamic Configuration RAM ................................................................198 21.3.3 Status and Control Module...................................................................................199 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 11 21.3.3.1 Status and Control Mode .....................................................................200 21.3.3.2 Control Register Operation ..................................................................202 21.3.3.3 Parallel Input/Output Mode ..................................................................203 21.3.3.4 Counter Mode ......................................................................................204 21.3.3.5 Sync Mode ...........................................................................................205 21.3.3.6 Status and Control Clocking.................................................................205 21.3.3.7 Auxiliary Control Register.....................................................................205 21.3.3.8 Status and Control Register Summary.................................................206 21.3.4 Reset and Clock Control Module .........................................................................206 21.3.4.1 Clock Control........................................................................................207 21.3.4.2 Reset Control .......................................................................................209 21.3.4.3 UDB POR Initialization .........................................................................211 21.3.5 UDB Addressing ..................................................................................................212 21.3.5.1 Working Register Address Space ........................................................212 21.3.5.2 Configuration Register Address Space ................................................214 21.3.5.3 UDB Configuration Address Space......................................................214 21.3.5.4 Routing Configuration Address Space .................................................214 21.3.6 System Bus Access Coherency ..........................................................................215 21.3.6.1 Simultaneous System Bus Access.......................................................215 21.3.6.2 Coherent Accumulator Access (Atomic Reads and Writes).................215 21.4 UDB Working Register Reference........................................................................................216 22. UDB Array and Digital System Interconnect 22.1 22.2 22.3 22.4 217 Features ...............................................................................................................................217 Block Diagram ......................................................................................................................217 How It Works ........................................................................................................................218 UDB Array System Interface ...............................................................................................220 22.4.1 UDB Array POR Initialization ...............................................................................220 22.4.2 UDB POR Configuration Sequence ....................................................................221 22.4.2.1 Quadrant Route Disable ......................................................................222 22.4.3 UDB Sleep and Power Control ...........................................................................222 22.4.4 UDB Register References and Address Mapping ...............................................222 23. Controller Area Network (CAN) 225 23.1 Features ...............................................................................................................................225 23.2 Block Diagram ......................................................................................................................226 23.3 CAN Message Frames .........................................................................................................226 23.3.1 Data Frames ........................................................................................................226 23.3.1.1 Standard Data Frame...........................................................................226 23.3.1.2 Extended Data Frame ..........................................................................227 23.3.2 Remote Frame.....................................................................................................228 23.3.3 Error Frame .........................................................................................................228 23.3.4 Overload Frame...................................................................................................228 23.4 Transmitting Messages in CAN ............................................................................................228 23.4.1 Message Arbitration.............................................................................................230 23.4.2 Message Transmit Process .................................................................................230 23.4.3 Message Abort ....................................................................................................231 23.4.4 Transmitting Extended Data Frames ...................................................................231 23.5 Receiving Messages in CAN ................................................................................................231 23.5.1 Message Receive Process ..................................................................................232 23.5.2 Acceptance Filter .................................................................................................232 23.5.2.1 Example ...............................................................................................232 23.5.3 DeviceNet Filtering ..............................................................................................234 12 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 23.6 23.7 23.8 23.9 23.5.4 Filtering of Extended Data Frames ......................................................................234 23.5.5 Receiver Message Buffer Linking ........................................................................235 Remote Frames ....................................................................................................................235 23.6.1 Transmitting a Remote Frame by the Requesting Node......................................236 23.6.2 Receiving a Remote Frame .................................................................................236 23.6.3 RTR Auto Reply ...................................................................................................236 23.6.4 Remote Frames in Extended Format...................................................................236 Bit Time Configuration ..........................................................................................................236 23.7.1 Allowable Bit Rates and System Clock (CLK_BUS) ............................................236 23.7.2 Setting Bit Rate TSEG1 and TSEG2 ...................................................................237 23.7.2.1 Example ...............................................................................................238 Error Handling and Interrupts in CAN ...................................................................................238 23.8.1 Types of Errors.....................................................................................................238 23.8.1.1 BIT Error...............................................................................................238 23.8.1.2 FORM Error..........................................................................................238 23.8.1.3 ACKNOWLEDGE Error ........................................................................238 23.8.1.4 CRC Error.............................................................................................238 23.8.1.5 STUFF Error.........................................................................................239 23.8.2 Error States in CAN..............................................................................................239 23.8.3 Interrupt Sources in CAN .....................................................................................239 Operating Modes in CAN......................................................................................................240 23.9.1 Listen Only Mode .................................................................................................240 23.9.2 Run/Stop Mode ....................................................................................................240 24. USB 241 24.1 Features................................................................................................................................241 24.2 Block Diagram ......................................................................................................................242 24.2.1 Serial Interface Engine (SIE) ...............................................................................243 24.2.2 Arbiter ..................................................................................................................244 24.2.2.1 SIE Interface Module............................................................................244 24.2.2.2 CPU Interface Block .............................................................................244 24.2.2.3 Memory Interface .................................................................................244 24.2.2.4 DMA Interface ......................................................................................244 24.2.2.5 Arbiter Logic .........................................................................................244 24.2.2.6 Synchronization Block ..........................................................................245 24.3 How it Works.........................................................................................................................245 24.3.1 Operating Frequency ...........................................................................................245 24.3.2 Operating Voltage ................................................................................................245 24.3.3 Transceiver ..........................................................................................................245 24.3.4 Endpoints .............................................................................................................245 24.3.5 Transfer Types .....................................................................................................246 24.3.6 Interrupts..............................................................................................................246 24.4 Logical Transfer Modes ........................................................................................................246 24.4.1 Store and Forward Mode .....................................................................................248 24.4.1.1 No DMA Access ...................................................................................248 24.4.1.2 Manual DMA Access ............................................................................249 24.4.2 Cut Through Mode ...............................................................................................251 24.4.3 Control Endpoint Logical Transfer........................................................................253 24.5 PS/2 and CMOS I/O Modes..................................................................................................254 24.6 Register List .........................................................................................................................255 25. Timer, Counter, and PWM 257 25.1 Features................................................................................................................................257 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 13 25.2 Block Diagram ......................................................................................................................257 25.3 How It Works ........................................................................................................................258 25.3.1 Clock Selection ....................................................................................................258 25.3.2 Enabling and Disabling Block ..............................................................................259 25.3.3 Input Signal Characteristics .................................................................................259 25.3.3.1 Enable Signal .......................................................................................260 25.3.3.2 Capture Signal .....................................................................................260 25.3.3.3 Timer Reset Signal...............................................................................262 25.3.3.4 Kill Signal .............................................................................................262 25.3.4 Operating Modes .................................................................................................263 25.3.4.1 Timer Mode – Free Run Mode .............................................................263 25.3.4.2 Gated Timer Mode ...............................................................................264 25.3.4.3 Pulse-width Modulator Mode................................................................268 25.3.4.4 One Shot Mode ....................................................................................271 25.3.5 Interrupt Enabling ................................................................................................271 25.3.6 Sleep Mode Behavior ..........................................................................................272 25.4 Register Listing.....................................................................................................................272 26. I 2C 273 26.1 Features ...............................................................................................................................273 26.2 Background Information .......................................................................................................274 26.2.1 I2C Bus Description .............................................................................................274 26.2.2 Typical I2C Data Transfer ....................................................................................274 26.3 How It Works ........................................................................................................................274 26.3.1 Bus Stalling (Clock Stretching) ............................................................................275 26.3.2 System Management Bus ...................................................................................275 26.3.3 Pin Connections ..................................................................................................275 26.3.4 I2C Interrupts .......................................................................................................275 26.3.5 Control by Registers ............................................................................................275 26.3.6 Operating the I2C Interface..................................................................................275 26.3.6.1 Slave Mode ..........................................................................................277 26.3.6.2 Master Mode ........................................................................................278 26.3.6.3 Multi-Master Mode ...............................................................................279 26.4 Hardware Address Compare ................................................................................................279 26.5 Wake from Sleep ..................................................................................................................279 26.6 Slave Mode Transfer Examples ...........................................................................................280 26.6.1 Slave Receive......................................................................................................281 26.6.2 Slave Transmit.....................................................................................................282 26.7 Master Mode Transfer Examples .........................................................................................283 26.7.1 Single Master Receive.........................................................................................283 26.7.2 Single Master Transmit........................................................................................284 26.8 Multi-Master Mode Transfer Examples.................................................................................285 26.8.1 Multi-Master, Slave Not Enabled .........................................................................285 26.8.2 Multi-Master, Slave Enabled ................................................................................286 27. Digital Filter Block (DFB) 287 27.1 Features ...............................................................................................................................287 27.2 Block Diagram ......................................................................................................................287 27.3 How It Works ........................................................................................................................288 27.3.1 Controller .............................................................................................................288 27.3.1.1 FSM RAM.............................................................................................289 27.3.1.2 Program Counter..................................................................................290 27.3.1.3 Control Store ........................................................................................290 14 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 27.3.1.4 Next State Decoder ..............................................................................290 Datapath ..............................................................................................................291 27.3.2.1 MAC .....................................................................................................292 27.3.2.2 ALU ......................................................................................................292 27.3.2.3 Shifter and Rounder .............................................................................292 27.3.3 Address Calculation Unit......................................................................................293 27.3.4 Bus Interface and Register Descriptions..............................................................293 27.3.4.1 Streaming Mode ...................................................................................293 27.3.4.2 Block Transfer Modes ..........................................................................294 27.3.4.3 Result Handling ....................................................................................295 27.3.4.4 Data Alignment.....................................................................................297 27.3.4.5 DMA and Semaphores .........................................................................297 27.3.4.6 DSI Routed Inputs and Outputs ...........................................................297 27.4 DFB Instruction Set...............................................................................................................298 27.5 Usage Model.........................................................................................................................301 27.3.2 Section F: Analog System 303 Top Level Architecture ...................................................................................................................304 28. Switched Capacitor/Continuous Time 305 28.1 Features................................................................................................................................305 28.2 Block Diagram ......................................................................................................................305 28.3 How it Works.........................................................................................................................307 28.3.1 Operational Mode of Block is Set.........................................................................307 28.4 Naked Opamp.......................................................................................................................307 28.4.1 Bandwidth/Stability Control ..................................................................................307 28.4.1.1 BIAS_CONTROL..................................................................................307 28.4.1.2 SC_COMP[1:0].....................................................................................308 28.4.1.3 SC_REDC[1:0] .....................................................................................308 28.5 Continuous Time Unity Gain Buffer ......................................................................................309 28.6 Continuous Time Programmable Gain Amplifier ..................................................................309 28.7 Continuous Time Transimpedance Amplifier........................................................................310 28.8 Continuous Time Mixer.........................................................................................................312 28.9 Sampled Mixer......................................................................................................................313 28.10 Delta Sigma Modulator .........................................................................................................315 28.10.1 First-Order Modulator, Incremental Mode ............................................................316 28.11 Track and Hold Amplifier ......................................................................................................317 29. Analog Routing 319 29.1 Features................................................................................................................................319 29.2 Block Diagram ......................................................................................................................319 29.3 How it Works.........................................................................................................................322 29.3.1 Analog Globals (AGs) ..........................................................................................322 29.3.2 Analog Mux Bus (AMUXBUS)..............................................................................322 29.3.3 Liquid Crystal Display Bias Bus (LCDBUS) .........................................................322 29.3.4 Analog Local Bus (abus)......................................................................................324 29.3.5 Switches and Multiplexers ...................................................................................324 29.3.5.1 Control of Analog Switches ..................................................................324 29.4 Analog Resource Blocks – Routing and Interface ................................................................326 29.4.1 Digital-to-Analog Converter (DAC).......................................................................327 29.4.2 Comparator ..........................................................................................................328 29.4.3 Delta Sigma Modulator (DSM) .............................................................................329 29.4.4 Switched Capacitor ..............................................................................................330 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 15 29.4.5 Opamp .................................................................................................................331 29.4.6 Low-Pass Filter (LPF) ..........................................................................................331 29.5 Low-Power Analog Routing Considerations .........................................................................332 29.5.1 Mitigating Analog Routes with Degraded Low-power Signal Integrity .................332 29.6 Analog Routing Register Summary ......................................................................................333 30. Comparators 335 30.1 Features ...............................................................................................................................335 30.2 Block Diagram ......................................................................................................................335 30.3 How it Works ........................................................................................................................336 30.3.1 Input Configuration ..............................................................................................336 30.3.2 Power Configuration ............................................................................................336 30.3.3 Output Configuration ...........................................................................................336 30.3.4 Hysteresis ............................................................................................................337 30.3.5 Wake Up ..............................................................................................................337 30.3.6 Comparator Clock................................................................................................337 30.3.7 Offset Trim ...........................................................................................................337 30.3.8 Register Summary ...............................................................................................339 31. Opamp 341 31.1 Features ..............................................................................................................................341 31.2 Block Diagram ......................................................................................................................342 31.3 How it Works ........................................................................................................................342 31.3.1 Input and Output Configuration ...........................................................................342 31.3.2 Power Configuration ............................................................................................342 31.3.3 Buffer Configuration.............................................................................................343 31.3.4 Register Summary ...............................................................................................343 32. LCD Direct Drive 345 32.1 Features ...............................................................................................................................345 32.2 LCD System Operational Modes ..........................................................................................345 32.3 LCD Always Active ...............................................................................................................346 32.3.1 Functional Description .........................................................................................347 32.3.1.1 LCD DAC .............................................................................................347 32.3.1.2 LCD Driver Block..................................................................................348 32.3.1.3 UDB......................................................................................................351 32.3.1.4 DMA .....................................................................................................351 32.4 LCD Low-Power Mode .........................................................................................................351 32.4.1 Functional Description .........................................................................................352 32.4.1.1 LCD Timer............................................................................................352 32.4.1.2 UDB......................................................................................................352 32.4.1.3 DMA .....................................................................................................353 32.4.1.4 LCD DAC and Driver: Low Power Feature...........................................353 32.4.2 Timing Diagram for LCD Low-Power Mode .........................................................355 32.5 LCD Usage Models ..............................................................................................................357 33. CapSense 359 33.1 Features ...............................................................................................................................359 33.2 Block Diagram ......................................................................................................................359 33.3 How It Works ........................................................................................................................360 33.3.1 Reference Driver .................................................................................................360 33.3.2 Low-pass Filter ....................................................................................................360 33.3.3 Analog Mux Bus ..................................................................................................360 33.3.4 GPIO Configuration for CapSense ......................................................................360 16 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 33.3.5 Other Resources..................................................................................................361 33.4 CapSense Delta Sigma Algorithm ........................................................................................362 34. Temperature Sensor 34.1 34.2 34.3 34.4 365 Features................................................................................................................................365 Block Diagram ......................................................................................................................365 How It Works ........................................................................................................................366 Command and Status Interface ............................................................................................366 34.4.1 Status Codes........................................................................................................366 34.4.2 Temperature Sensor Commands .........................................................................366 34.4.2.1 Get Temperature ..................................................................................366 34.4.2.2 Setup Temperature Sensor ..................................................................367 34.4.2.3 Disable Temperature Sensor ...............................................................368 35. Digital-to-Analog Converter 369 35.1 Features................................................................................................................................369 35.2 Block Diagram ......................................................................................................................369 35.3 How It Works ........................................................................................................................370 35.3.1 Current DAC ........................................................................................................370 35.3.2 Voltage DAC ........................................................................................................370 35.3.3 Output Routing Options .......................................................................................370 35.3.4 Making a Higher Resolution DAC ........................................................................371 35.4 Register List .........................................................................................................................372 36. Precision Reference 373 36.1 Block Diagram ......................................................................................................................373 36.2 How It Works ........................................................................................................................373 37. Delta Sigma Converter 377 37.1 Features................................................................................................................................377 37.2 Block Diagram ......................................................................................................................377 37.3 How It Works ........................................................................................................................378 37.3.1 Input Buffer ..........................................................................................................378 37.3.2 Delta Sigma Modulator .......................................................................................379 37.3.2.1 Clock Selection.....................................................................................380 37.3.2.2 Capacitance Configuration ...................................................................380 37.3.2.3 Gain Configuration ...............................................................................381 37.3.2.4 Power Configuration.............................................................................382 37.3.2.5 Other Configuration Options.................................................................386 37.3.2.6 Quantizer..............................................................................................386 37.3.2.7 Reference Options ...............................................................................386 37.3.2.8 Reference for DSM: Usage Guidelines ................................................389 37.3.3 Analog Interface...................................................................................................390 37.3.3.1 Conversion of Thermometric Code to Two’s Complement...................391 37.3.3.2 Modulation Input...................................................................................391 37.3.3.3 Clock Selection and Synchronization ...................................................391 37.3.4 Decimator.............................................................................................................391 37.3.4.1 Shifters .................................................................................................391 37.3.4.2 CIC Filter ..............................................................................................392 37.3.4.3 Post Processing Filter ..........................................................................392 37.3.5 ADC Conversion Time .........................................................................................393 37.3.6 Coherency Protection ..........................................................................................393 37.3.6.1 Protecting Writes (Gain/Offset) with Coherency Checking...................393 37.3.6.2 Protecting Reads (Output Sample) with Coherency Checking.............394 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 17 37.3.7 Modes of Operation .............................................................................................394 38. Successive Approximation Register ADC 397 38.1 Features ...............................................................................................................................397 38.2 How It Works ........................................................................................................................398 38.2.1 Input Selection .....................................................................................................398 38.2.2 Clock Selection ....................................................................................................398 38.2.3 Input Sampling.....................................................................................................398 38.2.4 Power Modes.......................................................................................................398 38.2.5 Reference Selection ............................................................................................398 38.2.6 Operational Modes ..............................................................................................399 38.2.7 SAR ADC Output .................................................................................................400 Section G: Program and Debug 401 Top Level Architecture ...................................................................................................................401 39. Test Controller 403 39.1 Features ...............................................................................................................................403 39.2 Block Diagram ......................................................................................................................403 39.3 Background Information .......................................................................................................404 39.3.1 JTAG Interface.....................................................................................................404 39.3.2 Serial Wire Debug Interface ................................................................................407 39.4 How It Works ........................................................................................................................409 39.4.1 JTAG Interface Implementation ...........................................................................409 39.4.2 SWD Interface Implementation............................................................................ 411 39.4.3 Acquiring the Debug Port ....................................................................................412 39.4.4 Cortex M3 DAP Access Criteria ..........................................................................412 39.4.5 Device Programming ...........................................................................................413 39.4.6 Boundary Scan ....................................................................................................413 39.4.7 Boundary Scan Pin Order....................................................................................413 40. Cortex-M3 Debug and Trace 415 40.1 Features ...............................................................................................................................415 40.2 How It Works ........................................................................................................................416 40.2.1 Test Controller (TC) .............................................................................................416 40.2.2 PSoC 5LP JTAG Instructions...............................................................................417 40.2.2.1 Debug Port and Access Port Registers................................................417 40.2.2.2 Test Controller Interface Pins...............................................................417 40.2.3 TRACEPORT.......................................................................................................417 40.3 Core Debug ..........................................................................................................................417 40.3.1 Enabling the Debug .............................................................................................417 40.3.2 Halting .................................................................................................................417 40.3.3 Stepping...............................................................................................................417 40.3.4 Accessing PSoC Memory and Registers.............................................................417 40.4 System Debug ......................................................................................................................418 40.4.1 Flash Patch and Breakpoint (FPB) Unit...............................................................418 40.4.2 Data Watchpoint and Trace (DWT)......................................................................419 40.4.3 Instrumentation Trace Macrocell (ITM) ................................................................419 40.4.4 Embedded Trace Macrocell (ETM)......................................................................419 40.5 Tracing Interface...................................................................................................................420 40.5.1 Single Wire Viewer ..............................................................................................420 40.5.1.1 Enabling SWV ......................................................................................421 40.5.1.2 Communicating with SWV....................................................................421 40.5.2 TRACEPORT.......................................................................................................421 18 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 40.5.2.1 Enabling TRACEPORT ........................................................................421 40.5.2.2 Communicating with TRACEPORT......................................................421 40.5.3 Using Multiple Interfaces Simultaneously ............................................................421 41. Nonvolatile Memory Programming 423 41.1 Features................................................................................................................................423 41.2 Block Diagram ......................................................................................................................423 41.3 How It Works ........................................................................................................................424 41.3.1 Commands...........................................................................................................424 41.3.1.1 Command Code Descriptions ..............................................................425 41.3.1.2 Command Failure Codes .....................................................................426 41.3.2 Register Summary ...............................................................................................426 41.3.3 Flash Protection Settings .....................................................................................427 Glossary 429 Index 445 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 19 20 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Section A: Overview This document encompasses the PSoC® 5LP family of devices. In conjunction with the device datasheet and PSoC 5LP Registers TRM, it contains complete and detailed information about how to design with the IP blocks that construct a PSoC 5LP device. This document describes the analog and digital architecture, and helps to better understand the features of the device. This section consists of the following chapters: ■ Introduction chapter on page 23 ■ Getting Started chapter on page 29 ■ Document Construction chapter on page 31 See the PSoC® 5LP Registers TRM (Technical Reference Manual) for complete register sets. Document Revision History Table 1-1. PSoC® 5LP Architecture TRM (Technical Reference Manual) Revision History Revision ** Issue Date 09/25/2012 Origin of Change VVSK Description of Change Initial version of the PSoC 5LP Architecture TRM Updated use of bypass capacitor for reference voltage (section 38.2.5 Reference Selection) Added information on the effect of changing pin modes (section 19.3.2 I/O Modes) *A 11/21.2012 VVSK Updated Tables 20-3 to 20-6 Added sections 18.1.2 Low-Voltage Reset and High-Voltage Reset and 18.1.6.1 Preservation of Reset Status; updates to 15.3.3. Voltage Monitoring Added information on accessing DAP with third-party tools (section 8.2.5 DPS[1:0] *B 06/18/2013 ANTO Added 18.1.2 Low-Voltage Reset and High- Voltage Reset and 18.1.6.1 Preservation of Reset Status; updates to 15.3.3. Voltage Monitoring. Added information on effect of changing pin modes in section 19.3.2 I/O Modes. Added information on accessing DAP with third-party tools in section 8.2.5. Added a note to sections 10.3 and 11.3. Updated section 25.3.4.1 (Period register setting to EN = 1). Updated Figure 14-1 and Tables 20-3 to 20-6. Modified the Datapath Top Level Diagram Updates to PHUB and DMAC chapters Updated Drive Modes diagram in the I/O System chapter on page 151. Corrected section 14.3.2.2 32.768 kHz Crystal Oscillator to mention the active mode operating current. *C 09/26/2013 ANTO Removed the comparator as a wakeup source from hibernate in section 16.5.2 Exiting Hibernate Mode. Corrected FTW register name in section 16.6.2 Fast Timewheel (FTW). Updated Successive Approximation Register ADC chapter on page 397 Clarified device behavior in multiple sections *D 07/02/2015 GJV Updated figures 6-2, 6-3, 18-1, 23-13, 25-4, and 28-1. Updated tables 4-2, 7-3, 37-1, and 41-2; added tables 18-2 and 21-6. Updated logo and copyright disclaimer. Added CAN initialization note to section 23.4.2 step #1. *E 10/14/2016 GJV Removed AltAct to Sleep transition from figure 16-1. Removed switches between OpAmp outputs and pins in Figure 29-11. Corrected Trace Port bitfields in Figure 40-4. Fixed broken link to PSoC 5LP Programming Specification. *F 05/31/2017 SHEA Updated logo and copyright information. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 21 Section A: Overview 22 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 1. Introduction With a unique array of configurable digital and analog blocks, the Programmable System-on-Chip (PSoC®) is a true systemlevel solution, offering a modern method of signal acquisition, processing, and control with exceptional accuracy, high bandwidth, and superior flexibility. Its analog capability spans the range from thermocouples (DC voltages) to ultrasonic signals. PSoC 5LP (CY8C58LP, CY8C56LP, CY8C54LP, CY8C52LP) families are fully scalable 32-bit PSoC platform devices that have these characteristics: ■ High-performance, configurable digital system that supports a wide range of communication interfaces, such as USB, I2C, and CAN ■ High-precision, high-performance analog system with up to 20-bit ADC, DACs, comparators, opamps, and programmable blocks to create PGAs, TIAs, mixers, and so on ■ Easily configurable logic array ■ Flexible routing to all pins ■ High-performance, 32-bit ARM Cortex-M3 core ■ PSoC Creator, an integrated development environment software This document describes PSoC 5LP devices in detail. Using this information, designers can easily create system-level designs, using a rich library of prebuilt components, or custom verilog, and a schematic entry tool that uses the standard design blocks. PSoC 5LP devices provide unparalleled opportunities for analog and digital bill of materials (BOM) integration, while easily accommodating last-minute design changes. For a discussion of the registers of the PSoC 5LP device, see Cypress document 001-82120, the PSoC 5LP Registers TRM. It lists all the registers in mapping tables in address order. 1.1 Top Level Architecture Figure 1-1 on page 24 shows the major components of PSoC 5LP devices. The PSoC 5LP device uses the 32-bit Cortex M3 core. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 23 Introduction Figure 1-1. Top Level Architecture for PSoC 5LP Devices Analog Interconnect Universal Digital Block Array (N x UDB) Clock Tree GPIOs 32.768 kHz (Optional) IMO UDB UDB UDB UDB UDB I2C Slave 16-Bit PRS UDB UDB UDB 8-Bit Timer 8-Bit SPI Logic 12-Bit SPI UDB UDB UDB UDB UDB Logic UDB UDB CAN 2.0 Nx Timer, Counter, PWM I2C Master/Slave FS USB 2.0 USB PHY D+ D- GPIOs UDB 16-Bit PWM Sequencer Quadrature Decoder 8-Bit Timer Xtal Osc SIO DIGITAL SYSTEM SYSTEM WIDE RESOURCES Usage Example for UDB 4 to 25 MHz (Optional) GPIOs Digital Interconnect 12-Bit PWM UART RTC Timer MEMORY SYSTEM WDT and Wake EEPROM SRAM EMIF FLASH CPU SYSTEM Cortex-M3 CPU Interrupt Controller Program, Debug Program Boundary Scan Power Management System LCD Direct Drive Digital Filter Block POR and LVD Sleep Power GPIOs SIOs Clocking System ANALOG SYSTEM ADCs N x SAR ADC + Nx Opamp – 3 per Opamp N x SC/CT Blocks (TIA, PGA, Mixer, etc.) Auxiliary ADC 1.8-V LDO Temperature Sensor SMP CapSense N x DAC + Nx DEL SIG ADC Nx CMP – GPIOs GPIOs Debug, Trace PHUB DMA ILO 1.71 to 5.5 V GPIOs SYSTEM BUS 0.5 to 5.5 V (Optional) 24 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Introduction 1.2 Features PSoC 5LP devices have these major components. See Figure 1-1 on page 24. slower, save power, or use its cycles to improve the performance of firmware algorithms. 1.3.4 Cache Controller ■ Cortex-M3 central processing unit (CPU) with a nested vectored interrupt controller and a high-performance DMA controller ■ Several types of memory elements including SRAM, flash, and EEPROM In PSoC 5LP devices, the flash cache also reduces system power consumption by reducing the frequency with which flash is accessed. The processor speed itself is configurable allowing for active power consumption tuned for specific applications. ■ System integration features, such as clocking, a featurerich power system, and versatile programmable inputs and outputs 1.4 ■ Digital system that includes configurable universal digital blocks (UDBs) and specific function peripherals, such as CAN and USB ■ Analog subsystem that includes configurable switched capacitor (SC) and continuous time (CT) blocks, up to 20-bit Delta Sigma converters, 8-bit DACs that can be configured for 12-bit operation, more than one SAR ADC, comparators, PGAs, and more ■ Programming and debug system through JTAG, serial wire debug (SWD), and single wire viewer (SWV) 1.3 1.3.1 CPU System Memory The PSoC nonvolatile subsystem consists of flash, bytewritable EEPROM, and nonvolatile configuration options. The CPU can reprogram individual blocks of flash, enabling boot loaders. An Error Correcting Code (ECC) can enable high-reliability applications. A powerful and flexible protection model allows you to selectively lock blocks of memory for read and write protection, securing sensitive information. The byte-writable EEPROM is available on-chip for the storage of application data. Additionally, selected configuration options, such as boot speed and pin drive mode, are stored in nonvolatile memory, allowing settings to become active immediately after power-onreset (POR). Processor The PSoC 5LP CPU subsystem is built around a 32-bit three stage pipelined ARM Cortex-M3 processor running up to 80 MHz. The PSoC 5LP instruction set is the same as the Thumb-2 instruction set available on standard Cortex- M3 devices. 1.3.2 Interrupt Controller The CPU subsystem includes a programmable Nested Vectored Interrupt Controller (NVIC), DMA (Direct Memory Access) controller, flash cache ECC, and RAM. The NVIC of PSoC 5LP devices provide low latency by allowing the CPU to vector directly to the first address of the interrupt service routine, bypassing the jump instruction required by other architectures. The PSoC 5LP interrupt controller also offers a few advanced interrupt management capabilities, such as interrupt tail chaining to improve stack management with multiple pending interrupts providing lower latency. 1.3.3 DMA Controller The DMA controller allows peripherals to exchange data without CPU involvement. This allows the CPU to run PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 25 Introduction 1.5 System Wide Resources The individual elements of system wide resources are discussed in these sections. 1.5.1 I/O Interfaces 1.5.3 Power Supply PSoC 5LP devices support extensive supply operating ranges from 1.7 V to 5.5 V, allowing operation from regulated supplies such as 1.8 ± 5%, 2.5 V ± 10%, 3.3 V ± 10%, 5.0 V ± 10%, or directly from a wide range of battery types. PSoC 5LP devices have three I/O types: 1.5.3.1 ■ General Purpose Input/Output (GPIO) – Every GPIO has analog I/O, digital I/O, LCD drive, CapSense®, flexible interrupt, and slew rate control capability. All I/Os have a large number of drive modes that are set at POR. PSoC 5LP devices also provide up to four individual I/O voltage domains through the VDDIO pins. ■ Special Input/Output (SIO) – The SIOs on PSoC 5LP devices allow setting VOH independently of VDDIO when used as outputs. When SIOs are in input mode, they are high impedance, even when the device is not powered or when the pin voltage goes above the supply voltage. This makes the SIO ideal for use on an I2C bus where the PSoC 5LP devices are not powered, even though other devices on the bus are powered. The SIO pins also have high-current sink capability for applications such as LED drive. The PSoC platform provides an integrated high-efficiency synchronous boost converter that is used to power the device from supply voltages as low as 0.5 V. This converter enables the device to power directly from a single battery or solar cell. You can employ the boost converter to generate other voltages required by the device, such as a 3.3-V supply for LCD glass drive. The boost output is available on the VBOOST pin, allowing other devices in the application to draw power from the PSoC device. ■ USB Input/Output (USBIO) – For devices with FullSpeed USB, the USB physical interface is also provided (USBIO). When not using USB, these pins can be used for limited digital functionality and device programming. 1.5.2 Sleep Modes The PSoC platform supports five low-power sleep modes, from the lowest current RAM retention mode (hibernation) to the full function active mode. A 1.0-A RTC mode runs the optional 32.768-kHz watch crystal continuously to drive the RTC timer that is used to maintain RTC. Power to all major functional blocks, including the programmable digital and analog peripherals, is controlled independently by firmware. This function allows low-power background processing when some peripherals are not in use. Internal Clock Generators PSoC devices incorporate flexible internal clock generators, designed for high stability and factory-trimmed for absolute accuracy. The internal main oscillator (IMO) is the master clock base for the system with 1% absolute accuracy at 3 MHz. The IMO can be configured to run from 3 MHz up to 48 MHz. Multiple clock derivatives are generated from the main clock frequency to meet application needs. PSoC 5LP devices provide a PLL to generate system clock frequencies up to the maximum operating frequency of the device (80 MHz). The PLL can be driven from the IMO, an external crystal, or an external reference clock. The devices also contain a separate, very low power internal low-speed oscillator (ILO) for the sleep and watchdog timers. The ILO provides two primary outputs, 1 kHz and 100 kHz. A 32.768kHz external watch crystal is also supported for use in realtime clock (RTC) applications. The clocks, together with programmable clock dividers, provide the flexibility to integrate most timing requirements. 26 1.5.3.2 Boost Converter 1.6 Digital System The digital subsystems of PSoC 5LP devices provide these devices their first half of unique configurability. The subsystem connects a digital signal from any peripheral to any pin through the Digital System Interconnect (DSI). It also provides functional flexibility through an array of small, fast, low-power universal digital blocks (UDBs). Each UDB contains Programmable Array Logic (PAL) and Programmable Logic Device (PLD) functionality, together with a small state machine engine to support a wide variety of peripherals. In addition to the flexibility of the UDB array, PSoC devices provide configurable digital blocks targeted at specific functions. These blocks include 16-bit timer/counter/PWM blocks, I2C slave/master/multi-master, Full Speed USB, and CAN 2.0b. See the device datasheet for a list of available specific function digital blocks. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Introduction 1.7 Analog System The PSoC analog subsystem provides the device the second half of its unique configurability. All analog performance is based on a highly accurate absolute voltage reference. 1.7.4 Digital-to-Analog Converters Four high-speed voltage or current DACs support 8-bit output signals at waveform frequencies up to 8 MHz and can be routed out of any GPIO pin. These DACs can be combined together to create a higher resolution 12-bit DAC. The configurable analog subsystem includes: ■ Analog muxes ■ Comparators ■ Voltage references ■ Opamps ■ Mixers ■ Transimpedance amplifiers (TIA) ■ Analog-to-digital converters (ADC) ■ Digital-to-analog converters (DAC) ■ Digital filter block (DFB) Higher resolution voltage DAC outputs are created using the UDB array to create a pulse width modulated (PWM) DAC of up to 10 bits, at up to 48 kHz. The digital DACs in each UDB support PWM, PRS, or Delta Sigma algorithms with programmable widths. 1.7.5 Additional Analog Subsystem Components All GPIO pins can route analog signals into and out of the device, using the internal analog bus. This feature allows the device to interface up to 62 discrete analog signals. In addition to the ADCs, DACs, and the DFB, the analog subsystem provides components such as multiple comparators, uncommitted opamps, and configurable switched capacitor/continuous time (SC/CT) blocks supporting transimpedance amplifiers, programmable gain amplifiers, and mixers. 1.7.1 1.8 Delta Sigma ADC The heart of the analog subsystem is a fast, accurate, configurable Delta Sigma ADC. With less than 100 µV offset, a gain error of ±0.1%, integral nonlinearity (INL) less than 1 LSB, differential nonlinearity (DNL) less than 0.5 LSB, and signal-to-noise ratio (SNR) better than 90 dB (Delta Sigma) in 16-bit mode, this converter addresses a wide variety of precision analog applications, including some of the most demanding sensors. 1.7.2 Successive Approximation Register ADC Another type of ADC seen on PSoC 5LP devices is the Successive Approximation Register (SAR) ADC. Featuring 12bit conversions at up to 1 Msps, it offers low nonlinearity, low offset errors, and an SNR better than 70 dB; it is well suited for a variety of higher-speed analog applications. Some PSoC devices offer both types of ADC and can have multiple instances of each. See the device datasheet for specific details. 1.7.3 Program and Debug JTAG (4-wire) or serial wire debugger (SWD) (2-wire) interfaces are used for programming and debug. The 1-wire single wire viewer (SWV) can also be used for “printf” style debugging. By combining SWD and SWV, you can implement a full debugging interface with just three pins. These standard interfaces enable debugging or programming the PSoC device with a variety of hardware solutions from Cypress or third party vendors. PSoC 5LP devices support on-chip break points, and an instruction and data trace memory for debug. The PSoC 5LP device offers many more advanced debugging features, such as Flash patch breakpoint capability to update instructions without reprogramming, fast “printf” style debugging using the Trace Port Interface Unit (TPIU) module, clock cycle counting capability, and various other features with Data Watchpoint and Trace (DWT) modules. JTAG also supports standard JTAG scan chains for board level test and chaining multiple JTAG devices. Digital Filter Block The ADC output can optionally feed the programmable digital filter block (DFB) via DMA without CPU intervention. The DFB can be configured to perform IIR and FIR digital filters and a variety of user defined custom functions. The DFB can implement filters with up to 64 taps. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 27 Introduction 28 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 2. Getting Started The quickest path to understanding any PSoC® device is to read the device datasheet and use PSoC Designer™ or PSoC Creator™ integrated development environments (IDE) software. This technical reference manual helps to understand the details of the PSoC 5LP integrated circuit and its implementation. For the most up-to-date ordering, packaging, or electrical specification information, refer to the individual PSoC device’s datasheet or go to http://www.cypress.com/psoc. 2.1 Support Free support for PSoC products is available online at http://www.cypress.com. Resources include Training Seminars, Discussion Forums, Application Notes, PSoC Consultants, TightLink Technical Support Email/Knowledge Base, and Application Support Technicians. Applications assistance can be reached at http://www.cypress.com/support/ or by phone at: 1-800-541-4736. 2.2 Product Upgrades Cypress provides scheduled upgrades and version enhancements for PSoC Creator free of charge. Upgrades are available from your distributor on CD-ROM, or download them directly from http://www.cypress.com under the Software option. Also provided are critical updates to system documentation under the Documentation tab. 2.3 Development Kits Development kits are available from Digi-Key, Avnet, Arrow, and Future. The Cypress Online Store contains development kits, C compilers, and the accessories you need to successfully develop PSoC projects. Go to the Cypress Online Store web site at http://www.cypress.com/shop/. Under Product Categories click PSoC (Programmable System-on-Chip) to view a current list of available items. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 29 Getting Started 30 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 3. Document Construction The following sections include these topics: ■ Section B: CPU System on page 35 ■ Section C: Memory on page 83 ■ Section D: System Wide Resources on page 107 ■ Section E: Digital System on page 173 ■ Section F: Analog System on page 303 ■ Section G: Program and Debug on page 401 3.1 Major Sections For ease of use, information is organized into sections and chapters that are divided according to device functionality. ■ Sections – Presents the top-level architecture, how to get started and conventions and overview information about any particular area that help inform the reader about the construction and organization of the product. ■ Chapter – Presents the chapters specific to some individual aspect of the section topic. These are the detailed implementation and use information for some aspect of the integrated circuit. ■ Glossary – Defines the specialized terminology used in this technical reference manual. Glossary terms are presented in bold, italic font throughout. ■ PSoC® 5LP Registers TRM (Technical Reference Manual) – Supply all device register details summarized in the technical reference manual. These are additional documents. 3.2 Documentation Conventions There are only four distinguishing font types used in this document, besides those found in the headings. ■ The first is the use of italics when referencing a document title or file name. ■ The second is the use of bold italics when referencing a term described in the Glossary of this document. ■ The third is the use of Times New Roman font, distinguishing equation examples. ■ The fourth is the use of Courier New font, distinguishing code examples. 3.2.1 Register Conventions Register conventions are detailed in the PSoC® 5LP Registers TRM (Technical Reference Manual). 3.2.2 Numeric Naming Hexadecimal numbers are represented with all letters in uppercase with an appended lowercase ‘h’ (for example, ‘14h’ or ‘3Ah’) and hexadecimal numbers may also be represented by a ‘0x’ prefix, the C coding convention. Binary numbers have an appended lowercase ‘b’ (for example, 01010100b’ or ‘01000011b’). Numbers not indicated by an ‘h’ or ‘b’ are decimal. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 31 Document Construction 3.2.3 Units of Measure Table 3-2. Acronyms (continued) Symbol This table lists the units of measure used in this document. Table 3-1. Units of Measure Symbol °C Unit of Measure degrees Celsius dB decibels fF femtofarads Hz Hertz k kilo, 1000 K kilo, 2^10 KB 1024 bytes, or approximately one thousand bytes Kbit kHz 1024 bits kilohertz (32.000) k kilohms MHz megahertz M megaohms µA microamperes µF microfarads µs microseconds µV microvolts µVrms microvolts root-mean-square mA milliamperes ms milliseconds mV millivolts nA nanoamperes ns nanoseconds nV nanovolts ohms pF picofarads pp peak-to-peak ppm parts per million SPS samples per second Unit of Measure BINC bit implemented no connection BOM bill of materials BR bit rate BRA bus request acknowledge BRQ bus request CAN controller area network CBUS comparator bus CI carry in CMP compare CMRR common mode rejection ratio CO carry out CPU central processing unit CRC cyclic redundancy check CT continuous time DAC digital-to-analog converter DAP debug access port on ARM Cortex™-M3 of PSoC 5LP DC direct current DFB digital filter block DI digital or data input DMA direct memory access DMAC direct memory access controller DNL differential nonlinearity DO digital or data output DSI digital signal interface ECO external crystal oscillator EEPROM electrically erasable programmable read only memory EMIF external memory interface FB feedback FSR full scale range GIE global interrupt enable GPIO general purpose I/O 2 sigma: one standard deviation I C inter-integrated circuit V volts ICE In-circuit emulator IDE integrated development environment ILO internal low-speed oscillator IMO internal main oscillator INL integral nonlinearity I/O input/output IOR I/O read IOW I/O write IRES initial power on reset IRA interrupt request acknowledge IRQ interrupt request ISR interrupt service routine ISSP In-system serial programming IVR interrupt vector read LFSR linear feedback shift register 3.2.4 Acronyms This table lists the acronyms that are used in this document Table 3-2. Acronyms Symbol ABUS AC ADC Unit of Measure analog output bus alternating current analog-to-digital converter API application programming interface APOR analog power-on reset BC BIFC 32 broadcast clock bit implemented functioning connection PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Document Construction Table 3-2. Acronyms (continued) Symbol Table 3-2. Acronyms (continued) Unit of Measure Symbol Unit of Measure LRb last received bit SPIS serial peripheral interconnect slave LRB last received byte SRAM static random-access memory LSb least significant bit SROM supervisory read only memory LSB least significant byte SSADC single slope ADC LUT lookup table SSC supervisory system call MISO master-in-slave-out SWD single wire debug MOSI master-out-slave-in SWV single wire viewer MSb most significant bit TC terminal count MSB most significant byte TD transaction descriptors NVIC nested vectored interrupt controller on Cortex-M3 of PSoC 5LP TIA transimpedance amplifier UDB universal digital block USB universal serial bus USBIO USB I/O VCO voltage controlled oscillator WDT watchdog timer WDR watchdog reset XRES_N external reset, active low PC program counter PCH program counter high PCL program counter low PD power down PGA programmable gain amplifier PHUB peripheral hub PICU port interrupt control unit PM power management PMA PSoC memory arbiter POR power-on reset PPOR precision power-on reset PRS pseudo random sequence PSoC® Programmable System-on-Chip PSRAM pseudo SRAM PSRR power supply rejection ratio PSSDC power system sleep duty cycle PVT process voltage temperature PWM pulse-width modulator RAM random-access memory RAS row address strobe RETI return from interrupt RO relaxation oscillator ROM read only memory RW read/write SAR successive approximation register SC switched capacitor SIE serial interface engine SIO special I/O SE0 single-ended zero SNR signal-to-noise ratio SOF start of frame SOI start of instruction SP stack pointer SPD sequential phase detector SPI serial peripheral interconnect SPIM serial peripheral interconnect master PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 33 Document Construction 34 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Section B: CPU System The PSoC 5LP CPU subsystem is built around a 32-bit three stage pipelined ARM Cortex-M3 processor running up to 80 MHz. This section includes the following chapters: ■ Cortex™-M3 Microcontroller chapter on page 37 ■ PHUB and DMAC chapter on page 55 ■ Interrupt Controller chapter on page 73 Top Level Architecture CPU System Block Diagram System Bus CPU SYSTEM Cortex M3 CPU Interrupt Controller MEMORY SYSTEM PROGRAM and DEBUG PHUB DMA PSoC 5LP CPU System Block Diagram System Bus CPU SYSTEM 8051 or Cortex M3 CPU Interrupt Controller PROGRAM and DEBUG MEMORY SYSTEM Cache Controller PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB DMA 35 Section B: CPU System 36 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 4. Cortex™-M3 Microcontroller The PSoC 5LP ARM Cortex-M3 core is a high performance, low-power 32-bit Central Processing Unit (CPU). It has an efficient Harvard 3-stage pipeline core, a fixed 4 GB memory map, and supports the 16/32-bit Thumb-2 instruction set. The Cortex-M3 also features hardware divide instructions and low-latency ISR (Interrupt Service Routine) entry and exit. The Cortex-M3 processor includes a number of other components that are tightly linked to the CPU core. These include a Nested Vectored Interrupt Controller (NVIC), a SYSTICK timer, and numerous debug and trace blocks. This section gives an overview of the Cortex-M3 processor. For further details please see the ARM Cortex-M3 Technical Reference Manual available at http://www.arm.com. Figure 5-1 shows a diagram of the Cortex-M3 and its interface to different blocks on the device. 4.1 Features ■ Three stage pipelining operating at 1.25 DMIPS/MHz. This helps to increase execution speed or reduce power. ■ Supports Thumb-2 instruction set: ❐ The Thumb-2 instruction set supports complex operations with both 16- and 32-bit instructions ❐ Atomic bit level read and write instructions ❐ Support for unaligned memory access ■ Improved code density, ensuring efficient use of memory. ■ Easy to use, ease of programmability and debugging: ❐ ■ Nested Vectored Interrupt Controller (NVIC) unit to support interrupts and exceptions: ❐ ■ Ensures easier migration from 8- and 16-bit processors Helps to achieve rapid interrupt response Extensive debug support including: ❐ Serial Wire Debug Port (SWD-DP), Serial Wire JTAG Debug Port (SWJ-DP) ❐ Break points ❐ Flash patch ❐ Instruction tracing ❐ Code tracing PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 37 Cortex™-M3 Microcontroller Figure 4-1. PSoC 5LP Cortex-M3 Block Diagram Interrupt Inputs Nested Vectored Interrupt Controller (NVIC) I- Bus JTAG/SWD D-Bus Embedded Trace Module (ETM) Instrumentation Trace Module (ITM) S-Bus Trace Pins: Debug Block (Serial and JTAG) Flash Patch and Breakpoint (FPB) Trace Port 5 for TRACEPORT or Interface Unit 1 for SWV mode (TPIU) Cortex M3 Wrapper C-Bus AHB 32 KB SRAM Data Watchpoint and Trace (DWT) Cortex M3 CPU Core AHB Bus Matrix Bus Matrix Cache 256 KB ECC Flash AHB 32 KB SRAM Bus Matrix AHB Bridge & Bus Matrix DMA PHUB AHB Spokes GPIO & EMIF Prog. Digital Prog. Analog Special Functions Peripherals The bus interfaces in the Cortex-M3 are based on AHB-Lite (Advanced High Performance Bus-Lite) and the APB (Advanced Peripheral Bus) protocols. The bus interfaces available in the Cortex-M3 are: ■ I-Code Bus for instruction fetches ■ D-Code Bus for data fetches ■ System Bus for instruction and data fetches in memory regions 0x20000000 to 0xDFFFFFFF and 0xE0100000 to 0xFFFFFFFF ■ External Private Peripheral Bus used to debug components ■ Debug Access Port used to connect the debug interface blocks such as SWJ-DP 38 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex™-M3 Microcontroller 4.2 How it Works The Cortex-M3 is a 32-bit processor with a 32-bit data path, 32-bit register, and a 32-bit memory interface. It supports both 16-bit and 32-bit instructions in the Thumb-2 instruction set. Because the Cortex-M3 does not support the ARM instruction set, it is not backward compatible with the ARM7 processor. The processor supports two operating modes: a single cycle 32-bit multiplication instruction, and hardware divide instructions. 4.2.1 Registers The Cortex-M3 has 16 32-bit registers (Figure 4-2). They are: ■ R0 to R12 - general purpose registers ❐ R0 to R7 – can be accessed by all instructions ❐ R8 to R12 – can be accessed by all 32-bit and some 16-bit instructions ■ R13 – Stack Pointer (SP). There are two stack pointers, with only one available at a time. The SP is always 32-bit word aligned; bits [1:0] are always ignored and considered to be ‘0’. ■ R14 – Link register. Stores the return program counter during function calls. ■ R15 – Program counter. This register can be written to control program flow. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 39 Cortex™-M3 Microcontroller Figure 4-2. Cortex-M3 Registers R0 General Purpose Register R1 General Purpose Register R2 General Purpose Register R3 General Purpose Register R4 General Purpose Register R5 General Purpose Register Low Registers R6 General Purpose Register R7 General Purpose Register R8 General Purpose Register R9 General Purpose Register High Registers R10 General Purpose Register R11 General Purpose Register R12 General Purpose Register R13 (MSP) 40 R13 (PSP) Main Stack Pointer (MSP), Process Stack Pointer (PSP) R14 Link Register R15 Program Counter PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex™-M3 Microcontroller 4.2.1.1 Special Registers The special registers can be accessed only using special instructions and cannot be used for normal data processing. CortexM3 supports three sets of special registers: Figure 4-3. Cortex-M3 Special Registers xPSR Program Status registers PRIMASK Special registers Interrupt Mask registers FAULTMASK BASEPRI CONTROL Control register Program Status Registers These registers consist of: ■ Application Program Status Register (APSR) ■ Interrupt Program Status Register (IPSR) ■ Execution Program Status Register (EPSR) These registers provide ALU flags (zero, carry), execution status, and current executing interrupt number. The three PSRs can be accessed separately or collectively, using the special instructions MSR and MRS. They can be collectively addressed as xPSR. Figure 4-4. Cortex-M3 Program Status Registers xPSR 31 30 29 28 27 26:25 24 23:20 19:16 15:10 9 8:0 N Z C V Q ICI/IT T -- -- ICI/IT - Exception Number Where: Interrupt Mask Registers ■ N – Negative Flag ■ ■ Z – Zero Flag PRIMASK – Used to disable all interrupts except the Nonmaskable Interrupt (NMI) and HardFault ■ C – Carry/Borrow Flag ■ FAULTMASK – Used to disable all interrupts except NMI ■ V – Overflow Flag ■ ■ Q – Sticky Saturation Flag BASEPRI – Used to disable interrupts of specified or lower priority levels. ■ ICI / IT – Interrupt-Continual Instruction (ICI) bits / IFTHEN instruction status bit ■ T – Thumb-2 Instruction. Always set to 1. Clearing this results in an exception ■ Exception Number – Indicates which exception the processor is currently handling PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F These registers are used by the NVIC to mask an interrupt or exception. 41 Cortex™-M3 Microcontroller Control Register This register controls the stack pointer selection and the privilege level of the processor. It has only two bits: CONTROL[0] When the code is in user level, it cannot access the debug resources and certain important registers. In addition to the privilege levels, the processor supports two types of operating modes: ■ Thread Mode – Thread mode is used by all normal applications. During the thread mode the Process Stack Pointer (PSP) is used. The thread mode can exist in both privileged level and user level. Switching from privileged level to user level can be done by just writing to the control register but the reverse cannot be done. When an exception occurs, the system is automatically taken to privileged level and at the exit of the exception it comes back to the user level. Restoring to the privileged level can be done only by going through an exception handler that programs the control register for the privileged mode. ■ Handle Mode – Handle mode is used by OS kernel and exception handlers. During this mode, the main stack pointer (MSP) is used. The handle mode can exist only in the privileged level. ‘0’ Privileged in Thread Mode ‘1’ User state in Thread mode CONTROL[1] ‘0’ Default stack is used ‘1’ Alternate stack is used 4.2.2 Operating Modes The Cortex-M3 supports two privilege levels: ■ Privileged – Code has no limit to resources ■ User – Code has some limits to the resources Privilege level can be controlled using the control register. Figure 4-5. Operating Modes Privilege Level User Privileged n/a Handle Mode Thread Mode Handle Mode: running an interrupt service routine Thread Mode: running background code 42 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex™-M3 Microcontroller Figure 4-6. Operating Mode Transitions Privileged Handle Mode Exception Entry / Exit User Thread Mode 4.2.3 Program Control Register Pipelining The three stage pipelining includes: ■ Fetch – The instruction is fetched from memory ■ Decode – Generating the addresses and branch prediction ■ Execute – Instruction execution based on the address and branches The branch prediction unit is enhanced so that it gives nearly no ALU usage penalty. Pipelining can give zero to two wait states when executing an instruction. 4.2.4 Exception Entry / Exit Thumb-2 Instruction Set The Cortex-M3 supports a wide range of 16- and 32-bit instructions. It does not support all ARM instructions, including: Privileged Thread Mode ■ Saturation ■ Miscellaneous Default Cortex-M3 supports unique instructions. The following table is a summary of the important instructions: Table 4-1. Cortex-M3 Unique Instructions Instruction Functionality MSR, MRS To access special registers IT IF-THEN instruction supporting up to 4 succeeding instructions CBZ, CBNZ Compare and then branch SDIV, UDIV Signed and Unsigned Divide REV, REVH, REVSH Reverse the byte order in data word, upper half word, lower half word, respectively RBIT Reverses bit order in a data word SXTB, SXTH, UXTB, UXTH Extend a byte or half word into a word BFC, BFI BFC - Clears any number of adjacent bits in any position BFI – Copies any number of bits from any register to another register to any mentioned location ■ Branch with link and exchange state ■ Switch endian UBFX, SBFX ■ Certain coprocessor instructions LDRD, STRD Transfer 2 words of data from or into 2 registers ■ Hint instructions TBB, TBH Table Branch Byte and Table Branch Halfword for branch tables ■ DSP instructions ■ Change process instructions The instruction includes these data processing operations: ■ Multiply and divide ■ Bit ■ Shift ■ Load store ■ Branch ■ Barrier ■ Exception generating ■ System PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Unsigned and signed bit field extract instructions The following sections detail some of the instruction types. For the entire summary of the instruction set, refer to the Cortex-M3 Technical Reference Manual available at http:// infocenter.arm.com/help/topic/com.arm.doc.ddi0337e/ DDI0337E_cortex_m3_r1p1_trm.pdf. 4.2.4.1 Data Processing Operations The Cortex-M3 provides many different instructions for data processing. A few basics are introduced here. Many data operation instructions can have multiple instruction formats. 43 Cortex™-M3 Microcontroller The Cortex-M3 supports arithmetic functions ADD, SUB (subtract), MUL (multiply), and UDIV/SDIV (unsigned and signed divide). The Cortex-M3 supports 32-bit multiply instructions and multiply accumulate instructions that give 64-bit results. These instructions support signed or unsigned values. Another group of data processing instructions are logical operations such as AND, ORR (or), EOR (exclusive OR), and rotate and shift functions. In some cases the rotate operation can be combined with other operations. Another group of data processing instructions is used for reversing data bytes in a register. These instructions are usually used for conversion between little endian and big endian data. The last group of data processing instructions is for bit field processing. Instructions such as BFC, BFI, SBFX, and UBFX are used to clear, set, and copy bits with sign extension or zero extension. 4.2.4.2 Load Store Operations One of the most basic functions in a processor is transfer of data. In the Cortex-M3, data transfers can be one of the following types: ■ Moving data between register and register ■ Moving data between memory and register ■ Moving data between special register and register ■ Moving an immediate data value into a register The command to move data between registers is MOV (move). For example, moving data from register R3 to register R8 looks like this: 4.2.4.3 Branch Operations The branch operations include: ■ Call and Unconditional branch instructions ■ Decision and Conditional branch instructions ■ Combined Compare and Conditional Branch ■ Conditional Branching using IT instructions The IT (IF-THEN) instruction block is very useful for handling small conditional code. It avoids branch penalties because there is no change to program flow. It can provide a maximum of four conditionally executed instructions with one condition check. 4.2.4.4 Instruction Barrier and Memory Barrier Instructions The Cortex-M3 supports a number of barrier instructions. These instructions are needed with complex memory systems. In some cases, if memory barrier instructions are not used, race conditions can occur. There are three barrier instructions in the Cortex-M3: ■ DMB (Data Memory Barrier) – Ensures that all memory accesses are completed before new memory access is committed. For example, when you do a data write followed immediately by a read on a dual port memory, if the memory write is buffered, the DMB instruction can be used to ensure the read gets the updated value. ■ DSB (Data Synchronization Barrier) – Ensures that all memory accesses are completed before the next instruction is executed ■ ISB (Instruction Synchronization Barrier) – Flushes the pipeline and ensures that all previous instructions are completed before executing new instructions MOV R8, R3 Another instruction can generate the negative value of the original data; it is called MVN (move negative). The basic instructions for accessing memory are Load and Store. Load (LDR) transfers data from memory to registers, and Store transfers data from registers to memory. The transfers can be in different data sizes (byte, half word, word, and double word). Multiple Load and Store operations can be combined into single instructions called LDM (Load Multiple) and STM (Store Multiple). ARM processors also support memory accesses with preindexing and post-indexing. Two other types of memory operation are stack PUSH and stack POP. The Cortex-M3 has a number of special registers. To access these registers, use the instructions MRS and MSR. 44 4.2.4.5 Saturation Operations The Cortex-M3 supports two instructions that provide signed and unsigned saturation operations: SSAT and USAT (for signed data type and unsigned data type, respectively). Saturation is commonly used in signal processing, for example, in signal amplification. The saturation operation does not prevent the distortion of the signal, but the amount of distortion is greatly reduced in the signal waveform. 4.2.5 SysTick Timer The SysTick timer is integrated with the NVIC and generates the SYSTICK interrupt. This interrupt can be used for task management in a real time system. The timer has a reload register with 24 bits available to use as a countdown value. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex™-M3 Microcontroller The timer can take an internal clock (the free running clock on the CM3 processor) or an external clock through the STCLK. In PSoC 5LP devices use one of three sources as, ILO (1 kHz), ILO_100 (100 kHz), or the SYSCLK (BUSCLK). 4.2.6 Debug and Trace: The Cortex-M3 provides a wide range of debugging components. The debug unit is tightly linked with the core. The important features of the debug and trace are: ■ Debug access to all memory and registers in the system including Cortex-M3 register bank when the core is running, halted, or held in reset. ■ Serial Wire Debug Port (SW-DP) and Serial Wire JTAG Debug Port (SWJ-DP) debug access. ■ Flash Patch and Breakpoint (FPB) unit for implementing breakpoints and code patches. ■ Data Watchpoint and Trace (DWT) unit for implementing watchpoints, data tracing, and system profiling. ■ Support for six breakpoints and four watchpoints. ■ Instrumentation Trace Macrocell (ITM) for support of printf style debugging. ■ Embedded Trace Macrocell (ETM) for instruction trace. ■ Trace Port Interface Unit (TPIU) for bridging to a Trace Port Analyzer (TPA). addresses, the unaligned accesses support data operations at odd addresses also. Unaligned accesses have limitations. Some instructions cannot support unaligned accesses. You can execute code from within the code, SRAM, or the external RAM space. The Cortex-M3 uses little-endian format. 4.3.1 Bus Interface to SRAM Memory The 64 KB of SRAM in PSoC 5LP is split into two 32 KB of SRAM. The SRAM can be accessed by the C-Bus, S-Bus, and the PHUB's DMA. The priority decoder gives a higher priority to the C-Bus in the upper 32 KB of SRAM, whereas the PHUB DMA takes a higher priority in the lower 32 KB of SRAM. The upper and lower halves of SRAM can be accessed simultaneously but with different buses. The Cortex-M3 supports a separate debug and trace interface. The debug interface uses the APB (Access Port Bus), which supports both JTAG and SWD. The trace interface uses the TPIU (Trace Port Interface Unit). For further details about the debug and trace feature, see the Test Controller chapter on page 403 and the Cortex-M3 Debug and Trace chapter on page 415. 4.3 Memory Map The Cortex-M3 has a linear 32-bit (4 GB) address space, as shown in Figure 4-7. See also the Memory Map chapter on page 105. The address space includes two bit-band alias regions, one for the SRAM space and the other for the Peripherals space. Accesses to a bit-band alias region affect individual bits in the corresponding bit-band region. For example, writing a 1 to address 0x22000000 sets bit 0 of address 0x20000000, and writing a 0 to address 0x42000004 clears bit 1 of address 0x40000000. Reading address 0x22000008 returns a 1 or 0, depending on the value of bit 2 of address 0x20000000. The processor supports unaligned accesses. Unlike aligned access where the data can be situated only at even PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 45 Cortex™-M3 Microcontroller Figure 4-7. Cortex-M3 Memory Map 0xE00FF000 0xE00FEFFF 0xE0042000 ROM Table External Private Peripheral Bus ETM 0xE0041000 0xFFFFFFFF TPIU 0xE0040000 Vendor Specific Private Peripheral Bus: Debug/External 0xE003FFFF Reserved 0xE000F000 Private Peripheral Bus: Internal NVIC 0xE0100000 0xE00FFFFF 0xE0040000 0xE003FFFF 0xE0000000 0xDFFFFFFF 0xE000DFFF Reserved 0xE0003000 FPB 0xE0002000 External Device DWT 0xE0001000 ITM 0xE0000000 1 GB 0xA0000000 0x9FFFFFFF 0x43FFFFFF Bit-Band Alias External RAM 0x42000000 32 MB 0x41FFFFFF 31 MB 0x40100000 0x40000000 1 MB 1 GB Bit-Band region Peripherals 0.5 GB 0x40000000 0x3FFFFFFF 0.5 GB 0x20000000 0x1FFFFFFF 0x23FFFFFF Bit-Band Alias 0x22000000 32 MB 0x21FFFFFF 31 MB 0x20100000 0x20000000 1 MB 4.4 Bit-Band region Exceptions The Cortex-M3 provides a feature-packed exception architecture that supports a number of system exceptions and external interrupts. Exceptions are numbered 1 to 15 for 46 0x60000000 0x5FFFFFFF SRAM Code 0.5 GB 0x00000000 system exceptions and 16 and above for external interrupt inputs. PSoC 5LP architecture supports 32 external interrupts. The exceptions are handled by the NVIC. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex™-M3 Microcontroller Most of the exceptions have programmable priority, and a few have fixed priority. Table 4-2 shows the list of exceptions available in the Cortex-M3: Table 4-2. PSoC 5LP Exceptions Interrupt Number Exception Type Priority Comment 1 Reset -3 (highest) Not programmable Reset 2 NMI -2 Not programmable Non-Maskable Interrupt 3 Hard Fault -1 Not Programmable All fault conditions if the corresponding handler is not enabled 4 Reserved NA – 5 Bus Fault Programmable Bus error occurs when AHB interface receives an error response from a bus slave (also called prefetch abort if it is an instruction fetch or data abort if it is a data access) 6 Usage Fault Programmable Exceptions due to program error 7 Reserved NA – 8 Reserved NA – 9 Reserved NA – 10 Reserved NA – 11 SVCall Programmable System Service Call 12 Debug Monitor Programmable Debug monitor (watchpoints, breakpoints, external debug request) 13 Reserved NA – 14 PendSV Programmable Pendable request for system device 15 SYSTICK Programmable System Tick Timer The value of the current running exception is indicated by the special register IPSR or from the NVIC's Interrupt Control State Register (the VECTACTIVE field). ■ Bus faults ■ Memory Management Faults ■ Usage Faults Interrupts are a subset of exceptions. So exceptions are handled the same way as an interrupt. The exception handler for each exception is stored in the interrupt vector table. The vector table begins with the exception handler and is followed by the interrupt service routine addresses. The vector table pointer is dynamically changeable. Also, if the vector table is in SRAM, then vectors can be dynamically changeable. ■ Hard Faults 4.4.1 When these types of faults (except vector fetches) take place, and if the corresponding exception handler is enabled and no other exceptions with the same or higher priority are running, the fault exception handler will be executed. If the exception handler is enabled but at the same time the core receives another exception handler/interrupt with higher priority, this fault exception handler will be pending and will be executed after the high priority exception/interrupt has completed its execution. Priority Definitions In the Cortex-M3, whether and when an exception can be carried out can be affected by the priority of the exception. A higher priority (smaller number in priority level) exception can preempt a lower priority (larger number in priority level) exception; this is the nested exception/interrupt scenario. From the above table, you can see that some of the exceptions (reset, NMI, and hard fault) have fixed priority levels. They are negative numbers to indicate that they are higher priority than other exceptions. Other exceptions have programmable priority levels. 4.4.2 Fault Exceptions A number of system exceptions are useful for fault handling. There are several categories of faults: PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F The faults can be enabled by setting the corresponding bits in the handler control and state register. The reason for a particular fault is updated in the corresponding status register (for example, BFSR register for bus fault, MFSR for memory management fault, UFSR for Usage Fault, HFSR for Hard Fault). These registers can be read to know the exact reason for fault. If the fault handler is not enabled or when the fault happens in an exception handler that has the same or higher priority than the current fault handler, the hard fault handler will be executed instead. Bus Faults 47 Cortex™-M3 Microcontroller Bus faults are produced when an error response is received during a transfer on the AHB interfaces. It can happen during prefetch, data read/write, or during stacking and unstacking operations. Memory Management Faults Memory management faults can be caused by certain illegal accesses, including the following: ■ Trying to execute code from non-executable memory regions ■ Writing to read-only regions ■ Access in the user state to a region defined as privileged access only Usage Faults Usage faults can be caused by a number of things, including the following: ■ Undefined instructions ■ Coprocessor instructions (the Cortex-M3 processor does not support a coprocessor, but it is possible to use the fault exception mechanism to run software compiled for other Cortex processors via coprocessor emulation) SVC SVC is for generating system function calls. It can be configured to generate an interrupt. This interrupt can be used for task management in a realtime system. SVC is generated using the SVC instruction. PendSV PendSV works with SVC in the OS. Although SVC (by SVC instruction) cannot be pended (an application calling SVC will expect the required task to be done immediately), PendSV can be pended and is useful for an OS to pend an exception so that an action can be performed after other important tasks are completed. PendSV is generated by writing ‘1’ to the NVIC PendSV pending register. A typical use of PendSV is context switching. SysTick Timer Exception The SysTick Timer exception takes the vector number 15. Cortex-M3 supports a 24-bit down counter. This timer is very useful to perform task management where the software can be handled inside the timer interrupt. The SYSTICK Timer can be used to generate interrupts. It has a dedicated exception type and exception vector. It makes porting operating systems and software easier because t he process is the same across different CortexM3 products. ■ Trying to switch to the ARM state (software can use this faulting mechanism to test whether the processor on which it runs supports ARM code; because the CortexM3 does not support the ARM state, a usage fault takes place if there is an attempt to switch) ■ Invalid interrupt return (link register contains invalid/ incorrect values) The SYSTICK Timer is controlled by four registers. Of the four registers, TICKINT is used to enable or disable the timer exception. ■ Unaligned memory accesses using multiple load or store instructions 4.5 It is also possible, by setting up certain control bits in the NVIC, to generate usage faults for: ■ Divide by zero ■ Any unaligned memory accesses Hard Faults The hard fault handler can be caused by: ■ ■ Usage faults, bus faults, and memory management faults if their handler cannot be executed. Bus faults during vector fetch (reading of a vector table during exception handling). 4.4.3 System Call Exceptions SVC (System Service Call) and PendSV (Pended System Call) are two exceptions targeted at software and operating systems. 48 Nested Vector Interrupt Controller (NVIC) The Nested Vectored Interrupt Controller, or NVIC, is an integral part of the Cortex-M3 processor. It is closely linked to the Cortex-M3 CPU core logic. Its control registers are accessible as memory-mapped devices. Besides control registers and control logic for interrupt processing, the NVIC also contains control registers for the SYSTICK Timer, and debugging controls. Following are the important features of the NVIC: ■ Supports 32 interrupts and 16 exceptions. ■ Configurable priority levels. ■ Dynamic reprioritization of interrupts. ■ Support for nested interrupts ■ Programmable interrupt vector ■ Supports tail-chaining and late arrival interrupts. This enables back-to-back interrupt processing without the PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex™-M3 Microcontroller overhead of state saving and restoration between interrupts. ■ Processor state automatically saved upon interrupt entry, and restored upon interrupt exit, with no instruction overhead. 4.5.1 Basic Interrupt Configuration Each external interrupt has several associated registers. ■ Enable and Clear Enable ■ Set Pending and Clear Pending ■ Priority Level ■ Active Status ■ Exception-masking registers (PRIMASK, FAULTMASK, and BASEPRI) ■ Vector Table Offset The interrupt enable and clear enable registers are 32-bit registers. They are used to enable/disable an interrupt. An interrupt that is waiting for the CPU execution sets the pending bit in the set pending register. After the interrupt is executed by the CPU, the interrupt is cleared automatically by setting the clear-pending register. The interrupts can take priorities 0 to 7. The priorities are configured using the 3-bit priority registers. They can be dynamically configured during run time. The Active Status register stores the details of the interrupt currently active. A bit set in this register indicates that the corresponding interrupt is currently active. An interrupt is called active if it is currently executed by the CPU or if it is already nested and put to the stack. After the interrupt execution is complete, the active status bit of the interrupt is automatically cleared. With PSoC 5LP devices, the addresses of the interrupt service routine are stored in the Interrupt vector table. The interrupt vector table can be located either in RAM or ROM. The position of the vector table is controlled using the Vector Table Offset register. The exception masking registers, PRIMASK, FAULTMASK and BASEPRI, are special registers used to mask the interrupts and exceptions. ■ PRIMASK – When set, all interrupts except NMI and Fault interrupts are masked ■ FAULTMASK – When set, all interrupts except NMI are masked ■ BASEPRI – Masks all interrupts at the specified priority and lower priorities PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 4.5.1.1 Example Procedures in Setting Up an Interrupt Here is a simple example procedure for setting up an interrupt: 1. Copy the Hard Fault and NMI handlers to a new vector table location if vector table relocation is required. (In simple applications, this might not be needed.) 2. The Vector Table Offset register should also be set up to get the vector table ready (optional). 3. Set up the interrupt vector for the interrupt. Because the vector table may have been relocated, read the Vector Table Offset register; then calculate the correct memory location for your interrupt handler. This step might not be needed if the vector is hardcoded in ROM. 4. Set up the priority level for the interrupt. 5. Enable the interrupt. 4.5.2 Nested Interrupts Nested interrupt support is built into the Cortex-M3 processor core and the NVIC. The nesting is done based on the priority of the interrupts. When the processor is handling an exception, all other exceptions with the same or lower priority will be blocked. When a high priority interrupt occurs, the low priority interrupt is nested and the high priority interrupt completes the execution. Because automatic hardware stacking and unstacking is done, nesting is done without risk of losing data in registers. Cortex-M3 uses the main stack to store the nesting interrupt details; therefore, ensure sufficient stack space is available. Reentrant exceptions are not supported in the Cortex-M3. 4.5.3 Tail-Chaining Interrupts The Cortex-M3 uses a number of methods to improve interrupt latency. Tail-chaining is one such method. When an exception takes place but the processor is handling another exception of the same or higher priority, the exception will be pended. When the processor has finished executing the current exception handler, instead of POP, the registers go back into the stack and PUSH it back in again, skipping the unstacking and the stacking. In this way the timing gap between the two exception handlers is greatly reduced. 4.5.4 Late Arrivals Another feature that improves interrupt performance is late arrival exception handling. When an exception takes place and the processor has started the stacking process, and if during this delay a new exception arrives with higher pre- 49 Cortex™-M3 Microcontroller emption priority, the late arrival exception will be processed first. For example, if Exception #1 (lower priority) takes place a few cycles before Exception #2 (higher priority), the processor will behave such that Handler #2 is executed as soon as the stacking completes. After this the Handler #1 will be executed. 4.5.5 Interrupt Latency The term interrupt latency refers to the delay from the start of the interrupt request to the start of interrupt handler execution. ■ In the Cortex-M3 processor, if the memory system has zero latency, and provided that the bus system design allows vector fetch and stacking to happen at the same time, the interrupt latency can be as low as 12 cycles. This includes stacking the registers, vector fetch, and fetching instructions for the interrupt handler. However, this depends on memory access wait states and a few other factors. ■ For tail-chaining interrupts, because there is no need to carry out stacking operations, the latency of switching from one exception handler to another exception handler can be as low as 6 cycles. ■ When the processor is executing a multi-cycle instruction such as divide, load double, or store double, the instruction can be abandoned and restarted after the interrupt handler completes. ■ To reduce exception latency, the Cortex-M3 processor allows exceptions in the middle of multiple load and store instructions (LDM/STM). If the LDM/STM instruction is executing, the current memory accesses will be completed, and the next register number will be saved in the stacked xPSR (ICI bits). After the exception handler completes, the multiple load/store will resume from the point at which the transfer stopped. 4.5.6 Faults Related to Interrupts Faults (bus fault, memory fault) can happen during the following stages of interrupt execution: ■ Stacking ■ Unstacking ■ Vector Fetches 50 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 5. PSoC 5LP Cache Controller The cache block is an Instruction cache only. It services instruction fetches from the CPU. It stores lines of code from the flash in its internal buffer for fast accesses made by the CPU at a later time. 5.1 Features ■ Instruction cache ■ Direct mapped ■ 1 KB, 4-way set associative cache memory ■ Registers for measuring cache hit/miss ratios ■ Error correction code (ECC) support ■ Error logging and interrupt generation ■ Designed to put flash into sleep automatically to save power 5.2 Block Diagram Figure 5-1 shows the system interaction with the cache block as well as the cache interfaces and data/instruction flow. Figure 5-1. Cache Interfaces CPU 1 2 Cache Control 5 RAM PHUB 4 Flash Interface 3 FLASH SPC EMIF 11 External Memory PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 51 PSoC 5LP Cache Controller Table 5-1. Cache Operational Interfaces Interface Function 1 CPU sends instruction fetch request through this interface to the cache and eventually receives back the instruction 2 When the CPU instruction fetch that gets a hit in the cache, it is retrieved from the cache memory (RAM) through this interface. 3 CPU instruction fetch (interface #1) that gets a miss in the cache is translated into one fetch request from the FLASH. The FLASH access time is much larger than the Cache RAM access time, up to 4 CPU clock cycles. 4 Instructions returned from the FLASH are cached through this interface for later CPU use. Note that requests from the PHUB interface are never cached. 5 The CPU can read and write data using this interface. The internal cache registers and RAM are also accessible and FLASH contents are readable using this interface through PHUB’s special register spoke. 5.3 Cache Enabling and Disabling To enable the cache, set the DISABLE bit (Bit 0) of CACHE.CC_CTL register t to 0. 5.4 Code Protection and Security The ECC block is responsible for error detection and correction. The cache gets the error status from the ECC block for requested fills from the flash. The error status gets logged into software visible registers in the cache. An uncorrectable error will prevent the fill data from being written into the cache RAM and causes entire line to be invalidated. ECC_ADDR[0:28] field of CACHE.ECC_CORR register gives the flash address where error was detected; this address field is valid only when INT_VALID field of this register is set to 1. Interrupt can also be generated on ECC correction by setting INT_ENB bit of CACHE.ECC_CORR register. If ECC correction fails, then the flash address where error happened can be obtained from CACHE.ECC_ERR register. 5.5.1 Measuring Cache Hits or Misses The CACHE.HITMISS register provides two 16-bit counters that count the number of cache hits and misses. To measure the cache performance, reset the HITMISS register to 0 at the start of the block of code to measure. Then the code is executed and at the end of the code under measurement, the HITMISS register should be read. The cache hit ratio can be computed asCache hit ratio = the number of cache hits (HITMISS[31:16])/Number of cache misses (HITMISS[15:0]) 5.6 Cache Induced Flash Lowpower Mode Flash is put to low-power mode when the cache predicts that a flash access is not needed in the near future, based on reaching a programmed number of sequential hits. This feature helps to reduce the overall power consumption of the device. The threshold value of sequential hits can be programmed in LP_MODE bits of CACHE.CC_CTL register. To put the FLASH into low-power mode immediately, LP_MODE bits should be set to 0. This should be done when executing code from SRAM. ECC data in flash will change every time a write is done to the corresponding flash row. When a flash region is read, the corresponding ECC data will be used for error checking. The error checking is dynamic and happens every time the cache reads from the flash; this means, the comparison is for the latest data written to flash. 5.7 5.5 Cache status on system reset: Invalidating the Cache Line Software can invalidate all cached data associated with an interface by setting the Flush bit (Bit 2) of CACHE.CC_CTL register. Invalidate takes effect in 1 cycle and affects all lines. 52 Sleep Mode Behavior When the device wakes up from low-power modes, all cache data and tags are invalidated. However, all the cache registers (where cache settings are made) maintain their state and are not reset. The cache will be refilled as the CPU begins fetching instructions. On reset, cache is invalidated and begins to fill with the first request from the CPU. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PSoC 5LP Cache Controller 5.8 Cache Limitations All instructions are assumed to be in the flash. There is no direct path from the cache to the external memory. Instructions from the external memory must be explicitly moved into the flash by software, before they can be used by the CPU. Cache coherency is the software's responsibility; no hardware mechanism exists to ensure coherency. If the software modifies the FLASH or memory contents, it also needs to invalidate the cache and ensure the new instruction is fetched into the cache. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 53 PSoC 5LP Cache Controller 54 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 6. PHUB and DMAC PSoC® 5LP devices use a high-performance bus for peripheral access and bulk data transfer. The high-performance bus and the associated central controller are known as the peripheral hub (PHUB). The PHUB is a programmable and configurable central bus backbone within a PSoC 5LP device that ties the various on-chip system elements together. It consists of multiple spokes; each spoke is connected to one or more peripheral blocks. The PHUB also includes a direct memory access controller (DMAC), which is used for data transfer. The DMAC supports multiple DMA channels. There are two bus masters (blocks that can initiate bus traffic) in PSoC 5LP devices. These are the DMAC and the CPU. An arbiter in the PHUB is responsible for arbitrating requests from the CPU and the DMAC. Upon receiving a request from the microcontroller or the DMAC, the PHUB relays the request to the appropriate peripheral spoke. 6.1 PHUB PHUB manages arbitration between the CPU and DMAC. 6.1.1 Features The PHUB has the following features: ■ Industry-standard Advanced Microcontroller Bus Architecture High-performance Bus (AMBA-HB) lite protocol ■ 8 spokes connected to various peripherals ■ 8-/16-/32-bit data-width support ■ Peripherals of various address widths connected to the same spoke ■ Includes programmable DMAC with 24 direct memory access (DMA) channels ■ Byte order and data width difference translation 6.1.2 Block Diagram Figure 6-1 on page 56 is the block diagram of the PHUB. The DMAC is also shown. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 55 PHUB and DMAC Figure 6-1. PHUB Block Diagram CPU PHUB CPU Interface CHn CSRs CHn Channel[n] CSRs Config/ Status Channel Arbitration CFGMEM TDMEM Spoke 0 Local Memory DMAC Local Spoke / PHUB Config/Status SRAM Spoke Arbitration Spokes to Peripherals 6.1.3 How It Works Table 6-1. Spoke Configuration Spoke Address Width (in bits) Data Width (in bits) 0 14 32 SRAM The PHUB connects to the peripherals using a spoke. There are eight spokes. Each spoke connects to one or more peripherals. Each spoke is configured for: 1 9 16 I/O interface, port interrupt control unit (PICU), external memory interface (EMIF) 2 19 32 PHUB local spoke, power management, clock, serial wire viewer (SWV), EEPROM ■ Address width – The address width of a spoke depends on the maximum number of addresses required for the peripherals connected to the spoke. 3 11 16 Delta-sigma ADC, analog interface 4 10 16 USB, CAN, fixed-function I2C, fixed-function timers Data width – The data width of a spoke can be 16 or 32 bits. Eight-bit data transfer can be performed on 16and 32-bit spokes. 5 11 32 Digital filter block (DFB) 6 17 16 UDB set 0 registers (including DSI, configuration, and control registers), UDB interface 7 17 16 UDB set 1 registers (including DSI, configuration, and control registers) The PHUB is used to connect the CPU to memory and peripherals, including SRAM, flash, EEPROM, analog subsystem, digital blocks, digital filter block, and others. ■ ■ Number of peripherals – This depends on the device architecture. Each spoke is usually connected to multiple peripherals. Table 6-1 shows the address width, data width, and peripherals connected to each spoke in the PSoC 5LP device. ■ Peripheral Names The peripherals connected to each spoke can have data widths longer than the spoke. For example, a DeltaSigma ADC can support up to 20-bit data although it is placed in the 16-bit spoke (spoke 03). In this case, the PHUB uses an internal FIFO to accommodate the width differences during data transfer. 56 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC ■ One peripheral can extend across multiple spokes. In this case, the peripheral will have different address spaces that are connected to each spoke. For example, Table 6-1 shows that UDB registers extend across two spokes. UDB registers can be accessed in 8-bit mode and also in 16-bit mode. In this case, the 8-bit mode access needs a different address space than the 16-bit mode access though they reside in the same spoke. 6.2 DMA Controller The DMA Controller (DMAC) transfers data between memory and peripherals. ■ Uses the PHUB for data transfer ■ Includes 24 DMA channels ■ Includes 128 transaction descriptors (TD) ■ Eight levels of priority per channel Peripherals of different data widths can be connected to a single spoke. ■ Transactions can be triggered by any digitally routable signal, the CPU, or another DMA channel An example of this is spoke 3, which is connected to the analog interface (digital-to-analog converter) and deltasigma ADC. The delta-sigma ADC can support up to 20-bit data, and the digital-to-analog converter register is 8-bit. ■ Transactions can be stalled or canceled ■ Each transaction can be from 1 to 64 KB ■ Large transactions can be broken into smaller bursts of 1 to 127 bytes with Intraspoke burst count restricted to 16. ■ Each channel can be configured to generate an interrupt at the end of transfer The spoke address width, data width, and peripherals are fixed in a device and cannot be changed. The spoke and the peripheral details affect the time required for data transfer. interspoke and intraspoke transfers take different amounts of time. ■ Supports byte swapping, for conversion between bigendian and little-endian formats ■ Handles data-width differences The effects of spoke data width, and interspoke and intraspoke transfer, on latency of data transfer are explained in 6.1.4 Arbiter. As shown in Figure 6-1 on page 56, the PHUB includes local memory to store configuration data. The local memories are called ■ ■ Spoke 0 is connected to SRAM. The CPU can access the SRAM without going through the PHUB. The DMAC accesses the SRAM through PHUB. 6.1.4 Arbiter The PHUB receives data read or write requests from either the CPU or the DMAC. The PHUB processes each request to determine which spoke and peripheral should be accessed, and then manages the data access. When the DMAC and CPU initiate transactions in the PHUB at the same time, the arbiter decides which request has priority. The priority can be configured for every spoke except spoke 0. Spoke 0 is accessed only by the DMAC because the CPU has a separate interface to SRAM. You can configure priority using the “spk_cpu_pri” bits in the PHUB_CFG register. When the CPU and DMAC access different spokes simultaneously, both accesses are independent and arbitration is not necessary. This enables a multiprocessing environment. The exception is the SRAM, which has direct access by the CPU and PHUB. In this case, there is no arbitration required for SRAM. This helps to reduce the SRAM latency access. The arbitration issues when the CPU and DMA want to access the same spoke simultaneously are detailed in further sections. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 6.2.1 Local Memory ■ Configuration memory (CFGMEM) ■ Transaction descriptor memory (TDMEM) The PHUB also includes a 16-byte FIFO for data handling during data transfers. The CGFMEM is used to store the DMA channel configuration data. There are two registers: CFGMEMn.CFG0 and CFGMEMn.CFG1 (where n can be from 0 to 23) for each channel. Each register is 32 bits, so the size of CFGMEM is 8 bytes × 24 channels = 192 bytes. The TDMEM is used to store the TD configuration data, which includes the number of bytes to transfer, source address, destination address, next TD, and other configuration data. Each TD has two registers: TDMEMn.ORIG_TD0 and TDMEMn.ORIG_TD1. Each register is 32 bits, so the size of TDMEM is 8 bytes × 128 TDs = 1 KB of memory. The local memory is accessed through the local spoke of the PHUB (see Table 6-1 on page 56). 6.2.2 How the DMAC Works The DMAC is one of the bus masters for PHUB. The DMAC can perform the following data transfers: 57 PHUB and DMAC ■ Memory to memory ■ Memory to peripheral ■ Peripheral to memory ■ Peripheral to peripheral The source engine selects the spoke to which the source peripheral is connected. When the spoke is available for data transfer, the data transfer from the source begins. ■ This phase selects the spoke on which the destination peripheral is available. When the spoke is available, the data collected in the source engine phase is transferred to the destination peripheral. Any DMA channel goes through the following phases to perform data transfers: ■ Arbitration phase ■ Fetch phase ■ Source engine phase ■ Destination engine phase ■ Write back phase Destination engine phase ■ Write back phase This phase is the completion phase were the TD and DMA channel configurations are updated after data transfer. Ideal conditions for data transfer are: The total time required for a DMA transfer depends on the time taken for each phase. The DMA transfer can be either an intraspoke DMA transfer or interspoke DMA transfer ■ Single requestor ■ CPU doesn't interrupt the fetch phase ■ Both source and destination spoke are readily available In an intraspoke transfer, the data transfer happens within the same spoke. This transfer makes use of the internal FIFO. ■ Source spoke and destination spoke are of same width ■ Source and destination address start at even addressing ■ Transfer count is a multiple of burst count ■ Burst count matches the spoke width ■ Arbitration phase The DMAC selects which DMA channel to process based on the priority. ■ ■ Fetch phase The number of bursts for transfer (N) = Transfer count Spoke width The DMAC fetches the TD and DMA channel details from the configuration registers. 6.2.2.1 Interspoke Transfers The timing diagram for an interspoke transfer under ideal conditions is shown in Figure 6-2. Source engine phase Figure 6-2. Interspoke Transfer Cycle Timing Bus Clock DMA request latch phase Arbitration Phase Fetch Phase Source Engine Phase Destination Engine Phase Control Cycles Data Cycles Control Cycles Data Cycles Control 1 Control 2 Control 3 Control 4 Control N Data 1 Data 2 Data 3 Data N-1 Data N Control 1 Control 2 Control N-2 Control N-1 Control N Data 1 Data N-3 Data N-2 Data N-1 Data N Write Back Phase The total number of cycles for data transfer in the case of interspoke DMA transfers is the sum of cycles required for each phase. Total cycle time = Arbitration phase time (1) + Fetch phase (1) + Source Engine phase (N + 3) + Destination engine phase (0, because it happens in parallel with the source engine phase) + Write back phase (1) 58 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC Total cycle time = N + 6 cycles (where N = Transfer count Spoke width) Example You want to move five samples of 16-bit ADC data to memory. Notes ■ The ADC (decimator) is connected to spoke 3 which is a 16-bit spoke. ■ Memory is in Spoke 0, which is a 32-bit spoke) The DMA configuration includes: ■ DMA channel burst count (configured in CFGMEMn.CFG0) = 2 ■ TD transfer count (configured in TDMEMn.ORIG_TD0) = 2 bytes × 5 samples = 10 ■ TD configuration includes an Increment Destination Address to copy data to an array in the memory (configured in TDMEMn.ORIG_TD0) ■ N = Transfer count Spoke width = 10 2 = 5 For more information about the DMA configuration, refer to the PHUB registers in the PSoC 5LP Registers TRM. The source engine phase needs N + 3 cycles = 8 cycles. Total cycle time required for interspoke transfer is N + 6 = 5 + 6 = 11 cycles. 6.2.2.2 Intraspoke Transfer The timing diagram for intraspoke transfer under ideal conditions is shown in Figure 6-3. Figure 6-3. Intraspoke Transfer Cycle Timing Bus Clock DMA request latch phase Arbitration Phase Fetch Phase Source Engine Phase Destination Engine Phase Control Cycles Data Cycles Control 1 Control 2 Control 3 Control N Data 1 Data 2 Data N-1 Data N Control Cycles Data Cycles Control 1 Control 2 Control 3 Control N-1 Control N Data 1 Data 2 Data N-2 Data N-1 Data N Write Back Phase The total number of cycles for data transfer in the case of intraspoke DMA transfer is the sum of the cycles required for each phase. Total cycle time = Arbitration phase time (1) + Fetch phase (1) + Source engine phase (N + 1) + Destination engine phase (N + 1) + Write back phase (1) Total cycle time = 2N + 5 cycles (where N = Transfer count Spoke width) Intraspoke DMA transfer burst count should be limited to 16. In intraspoke DMA transfers, because the source and destination reside in the same spoke, the 16-byte internal FIFO of the PHUB is used as an intermediate buffer. When the FIFO is full, the PHUB waits for the FIFO to be emptied and the destination engine to read the data, and then fills the next set of data. This is the reason why the destination engine phase cannot happen in parallel with the source engine phase. Example PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 59 PHUB and DMAC You want to move four 32-bit data words from one SRAM location to another SRAM location. Notes ■ SRAM lies in spoke 0, which is a 32-bit spoke. ■ In this case, both source and destination is SRAM. The DMA configuration includes: ■ Burst count (configured in CFGMEMn.CFG0) 4 ■ Transfer count (configured in TDMEMn.ORIG_TD0) 4 bytes × 4 words = 16 ■ TD configuration includes increment source address and increment destination address to copy data from one array to another (configured in TDMEMn.ORIG_TD0) ■ N = Transfer count Spoke width = 16 4 = 4 The source and destination engine phase needs 2N + 2 cycles = (2 × 4) + 2 cycles = 10 cycles Total cycle time required for intraspoke transfer is 2N + 5 = (2 × 4 + 5) = 13 cycles 6.2.2.3 Handling Multiple DMA Channels The DMAC can perform phases in parallel. This helps to reduce the latency for executing data transfer. When multiple channels need to execute, the channels can be pipelined. Figure 6-4 shows processing of two DMA channels that were requested at the same time. The figure shows only the interspoke transfer. The same is applicable also for intraspoke transfer. Figure 6-4. Multiple DMA Channel Processing Bus Clock Arbitration phase for Channel 1 Fetch phase for Channel 1 Arbitration phase for Channel 2 Command Data Control Data Control Burst = 1 Burst = 2 Burst = N Burst = 1 Burst = 2 Burst = N Source Engine Phase for Channel 1 Destination Engine Phase for Channel 1 Fetch phase for Channel 2 Write back Phase for Channel 1 Command Data Control Data Control Burst = 1 Burst = 2 Burst = N Burst = 1 Burst = 2 Burst = N Source Engine Phase for Channel 2 Destination Engine Phase for Channel 2 Write Back phase for Channel 2 6.2.2.4 DMA Channel Priority Each channel can take a priority from 0 to 7 with 0 being the highest priority. The DMAC supports two different methods to handle the priority: simple priority, and grant allocation fairness algorithm. The priority handling method can be changed by writing to register PHUB.CFG bit “simple_pri” (bit 23). 60 ■ Simple Priority: This method handles the channels like any normal priority algorithm where a high priority channel can interrupt a low priority channel ■ Grant allocation Fairness algorithm: In this method, the channel 0 and 1 take highest priority and no other priority can interrupt the channels with priority 0 and 1. A DMA Channel of priority 0 and priority 1 occupy the bus 100%. Rest of the priorities share the bus based on the number of channels requested at that time. Because priority 0 has higher priority than 1, priority 0 can interrupt priority 1. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC In both the cases, a DMA channel of low priority can be interrupted by a high priority channel only during the source engine phase Under ideal conditions the Arbitration phase takes one cycle. Examples using the Grant allocation Fairness Algorithm Scenario 1 DMAC is free. Channel A with Priority 0 comes Figure 6-5. Priority 0 and Idle DMAC Bus Clock Channel A Priority 0 Arbitration Data Transfer Fetch 100 % Bus Use Write Back Scenario 2 DMAC is free. Channel B with Priority 1 is executing. Channel A with Priority 0 comes Figure 6-6. Priority 0 and Priority 1 Bus Clock Burst 1 Channel B Priority 1 Arbitration Fetch Burst 2 Burst N Burst 3 Data Transfer Data Transfer Data Transfer Data Transfer Request for Channel A (Priority 0) arrives Channel B resumes Burst 1 Channel A Priority 0 Arbitration Fetch Data Transfer Burst M 100 % Bus Use before Interruption and after high Priority Channel Completion Write Back Channel B completed Channel A completed Data Transfer 100 % usage of bus Write Back Scenario 3 DMAC is free. Channel B with Priority 2 is executing. Channel A with Priority 0/1 comes Figure 6-7. Priority 0/1 and Other Low Priority Bus Clock Burst 1 Channel B Priority 2 Arbitration Fetch Burst 2 Burst 3 Data Transfer Data Transfer Data Transfer Request for Channel A (Priority 0/ 1) arrives Channel B resumes Burst 1 Channel A Priority 0/ 1 Arbitration Fetch Data Transfer Burst M Data Transfer Burst N Data Transfer Write Back 100 % Bus Use before Interruption and after High Priority Channel Completion Channel B completed Channel A completed Write Back 100 % Bus Use Scenario 4 DMAC is free. Channel B with Priority 3 is executing. Channel A with Priority 2 comes PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 61 PHUB and DMAC Figure 6-8. Lower Priority Channels with Grant Allocation Bus Clock Burst 2 Burst 1 Channel B Priority 3 Arbitration Burst 3 Data Transfer Data Transfer Fetch Burst 4 Data Transfer Data Transfer The sharing of bus goes on until either of the channels completes the data transfer Channel A and B share the bus Request for Channel A (Priority 2) Arrives Burst 1 Channel A Priority 2 Arbitration Data Transfer Fetch The channels with priorities 2-7 are given access according to Table 6-2. Burst 2 Burst 3 Data Transfer Burst 4 Data Transfer Data Transfer Burst 5 Data Transfer Channel A gets more share of the bus because of it’s priority Because there are as many 24 DMA channels but only 8 priority levels, there can be multiple channels taking the same priority levels. Table 6-2. Priority Levels and Bus Allocation Priority Level DMAC uses the Round Robin method to handle DMA Channels with same priority. In case of Round Robin algorithm, the DMA channel which was not executed recently takes a higher priority. The execution of same priority DMA channels when round robin algorithm is enabled depends on Bus Allocation Percentage 2 50 3 25 4 12.5 5 6.3 6 3.1 7 1.5 When DMA channels of varied priority request for DMAC at a time, 100 percent of bus bandwidth will be allocated for channels of priority 0 or 1. ■ The last time when the channel was enabled ■ If the last time is the same for 2 channels, then DMA Channel with lower number takes higher priority Table 6-2 applies only if DMA channels with priorities 2 to 7 request simultaneously. Otherwise, the DMA channel with higher priority is given more access than Table 6-2 shows. Figure 6-9 shows a channel priority wheel that describes how the next 63 requests are handled if all channels with priorities 2 to 7 request simultaneously. If a channel with priority 2 to 7 is NOT requesting, the slots of the missing channel priority are used by the channel with the highest priority. In that case, channels with higher priority get more access than Figure 6-9 shows. Figure 6-9. DMA Channel Priority Wheel 2 5 2 2 4 2 3 2 3 2 6 2 3 3 2 4 2 4 3 2 Channel Request 2 3 5 2 2 2 3 3 2 4 2 DMA Channel priority wheel 2 4 2 3 2 3 7 2 2 5 3 2 2 3 4 2 4 2 2 3 62 2 2 6 2 3 2 4 2 3 3 2 5 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC Figure 6-10. Round Robin Scheduling All Channels have the same priority X Ch 5 Ch 1 Ch 2 Ch 3 Ch 4 Ch 4 Ch 5 Ch 3 Ch 2 Last executed time (t) Ch 1 Order of execution 6.2.2.5 DMA Latency in case of Nonideal Conditions The previous section explained the latency in case of ideal condition. But in real time, the ideal condition rarely exists. This section explains the latency calculation in case of nonideal conditions. The latency calculation in case of nonideal conditions cannot be explained using formula as against the ideal condition. source or destination spoke may be already used by CPU or another DMA channel When source and destination spoke is already in use, the PHUB does the arbitration. The following flow chart shows the arbitration mechanism. Multiple Requestors In real time system the PHUB will be requested by multiple channels and by CPU also. If there are multiple DMA channels sending request at the same time, the arbitration phase will take 2 cycles instead of the ideal 1 cycle CPU Interrupts with Fetch Phase The fetch phase ideally takes only 1 cycle for the PHUB to access the configuration registers through the PHUB local spoke. When CPU interrupts the fetch phase, the latency depends on when the CPU releases the configuration registers. Typically CPU takes 2 cycles for the access of configuration registers. Also, there might be some high priority DMA channel in the Fetch phase. These scenarios will also add to the DMA Channel execution latency. Source and Destination Spokes in Use The source and destination for a particular DMA Channel should be free for the channel to use it. In real time, a PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 63 PHUB and DMAC Figure 6-11. DMA Channel Arbitration Assume Channel A is the DMA Channel trying to access the spoke Is CPU using the spoke Yes Channel B using spoke Is it a CPU Priority spoke No No Current burst for CPU is completed CPU process is interrupted Does Channel B have high Priority Yes Yes Latency depends on CPU processing time Has the CPU released the spoke? No Latency depends on burst length of the other DMA channel No Is the burst completed for Channel B Yes Yes DMA channel accesses the spoke Current burst for Channel B is completed Channel B is interrupted The Channel A accesses the spoke DMA channel accesses the spoke Channel A accesses the spoke DMA Channel completes transfer Channel A completes transfer Spoke released for the DMA This latency is not measurable and depends on the real time situation where same spoke can be accessed by multiple resources. Source and destination peripherals are not Ready When the source or the destination peripheral is not ready to send or receive data, then the DMA channel has to wait till it is ready. In case of source peripheral not ready, the DMA channel will wait for the source peripheral to become ready In case of destination peripheral not ready, the DMA channel will use the 16 byte FIFO of the PHUB. It reads the data from the source and fills it in the FIFO till the destination peripheral is ready. Thus the internal 16 byte FIFO is used during intra-spoke transfer and also during the conditions where the source and destination peripherals are no ready. The spoke widths play a very important role in latency. There are chances that the source spoke might be smaller than the destination spoke and vice versa. In this case the burst count also plays an important role. Let's see some examples for this condition Scenario 1 (Interspoke: 16 bit spoke to 32 bit spoke; Burst of 2) ■ Source: 16 bit spoke (ADC) ■ Destination: 32 bit spoke (DFB) ■ Burst count: 2 (for 16 bit ADC data) Source and destination spoke are of different width 64 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC Figure 6-12. Data Transfer between 16-bit and 32-bit Spoke Bus clock Peripheral A (16 bit data) 16 bit spoke 32 bit spoke 2 Bytes 2 Bytes Peripheral B 2 Bytes Burst Count = 2 Scenario 2 (Interspoke: 16 bit spoke to 32 bit spoke; Burst of 4) ■ Source: 16 bit spoke (ADC) ■ Destination: 32 bit spoke (DFB) ■ Burst count: 4 (for 20 bit ADC data) Figure 6-13. Data Transfer Between 16 bit and 32 bit Spoke Bus clock Peripheral A (32 bit data) 16 bit spoke 32 bit spoke Peripheral B 2 Bytes SourceAddr++ Burst Count = 4 2 Bytes Source address incremented by source spoke width to read the next 2 bytes of data Scenario 3 (Interspoke: 32 bit spoke to 16 bit spoke; Burst of 4) ■ Source: 32 bit spoke (Memory) ■ Destination: 16 bit spoke (UDB peripheral) ■ Burst count: 4 Figure 6-14. Data Transfer Between 16 bit and 32 bit Spoke Bus clock Peripheral A (32 bit data) 32 bit spoke 16 bit spoke Peripheral B (32 bit data) 2 Bytes DestAddr++ Burst Count = 4 2 Bytes Destination address incremented by destination spoke width to write the next 2 bytes of data Scenario 4 (Interspoke: 16 bit spoke to 16 bit spoke; Burst of 2) ■ Source: 16 bit spoke PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 65 PHUB and DMAC ■ Destination: 16 bit spoke ■ Burst count: 2 Figure 6-15. Data Transfer Between Two 16 bit Spoke Bus clock Peripheral A (16 bit data) 16 bit spoke 16 bit spoke 2 Bytes Peripheral B (16 bit data) 2 Bytes Burst Count = 2 Scenario 5 (Interspoke: 16 bit spoke to 16 bit spoke; Burst of 4) ■ Source: 16 bit spoke ■ Destination: 16 bit spoke ■ Burst count: 4 Figure 6-16. Data Transfer Between Two 16 bit Spoke Bus clock Peripheral A (32 bit data) 16 bit spoke 16 bit spoke 2 Bytes Burst Count = 4 Peripheral B (32 bit data) SourceAddr++, DestAddr++ 2 Bytes Source and Destination address incremented by their spoke widths to read and write the next 2 bytes of data Scenario 6 (Intraspoke: 16 bit spoke; Burst of 1) ■ Source and destination: Same spoke (16 bit) ■ Burst count: 1 66 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC Figure 6-17. Intraspoke Data Transfer Peripheral A (8 bit data) 16 bit spoke Peripheral B (8 bit data) Bus clock 1 Byte to PHUB FIFO Burst Count = 1 PHUB FIFO to Destination Data read and write using intermediate PHUB FIFO Scenario 6 (Intraspoke: 16 bit spoke; Burst of 2) ■ Source and destination: Same spoke (16 bit) ■ Burst count: 2 Figure 6-18. Intraspoke Data Transfer Peripheral A (16 bit data) 16 bit spoke Peripheral B (16 bit data) Bus clock 2 Bytes to PHUB FIFO Burst Count = 2 PHUB FIFO to Destination Data read and write using intermediate PHUB FIFO PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 67 PHUB and DMAC Source and destination address do not have even addressing The address of the source and destination play a very important role in deciding the latency. The AHB protocol supports reading from even addresses. Use this notation for a 32 bit spoke. Figure 6-19. Addressing in 32 bit Spoke Address n Byte 0 Byte 1 Byte 2 Byte 3 Address n + 1 Byte 0 Byte 1 Byte 2 Byte 3 Figure 6-20. Addressing in 16 bit Spoke Address n Byte 0 Byte 1 Address n + 1 Byte 0 Byte 1 Scenario 1: 32 bit spoke, Burst count of 4, Address begins at Byte 1 Figure 6-21. Odd Addressing in 32-Bit Spoke Bus Clock Byte 1 Data Read cycles for Burst = 4 Byte 2 and 3 Byte 0 of Addr + 1 As seen from the above figure, when the even addressing is not met, the bus cycle increases. In ideal condition where the address begins at Byte 0, a single cycle is sufficient to read all the 4 bytes. Scenario 2: 16 bit spoke, Burst count of 2, Address begins at Byte 1 Figure 6-22. Odd Addressing In 16 bit Spoke Bus Clock Byte 1 Data Read cycles for Burst = 2 Byte 0 of Addr + 1 6.2.2.6 68 Request per Burst Bit PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC The data to be transferred can be split into multiple burst each of same size. This feature is useful under the following situations: ■ When the user doesn't want to hog the bus with a single channel which has huge data to transfer ■ When the user needs to control the transfer times The “Request per bit” is bit 7 in CFGMEMn.CFG0 register. This bit is available for individual channel. When this bit is set, the DMA needs a request to transfer the next burst of data. When this bit is set, the DMA channel should go through the whole process from Arbitration phase till Write back phase for every burst. Thus the “Request per bit” parameter will significantly increase the transfer time 6.2.2.7 Work Sep Bit The “work_sep” bit is bit 5 of the CHn.BASIC_CFG register. This bit is available for individual channel. When this bit is cleared, a TD mapped to that particular DMA channel cannot restore its initial configuration after the data transfer. The TD will retain its last source address, destination address and transfer count details at the end of transfer. When this bit is set, a TD mapped to that particular DMA channel restores its initial configuration after the data transfer. This is very useful when the TD should be repeated. When the “work_sep” bit is set, DMA uses a separate processing area to store the TD configuration details. 6.3 DMA Transaction Modes The DMA channels can be chained to perform complex operation. Similarly TDs can be nested or chained to perform complex operations. Chaining of TDs is done using the bit “next_td_ptr” in TDMEMn.ORID_TD0 register. This flexibility of the DMA channel and TD helps to create both simple and complex cases General use cases might include the following types 6.3.1 6.3.2 Auto Repeat DMA A static pattern is repetitively read from system memory and written to a peripheral. This is done with a single TD that chain to itself. Figure 6-24. Auto Repeat DMA DMA Channel A 6.3.3 TD A Ping Pong DMA Double buffering is used to allow one buffer to be filled by one client, while another client is consuming the data previously received in the other buffer. In its simplest form, this is done by chaining two TDs together where each TD calls the opposite TD when complete. Figure 6-25. Ping Pong DMA DMA Channel A TD A TD B 6.3.4 Circular DMA This is similar to ping pong DMA except that it contains more than two buffers. In this case, there are multiple TDs where after the last TD is complete it chains back to the first TD. Figure 6-26. Circular DMA DMA Channel A TD A Simple DMA TD B A single TD is used to transfer data between two peripherals or memory locations. TD C Figure 6-23. Simple DMA Transfer TD D DMA Channel A TD A 6.3.5 Indexed DMA An external master requires access to locations on the system bus as if those locations were shared memory. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 69 PHUB and DMAC Example: If a peripheral was configured as an SPI or I2C slave where an address is received by the external master, that address becomes an index or offset into the internal system bus memory space. This is accomplished with an initial “address fetch” TD that reads the target address location from the peripheral and writes that value into a subsequent TD in the chain. This causes the TD chain to be modified during the process. When the “address fetch” TD completes, it can move onto the next TD, which has the new address information embedded in it. This TD carries out the data transfer with the address location requested by the external master. Figure 6-27. Indexed DMA Index DMA Channel A TD A TD B TD C of data phase TDs) can begin (potentially using scatter gather). After the data phase TDs finish, a status phase TD can be invoked that reads some memory mapped status information from the peripheral and copies it to a location in system memory specified by the CPU for later inspection. Multiple sets of configuration/data/status phase sub-chains can be strung together to create larger chains that transmit multiple packets in this way. A similar concept exists in the opposite direction for the reception of the packets. 6.3.8 Nested DMA One TD can modify another TD, as the TD configuration space is memory mapped, just as any other peripheral. Example: A first TD loads a second TDs configuration and then calls the second TD. The second TD moves data as required by the application. When complete, the second TD calls the first TD, which again updates the second TDs configuration. This process repeats as often as necessary. TD D TD E 6.3.6 Scatter Gather DMA Multiple noncontiguous sources or destinations are required to effectively carry out an overall DMA transaction. Example: A packet can be required to be transmitted off of the device and the packet elements, including the header, payload, and trailer exist in various non-continuous locations in memory. Scatter-gather DMA allows the segments to concatenate together by using multiple TDs in a chain that gathers data from multiple locations. A similar concept applies for the reception of data onto the device. Certain parts of the received data may need to be scattered to various locations in memory for software- processing convenience. Each TD in the chain specifies the location for each discrete element in the chain. 6.3.7 Packet Queuing DMA This is similar to scatter gather DMA, but it specifically connotes packet protocols whereby there can be separate configuration, data, and status phases associated with sending or receiving a packet. Example: To transmit a packet, a memory mapped configuration register can be written inside a peripheral specifying the overall length of the ensuing data phase. This configuration information can be setup by the CPU anywhere in system memory and copied with a simple TD to the peripheral. After the configuration phase, a data phase TD (or a series 70 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F PHUB and DMAC 6.4 Register List Table 6-3. PHUB and DMA Register List Register Name PHUB_CFG Comments PHUB General Configuration register Features Specifies prune_clock delay, number of wait states, allocation fairness algorithm, priority, priority spoke, CPU_CLOCK_EN setting PHUB detects the following errors: PHUB_ERR PHUB Error Detection register 1. Bus Timeout 2. Unpopulated address access 3. Peripheral AHB ERROR response If the error was detected as a result of a CPU access then PHUB will send an AHB ERROR response to the CPU. If the error was detected as a result of either a CPU or DMA access then PHUB will set the corresponding bit in the following ERR register. PHUB_ERR_ADDR PHUB Error Address register Contains the address that caused an error to trigger PHUB_CH[0..23]_BASIC_CFG Channel Basic Configuration register Sets basic channel configurations in gates inside PHUB PHUB_CH[0..23]_ACTION Channel Action register Sets action for each channel PHUB_CH[0..23]_BASIC_STATUS Channel Basic Status register Provides status information in gates inside PHUB PHUB_CFGMEM[0..23]_CFG0 PHUB Channel Configuration register 0 Each channel has some configuration information stored in RAM. This configuration information is called CHn_CFG0/1. PHUB_CFGMEM[0..23]_CFG1 PHUB Channel Configuration register 1 CHn_CFG0/1 are stored in CFGMEM at {CH_NUM[5:0], 000}. PHUB_TDMEM[0..127]_ORIG_TD0 PHUB Original Transaction Descriptor 0 PHUB_TDMEM[0..127]_ORIG_TD1 PHUB Original Transaction Descriptor 1 Each channel has a TD chain (as short as one TD in length) that provides instructions to the DMAC for carrying out a DMA sequence for the channel. The TD chain is comprised of one or more CHn_ORIG_TD0/1 TDs. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F DMAC accesses the CHn_ORIG_TD0/1 chain from TDMEM and the address in TDMEM of the current TD in the chain is {TD_PTR[7:0], 000}. 71 PHUB and DMAC 72 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 7. Interrupt Controller The Interrupt Controller provides the mechanism for hardware resources to change the program address to a new location independent of the current execution in the main code. The interrupt controller also handles continuation of the interrupted code being executed after the completion of the interrupt service routine. 7.1 Features The following are features of the interrupt controller: ■ Supports 32 interrupt lines ■ Programmable interrupt vector ■ Configurable priority levels from 0 to 7 ■ Support for dynamic change of priority levels ■ Support for individual enable/ disable of each interrupt ■ Nesting of interrupts ■ Multiple sources for each interrupt line (can be either fixed function, UDB, or from DMA) ■ Supports both level trigger and pulse trigger ■ Tail chaining, late arrivals and exceptions are supported in PSoC 5LP devices 7.2 Block Diagram Figure 7-1 is a block diagram of the interrupt controller. Figure 7-1. Interrupt Controller Block Diagram Interrupt Signals 0 1 2 16-bit Interrupt Vector Address (IAV) Interrupt Controller Interrupt Request (IRQ) CPU Acknowledgment for Interrupt Entry (IRA) 31 Acknowledgment for Interrupt Exit (IRC) PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 73 Interrupt Controller 7.3 How It Works The interrupt controller supports 32 interrupt signals. The interrupt signal can come from one of the three sources (see Figure 7-2): ■ Fixed function block ■ DMA channels ■ UDB blocks The interrupt signal routing is very flexible with PSoC 5LP architecture. The interrupt lines pass through a multiplexer. The mux selects one among the following: Fixed function IRQ (Interrupt request), UDB IRQ with level, UDB IRQ with Edge, and DMA IRQ. The IDMUX.IRQ_CTL register is used to configure the mux for the IRQ selection. Figure 7-2. Interrupt and DMA Processing in the IDMUX Fixed Function IRQs 0 1 Interrupt Controller UDB IRQs 2 UDB Array Edge Detect 3 UDB DRQs DMA termout (IRQs) 0 Fixed Function DRQs DMA Controller 1 Edge Detect The interrupt controller unit prioritizes and sends the request to the CPU for execution. The list of interrupt sources and the corresponding interrupt number is available in the device datasheet. 2 Table 7-1. Bit Status During Read and Write Register Operation Bit Value Comment 1 To enable the interrupt 0 No effect 1 Interrupt is enabled 0 Interrupt is disabled 1 To disable the interrupt 0 No effect 1 Interrupt is enabled 0 Interrupt is disabled Write 7.3.1 Enabling Interrupts The interrupt controller provides features to enable and disable individual interrupt lines. The Enable register (SETEN) and the Clear Enable register (CLREN), respectively, enable and disable the interrupt lines. Each bit in the register corresponds to an interrupt line; these registers enable and disable interrupts and read the enable status of interrupts. The register that is updated latest (SETEN or CLREN register) determines the interrupt enable status. Table 7-1 shows the status of bits during read and write. 74 SETEN Read Write CLREN Read PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Interrupt Controller 7.3.2 Pending Interrupts 7.3.3 When the interrupt controller receives the interrupt signal, it sets the pending bit. “Set Pending register” (SETPEND) and the “Clear Pending register” (CLRPEND) also allow the pending bit to be set and cleared through software. Each bit in the register corresponds to an interrupt line. The pending bit status can be read by reading these registers. For both pulse/level interrupts, the pending bit is cleared immediately upon receiving the acknowledgement from the CPU on interrupt entry (IRA). For pulse interrupts, the pending bit can be set again by arrival of a new pulse interrupt on the same line after the IRA. But for level interrupt, the interrupt controller checks the status of the interrupt line when it receives the acknowledgement from the CPU on interrupt exit (IRC). During that time, if the interrupt line is still asserted, the pending bit is reset. If there is no assertion on the interrupt line, the pending bit remains in cleared state. The interrupt controller provides a priority handling feature to help a user assign priority for each interrupt. Characteristics of this feature are as follows: ■ Eight levels of interrupt priorities from 0 to 7. ■ Priority level 0 is highest and level 7 is lowest. ■ Priority levels set using the Interrupt Priority Registers PRI_[x]. ■ Support of dynamic configuration of priority levels – A change of priority level of an interrupt on the fly does not affect the current execution of the same interrupt; it takes effect for the next assertion. Priority handling is very important in the following cases: ■ Case 1 – If an interrupt (INT B) is asserted when another interrupt (INT A) is being executed, there are three possibilities with unique handling sequences: ❐ Operation If INT A has lower priority than INT B: 1.INT A is stopped at the point of execution. Table 7-2. Pending Bit Status Register Interrupt Priority Bit Value 2.The details of INT A are pushed to the stack, and INT B begins to execute. Comment 3.After the execution of INT B, INT A execution is resumed from the point of its interruption. 1 To put an interrupt to pending 0 No effect 1 Interrupt is pending 0 Interrupt is not pending 1.INT B has to wait until INT A is executed. 1 To clear a pending interrupt 0 No effect 2.After the execution of INT A, INT B can start execution. 1 Interrupt is pending 0 Interrupt is not pending Write SETPEND ❐ Read Write CLRPEND ❐ Read The pending register can also be written by software. When the software writes a 1 to the pending bit, it activates the interrupt. When software clears the pending bit, the interrupt does not occur. When the software request to clear a pending bit and hardware request to set the pending bit occurs simultaneously, the hardware request takes the higher priority. Setting of the pending bit when the same bit is already set results in only one execution of the interrupt. The pending bit can be updated regardless of whether or not the corresponding enable bit is set. If the enable bit is not set, the interrupt line will be pended until the interrupt is enabled, unless the user clears the bit. It is advisable to check the state of the pending bit before enabling the interrupt. The choice is left to the user, of whether to set the pending bit before or after the enable bit is set, for enabling the corresponding interrupt. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F If INT A has higher priority than INT B: If INT A and INT B have equal priority: 1.If INT A is being executed; INT B has to wait until INT A is executed. After the execution of INT A, INT B can start execution. 2.If INT B is being executed; INT A has to wait until INT B is executed. After the execution of INT B, INT A can start execution. ■ Case 2 – During the simultaneous occurrence of interrupts: ❐ If INT A has lower priority than INT B, then INT B wins arbitration and begins to execute. ❐ If INT A has higher priority than INT B, then INT A wins arbitration and begins to execute. ❐ If INT A and INT B have equal priority, then the interrupt with the lower index number wins arbitration and begins to execute. 75 Interrupt Controller 7.3.4 Level versus Pulse Interrupt The interrupt controller supports both Level and Pulse interrupts. The interrupt controller includes the Pulse detection logic, which detects the rising edge on the interrupt line. The pulse detection logic pends the interrupt bit whenever it detects the rising edge. The interrupt controller detects any assertion in the interrupt signal and executes the interrupt as follows: ■ ■ Level Interrupt – With level interrupts, the interrupt request bit in the corresponding peripheral register must be cleared by the firmware inside the interrupt service routine. If the interrupt request bit in the peripheral register is set, it results in a level high signal on the interrupt line. At the interrupt exit, if the interrupt request bit is set in the peripheral register, the interrupt pending bit is set again and the interrupt is processed again if it is enabled. Pulse Interrupt – A pulse occurs at the interrupt line. The low to high edge of the pulse sets the pending bit and the corresponding interrupt is executed. If the pulse occurs while the pending bit is already set, the second pulse has no effect, because the pending bit is already set. The Pending bit is automatically cleared by the interrupt controller at ISR entry. However, if the pulse comes while the interrupt is currently active, the interrupt pending bit is set again, and the interrupt is executed again. 7.3.5 Interrupt Execution The interrupt controller controls both Level and Pulse interrupt in the following sequence: 1. Interrupt execution corresponding to the interrupt signal requires the interrupt to be enabled (assuming priority and interrupt vector address are programmed already). 2. When an assertion occurs in the interrupt signal, the pending bit corresponding to the interrupt number is set in the pending register, indicating that the interrupt is waiting for its execution. 3. The Priority Decoding unit reads the priority and determines when the interrupt can be executed. 4. The interrupt controller sends the interrupt request to the CPU, along with the interrupt vector address for execution. 5. The CPU receives the request. 6. Interrupt Entry (IRA) – The CPU acknowledges the interrupt entry. The next assertion in the same interrupt line can be detected only after the interrupt entry. Any assertions before that are ignored. The interrupt controller clears the pending bit upon receiving the acknowledgement. 7. The current interrupt number and its priority are pushed to the interrupt controller stack by the interrupt controller. 8. Interrupt Exit (IRC) – When interrupt execution is completed, the processor is free to address the next request. The CPU acknowledges the interrupt exit. At the interrupt exit, the interrupt context (that is, interrupt number and priority) is popped from the stack. Figure 7-3 lists the basic operations during an interrupt signal assertion and its handling. 76 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Interrupt Controller Figure 7-3. Interrupt Signal Assertion and Handling Assertion on the interrupt Line Set of Pending Bit Wait until all high priority interrupts finish Send to Priority Decoding Unit No Is this the highest priority Interrupt? Yes Send request to CPU (with IRQ and IVA) CPU accepts request IRA sent by CPU (INTC pushes Interrupt details to its stack) Pending bit cleared for Pulse and Level Interrupt Interrupt execution Interrupt Exit acknowledgment (IRC) from CPU (INTC pops interrupt details from the stack) 7.4 PSoC 5LP Interrupt Controller Features Because PSoC 5LP architecture is based on the Cortex-M3 core, it has additional features supported by the Cortex-M3 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Interrupt line is checked in case of Level interrupt. If interrupt line has a high level, the pending bit is sent again. core. In PSoC 5LP devices, the interrupt controller is a part of the Cortex-M3 core. For more detailed information about the PSoC 5LP Interrupt Controller, see the ARM Cortex-M3 Technical Reference Manual available at http:// www.arm.com. 77 Interrupt Controller 7.4.1 Active Interrupts An active interrupt is the one being executed currently. The interrupt priority and interrupt number of the active interrupt are stored in the CPU stack. Whenever an interrupt begins to execute, the interrupt priority and number are pushed to the stack. The contents of the stack can be read to find the Active Interrupt details. With PSoC 5LP devices, the CPU stack is used. There are two stacks accessed using two different stack pointers: The Process Stack Pointer (PSP) and the Main Stack Pointer (MSP). Cortex-M3 can be configured to use two stacks. When it is configured to use both the stacks, the first interrupt uses the PSP or the MSP to store interrupt details, depending on which is currently active. The stack grows downwards. A nested interrupt uses only MSP to store the details. When it is not configured to use two stacks, only the MSP is used. PSoC 5LP devices also support an ACTIVE register to store the active status of the interrupt. Its characteristics are: ■ Each bit in the register indicates the active state of the corresponding interrupt. ■ When the bit is set to 1 in the ACTIVE register, the interrupt is active. When the bit is set to 0, the interrupt is currently inactive. ■ When the current running interrupt is suspended due to a high priority interrupt, the state of the current running interrupt is maintained as “Active” because it continues its execution after execution of the high priority interrupt. ■ The active state of the bit is cleared only after execution of the interrupt. PSoC 5LP devices also supports exceptions other than interrupts. The ACTIVE bits correspond only to interrupts and not to exceptions. The active status details of exceptions are stored in the Exception Status register. Exception Status registers are not only used to read the active status but also to enable exceptions. 7.4.2 The configuration controls how you use PSP and MSP. If both stacks are used, the Process Stack Pointer or Main Stack Pointer, which ever is currently active, is used by the first interrupt. All other nested interrupts use only the MSP. If only one stack is configured for use, the interrupt details are stored in the MSP. The sequence is: 1. When the high priority interrupt comes during the execution of the low priority interrupt, the interrupt controller sends a request to the CPU and low priority interrupt execution is stopped by the CPU at that point. 2. The details, such as instruction pointer and other general purpose registers for the low priority interrupt, are pushed to the stack. (The stack used depends on nesting. It can be either MSP or PSP as explained previously). 3. The number of nesting supported depends on the availability of stack space. Because system stack is used, the user should ensure that sufficient stack space is available. Insufficient stack space causes undetermined results. After the stack push for the low priority is done, the details of the current active interrupt (high priority interrupt) is stored in the CPU stack. The high priority interrupt executes. 4. After the higher priority interrupt has executed, the interrupt details of the high priority interrupt are popped from the stack. Following this, the details of the low priority interrupt (PC and other register details) are popped from the stack. The low priority interrupt continues its execution from the point of suspension. 5. Because the push and pop of stack is handled by the hardware, there is minimum latency; no instruction is involved in the operation. Figure 7-4 on page 79 shows a timing diagram of the register states during the nesting operation. Interrupt Nesting Nesting of an interrupt occurs when a high priority interrupt is asserted during a low priority interrupt execution. With PSoC 5LP architecture, only the CPU stack is available to store all nesting interrupt details. ■ Current interrupt number, current interrupt priority ■ Program counter, PSR, R0 to R3, R12 and LR ■ Depending on the application, other registers from R4 to R11 The CPU stack grows down while the CPU handles push and pop. 78 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Interrupt Controller Figure 7-4. Register Timing During Nesting Assertion of INT A Assertion of INT B Assertion of INT C Execution of INT A Stacking of INT A Register state before all executions PSP MSP XX XX Execution of Continuation Continuation INT C of INT B of INT A Execution of INT B Stacking of INT B Un-stacking of INT A Register state after Stacking of INT A and execution of INT B Register state during execution of INT A PSP Un-stacking of INT B MSP PSP XX INT A Details INT A Details Register state after Stacking of INT B and execution of INT C MSP PSP INT B details INT A Details MSP INT C details INT B details and other registers Other register details of INT A Other register details of INT A Register state after un-Stacking of INT B and execution of INT B PSP INT A Details Register state after un-Stacking of INT A and execution of INT A MSP INT B details PSP MSP INT A Details XX Register state after all executions PSP MSP XX XX Other register details of INT A In Figure 7-4, INT A is suspended, and the high priority interrupt INT B is executed. During nesting, the INT A is pushed to the stack. During execution of INT B, INT C occurs. So INT B is pushed, and INT C is executed. After INT C is executed, INT B is popped and executed. After INT B is executed, the stack is popped. When an interrupt begins to execute, interrupt information is stored in the stack; when it completes, the stack is popped. The use of both PSP and MSP is shown. It is assumed that PSP is active during the first interrupt and that the first active interrupt uses the PSP. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 79 Interrupt Controller 7.4.3 Interrupt Vector Addresses PSoC 5LP architecture has a feature that allows a user to specify the interrupt service routine for every interrupt line. The call of the interrupt service routine corresponding to an interrupt line is not a branch instruction. The address of the interrupt service routine is stored in the vector table, which results in the direct call of the routine. This method of execution prevents latency in the call of the interrupt service routine. When interrupt assertion occurs, the following sequence occurs: 1. The address of the interrupt service routine is taken from the interrupt vector table and is executed. 2. The list of interrupt vector addresses is stored in the vector table. The interrupt service routine address is programmable and is stored in the vector table. The vector table is a location in the memory and has a base address; the other vector addresses are accessed as offset from the base address. By default, the vector table is at location 0x00 in the ROM. The base address of the vector table can be changed; the vector table can be moved, either in the ROM itself or to the RAM. Each vector address is 32 bits long; when moving the vector table, the user should ensure that there is enough space to hold the supported 4-byte addresses for the 32 interrupt lines 3. PSoC 5LP devices contain the Vector Table Offset register that contains two data: Position of vector table in ROM/RAM. Offset value from the start ROM or RAM region. This offset value acts as the base address for the vector table. 4. When the vector table is moved, the boot image should contain the stack pointer value, Reset vector, NMI vector, and hard Fault vector, because these are required for the beginning of execution of code. 7.4.4 Tail Chaining Tail chaining is the process used to reduce interrupt latency. When a new interrupt assertion occurs at the same time as another interrupt being executed with the same or higher priority, the following sequence occurs: 1. The new interrupt with a lower priority is pended. 2. After the current interrupt is executed, the details of the current interrupt in the stack are not popped. 3. The details of the new interrupt are pushed to the stack and the new interrupt begins its execution. 4. After the execution of the new interrupt, details of the new interrupt and the previous interrupt are popped from the stack. Because stacking and unstacking are avoided between the two, interrupts, latency is greatly reduced. Tail chaining can save a maximum of six cycles. 7.4.5 Late Arrival Interrupts A late arrival interrupt occurs when another interrupt is being pushed to the stack for execution. Another feature reduces interrupt latency by handling such late arrival interrupts. The following sequence describes the process: 1. A low priority interrupt is asserted. 2. The details of the low priority interrupt are being pushed to the stack, when a high priority interrupt assertion happens. 3. After the stacking of the low priority interrupt, the high priority interrupt is stacked and executed, instead of the low priority interrupt. 4. After execution of the high priority interrupt, the low priority interrupt is executed. 5. Because the vector address is 32 bits long, the LSB is filled with 0x01, and the MSB contains the corresponding 24-bit ISR address to be executed. The presence of 0x01 in the LSB indicates Thumb instructions. 6. During the interrupt signal assertion, the address of the interrupt service routine (the Interrupt Vector Address (IVA)) is retrieved from this table and given to the CPU for execution of the interrupt. 7. Because PSoC 5LP devices also support exceptions, the vector table has the address corresponding to the 15 exceptions followed by the interrupt service routine addresses. 80 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Interrupt Controller 7.4.6 Exceptions PSoC 5LP architecture supports 15 different exceptions, as shown in Table 7-3. These exceptions are used to handle fault conditions that can occur in the system. Exceptions can have fixed priority or configurable priority. Exceptions are handled in the same manner as interrupts. The State register is used to enable or disable exceptions. Table 7-3. PSoC 5LP Exceptions Interrupt Number Exception Type Priority Comments 1 Reset -3 (highest) not programmable Reset 2 NMI -2 not programmable Non-Maskable Interrupt 3 Hard Fault -1 not programmable All fault conditions if the corresponding handler is not enabled 4 Reserved NA – 5 Bus Fault Programmable Bus error; occurs when AHB interface receives an error response from a bus slave (also called prefetch abort if it is an instruction fetch or data abort if it is a data access) 6 Usage Fault Programmable Exceptions due to program error 7 Reserved NA -- 8 Reserved NA -- 9 Reserved NA -- 10 Reserved NA -- 11 SVCall Programmable System Service Call 12 Debug Monitor Programmable Debug monitor (watchpoints, breakpoints, external debug request) 13 Reserved NA -- 14 PendSV Programmable Pendable request for system device 15 SYSTICK Programmable System Tick Timer 7.4.7 Interrupt Masking PSoC 5LP architecture supports special methods to mask interrupts and exceptions, preventing them from execution. Any new assertions in the interrupt lines are detected and pended until the interrupts are unmasked. Masking of interrupts is different from enabling or disabling. When masked, the interrupt is blocked for some time, even though it is enabled. This feature is useful when it is necessary to protect some critical section of code. When interrupts are masked, pending interrupts are not executed, even though the interrupts are enabled in the enable register. The interrupts are executed only when masking is cleared. PSoC 5LP devices have special registers to provide masking facilities, including: ■ PRIMASK – When the bit in the PRIMASK register is set, all interrupts and exceptions except NMI and Hard fault are blocked. ■ FAULTMASK – When the bit in the FAULTMASK register is set, all interrupts and exceptions except NMI are blocked. ■ BASEPRI – When interrupts below a certain priority level must be masked, the priority number can be specified in the BASEPRI register. All interrupts with a priority number equal to or less than the priority level specified in the BASEPRI register are masked. 7.5 Interrupt Controller and Power Modes The CPU core (Cortex-M3) can execute even when the power or clock for the Interrupt Controller is switched off. In this case, care should be taken during entry/ exit into differ- PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 81 Interrupt Controller ent low-power modes (alternate active, sleep and hibernate). On PSoC 5LP, an interrupt signal coming from a wakeup source should not be passed through the "Edge Detect" logic shown in Figure 7-2. The interrupt signal should be passed directly to the interrupt controller. This is a requirement only for sleep and hibernate power mode wakeup sources. Alternate active mode wakeup sources can have their interrupt signals either passing directly to the interrupt controller, or through the edge detect logic. Follow these steps before switching off the Interrupt Controller clock. 1. Clear all pending interrupts and disable all interrupts in Interrupt Controller. 2. NOP. 3. Disable the Global Interrupt bit. 4. Turn OFF the clock for Interrupt Controller in the CLOCK_EN bit in the INTC.CLOCK_EN register. It is preferred not to operate any Interrupt related functions when the clock to the interrupt controller is not available. When an Interrupt Service routine is executed by the CPU when the clock to the interrupt controller is switched off, the CPU should make sure the clock for the Interrupt Controller is re-enabled before the exit from the ISR (to process the IRC signal). If this is not taken care, it will lead to undefined behavior. When returning from the lower power mode or wants to continue in the alternate active mode, follow these steps: 1. Clock must be available to Interrupt Controller 2. Enable the Global interrupt bit 3. Enable the required interrupts in the Interrupt Controller The CPU can run when the interrupt controller clock is switched off only during active and alternate active modes. When the user wants to switch from alternate active to Active mode when the Interrupt controller clock is switched off. a. Follow the steps mentioned above to switch off the clock for the Interrupt controller b. Now the CPU can run any code that doesn't involve the Interrupt functionality. c. Switch to the active state whenever required d. To switch to active mode only on wake up on interrupt, then the CPU should keep polling the PM.MODE_CSR register to find when the system should switch to active mode. e. When switching back to active mode, follow the procedures mentioned above for switching from lowpower mode to active mode. 82 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Section C: Memory The PSoC® nonvolatile subsystem consists of flash, byte-writable EEPROM, and nonvolatile configuration options. The CPU can reprogram individual blocks of flash, enabling boot loaders. An Error Correcting Code (ECC) can enable high reliability applications. A powerful and flexible protection model allows the user to selectively lock blocks of memory for read and write protection, securing sensitive information. The byte-writable EEPROM is available on-chip for the storage of application data. Additionally, selected configuration options, such as boot speed and pin drive mode, are stored in nonvolatile memory, allowing settings to become active immediately after power on reset (POR). This section encompasses the following chapters: ■ Nonvolatile Latch chapter on page 85 ■ SRAM chapter on page 89 ■ Flash Program Memory chapter on page 93 ■ EEPROM chapter on page 95 ■ EMIF chapter on page 97 ■ Memory Map chapter on page 105 ■ Cache chapter on page 147 Top Level Architecture (Block diagram here taken from main block diagram in Introduction.) Memory Block Diagram System Bus MEMORY SYSTEM EEPROM SRAM CPU SYSTEM EMIF FLASH PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 83 Section C: Memory 84 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 8. Nonvolatile Latch A Nonvolatile Latch (NVL or NV latch) is an array of programmable, nonvolatile memory elements whose outputs are stable at low voltage. It is used to configure the device at Power on Reset. Each bit in the array consists of a volatile latch paired with a nonvolatile cell. On POR release nonvolatile cell outputs are loaded to volatile latches and the volatile latch drives the output of the NVL. 8.1 Features NV latches include: ■ A 4x8-bit NV latch for device configuration ■ A 4x8-bit Write Once NV latch for device security 8.2 Device Configuration NV Latch Device configuration NV latches allow configuration of PSoC® device parts before the CPU reset is released. For example, the user may configure each I/O port to be in one of four drive modes before CPU reset is released. Device configuration NV latch values have lower endurance and must be written in a narrower temperature window. Programming temperature range and endurance are traded off to meet the low voltage and wide temperature requirements. For endurance, retention, and temperature specs for NV latches see the specific device datasheet. The Device Configuration NV Latch register map is shown in Table 8-1. Table 8-1. Device Configuration Register Map Register Address 7 6 5 4 3 0x00 PRT3RDM[1:0] PRT2RDM[1:0] PRT1RDM[1:0] 0x01 PRT12RDM[1:0] PRT6RDM[1:0] PRT5RDM[1:0] 0x02 XRESMEN DEBUG_EN 0x03 8.2.1 Reserved DIG_PHS_DLY[3:0] ECCEN 2 1 0 PRT0RDM[1:0] PRT4RDM[1:0] PRT15RDM[1:0] DPS[1:0] CFGSPEED PRTxRDM[1:0] Port Reset Drive mode NVL bits enable selection of one of four drive modes to be in effect between the release of POR and the configuration of the device by user firmware. These four drive modes are a subset of the drive modes available by writing to the port drive mode registers. See the I/O System chapter on page 151 for more details. The following is a summary of the four NVL drive mode settings: ■ 00b – High impedance analog ■ 01b – High impedance digital ■ 10b – Resistive pull up ■ 11b – Resistive pull down PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 85 Nonvolatile Latch 8.2.2 XRESMEN GPIO pin (P1[2]) may be configured as an external reset (XRES) pin. The configuration of that pin is controlled with this NVL bit: 8.2.6 ECCEN ■ 0 – GPIO For devices that support an Error Correcting Code (ECC) in the flash, this NVL bit is used to set whether ECC is enabled. See the Flash Program Memory chapter on page 93 for more details. ■ 1 – XRES ■ 0 – ECC disabled ■ 1 – ECC enabled 8.2.3 DEBUG_EN The Debug Enable bit allows access to the on-chip debugger and allows programming, either in JTAG or SWD mode, without having to acquire the device in test mode. JTAG or SWD can be selected by the Debug Port Select (DPS) bits. When DEBUG_EN is not set, it is required to enter test mode to gain debugger access and enable device programming. ■ 0 – Debug Disabled (no debugger access except after test acquire) ■ 1 – Debug Enabled (debugger access with or without test acquire) 8.2.4 CFGSPEED The Configuration Speed NVL bit determines if the IMO defaults to a fast or slow speed. See the Clocking System chapter on page 109 for more details. This configuration is intended to balance the need for rapid boot and configuration against peak power consumption. ■ 0 – Slow (12 MHz IMO frequency) ■ 1 – Fast (48 MHz IMO frequency) 8.2.5 DPS[1:0] Debug Port Select NVL bits allow the user to select a debugging port interface that is active after POR is released. If the debug port’s disabled setting is used, the acquire functions of the test controller must be used to activate the debug port. See the Test Controller chapter on page 403 for more details. These NVL bits do not enable the debugger logic; they enable only the physical interface. The only way to enable the debug logic is for the user's firmware or configuration to write the debugger enable bit. ■ 00b – 5-wire JTAG ■ 01b – 4-wire JTAG ■ 10b – SWD (single wire debug) ■ 11b – Debug ports disabled For programming and debugging using third-party tools, the Debug Port Select should be configured for either the SWD or JTAG settings as applicable. Do not select the ‘Debug ports disabled’ setting while programming or debugging using the third-party tools. 86 8.2.7 DIG_PHS_DLY[3:0] This bit selects the digital clock phase delay in 1 ns increments. See the Clocking System chapter on page 109 for more details, ■ 0x00 – Clock disabled ■ 0x01 – 2.5 ns delay ■ 0x02 – 3.5 ns delay ■ … ■ 0X0A – 11.5 ns delay ■ 0x0B – 12.5 ns delay ■ 0x0C – Clock disabled ■ 0X0D – Clock disabled ■ 0X0E – Clock disabled ■ 0X0F – Clock disabled 8.3 Write Once NV Latch The Write Once (WO) latch is a type of nonvolatile latch. The cell itself is an NVL with additional logic wrapped around it. Each WO latch device contains 4 bytes (32 bits) of data. The wrapper outputs a 1 if a super-majority (28 of 32) of its bits match a pre-determined pattern (0x50536F43) and it outputs a 0 if this majority is not reached. When the output is 1, the Write Once NV latch locks the part out of Debug and Test modes; it also permanently gates off the ability to erase or alter the contents of the latch. Matching of all bits is intentionally not required, so that single (or few) bit failures do not deassert the WO latch output. The state of the NV latch bits after wafer processing is truly random with no tendency toward 1 or 0. The WOL only locks the part when the correct 32-bit key (0x50536F43) is loaded into the NVL's volatile memory, programmed into the NVL's nonvolatile cells, and the part is reset. The output of the WOL is only sampled on reset and used to disable the access. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Nonvolatile Latch This precaution prevents anyone from reading, erasing, or altering the content of the internal memory. If the device is protected with a WO latch setting, Cypress cannot perform failure analysis and, therefore, cannot accept RMAs from customers. The WO latch can be read via the SWD to electrically identify protected parts. 8.5 Sleep Mode Behavior NV latches remain powered up during sleep, but they stay in an idle state, not allowing any direct reads or writes. During sleep, the outputs of the NVLs remain stable. The user can write the key in WOL to lock out external access only if no flash protection is set. However, after setting the values in the WOL, a user still has access to the part until it is reset. Therefore a user can write the key into the WOL, program the flash protection data, and then reset the part to lock it. See the Flash, Configuration Protection chapter on page 167 for details on flash protection. 8.4 Programming NV Latch The volatile latch is intended to be initialized from a nonvolatile memory cell at POR release. NV Latches are configured by writing to the volatile cells of the array and then programming the volatile cell data into the nonvolatile cells (Write Nonvolatile Cell Mode). See the Nonvolatile Memory Programming chapter on page 423 for more details on NV latch programming sequence. NVL programming is done through a simple command/status register interface. Commands and data are sent as a series of bytes to either SPC_CPU_DATA or SPC_DMA_DATA, depending on the source of the command. Response data is read via the same register to which the command was sent. The following commands are used to program NVLs: ■ Command 0x00 – Load Byte Loads a single byte of data into the volatile cells at the given address. ■ Command 0x10 – Read Byte Reads a single byte of data from volatile cells at the given address. ■ Command 0x06 – Write User NVL Writes all nonvolatile cells in a User NVL with the corresponding values in its volatile latches. ■ Command 0x03 – Read User NVL Reads a single byte of data from nonvolatile cells at the given address. Note that when this command is executed, all of the bytes are transferred from nonvolatile cells to the volatile cells of the array. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 87 Nonvolatile Latch 88 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 9. SRAM PSoC® 5LP devices include on-chip SRAM. 9.1 Features PSoC 5LP SRAM has these features: ■ Organized as up to 16 blocks of 4 KB each, for CY8C55 family. ■ Code can be executed out of portions of SRAM, for CY8C55 family. ■ 8-, 16-, or 32-bit accesses. ■ Zero wait state accesses. ■ Arbitration of SRAM accesses by the CPU and the DMA controller. ■ Different blocks can be accessed simultaneously by the CPU and the DMA controller. 9.2 Block Diagram PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 89 SRAM Figure 9-1 shows CY8C55 family SRAM accesses. Figure 9-1. CY8C55 Family SRAM Accesses Cortex-M3 CPU 32 32 PHUB SRAM Peripheral Peripheral Figure 9-2 shows internal SRAM organization for the CY8C55 family. Figure 9-2. CY8C55 Family SRAM Organization SRAM Cortex-M3 CPU CPUIF PHUB PHUBIF SRAM BANK0 (32 KB) Lower SRAM SRAM BANK1 (32 KB) Upper SRAM 90 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F SRAM 9.3 How It Works The CY8C55 family has up to 64 KB SRAM implemented as sixteen 4 KB blocks. All 64 KB are accessible by the CortexM3 CPU and by the PHUB DMA controller in normal operation. The SRAM is further organized as two 32 KB memory banks, centered at address 0x20000000. This allows access to both SRAM banks with either the c-Bus (CortexM3 I and D buses) or the s-Bus (Cortex-M3 system bus). Code can be executed from all SRAM below address 0x20000000. The PHUB can use SRAM as a DMA source or target. All data paths to SRAM are 32 bits wide. The CPU has a direct connection to SRAM without going through the PHUB. In addition to faster SRAM access by the CPU, this allows for simultaneous accesses to SRAM by both the CPU and the PHUB DMA controller, because SRAM is physically implemented as multiple separate blocks. If the CPU and the PHUB are accessing separate blocks, they both have simultaneous unimpeded access. In case of contention, the following applies: ■ CY8C55 family – In most cases, the Cortex-M3 CPU has priority over the PHUB for all SRAM. The SRAM responds to CPU and PHUB accesses with zero wait states for both reads and writes as long as the access does not lose priority arbitration. Arbitration is done on a cycle-by-cycle basis at the time of SRAM access. The losing master is held off until the winning master has finished accessing the SRAM block; the losing master gains access on the cycle immediately after. SRAM data is maintained during all low-power and sleep modes. At reset, the SRAM contents are not initialized; they power up as unknown values. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 91 SRAM 92 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 10. Flash Program Memory PSoC® 5LP include on-chip flash memory. Additional flash is available for either error correction bytes or data storage. 10.1 Features PSoC 5LP flash memory has the following features: ■ Organized in rows, where each row contains 256 data bytes plus 32 bytes for either error correcting codes (ECC) or data storage. ■ For PSoC 5LP architecture: CY8C55 Family, organized as either one block of 128 or 256 rows, or as multiple blocks of 256 rows each. ■ Stores CPU program and bulk or nonvolatile data ■ For PSoC 5LP architecture: CY8C55 Family, 8-, 16-, or 32-bit read accesses. ■ Programmable with a simple command / status register interface (see Nonvolatile Memory Programming chapter on page 423). ■ Four levels of protection (see Nonvolatile Memory Programming chapter on page 423 and Flash, Configuration Protection chapter on page 167). 10.2 Block Diagram Figure 10-1 is a block diagram of the flash programming system. Figure 10-1. Flash Block Diagram Test Controller (TC) Debug on-Chip (DOC) CPU PHUB EEPROM PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Flash Programming Interface NVL 93 Flash Program Memory 10.3 How It Works Flash memory provides nonvolatile storage for firmware, device configuration data, bulk data storage, ECC data, factory configuration data, and protection information. Flash memory contains two regions – a main region, and a much smaller, extended region. All user data is stored in the main region, including ECC data. Factory configuration and user-defined protection data are stored in the extended region, also known as the hidden rows of flash. For each row, protection bits control whether the flash can be read or written by external debug devices and whether it can be reprogrammed by a boot loader. For more information see the Nonvolatile Memory Programming chapter on page 423 and Flash, Configuration Protection chapter on page 167. Flash can be read by both the CPU and the DMA controller. Flash is erased in 64-row sectors or in its entirety, and it is programmed in rows. Erase and programming operations are done by a programming system, using a simple command/status register interface. For more information see the Nonvolatile Memory Programming chapter on page 423. Note It can take as much as 20 milliseconds to write to EEPROM or flash. During this time the device should not be reset; otherwise, unexpected changes may be made to portions of EEPROM or flash. The reset sources (see Reset Sources on page 143) include XRES pin, software reset, and watchdog; make sure that these are not inadvertently activated. Also, configure the low-voltage detect circuits to generate an interrupt instead of a reset. Note When writing FLASH on PSoC 5LP devices, it is possible for data in the instruction cache to become stale, thus the cache data does not correlate to the data just written to FLASH. A call to CyFlushCache() is required to invalidate the data in cache and force fresh information to be loaded from FLASH. 10.4 Flash Memory Access Arbitration Flash memory can be accessed either by the cache controller or the nonvolatile memory programming interface (system performance controller (SPC)). Cache controller can perform only flash read operations while the SPC can perform both read and write operations on the flash memory. There is an internal arbitration mechanism to facilitate flash memory access by both the cache and the SPC. Flash memory is organized as flash arrays. PSoC 5LP can have up to four flash arrays, where each flash array size can be 94 up to 64 KB. The SPC and the cache controller cannot simultaneously access the flash memory locations. If the cache controller tries to access the flash at the same time as the SPC, then it must wait until the SPC completes its flash access operation. The CPU, which accesses the flash memory through the cache controller, is also halted until the cache is filled with the code to be executed from the flash memory. Similarly, if SPC tries to access the flash array at the same time as the cache controller, then it must wait until the cache controller completes its access operation. 10.5 ECC Error Detection and Interrupts The ECC detects conditions that may interfere with software operation. The information is logged into individual interrupt registers that become latched until the software clears the corresponding valid bit. All interrupt sources within the ECC are passed through a mask condition; then, they are reduced into a single interrupt request to the Interrupt Controller unit. When the software is notified about an existing interrupt in the ECC, the following sequence occurs: 1. The software reads the Interrupt Status register CACHE_INT_SR that provides the valid bits of all interrupts in a single read operation. 2. The software examines individual interrupt registers for more log information (CACHE_INT_LOG[0..5]). 3. Stored log information is cleared on read of registers. 4. After clearing of log information, the status register (CACHE_INT_SR) is automatically cleared, because it is a collection of valid bits of the log registers. Logging is always enabled; reporting may be disabled through the Interrupt Mask Register (CACHE_INT_MSK). The following conditions are detected by the hardware and logged as potential interrupt sources: ■ ECC – Single Bit – A single bit error was encountered during a fill operation and was fixed. ■ ECC – Multi Bit – A multi-bit error was encountered during a fill operation, but it cannot be corrected. ■ Attempted Flash Write – If a write to flash through the PHUB is attempted. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 11. EEPROM PSoC® 5LP devices have on-chip EEPROM memory. This family offers devices that range from 512 bytes to 2 kilobytes. 11.1 Features PSoC 5LP EEPROM memory has the following features: ■ Organized in rows, where each row contains 16 bytes ■ Organized as one block of 32, 64, or 128 rows, depending on the device ■ Stores nonvolatile data ■ Write and erase using SPC commands ■ Byte read access by CPU or DMA using the PHUB ■ Programmable with a simple command/status register interface (see Nonvolatile Memory Programming chapter on page 423) 11.2 Block Diagram There is no block diagram associated with EEPROM. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 95 EEPROM 11.3 How It Works EEPROM memory provides nonvolatile storage for user data. EEPROM write and erase operation is done using SPC commands. It may be read by both the CPU and the DMA controller, using the PHUB. All read accesses are 8-bit. If a PHUB access is attempted while the SPC is in control of EEPROM, a System Fault Interrupt is generated to the interrupt controller and the bit EEPROM_error is set in SPC_EE_ERR[0]. When set, this bit remains set until it is read from the PHUB. EEPROM can be taken in and out of sleep mode by setting the bit EE_SLEEP_REQ in SPC_FM_EE_CR[4], as shown in Table 11-1. Before a PHUB access of EEPROM is done, set the firmware EEPROM request bit AHB_EE_REQ in SPC_EE_SCR[0], then poll for the EEPROM acknowledge bit EE_AHB_ACK in SPC_EE_SCR[1] to be set. Before a PHUB access of EEPROM is done, firmware should set the EEPROM request bit AHB_EE_REQ in SPC_EE_SCR[0], then poll for the EEPROM acknowledge bit EE_AHB_ACK in SPC_EE_SCR[1] to be set. It is also possible to check the current sleep status of the EEPROM by reading the bit EE_AWAKE in SPC_FM_EE_CR[5], as shown in Table 11-2. Table 11-1. Bit Settings for EE_SLEEP_REQ in SPC_FM_EE_CR[4] Setting Description 0 (default) Wake up EEPROM 1 Put EEPROM to sleep Table 11-2. Bit Settings for EE_AWAKE in SPC_FM_EE_CR[5] Setting Description 0 EEPROM is asleep 1 (default) EEPROM is awake EEPROM is erased in 64-row sectors, or in its entirety, and is programmed in rows. Erase, programming and read operations are done by a programming system using a simple command/status register interface. For more information see Nonvolatile Memory Programming chapter on page 423. Contention priority between the cache controller and the programming system can be controlled by the bit EE_Priority, in SPC.FM_EE_CR[1], as shown in Table 11-3. Table 11-3. Bit Settings for EE_Priority, in SPC.FM_EE_CR[1] Setting Description 0 (default) PHUB has priority 1 Programming system has priority Note It can take as much as 20 milliseconds to write to EEPROM or flash. During this time the device should not be reset; otherwise, unexpected changes may be made to portions of EEPROM or flash. The reset sources (see Reset Sources on page 143) include XRES pin, software reset, and watchdog; make sure that these are not inadvertently activated. Also, configure the low-voltage detect circuits to generate an interrupt instead of a reset. 96 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 12. EMIF PSoC® 5LP architecture provide an external memory interface (EMIF) for connecting to external memory devices and peripheral devices. The connection allows read and write access to the devices. The EMIF operates in conjunction with UDBs, I/O ports, and other PSoC 5LP components to generate the necessary address, data, and control signals. The EMIF does not intercept address data between the PHUB and the I/O ports. It only generates the required control signals to latch the address and data at the ports. The EMIF generates a clock to run external synchronous and asynchronous memories. It can generate four different clock frequencies, which are the bus clock divided by 1, 2, 3, or 4. 12.1 Features The EMIF supports four types of external memory: synchronous SRAM, asynchronous SRAM, cellular RAM/PSRAM, and NOR flash. External memory can be accessed via the ARM Cortex-M3 external RAM space; up to 24 address bits can be used. The memory can be 8 or 16 bits wide. 12.2 Block Diagram Figure 12-1 is the EMIF block diagram. PSoC 5LP CPU Address Ports Address PHUB 24 Port Logic DMAC AHB Bus Data Ports Read / Write Data 16 DSI 6 AD DQ CLK WRn CEn ADSCn OEn ZZ_ 16 Control Port no_udb mode 6 udb mode EMIF EM_Clock EM_WRn EM_CEn EM_ADSCn EM_OEn EM_Sleep Xmem_wr Xmem_rd Udb_Ready 24 Addr[23:0] Data[15:0] External Memory Figure 12-1. EMIF Block Diagram UDB 3 Custom UDB Logic PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 97 EMIF 12.3 How It Works The address component of the EMIF uses up to three I/O ports. The I/O ports used for external memory address are selected by configuring the 3-bit portEmifCfg field in the PRT*_CTL register. The register can be configured so that the port is selected as either the most significant byte, the middle byte, or the least significant byte of the address. (See the I/O System chapter on page 151 for details of the PRT*_CTL register.) The data component of the EMIF uses one or two I/O ports. The I/O port or ports used for external memory data are selected by configuring the 3-bit portEmifCfg field in the PRT*_CTL register. The register can be configured so that the port is selected as either the most significant byte or the least significant byte of the data. (See the I/O System chapter on page 151 for details of the PRT*_CTL register.) 12.3.2 External Memory Support Table 12-2 on page 101 shows how different external memory types can be connected to the PSoC 5LP devices. Address lines use up to three I/O ports. Data lines use one or two ports, depending on whether the external memory is x8 or x16. Control lines use 3 to 6 pins on one I/O port. Spare pins on the address and data ports are not available for any other purpose. Spare pins on the control port are available for other purposes. The control component of the EMIF uses a single I/O port. The I/O port used for external memory control is selected by configuring the 3-bit portEmifCfg field in the PRT*_CTL register. The I/O port must be further configured by setting the byPass bit in the PRT*_BYP register. This allows the EMIF to drive the pins. The control signals are sent from the EMIF to the I/O port over the digital signal interface (DSI). 12.3.1 List of EMIF Registers This table lists EMIF registers. Table 12-1. EMIF Registers Register Usage EMIF_NO_UDB Controls whether a synchronous or asynchronous RAM is supported, versus a custom memory interface requiring additional UDB logic. EMIF_RP_WAIT_STATES Number of additional wait states used in a read operation. EMIF_MEM_DWN Puts the external memory into a power down state. EMIF_MEMCLK_DIV Sets the clock divider for the external memory clock frequency, which can equal the bus clock frequency divided by 1, 2, 3 or 4. Note that the external memory clock frequency cannot exceed 33 MHz. EMIF_CLOCK_EN Enables/disables the clock for the EMIF block, effectively turning the block on or off. EMIF_EM_TYPE Controls whether to generate control signals for a synchronous or asynchronous SRAM in NO_UDB mode. EMIF_WP_WAIT_STATES Number of additional wait states used in a write operation. 98 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F EMIF Figure 12-2. Synchronous SRAM [7:0] ADDR LO Port A0–A7 [0] A16 ADDR Hi Port [7:1] Unused [7:0] Data LO Port PSoC 5LP D0–D7 [7:0] Data Hi Port D8–D15 CE - CE1 OE Control Port OE WE - GW ADSC - ADSC EM-Clock CLK EM-Sleep 2 Synchronous A8–A15 SRAM (such as CY7C1342H) [7:0] ADDR MID Port ZZ Vddd Spare ADSA ADV CE2 BWE CE3 BWA BWB Mode PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 99 EMIF Figure 12-3. Asynchronous SRAM [7:0] ADDR LO Port A0 – A7 [7:0] ADDR MID Port A8 – A15 [1:0] A16, A17 [7:2] [7:0] Data LO Port [7:0] Data Hi Port Control Port D8 – D15 CE - CE OE - OE WE - WE ADSC - Unused EM-Clock Unused EM-Sleep 2 100 SRAM or Flash (such as CY7C1041D) PSoC D0 – D7 Asynchronous ADDR Hi Port Unused Spare PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F EMIF Table 12-2. External Memory Connections to PSoC 5LP Devices Synchronous SRAM Ex: CY7C1342H Asynchronous SRAM Ex: CY7C1041D Pseudo SRAM Ex: CYK256K16MCCB NOR Flash Ex: Intel 28F800C3 3 I/O PORTs A0 - A16 A0 - A17 A0 - A17 A0 - A18 2 I/O PORTs D0 - D15 PSoC 5LP Connection D0 - D15 D0 - D15 D0 - D15 1 I/O PORT pin: EM_CE CE1 CE CE CE 1 I/O PORT pin: EM_OE OE OE OE OE 1 I/O PORT pin: EM_WE GW WE WE WE 1 I/O PORT pin: EM_ADSC ADSC 1 I/O PORT pin: EM_CLOCK CLK 1 I/O PORT pin: EM_SLEEP ZZ RPa tie high ADSP WP tie high ADV tie high CE2 tie high BWE tie low CE3 tie low BWA BHE BHE tie low BWB BLE BLE tie low MODE a. RP is opposite polarity from the ZZ signal on the synchronous SRAM. Either add an inverter to the EM_SLEEP signal or program the EMIF_MEM_DOWN register with the opposite polarity. 12.3.3 Sleep Mode Behavior All EMIF registers keep their value during sleep mode. The MEM_DWN register controls external memory sleep mode; the external control signal ZZ is asserted or deasserted. If an external memory access happens when MEM_DWN is set, ZZ is not asserted until after the current transfer is completed. ZZ is deasserted when the MEM_DWN register is cleared; it then takes two external memory clock cycles for the memory to wake up. To completely turn off the EMIF block, clear the CLOCK_EN register. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 101 EMIF 12.4 EMIF Timing The EMIF is clocked by bus clock – the same signal that clocks the CPU and the PHUB. Within the EMIF block, the bus clock can be divided by 1, 2, 3, or 4; the output is the EM_CLOCK signal to the external memory IC. The following table shows the number of PHUB wait states generated by the EMIF depending on how much the input clock is divided. Table 12-3. PHUB Wait States Generated by EMIF EM_CLOCK = Read Wait States Write Wait States 1 1 2 2 3 4 3 5 6 4 7 8 Bus Clock Divided By The EMIF.WAIT_STATES register can also be used to add up to seven more wait states. An important limitation is that the maximum I/O rate of PSoC 5LP GPIO pins is 33 MHz. This makes the maximum frequency of EM_CLOCK 33 MHz. The following table shows limitations of EM_CLOCK frequency relative to the bus clock: Table 12-4. Limitations of EM_CLOCK Relative to Bus Clock Bus Clock Frequency EM_CLOCK = Bus Clock Divided By < 33 MHz 1, 2, 3, or 4 33 - 66 MHz 2, 3, or 4 > 66 MHz 3 or 4 The maximum frequency of the bus clock is 80 MHz for PSoC 5LP devices. In most cases, EMIF_MEMCLK_DIV must be used to divide EM_CLOCK to a frequency less than or equal to 33 MHz. Given the above restriction on EM_CLOCK frequency, and the relation of EM_CLOCK to EM_ADSC-, EM_CE-, and EM_WE-, it can be seen that the minimum pulse widths of these signals is 30.3 ns. Figure 12-4. Synchronous Write Cycle Timing EM_Clock EM_CEn Address EM_Addr EM_OEn EM_Data Data EM_ADSCn 102 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F EMIF Figure 12-5. Synchronous Read Cycle Timing EM_Clock EM_CEn EM_Addr Address EM_OEn EM_Data Data EM_ADSCn Figure 12-6. Asynchronous Write Cycle Timing EM_Addr Address EM_CEn EM_OEn EM_WEn EM_Data Data Figure 12-7. Asynchronous Read Cycle Timing EM_CEn EM_Addr Address EM_OEn EM_WEn EM_Data PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Data 103 EMIF 12.5 Using EMIF with MemoryMapped Peripherals The EMIF can also be used with external peripheral devices that have a bus interface similar to asynchronous memory devices, that is, they address, data, CE-, WE-, and OE-. The speed of the interface must be considered in the same manner as described above. The maximum data bus size is 16 bits, and the minimum address bus size is 8 bits. If multiple external memory and peripheral devices are used, address decoding to the multiple device selects may become complex and must be given careful consideration. 12.6 unaligned 16-bit or 32-bit transfers to an external memory, as the processor may convert these into multiple 8 bit aligned accesses. However, 32 or 16-bit aligned transfers are handled correctly by the processor and PHUB. 12.6.4 8-bit Memory Transfers DMA Transfers: For DMA transfers to/from an 8 bit external memory, the burst count should always be 1, irrespective of the transfer count. For example, if the burst count is set as 2 to transfer two bytes to external memory, the PHUB will try to do a 16-bit transfer in a single burst instead of breaking the transfer down into two individual transfers with the 8-bit memory. Additional Configuration Guidelines The PHUB assumes all peripherals including external memory are byte addressable. Port logic is natively 16 bits wide, so care must be taken when setting up communication with either an 8 or 16 bit external memory. The following section describes some guidelines to configure the port pins and set up the memory access methods (either CPU or DMA) for optimal performance. 12.6.1 Address Bus Configuration Configure three of the available ports as output EMIF address ports. Because PHUB peripherals are byte addressable regardless of the external memory data bus size, up to 2^24 bytes of external memory can be accessed. If an 8-bit memory is used, up to 24-bit address lines can be directly connected to the memory. If a 16-bit memory is used, the LSB address line (A0) of the memory chip should be connected to the second address line (A1) of the PSoC and the LSB address line (A0) of the PSoC should be ignored. This is because the PHUB increments the address by 2 while doing 16-bit transactions. 12.6.2 Data Bus Configuration For 16 bit memories, two ports should be configured as bidirectional EMIF data ports. For 8bit memories, only one port should be configured as a bidirectional EMIF data port. 12.6.3 16-bit Memory Transfers DMA Transfers: For DMA transfers to/from 16bit external memory, odd burst counts are not supported because 8 bit transfers are not supported on a 16bit interface. CPU Transfers: With the 32 bit ARM M3 processor in PSoC 5LP, 16-bit memory can be directly accessed by the CPU. The only limitation here is the PSoC 5LP cannot initiate 8 bit transfers to 16-bit memories and should not initiate 104 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 13. Memory Map All PSoC® 5LP memory (flash, EEPROM, Nonvolatile Latch, and SRAM) and all registers are accessible by the CPU, DMA controller, and in most cases by the debug systems. This chapter contains an overall map of the addresses of the memories and registers. 13.1 Features The PSoC 5LP memory map has the following features: ■ ARM Cortex-M3 32-bit linear address space, with regions for code, SRAM, peripherals, external RAM, and CPU internal registers. ■ Flash is mapped to the Cortex-M3 code region. ■ Half of SRAM is mapped to the code region, the other half to the SRAM bitband region. ■ SRAM mapped to the code region is also accessible by DMA in the SRAM bitband region. ■ External memory (see the EMIF chapter on page 97) is mapped to the external RAM region. ■ All other memories, and all registers, are accessed in the Cortex-M3 peripheral bitband region. 13.2 Block Diagram There is no block diagram associated with the memory map. 13.3 How It Works The PSoC 5LP memory maps are detailed in the following sections. For additional information see the PSoC® 5LP Registers TRM (Technical Reference Manual). PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 105 Memory Map 13.3.1 PSoC 5LP Memory Map The ARM Cortex-M3 has a fixed address map allowing access to peripherals using simple memory access instructions. The 32-bit (4 GB) address space is divided into the regions shown in Table 13-1. Note that code can be executed from the code, SRAM, and external RAM regions. Table 13-1. PSoC 5LP Memory Map Address Range 0x00000000 – 0x1FFFFFFF Size Use 0.5 GB Program code. Includes the exception vector table at power up, which starts at address 0 0x20000000 – 0x3FFFFFFF 0.5 GB SRAM. This includes a 1 MByte bit-band region starting at 0x20000000, and a 32 Mbyte bit-band alias region starting at 0x22000000. 0x40000000 – 0x5FFFFFFF 0.5 GB Peripherals. This includes a 1 MByte bit-band region starting at 0x40000000, and a 32 Mbyte bit-band alias region starting at 0x42000000. 0x60000000 – 0x9FFFFFFF 1 GB External RAM 0xA0000000 – 0xDFFFFFFF 1 GB External peripherals 0xE0000000 – 0xFFFFFFFF 0.5 GB Internal peripherals, including the NVIC and debug and trace modules The PSoC 5LP address map is shown in Table 13-2. For more information see the Cortex-M3 chapter. Table 13-2. PSoC 5LP Address Map Address Range 0x0000 0000 – 0x0003 FFFF Purpose Up to 256 KB Flash 0x1FFF 8000 – 0x1FFF FFFF Up to 32 KB SRAM in code region 0x2000 0000 – 0x2000 7FFF Up to 32 KB SRAM in SRAM region 0x2000 8000 – 0x2000 FFFF Alias of address range 0x1FFF 8000 – 0x1FFF FFFF, accessible by DMA 0x4000 4000 – 0x4000 42FF Clocking, PLLs, and oscillators 0x4000 4300 – 0x4000 43FF Power management 0x4000 4500 – 0x4000 45FF Ports interrupt control 0x4000 4700 – 0x4000 47FF Flash programming interface 0x4000 4900 – 0x4000 49FF I2C controller 0x4000 4E00 – 0x4000 4EFF Decimator 0x4000 4F00 – 0x4000 4FFF Fixed timer/counter/PWMs 0x4000 5000 – 0x4000 51FF General purpose I/Os 0x4000 5300 – 0x4000 530F Output port select register 0x4000 5400 – 0x4000 54FF External memory interface control registers 0x4000 5800 – 0x4000 5FFF Analog subsystem interface 0x4000 6000 – 0x4000 60FF USB controller 0x4000 6400 – 0x4000 6FFF UDB configuration 0x4000 7000 – 0x4000 7FFF PHUB configuration 0x4000 8000 – 0x4000 87FF EEPROM 0x4000 A000 – 0x4000 A400 CAN 0x4000 C000 – 0x4000 C800 Digital filter block 0x4001 0000 – 0x4001 FFFF Digital interconnect configuration 0x4800 0000 – 0x4800 7FFF Flash ECC bytes 0x6000 0000 – 0x60FF FFFF External Memory Interface (EMIF) 0xE000 0000 – 0xE00F FFFF Cortex-M3 PPB registers, including NVIC, debug, and trace 106 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Section D: System Wide Resources The System Wide Resources section details three types of I/O, internal clock generators, power supply, boost converter, and sleep modes. This section contains these chapters: ■ Clocking System chapter on page 109 ■ Power Supply and Monitoring chapter on page 125 ■ Low-Power Modes chapter on page 135 ■ Watchdog Timer chapter on page 141 ■ Reset chapter on page 143 ■ Auxiliary ADC chapter on page 179 ■ I/O System chapter on page 151 ■ Flash, Configuration Protection chapter on page 167 Top Level Architecture System Wide Resources Block Diagram SYSTEM WIDE RESOURCES WDT and Wake RTC Timer ILO System Bus IMO Clock Tree Xtal Osc Clocking System POR and LVD Sleep Power 1.8V LDO SMP Power Management System PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 107 Section D: System Wide Resources 108 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 14. Clocking System The clocking system provides clocks for the entire device. It allows the user to trade off current, frequency, and accuracy. A wide range of frequencies can be generated, using multiple sources of clock inputs combined with the ability to set divide values. 14.1 Features The clock system includes these clock resources: ■ Four internal clock sources increase system integration: ❐ 3 to 74.7 MHz internal main oscillator (IMO) ±1% at 3 MHz ❐ 1 kHz, 33 kHz, 100 kHz internal low-speed oscillator (ILO) outputs ❐ 48 MHz clock doubler output for USB, sourced from IMO, MHz External Crystal Oscillator (MHzECO), and Digital System Interconnect (DSI) ❐ 24 to 80 MHz fractional Phase-Locked Loop (PLL) sourced from IMO, MHzECO, and DSI ■ Clock generated using a DSI signal from an external I/O pin or other logic ■ Two external clock sources provide high precision clocks: ❐ 4 to 25 MHz External Crystal Oscillator (MHzECO) ❐ 32.768 kHz External Crystal Oscillator (kHzECO) for real-time clock (RTC) ■ Dedicated 16-bit divider for bus clock ■ Eight individually sourced 16-bit clock dividers for the digital system peripherals ■ Four individually sourced 16-bit clock dividers with skew for the analog system peripherals ■ IMO has a USB mode that synchronizes to USB host traffic, requiring no external crystal for USB. (USB equipped parts only) PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 109 Clocking System 14.2 Block Diagram Figure 14-1 gives a generic view of the Clocking System in PSoC 5LP devices. Figure 14-1. Clocking System Block 4 MHz - 25 MHz ECO External I/O or DSI 0 MHz - 33 MHz 8-bit Clock Divider clk_pll Digital (User) Clock Mux and 16-Bit Divider ... clk_sync_d 1 clk_imo 2 clk_xtal 3 clk_ilo 4 5 clk_pll clk_32k 6 clk_dsi_glb The components of the clocking system block diagram are defined as follows: ■ Internal main oscillator (IMO) ■ Internal Low-speed Oscillator (ILO) ■ A 4 to 25 MHz External Crystal Oscillator (MHzECO) ■ A 32 kHz External Crystal Oscillator (kHzECO) ■ Digital System Interconnect (DSI) signal, which can be derived from the clocks developed in UDBs or off-chip clocks routed through pins ■ A PLL to boost the clock frequency of some select internal and external sources ■ Five types of clock outputs: ❐ Digital clocks ❐ Analog clocks ❐ Special purpose clocks ❐ System clock ❐ USB clock Clock Sources Clock sources for the device are classified as internal oscillators and external crystal oscillators. There is an option of using a PLL or a frequency doubler to derive higher fre- 110 7 dsi_a[n] clk_sync_a[n] Analog (User) Clock Mux and 16-Bit Divider s k e w ... x8 Bus Clock Divider 16-Bit clk_sync_d 0 USB Clk Mux + Div2 clk_ilo Master Clock Mux 24 - 80 MHz PLL 7 14.3 1/33/100 kHz ILO clk_imo2x clk_imo clk_pll clk_dsi_glb clk_32k clk_xtal clk_dsi_glb clk_imo clk_imo2x 48 MHz Doubler dsi_d[n] 32.768 kHz ECO dsi_clkin 3 MHz – 74.7 MHz IMO x4 quency outputs from existing clocks. Signals can be routed from the DSI and used as clocks in the clock trees. 14.3.1 Internal Oscillators PSoC devices have two internal oscillators: the internal main oscillator (IMO) and the internal low-speed oscillator (ILO). 14.3.1.1 Internal Main Oscillator The IMO operates with no external components and outputs a stable clock, clk_imo, at a variety of user-selectable frequencies: 3, 6, 12, 24, 48, 62.6, and 74.7 MHz. Frequencies are selected using the register FASTCLK_IMO_CR[2:0]. The clock accuracy is 1% typical at 3 MHz and it varies with frequency. See the device datasheet for IMO accuracy specification. Clock Doubler The block has one additional clock output. A doubled clock, IMOCLKX2 outputs a clock at twice the frequency of the input clock. The doubler is enabled by register bit FASTCLK_IMO_CR[4]. The doubler can also take clock inputs (XCLK) other than IMO and have a DSI or MHzECO as input. This feature is enabled by the bit FASTCLK_IMO_CR[5]. The DSI / MHzECO can be selected in the CLKDIST_CR[6] register bit. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Clocking System The clock distribution register CLKDIST_CR[5:4] is responsible for selecting between IMO or IMO × 2 outputs. Figure 14-2 is a summary block diagram of the IMO. Note The output of the clock doubler should only be used for clocking the USB block. It should not be used to clock any other peripherals in the device. Fast-Start IMO (FIMO) An alternate mode of the IMO is available for fast start-up out of sleep modes. This fast-start IMO (FIMO) mode provides a clock output within 1 µs after exiting the power down state. The fast-start IMO uses a special fast bias circuit that is stable more quickly than the high accuracy bias that is used during normal operation. This fast bias is less accurate than the normal bias, resulting in a less accurate clock frequency. The normal, high-accuracy bias is always used when running user code. During the transition from FIMO to regular IMO, glitches can occur if the frequency selection for the two configurations are not the same. Stated explicitly, at the transition, FASTCLK_IMO_CR[2:0] should match PWRSYS_WAKE_TR1[2:0]. NVL Frequency Selection Upon entering the boot phase of startup, the IMO frequency and a portion of its trim are set using values stored in user NVLs. This allows the user to select a faster clock frequency for a portion of device startup. The top two bits of IMO trim stored in the IMO_TR2 register are populated from the NVL register MNVL_FIMO_TRIM[1:0]. The frequency selection bits in register FASTCLK_IMO_CR[2:0] have their most significant bit populated using NVL register CNVL_CFGSPEED. The NVL register will set the frequency to 12 MHz when set to 0, and 48 MHz when set to 1. This NVL selection will be overwritten during firmware startup with a more complete frequency selection and trim. Note 48-MHz startup should not be selected in devices with a maximum operating frequency rating below 48 MHz. Figure 14-2. IMO Block Diagram FASTCLK_IMO_CR[2:0] dsi_clkin clk_eco_Mhz FASTCLK_IMO_CR[5] Osc (3/6/12/ 24/48/ 62.6/74.7 MHz) IMO SRC MUX IMOCLK XCLK CLK MUX Doubler IMOCLK X 2 FASTCLK_IMO_CR[4] IMO OUT MUX clk_imo CLKDIST_CR[5:4] CLKDIST_CR[6] 14.3.1.2 Internal Low-Speed Oscillator The ILO produces two primary independent output clocks with no external components and with very low power consumption. These two outputs operate at nominal frequencies of 1 kHz and 100 kHz. The two clocks run independently, are not synchronized to each other, and can be enabled or disabled together or independently. The 1 kHz clock is typically used for a background central timewheel and also for the watchdog timer. The 100 kHz clock can provide a low-power system clock, or it can be used to time intervals such as for sleep mode entry and exit. A third 33kHz clock output is available — a divide-by-3 of the 100 kHz output. In addition to the multiplexed output that can enter the clock distribution, the output clocks route to the following functions: PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F ■ clk_ilo1K – to the central timewheel (also called the sleep timer) and watchdog timer. See the Low-Power Modes chapter on page 135 for more details. ■ clk_ilo100K – to the fast timewheel. This oscillator operates at very low current and is, therefore, the best fit for use in low-power modes. The two sources, 1 kHz and 100 kHz, can be enabled and disabled, using the SLOWCLK_ILO_CR0[1] and SLOWCLK_ILO_CR0[2], respectively. SLOWCLK_ILO_CR0[5] enables the divide by 3 to create the 33 kHz output. The out puts from the ILO can be routed to the clock distribution network. CLKDIST_CR[3:2] is responsible for this selection. Figure 14-3 is a summary block diagram of the ILO. There are dedicated routes for some of the clock outputs that are not shown in the figure. 111 Clocking System Figure 14-3. ILO Block Diagram SLOWCLK_ILO_CR0[1] clk_ilo1K 1 kHz Osc Divide by 3 BIAS SLOWCLK_ILO_CR0[2] SLOWCLK_ILO_CR0[5] 100 kHz Osc External Oscillators PSoC devices have two external crystal oscillators: the MHz Crystal Oscillator (MHzECO) and the 32.768 kHz Crystal Oscillator (kHzECO). 14.3.2.1 clk_ilo100K CLKDIST_CR[3:2] The ILO clocks are all disabled in the Hibernate mode. SLOWCLK_ILO_CR0[4] is the power down mode bit governing the wakeup speeds of the device. Setting the bit slows down the startup, but it provides a low-power operation. 14.3.2 ILO Out Mux clk_ilo33K MHz Crystal Oscillator The 4-25 MHz external crystal oscillator MHzECO circuit provides for precision clock signals. The block supports a variety of fundamental mode parallel resonance crystals. When used in conjunction with the on-chip PLL, a wide range of precision clock frequencies can be synthesized, up to 80 MHz. The crystal pins are shared with a standard I/O function (GPIO / LCD / Analog Global), which must be tristated to operate the crystal oscillator with an attached external crystal. The crystal output routes to the clock distribution network as a clock source option, and it can also route through the IMO doubler to produce doubled frequencies, if the crystal frequency is in the valid range for the doubler. The oscillator allows for a wide range of crystal types and frequencies. Startup times vary with frequency and crystal quality. The xcfg bits of the FASTCLK_XMHZ_CFG0[4:0] register are used to match the oscillator settings to the crystal. The oscillator can be enabled by FASTCLK_XMHZ_CSR[0]. Figure 14-4 is a block diagram of the MHzECO. Figure 14-4. MHzECO Block Diagram External Components Xop 4-25 MHz Crystal Oscillator Load 4-25 MHz Capacitors Crystal clk_eco_MHz Xip Figure 14-5. MHzECO Oscillator Fault Recovery FASTCLK_XMHZ_CSR[6] FASTCLK_XMHZ_CSR[7] Xerr and Xprot Xosc Out clk_eco_MHz MHz XOSC clk_imo 112 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Clocking System Fault Recovery The block contains an option to detect crystal oscillator failure. Clock failure is detected by comparing the amplitude of the XIn signal to a user-selectable voltage. This voltage is selected using the "vref_sel_wd" bits in the register FASTCLK_XMHZ_CFG1. The clock failure can occur due to environmental conditions (such as moisture) that affect the crystal and cause oscillators to stop. Clock failure status is indicated by the clock error status bit (FASTCLK_XMHZ_CSR[7]). If the FASTCLK_XMHZ_CSR[6] bit is set, the fault recovery option is enabled. In this case, when the crystal oscillator fails, the crystal oscillator output is driven low. The IMO is enabled (if it is not already running), and the IMO output routes through the crystal oscillator output mux. It takes six IMO cycles after error signal assertion for the IMO to appear outside of the block. In this way, the system can continue to operate through a crystal fault. This functionality is illustrated in Figure 14-5. Low-power Operation The MHz crystal oscillator does not operate in the SLEEP and HIBERNATE modes. This means that you need to disable the oscillator to enter SLEEP and HIBERNATE modes. The 32 kHz crystal oscillator can be kept active, for precise timing (RTC), in SLEEP mode. If the MHz crystal oscillator is not disabled when the device is put into any of these modes, the mode entry is skipped, and the code continues to execute in active mode. Because this clock must be disabled to enter SLEEP mode, a typical approach is to switch clock trees to the IMO source and then disable the crystal oscillator (and the PLL also, if it is on). Then SLEEP or HIBERNATE mode can be entered. After waking up from a sleep mode, the crystal oscillator can be reenabled and used as a clock source when stable. 14.3.2.2 32.768 kHz Crystal Oscillator The 32.768 kHz external crystal oscillator kHzECO circuit produces a precision timing signal at very low power. The circuit uses an inexpensive external 32.768 kHz crystal and associated load capacitors that can be used to produce a real time clock. Current consumption can be much less than 1 µA. Figure 14-6. kHzECO Block Diagram External Components Xo 32 kHz Crystal Oscillator Load 32 kHz Capacitors Crystal clk_eco_kHz Xi Low-power Operation The oscillator operates at two power levels, depending on the state of the LPM bit (SLOWCLK_X32_CR[1]) and the device sleep mode status. In Active mode, by default, the oscillator is configured for high-power mode, which consumes 1-3 µA and minimizes sensitivity to noise. If the LPM mode is set for a low-power mode, the oscillator goes into low power only when the device goes to SLEEP/HIBERNATE. If LP_ALLOW (SLOWCLK_X32_CFG[7]) is set, the oscillator enters low-power mode immediately when the LPM bit is set. When enabled, the oscillator does not stabilize instantly, and requires some time to oscillate consistently. The ANA_STAT (SLOWCLK_X32_CR[5]) bit indicates whether oscillation is stable after measuring the waveform’s amplitude. The oscillator must always be started in high power mode to avoid excessively long startup delays. Real Time Clock One of the major uses of the kHzECO oscillator is for RTC implementation. The block level illustration of the RTC implementation is shown in Figure 14-7. The RTC timing is derived from the 32 kHz external crystal oscillator, as shown in Figure 14-7. Therefore, for the functioning of the RTC, the 32 kHz external crystal must be enabled through the register SLOWCLK_X32_CR[0]. The generated 32 kHz is divided to achieve a one pulse per second. The register PM_TW_CFG2[4] enables one pulse per second functionality. By enabling the bit PM_TW_CFG2[5], the RTC generates an interrupt every second. The interrupt is routed through the DSI and is brought out as an interrupt. See the UDB Array and Digital System Interconnect chapter on page 217 for more details on usage. RTC functionality is available for use in all power modes except the Hibernate mode. This clock routes to the clock distribution network as an input clock source and also to the RTC timer. This oscillator is one of the clock sources available to the clock distribution logic. The kHzECO is enabled and disabled by the register SLOWCLK_X32_CR[0]. Figure 14-6 is a block diagram of the kHzECO. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 113 Clocking System Figure 14-7. RTC Implementation clk_eco_kHz 32 kHz ECO To Clock Distribution Generates a one-pps Interrupt Divide by 32768 En 32 kHz Crystal PM_TW_CFG2[4] PM_TW_CFG2[5] SLOWCLK_X32_CR[0] 14.3.3 Oscillator Summary A summary of the oscillator output frequency ratings is listed in Table 14-1. Table 14-1. Oscillator Summary Fmin Fmax IMO Source 3 MHz 74.7 MHz ILO 1 kHz 100 kHz MHzECO 4 MHz kHzECO PLL 14.3.4 25 MHz 32.768 kHz 24 MHz 80 MHz DSI Clocks Signals can be routed from the Digital Signal Interconnect (DSI) and used as clocks in the clock trees. The sources of these clocks include: ■ Clocks developed in UDBs ■ Off-chip clocks routed through pins ■ Clock outputs from the clock distribution; fed directly back into the network through the routing fabric 14.3.5 Phase-Locked Loop trees. Note that when a PLL parameter is changed or the PLL is enabled or disabled, it takes four bus clock cycles (50 µs) for the corresponding status to be reflected in the FASTCLK_PLL_SR[0] status bit. Additionally, the FASTCLK_PLL_SR[0] bit is not updated while the PLL is disabled. This delay must be incorporated in the firmware before reading the status bit. The PLL's charge pump current (Icp) can be configured using bits 6:4 of register FASTCLK_PLL_CFG1. This bitfield must be set to 0x01 (2 µA) when the output frequency 67 MHz. It must be set to 0x02 (3 µA) when the output frequency is > 67 MHz. The PLL takes inputs from the IMO, the crystal oscillator MHzECO, or the DSI, which can be an external clock. Low-power Operation The PLL must be disabled before going into SLEEP/HIBERNATE mode. This allows clean entry into SLEEP/HIBERNATE and wakeup. The PLL can be reenabled after wakeup and when it is locked; then it can be used as a system clock. The device is designed not to go into SLEEP/HIBERNATE mode if the PLL is enabled when mode entry is attempted. (Execution continues without entering SLEEP/HIBERNATE mode in this case.) The on-chip Phase-Locked Loop (PLL) can be used to boost the clock frequency of the selected clock input (that is, IMO, MHzECO, and DSI clock) to run the device at maximum operating frequency. The PLL can synthesize clock frequencies in the range of 24 – 80 MHz. Its input and feedback dividers allow fine enough resolution to create many desired system clock frequencies. The PLL output routes to the clock distribution network as one of the possible input sources. The PLL is shown in Figure 14-8. The PLL uses a 4-bit input divider Q (FASTCLK_PLL_Q) on the reference clock and an 8-bit feedback divider P (FASTCLK_PLL_P). The outputs of these two dividers are compared and locked, resulting in an output frequency that is P/Q times the input reference clock. The PLL achieves frequency lock in less than 250 µs, and provides a bit that shows lock status (FASTCLK_PLL_SR[0]). When lock is achieved, the PLL output clock can be routed into the clock 114 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Clocking System Figure 14-8. PLL Block Diagram clk_imo clk_eco_Mhz dsi_clkin CLK MUX Q-Divider 4 Bits (1-16) FASTCLK_PLL_Q UP PFD Lock Detect FASTCLK_SR[0] DOWN Filter and VCO clk_pll To Clock Distribution P-Divider 8 Bits (4-256) FASTCLK_PLL_P PLL PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 115 Clocking System 14.4 Clock Distribution All of the clock sources discussed are distributed into the various domains of the device through clock distribution logic. Figure shows a block diagram of the clock distribution system. Figure 14-9. Clock Distribution System dsi_clkin clk_imo2x 4-25 MHz XTAL 33 kHz Watch XTAL clk_xtal IMO 3 - 74.7 MHz USB CLK MUX PLL Mux Prescale Master Clock Mux PLL clk_imo 8-bit divider clk_pll clk_usb 48 MHz dsi_gp clk_sync clk_spc 36 MHz ILO 1, 33, 100 kHz dsi_g 11 clk_sync_d DigitalPhaseMux clk_d2 clk_d_ff2 clk_d3 clk_d_ff3 Ana3 Phase Mux clk_ sync_a3 Ana2 Phase Mux clk_ sync_a2 16-Bit Divide resync dsi_a0 dsi_d4 16-Bit Divide clk_d4 clk_d_ff4 Ph- Sel resync 16-Bit Divide clk_d5 16-Bit Divide clk_d_ff5 dsi_d6 16-Bit Divide resync Ph-Sel clk_d6 clk_d_ff6 dsi_a2 16-Bit Divide resync dsi_d7 clk_d7 dig-resync dsi_a3 16-Bit Divide clk_d_ff7 Ph- Sel All of the clocks available in the device are routed across the device through digital and analog clock dividers. There are certain peripherals that require specific clock source for its operation. For example, Watchdog Timer (WDT) requires clk_ sync_a1 dig-resync 16-Bit Divide Ph-Sel Ana1 Phase Mux dig-resync dsi_a1 dsi_d5 116 resync s8misc_delay_top clk_d_ff1 resync 16-Bit Divide clk_d1 resync dsi_d3 Phase mod resync 16-Bit Divide 16-Bit Divide resync dsi_d2 resync 16-Bit Divide clk_d_ff0 resync dsi_d1 clk_d0 resync 16-Bit Divide resync clk_bus dsi_d0 dig-resync Ana0 Phase Mux clk_ sync_a0 clk_a0 clk_ad0 clk_a1 clk_ad1 clk_a2 clk_ad2 clk_a3 clk_ad3 internal low-speed oscillator (ILO). In such cases, the corresponding clock source is directly routed to the peripheral. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Clocking System The clock distribution can be considered to be a combination of the following clock trees. ■ System clocks ■ Digital clocks ■ Analog clocks ■ USB clock The clock distribution provides a set of eight dividers for the digital clock tree and four analog clock dividers for the analog clock tree. All of the clock sources come as input options for all of the clock dividers through eight input mux. Also, the divider outputs are synchronized to their respective domain clocks. A Master Clock Mux is available for distributing the sync clocks. There are options to provide delay on the digital sync clock. All eight digital dividers are synchronized to the same digital clock, but each of the analog clock divider outputs can be synchronized to analog clocks of different delays. The clock distribution also is responsible for the generation of the major clock domains in the device, such as the System clock, bus clock, and others. 14.4.1 Master Clock Mux The Master Clock Mux, shown in Figure 14-10, selects one clock from among the PLL, selected IMO output, the MHz crystal oscillator, and the DSI input (dsi_clkin). This clock source feeds the phase mod circuit to produce skewed clocks that are selected by the digital and analog phase mux blocks. The Master Clock Mux provides the re-sync clocks for the network: clk_sync_dig and the analog system clocks, clk_sync_a. The master clock must be configured to be the fastest clock in the system. The master clock also provides a mechanism for switching the clock source for multiple clock trees instantaneously, while maintaining clock alignments. For systems that must maintain known clock relationships, clock trees select the clk_sync_dig (or clk_sync_a*) clock as their input source. Therefore, when the source is changed (for example, when moving from the IMO source initially to a new PLL- synthesized frequency), all clocks change together through the Master Clock Mux output. The Master Clock Mux contains an 8-bit divider to generate lower frequency clocks, (CLKDIST_MSTR0[7:0]). It outputs an approximately 50% clock. Figure 14-10. Master Clock Mux CLKDIST_MSTR1_SRC_SEL[1:0] Divide-by-1 clk_pll clk_imo clk_eco_MHz 8-Bit Divider (1-256) D Q CLKDIST_MSTR0 clk_sync dsi_clkin 14.4.2 USB Clock The USB clock domain is unique because it can operate largely asynchronously from the main clock network. The USB logic contains a synchronous bus interface to the device while being able to run on a potentially asynchronous clock to process USB data. For full speed USB, the clock must have an accuracy of ±0.25%. The USB Clock Mux, shown in Figure 14-11, provides the clock to the USB logic. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 117 Clocking System Figure 14-11. USB Clock Mux CLKDIST_UCFG_SRC_SEL[1:0] IMOCLK IMOCLKX2 clk_usb PLL Divide-by-2 DSI dsi_glb_div[0] CLKDIST_UCFG_DIV2 The USB clock mux selects the USB clock from these clock sources. ■ imo1x (these options are available inside the IMO block): ❐ ■ ■ ■ 48 MHz DSI clock subjected to the accuracy of the source of the clock imo2x (these options are available inside the IMO block): ❐ 24 MHz crystal with doubler ❐ 24 MHz IMO with doubler with USB lock ❐ 24 MHz DSI input with doubler clk_pll: ❐ Crystal with PLL to generate 48 MHz ❐ IMO with PLL to generate 48 MHz ❐ DSI input with PLL to generate 48 MHz DSI input: ❐ 48 MHz In this situation, any of the choices can produce a valid 48 MHz clock for the USB. If the internal main oscillator is selected, it must be run with the oscillator locking function enabled, in which case it self tunes to the required USB accuracy when USB traffic arrives at the device. 14.4.3 USB Mode Operation This device works with an automatic clock frequency locking circuit for USB operation. This design allows for small frequency adjustments based on measurements of the incoming USB timing (frame markers) versus the IMO clock rate. With this clock locking loop, the clock frequency can stay within spec for the USB Full Speed mode (±0.25% accurate). The IMO must be operated at 24 MHz for proper clock locking, with the doubler supplying 48 MHz for USB logic. The USB locking feature for the IMO can be enabled by the register bit FASTCLK_IMO_CR[6]. Alternately, a 24 MHz crystal controlled clock doubled to 48 MHz can be supplied for Full Speed USB operation. Other crystal frequencies, such as 4 MHz can be used with the PLL to synthesize the necessary 48 MHz. Valid frequency for the PLL output, in this case, is 48 MHz. The DSI signal, dsi_glb_div[0], provides another DSI signal choice in addition to the clk_imo option above. As with the PLL, this clock must have USB accuracy and be 48 MHz. Clock Dividers Clock dividers form the main part of the clock distribution module and are used to divide and synchronize clock domains. Various clock sources and divider modes may be used together to generate many frequencies with some control over the duty cycle, as depicted in Figure 14-12. 118 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Clocking System Figure 14-12. Divider Implementation Divide-by-4 Example (M=3) 3 2 1 0 3 2 1 0 Source Clock Single Cycle Mode, Std. Phase 50% Mode, Std. Phase Single Cycle Mode, Early Phase 50% Mode, Early Phase Start (enable) Divide-by-5 Example (M=4) 4 3 2 1 0 4 3 2 1 0 Source Clock Single Cycle Mode, Std. Phase 50% Mode, Std. Phase 3 2 Single Cycle Mode, Early Phase 50% Mode, Early Phase Start (enable) The divider automatically reloads its divide count after reaching the terminal count of zero. The divider count is set in the register CLKDIST_DCFG[0..7]_CFG0/1 for digital dividers and CLKDIST_ACFG[0..3]_CFG0/1 for analog dividers. The counter is driven by the clock source selected from an 8-input mux, and the source selection is done in the register CLKDIST_DCFG[0..7]_CFG2[2:0] for digital dividers and CLKDIST_ACFG[0..7]_CFG2[2:0] for analog dividers. There are two divider output modes: single-cycle pulse and 50% duty cycle. In either output mode, a divide value of 0 causes the divider to be bypassed, giving a divide by 1. In this case, the input clock is passed to the output after a resync, if the sync option is selected (see Clock Synchronization on page 120). For a load value of M, the total period of the output clock is N = M + 1 cycles (of the selected input clock). For example, a load value of 4 gives a 5-cycle long output clock period. Divider outputs can each be configured to give one of four waveforms, as described below. 14.4.3.1 Single Cycle Pulse Mode In Single Cycle Pulse mode, by default, the divider generates a single high pulse clock at either the cycle after the terminal (zero) count or the half-count, and is otherwise low. This produces an output clock that is high for one cycle of the input clock, resulting in a 1-of-N duty cycle clock. This is illustrated in Figure 14-12. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 14.4.3.2 50% Duty Cycle Mode In 50% Duty Cycle mode, the output produces a clock that has an approximate 50% duty cycle, depending on whether the total number of counter cycles is even or odd. The 50% clock rising edge occurs at the equivalent rising edge location of the 1/N clock. For a count of M, there are N = M + 1 input clock cycles in the divider period. If M is odd, the total cycle count N is even, allowing for a nominal 50% duty cycle. The clock is high for the first (M + 1)/2 cycles, and then goes low for the remaining (M + 1)/2 cycles. If M is even, the total cycle count is odd, which means that the output clock is high longer than it is low (in standard phase mode). Specifically, it is high for the first (M/2) + 1 cycles and then low for the remaining M/2 cycles. This is illustrated in Figure 14-12 on page 119 for M = 3 and M = 4. The CLKDIST_DCFG[x]_CFG2[4] or CLKDIST_ACFG[x]_CFG2[4] bit in the configuration register for each clock output can be set high to provide the 50% duty cycle mode. An exact 50% duty cycle cannot be guaranteed in all cases, as it depends on the phase and frequency differences between the output clock and the sync clock. 119 Clocking System 14.4.3.3 Early Phase Option In addition to the two duty cycle choices, the outputs can be phase shifted to either go high after the terminal count, or at the half-period cycle. The default is referred to as Standard phase, with the rising edge of the output after the terminal count. The other option is called the Early Phase because the output can be shifted earlier in time to an approximate count, which is one-half of the divide value. The CLKDIST_DCFG_CFG2[5] or CLKDIST_ACFG_CFG2[5] bit in the configuration register for each clock output can be set high to give the Early Phase mode, with the rising edge near the half count. Analog clock dividers are similar in their architecture to digital dividers. However, they have an extra resync circuit to synchronize the analog clock to the digital domain clocks. Therefore, each of the analog dividers also has an output synchronized with the digital domain. This clock is synchronized to the output of the digital phase mux. The digital synchronized analog divider output is called clk_ad. This divider is useful for clean communication between analog and digital domain. 14.4.4 Clock Synchronization Each of the four analog dividers can be synchronized to four distinct phase shifted clocks. The phase on the respective analog dividers sync clocks can be provided in the PHASE_DLY field (CLKDIST_ACFG[x]_CFG3[3:0]). The analog clocks become synchronized when the SYNC bit is set (CLKDIST_ACFG[x]_CFG2[3]). These divided clocks synchronized to the analog clocks are called clk_a. The output of each clock tree provides for selection of one of four output clocks: ■ Resynchronized clock – A clock running at a maximum rate of clk_sync/2 is resynchronized by the phase delayed clk_sync. This output is activated by setting the sync bit. ■ Phase delayed clk_sync (such as clk_sync_dig) – The clock tree runs at the same rate as clk_sync, but just outputs this clock with proper phase delay. Note that the input clock source is ignored in this case. The output buffer is designed to match the final sync flop delay. ■ Unsynchronized divided clock – This produces an asynchronous clock, subject to the limitations described in Asynchronous Clocks on page 122. This mode is applicable when the sync bit is reset and the divider has a nonzero divide value. ■ Bypassed clock source – This routes the clock trees selected source to the output without going through the divider. This happens when the divider value is set to 0 and sync bit is reset. As in the previous case, this also produces an asynchronous clock. All digital and analog divider outputs can be synchronized to the clk_sync_dig signals (CLKDIST_DCFG[x]_CFG2[3] or CLKDIST_ACFG[x]_CFG2[3]), as shown in Figure 14-13. Each digital divider can be synchronized to the digital phase mux output by setting the sync bit (CLKDIST_DCFG[x]_CFG2[3]). The phase delay for the digital divider is based on the phase shift field of Nonvolatile Latch (NVL) bits DIG_PHS_DLY[3:0]. Note The clock divider output used by fixed-function blocks has not been delayed to match the clock output used by the UDB array and are therefore asynchronous. To avoid STA warnings and other potential problems, always put a Sync Component between fixed blocks and UDB Components. Figure 14-13. Resync Option Diagram clkout _ sel clock source Divider D Q D clk_sync_d (or clk_sync_a0-a3) Q D Q Clock tree output Asynchronous clocks (limited use) 14.4.5 Phase Selection and Control To keep the environment quiet in the analog processing domain, a phase difference must exist between the analog and digital system clocks. For this reason, in PSoC devices, a delay chain circuit provides taps to control the phase for the digital and analog clocks. This delay chain provides up to a 10 ns phase adjustment with nominal steps of 0.5 ns. The phase shifter is shown in Figure 14-14. 120 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Clocking System Figure 14-14. Phase Shifter Clk_sync Phase Delay Chain 1.5 ns Clk_sync_a0 1 ns 0.5 ns Ana0 Ana2 CLKDIST_ACFG[0]_CFG3[3:0] Clk_sync_a1 0 ns Clk_sync_a2 CLKDIST_ACFG[2]_CFG3[3:0] Ana1 Ana3 CLKDIST_ACFG[1_CFG3[3:0] Clk_sync_a3 CLKDIST_ACFG[3]_CFG3[3:0] Digital Phase DIG_PHS_DLY[3:0] Clk_sync_dig The phase shifter consists of a chain of (nominally) 0.5 ns buffers connected in cascade, with the output of each buffer ported out of the circuit (21 outputs). The input to this chain is clk_sync from the master clock divider. Five 5-bit muxes select the sync clock to drive the resync circuits. One is clk_sync_dig for the digital clock dividers (clk_bus and all digital clock dividers). The other four are independent delay selections, one for each analog divider. The selected phase value is defined in NVL bits for the digital and ACFG[n]_CLKDIST_ACFG_CFG3}_PHASE_DLY for the analog clocks. The clk_sync_dig phase shift selection must be applied at power up through NVL settings, because changing its value can cause clock glitching; the clk_bus clock should not be stopped for such a change. The analog phase shift selections can be made dynamically, because their output clocks can be disabled during any phase shift change. Outputs in the delay chain may have increased jitter. The expectation is that, in systems that need a low-jitter analog clock, the undelayed output (first tap) is selected because it has the lowest jitter. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 14.4.6 Divider Update To allow for clean updates of the dividers while running, and to align the starting point for a group of dividers, a load enable mechanism is provided. When a clock is running, it automatically reloads its count value on the terminal count. If a new value is loaded during countdown of the counter, this new value is loaded at the end of the count, and the next output clock period uses the new value. Because the divide value is 16 bits, there is a possibility that, when updating this register with two 8-bit writes, the full update might not complete when the terminal count occurs. This leads to an unexpected period being reloaded. To avoid this problem, a 16-bit shadow value (contained in registers {CLKDIST_WRK0*} and {CLKDIST_WRK1*}) allows atomic loads of the dividers, so the 16-bit dividers can be safely updated dynamically (while running). The shadow value can be loaded with two separate 8-bit operations. The mask registers ({CLKDIST_DMASK*} and {CLKDIST_AMASK*}) allow the user to select the target dividers for this shadow value. When the load bit, {CLKDIST_LD}_LOAD, register is written with a 1, all dividers selected in the mask registers have their period count updated to the shadow value. (If the divider is not enabled, it 121 Clocking System is safe to do partial writes directly to the divider period register without using the shadow register.) To align clocks, the mask registers are used again, but this time, they select dividers for auto-alignment. When the {CLKDIST_LD}_SYNC_EN bit register is written with a 1, all dividers selected in mask registers start (or re-start) together. If the dividers are already enabled, they immediately reload and continue counting from this value. If they are not enabled, writing the SYNC_EN bit also sets any corresponding enable bits in the divider enable registers ({PM_ACT_CFG*}), and the dividers begin counting. Writing a 1 to both of the {CLKDIST_LD}_LOAD and {CLKDIST_LD}_SYNC_EN bits can combine these two operations. This causes all selected dividers to load the shadow register value into their count value, to set all selected divider register enables (if not already enabled), and then to start (or restart) with this setting. The sync loading feature is not supported for clocks that are asynchronous to clk_bus. For instance, an external clock coming from the DSI that is not generated from clk_bus cannot have its divide value changed on the fly reliably. Glitching or transient improper divider loads may occur in this scenario. The clock hardware does not support changing the clock divider from N > 1 to N = 1 when the clock is enabled because it can produce illegal clock pulses. If the clock divider must be changed to N = 1, the respective clock divider en_clk_d bit in the PM_ACT_CFG2 register must first be disabled before making the change and then reenabled after the clock divider change is complete. 14.4.7 Power Gating of Clock Outputs Clock trees may be gated off (disabled). These gating signals come from the power manager, which contains a register, {PM_ACT_CFG1, PM_ACT_CFG2*}, to allow user selection of trees to enable or disable. When a clock tree is disabled, its divider is reset so that when reenabled, it reloads its count value. That is, the divider counters do not pause and hold their counts when disabled; they always start over with the latest configured divide count when reenabled. 14.4.8 14.4.9 Asynchronous Clocks Generally, all clocks used in the device must be derived from the same source, or synchronized to the main clk_sync clock. However there are possible exceptions: ■ A signal that comes on-chip routes through a GPIO, routes to the UDB array, interacts only with self-contained UDB functions, and routes out of the device. ■ Similar to the previous, but the signal routes to the interrupt controller instead of off-chip. The interrupt controller is able to handle arbitrarily phased events. ■ USB operation with the IMO locking to USB traffic. Although unlikely, in this case, the rest of the device may run off of a different clock, because the USB circuitry contains its own clk_bus synchronous interface, even if its USB clock is not synchronous. 14.5 Low-Power Mode Operation During sleep modes, clock network outputs are gated off, and most clock sources are disabled automatically by the power manager. The low frequency (kHz) clocks may still run, and various clocks are configured by the power manager to support wakeup and buzz modes. See the LowPower Modes chapter on page 135 for more details. The system will not go into a sleep mode if either the MHz crystal oscillator or the PLL are enabled. If either of these clocks are enabled, the part will simply continue execution without entering a sleep mode. Therefore, to enter a sleep mode when using either the MHz crystal oscillator or PLL, the user must configure the part to run from the IMO and then disable those clock sources. When entering and exiting low-power modes, the IMO should be set to 12 MHz with a post divide of 1. To achieve robust clocking into and out of sleep and hibernate modes, the clocks and clock dividers must be sequenced in firmware. This will also meet the wake up time specifications. PSoC Creator provides APIs to do this sequencing both before entering and after exiting low-power modes. System Clock The System Clock is derived from the clk_sync_dig, which is a phase shifted version of clk_sync. The System Clock, also named clk_bus, is the clock that drives the PHUB and associated bus logic. This must be the fastest synchronous clock that outputs to the system. There is an option for a 16-bit divider on the clk_sync_dig to generate the clk_bus CLKDIST_BCFG1/2. This also has the same resynchronization options as the other digital dividers. 122 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Clocking System 14.6 Clock Naming Summary Table 14-2 lists clock signals and their descriptions. Table 14-2. Clock Signals Clock Signal Description clk_sync_d Synchronization clock from the Master clock mux used to synchronize the dividers in the distribution dsi_clkin Clocks that are taken as input into the clock distribution from DSI clk_bus Bus clock for all peripherals clk_d[0:7] Output clock from the seven digital dividers clk_ad[0:3] Output clock from the four analog dividers synchronized to the digital domain clock clk_a[0:3] Output clock from the four analog dividers synchronized to the analog synchronization clock clk_usb Clock for USB block clk_imo2x Output of the doubler in the IMO block clk_imo IMO output clock clk_ilo1k 1 kHz output from ILO clk_ilo100k 100 kHz output from ILO clk_ilo33k 33 kHz output from ILO clk_eco_ kHz 32.768 kHz output from the kHz ECO clk_eco_ MHz 4-25 MHz output of the MHz ECO clk_pll PLL output dsi_glb_div DSI global clock source to USB block PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 123 Clocking System 124 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 15. Power Supply and Monitoring PSoC® 5LP devices have separate external analog and digital supply pins, labeled Vdda and Vddd, respectively. The devices have two internal 1.8 V regulators that provide the digital (Vccd) and analog (Vcca) supplies for the internal core logic. The output pins of the regulators (Vccd and Vcca) have very specific capacitor requirements that are listed in the datasheet. 15.1 Features These regulators are available: ■ Analog regulator for the analog domain supply ■ Digital regulator for the digital domain supply ■ Sleep regulator for the sleep domain ■ I2C regulator to power the I2C logic ■ Hibernate regulator to supply keep-alive power for state retention during hibernate mode PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 125 Power Supply and Monitoring 15.2 Block Diagram The power system consists of separate analog, digital, and I/O supply pins, labeled Vdda, Vddd, and Vddiox, respectively. It also includes two internal 1.8-V regulators that provide the digital (Vccd) and analog (Vcca) supplies for the internal core logic. The output pins of the regulators and the Vddio pins must have capacitors connected, as shown in Figure 15-1. The power system also contains a sleep regulator, an I2C regulator, and a hibernate regulator. Vdda must be greater than or equal to all other power supply pins (Vddd, Vddios) in PSoC LP. This power supply condition is required for the proper ON/OFF condition of the analog switches inside the device, and also for the implementation of the internal level switching logic when signals transition between multiple supply voltage domains. Figure 15-1. Power Domain Block Diagram 1 µF Vddio2 Vddd Vddd I/O Supply Vssd Vccd Vddio2 Vddio0 0.1 µ F 0.1 µF I/O Supply Vddio0 0.1 µF I2C Regulator Sleep Regulator Digital Domain Vdda Vdda Vcca Analog Regulator Digital Regulators Vssb 0.1 µF 1 µF . Vssa Analog Domain 0.1 µF I/O Supply Vddio3 Vddd Vssd I/O Supply Vccd Vddio1 Hibernate Regulator 0.1 µ F 0.1 µ F Vddio1 126 Vddd Vddio3 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Power Supply and Monitoring 15.3 How It Works The regulators shown in Figure 15-1 power the various domains of the device. All regulators, except the analog regulator, draw their input power from the Vddd pin supply. 15.3.1 Regulator Summary Digital and analog regulators are active during the active or alternate device active modes.They go into a low-power mode of operation in sleep or hibernate mode. The sleep and hibernate regulators are designed to fulfill power requirements in the low-power modes of the device. 15.3.1.1 Internal Regulators For external supplies from 1.8 V to 5.5 V, regulators are powered and the supply is provided through the Vddd/ Vdda pins. An external cap of ~1 µF is connected to the Vccd and Vcca pins. For the 1.71 V < Vcc < 1.89 V external supply, power up the device with Vccd/Vcca pins. In this mode, short the Vddd pin to Vccd and short the Vdda pin to Vcca. The internal regulator remains powered by default. After power up, disable the regulators, using the register PWRSYS.CR0 to reduce power consumption. 15.3.1.2 Sleep Regulator The sleep regulator supplies power to these circuits during the device sleep mode. ■ 32 kHz ECO ■ ILO ■ RTC Timer ■ WDT ■ Central Timewheel (CTW) ■ Fast Timewheel (FTW) 15.3.1.3 Hibernate Regulator The hibernate regulator, whose output is called Keep-Alive power (VpwrKA), powers domains of the device responsible for the state retention in hibernate mode. The VpwrKA is shorted to the active domain during active mode. 15.3.2 Boost Converter PSoC devices also have a boost converter that accepts an input voltage supplied by a battery or other source; it produces a selectable, higher output voltage than the input voltage. Applications that use a supply voltage of less than 1.71 V, such as solar panels or single cell battery supplies, may use the on-chip boost converter to generate a minimum of 1.8-V supply voltage. The boost converter may also be PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F used in any system that requires a higher operating voltage than the supply provides, such as driving a 5.0-V LCD glass in a 3.3-V system. With the addition of an inductor, Schottky diode, and capacitors, it produces a selectable output voltage sourcing enough current to operate the PSoC and other onboard components. The boost converter accepts an input voltage VBAT from 0.5 V to 3.6 V, and can start up with VBAT as low as 0.5 V. The converter provides a user-configurable output voltage of 1.8 V to 5.0 V (VOUT) in 100-mV increments by setting BOOST_CR0[4:0]. VBAT is typically less than VOUT; if VBAT is greater than or equal to VOUT, then VOUT will be slightly less than VBAT due to resistive losses in the boost converter. The block can deliver up to 50 mA (IBOOST) depending on configuration to both the PSoC device and external components. The sum of all current sinks in the design including the PSoC device, PSoC I/O pin loads, and external component loads must be less than the IBOOST specified maximum current. Four pins are associated with the boost converter: VBAT, VSSB, VBOOST, and IND. The boosted output voltage is sensed at the VBOOST pin and must be connected directly to the chip's supply inputs: VDDA, VDDD, and VDDIO, if used to power the PSoC device. The boost converter requires four components in addition to those required in a non-boost design, as shown in Figure 15-2. A 22-uF capacitor (CBAT) is required close to the VBAT pin to provide local bulk storage of the battery voltage and provide regulator stability. A diode between the battery and VBAT pin should not be used for reverse polarity protection because the diodes forward voltage drop reduces the VBAT voltage. Between the VBAT and IND pins, an inductor of 4.7 uH, 10 uH, or 22 uH is required. The inductor value can be optimized to increase the boost converter efficiency based on input voltage, output voltage, temperature, and current. Inductor size is determined by following the design guidance located in this section and the device datasheets electrical specifications. The inductor must be placed within 1 cm of the VBAT and IND pins and have a minimum saturation current of 750 mA. Between the IND and VBOOST pins, a Schottky diode must be placed within 1 cm of the pins. The Schottky diode should have a forward current rating of at least 1.0 A and a reverse voltage of at least 20 V. A 22-uF bulk capacitor (CBOOST) must be connected close to VBOOST to provide regulator output stability. It is important to sum the total capacitance connected to the VBOOST pin and ensure the maximum CBOOST specification is not exceeded. All capacitors must be rated for a minimum of 10 V to minimize capacitive losses due to voltage de-rating. 127 Power Supply and Monitoring Figure 15-2. Application of Boost Converter Powering PSoC Device PSoC VDDA External Load VDDD VDDD 0.1 µF 1.0 µF 0.1 µF 1.0 µF 0.1 µF 1.0 µF VBOOST Schottky, 1A IND 4.7 µH 10 µH 22 µH VBAT VDDIO0 Boost VDDIO2 Logic VDDIO1 22 µF VSSB VDDIO3 0.1 µF 0.1 µF 0.1 µF 0.1 µF 0.5–3.6 V VSSA 22 µF VSSD All components and values are required The boost converter may also generate a supply that is not used directly by the PSoC device. An example of this use case is boosting a 1.8-V supply to 4.0 V to drive a white LED. If the boost converter is not supplying the PSoC devices VDDA, VDDD, and VDDIO, it must comply with the same design rules as supplying the PSoC device, but with a change to the bulk capacitor requirements. A parallel arrangement 22 uF, 1.0 uF, and 0.1 uF capacitors are all required on the Vout supply and must be placed within 1 cm of the VBOOST pin to ensure regulator stability. 128 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Power Supply and Monitoring Figure 15-3. Application of Boost Converter not Powering PSoC Device VOUT External Load PSoC VDDA VDDD 22 µF 1.0 µF 0.1 µF VDDD VBOOST Schottky, 1A IND 4.7 µH 10 µH 22 µH VBAT VDDIO0 VDDA, VDDD, and VDDIO connections per section 6.2 Power System. Boost VDDIO2 Logic VDDIO1 22 µF VDDIO3 VSSB 0.5–3.6 V VSSA VSSD All components and values are required The boost converter is enabled or disabled by the register bit, BOOST_CR1[3]. The device provides the option of changing the boost output voltage by writing into the BOOST_CR0[4:0] register. By default, at startup the boost converter is enabled and configured for a 1.8-V output. If the boost converter is not used in a given application, tie the VBAT, VSSB, and VBOOST pins to ground and leave the IND pin unconnected. 15.3.2.1 Operating Modes The boost converter can be operated in two different modes: active and standby selected by the BOOST_CR0[6:5] register. Active mode is the normal mode of operation where the boost regulator actively generates a regulated output voltage. In active mode, the switching frequency is set to 400 kHz using an oscillator integrated into the boost converter and is not synchronized to any other clock. The output voltage is continuously monitored and supervisory data provided in BOOST_SR[4:0]. This register provides supervisory data against the output voltage selected. In standby mode, the boost oscillator and most boost functions are disabled, thus reducing power consumption of the boost circuit. In standby mode, only minimal power is provided, typically < 5 µA to power the PSoC device in sleep mode. The processor can determine when to use the thump PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F bit BOOST_CR0[7] to switch the transistor on for a 1-µs pulse when the voltage falls below the nominal voltage. In boost standby mode, the external 32-kHz crystal can be used to trigger inductor boost pulses on the rising and falling edge of the clock when the output voltage is less than the configured value. This is called automatic thump mode (ATM). To enable ATM, set the BOOST_CR2[0] bit and select the external 32-kHz crystal clock by setting BOOST_CR1[1:0] to 0x3. The boost typically draws 250 µA in active mode and 25 µA in standby mode. The boost operating modes must be used in conjunction with chip power modes to minimize total power consumption. Table 15-1 lists the boost power modes available in different chip power modes. Table 15-1. Chip and Boost Power Modes Compatibility Chip Power Modes Boost Power Modes Active or alternate active mode Boost must be operated in its active mode. Sleep mode Boost can be operated in either active or standby mode. In boost standby mode, the chip must wake up periodically for chip active-mode refresh or thump and return to sleep. Hibernate mode Boost can only be operated in its active mode. However, it is recommended not to use boost in chip hibernate mode due to high current consumption in boost active mode. 129 Power Supply and Monitoring 15.3.2.2 Status Monitoring Status monitoring for input and output voltages of the boost converter is available in the status register BOOST_SR. Output Voltage Monitor - The BOOST_SR[4:0] register provides status of the output voltage against the set nominal output voltage. Bit 4: ov - Above overvoltage threshold (nominal + 50 mV). Bit 3: vhi - Above high regulation threshold (nominal +25 mV). Bit 2: vnom - Above nominal threshold (nominal). Bit 1: vlo - Below low regulation threshold (nominal -25 mV). Bit 0: uv - Below under-voltage limit (nominal -50 mV). The boost converter generates a power manager interrupt when an under-voltage event occurs. This interrupt can be configured to wake the chip from Alternate Active or Sleep mode. If an under-voltage event occurs, BOOST_SR2[0] will be set to '1' until the register is read or a reset event occurs. To propagate the under-voltage condition to the interrupt controller to trigger an ISR, the BOOST_CR4[0] bit must be set. 15.3.2.3 Boost Firmware Requirements To ensure boost inrush current is within specification at startup, the Enable Fast IMO During Startup value must be unchecked in the PSoC Creator IDE. The Enable Fast IMO During Startup option is found in PSoC Creator in the design-wide resources (cydwr) file System tab. Unchecking this option configures the device to run at 12 MHz versus 48 MHz during startup while configuring the device. The slower clock speed results in reduced current draw through the boost circuit during the critical startup phase. When the bus clock is configured to a frequency greater than 24 MHz, the boost configuration registers BOOST_CR0, BOOST_CR1, BOOST_CR2, and BOOST_CR3 - must be read with two consecutive read operations, discarding the result from the first read. It is not allowed to access any other boost configuration register between two reads of a boost configuration register, but it is acceptable to access non-boost registers. To avoid this situation, boost registers should not be read in both main code as well as an ISR, or interrupts should be disabled when reading these four registers. This requirement of reading the control registers twice is to avoid a timing issue in accessing these registers. Writing of the boost registers can however occur at any clock rate with a single write instruction. On startup, the boost regulator uses an internal low precision reference to ensure the device can boot up to at least 1.71 V. To meet datasheet specifications, the boost must 130 switch to the devices precision voltage reference that is external to the boost block. The external precision reference is selected by setting the BOOST_CR2[3] bit. The external reference is automatically selected in the device boot code generated by the PSoC Creator IDE. 15.3.2.4 Boost Design Process Correct operation of the boost converter requires specific component values determined for each design's unique operating conditions. The CBAT capacitor, inductor, Schottky diode, and CBOOST capacitor components are required with the values specified in the datasheet electrical specifications. The only variable component value is the inductor LBOOST, which is sized primarily for correct operation of the boost across operating conditions and secondarily for efficiency. Additional operating region constraints exist for VOUT, VBAT, IOUT, and TA. Follow these steps to determine the boost converter operating parameters and LBOOST value. 1. Choose desired VBAT, VOUT, TA, and IOUT operating condition ranges for the application. 2. Determine if VBAT and VOUT ranges fit the boost operating range based on the TA range over VBAT and VOUT chart. If the operating ranges are not met, modify the operating conditions or use an external boost regulator. 3. Determine if the desired ambient temperature (TA) range fits the ambient temperature operating range based on the TA range over VBAT and VOUT chart. If the temperature range is not met, modify the operating conditions and return to step 2, or use an external boost regulator. 4. Determine if the desired output current (IOUT) range fits the output current operating range based on the IOUT range over VBAT and VOUT chart. If the output current range is not met, modify the operating conditions and return to step 2, or use an external boost regulator. 5. Find the allowed inductor values based on the LBOOST values over VBAT and VOUT chart. 6. Based on the allowed inductor values, inductor dimensions, inductor cost, boost efficiency, and VRIPPLE, choose the optimum inductor value for the system. Typical values for boost efficiency and VRIPPLE are provided in the Efficiency vs VBAT and VRIPPLE vs VBAT charts. In general, if high efficiency and low VRIPPLE are most important, then the highest allowed inductor value should be used. If low inductor cost or small inductor size are most important, then one of the smaller allowed inductor values should be used. If the allowed inductor(s) efficiency, VRIPPLE, cost, or dimensions are not acceptable for the application, then an external boost regulator should be used. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Power Supply and Monitoring 15.3.3 Voltage Monitoring The device has two circuits for detecting voltages that deviate from the selected threshold on the external digital / analog supplies: ■ Low-Voltage Interrupt (LVI) – The LVI circuit generates an interrupt when it detects a voltage below the set value. ■ High-Voltage Interrupt (HVI) – The HVI circuit generates an interrupt when it detects a voltage above the set value. The basic block diagram of voltage monitoring is shown in Figure 15-4. Figure 15-4. Voltage Monitoring Block Diagram Vdda Vddd RESET_CR1[3] Analog LVI (ALVI) Triplevel = RESET_CR0[7:4] Digital LVI (DLVI) AHVI RESET_CR1[1] Triplevel = RESET_CR0[3:0] RESET_CR1[2] RESET_CR1[0] Interrupt Controller 15.3.3.1 Low-Voltage Interrupt The LVI circuit generates an interrupt when it detects a voltage below the set value. These low-voltage monitors are off by default, but the trip level for the LVI can be set in the register RESET_CR0 from 1.7 V to 5.45 V in steps of 250 mV. The LVI circuit has a persistent status register bit in RESET_SR0 that is set until cleared by the user by reading from the register. Note that the LVI status bits in RESET_SR0 will be reset to ‘0’ when a device reset occurs due to a POR, LVI, or HVI condition. This bit is useful only when the LVI is configured as an interrupt source because an LVI reset also clears this bit. This bit is set whenever the voltage goes below the set value. There is distinct monitoring for low voltage on the analog and digital supply. The analog low-voltage interrupt (LVIA), enabled by RESET_CR1[1] and RESET_CR0[7:4], sets the LVIA threshold. The digital low-voltage interrupt (LVID), enabled by RESET_CR1[0] and RESET_CR0[3:0], sets the LVID threshold. Apart from this, when the voltage monitoring is enabled and the corresponding PRES bit is also enabled in RESET_CR3[7:6], the low-voltage condition triggers a corresponding reset. Both the LVIA and LVID resets are enabled by default. Note that the LVI reset will continuously occur as long as the LVI voltage condition persists. The user PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F code configures the LVI for reset. When low-voltage condition occurs after this configuration is done, the device is reset once and the user code starts executing from flash address zero. Again, when the CPU reaches the code that configures LVI for reset and the voltage is still low, the device will be reset again. This continuous cycle of device reset occurs until the low-voltage condition is no longer present. The interrupt is generated only when the corresponding bit in the RESET_CR1 register is set and the corresponding bits in RESET_CR3[7:6] cleared. Even if the interrupt output is not used to generate a processor interrupt, the status registers are updated by the circuit whenever LVI functions are enabled. In addition, the real-time status of each LVI circuit is available and captured in a real-time status register bit in RESET_SR2, so you can determine if an under/over voltage condition is still in effect. Similar to the reset condition, the LVI interrupt is continuously triggered until the voltage goes above the low-voltage trip point. The low-voltage detect (LVD) events, comprising LVIA and LVID, can also be used to cause a device to wake up from sleep or standby modes. When the Vddx (Vdda and/or Vddd) drops below the threshold, an interrupt is generated on wakeup. The interrupt is generated due to the LVD status 131 Power Supply and Monitoring bits being set in the RESET_SR0 register. In sleep mode, if the Vddx drops to the LVI trip point threshold, the device will wake up but may not always generate an interrupt due to the LVD status bits not being set. There is no effect on CPU or other subsystem performance. See the device datasheet for information on the voltage threshold settings. 15.3.3.2 High Voltage Interrupt The HVI circuit generates an interrupt when it detects a voltage above the fixed, safe operating value of 5.75 V on the external analog supply. There is just one HVI for both analog and digital supplies. The selection between monitoring the digital or analog supply is done by the RESET_CR1[3] bit, the default selection is for the Vdda supply. These high-voltage monitors are off by default, but this feature can be enabled in the register RESET_CR1[2]. The HVI circuit has a persistent status register bit in RESET_SR0 that is set until it is cleared by the user by reading or writing to the register. Note that the HVI status bits in RESET_SR0 will be reset to ‘0’ when a device reset occurs due to a POR, LVI, or HVI condition. This bit is useful only when the HVI is configured as an interrupt source because an HVI reset also clears this bit. This bit is set when the analog voltage value goes beyond the threshold value. Note that the HVI reset will continuously occur as long as the HVI voltage condition persists. The user code configures the HVI for reset. When high-voltage condition occurs after this configuration is done, the device is reset once and the user code starts executing from flash address zero. Again, when the CPU reaches the code that configures HVI for reset and the voltage is still high, the device will be reset again. This continuous cycle of device reset occurs until the high-voltage condition is no longer present. The interrupt is generated only when the corresponding bit in the register RESET_CR1[2] is unmasked. Even if the interrupt output is not used to generate a processor interrupt, the status registers are updated by the circuit whenever HVI functions are enabled. In addition the real-time output of each HVI circuit is available and captured in a realtime register bit in RESET_SR2, so you can determine if an overvoltage condition is still in effect. Similar to the LVIA/ LVID events, HVIA event is also available in active and standby modes. The HVIA interrupt can return the chip to active mode from standby mode. Similar to the reset condition, the HVI interrupt is continuously triggered until the voltage goes below the high-voltage trip point. 15.3.3.3 Processing a Low/High Voltage Detect Interrupt Both LVI and HVI circuits cause the same interrupt output signal, which is made available to the Interrupt Controller. 132 Further execution of the interrupt depends on the enable status for the interrupt line in the Interrupt Controller. After the interrupt occurs, the user code can interrogate status registers to determine which LVI or HVI circuit detected an under- or over-voltage condition. The actual interrupt output (LVD) is an OR function of the three persistent status register bits corresponding to LVI-D, LVI-A, and HVI. Therefore, to clear the interrupt, the ISR must clear these three register bits. The LVI and HVI interrupts are prone to a glitch when they are enabled. Exercise caution in the firmware to avoid any interrupt generated by the voltage detection circuitry at the moment when voltage detection is being enabled. One way to achieve this is by disabling the LVD interrupt before enabling the voltage detection and enabling it after some time, which avoids the potential glitch caused while enabling. During sleep mode, LVI and HVI circuits may be buzzed (periodically activated). If an interrupt occurs during buzzing, the system will first go through its wakeup sequence; then the interrupt is recognized and serviced. With the LVI configured as an interrupt, if the low-voltage condition and a soft reset (such as software reset, watchdog reset, segment reset) occur simultaneously, there is a chance that the low-voltage condition persists when the device resets due to the soft reset source. This will result in the low-voltage condition causing a hard reset as well. If a hard reset occurs, it results in the clearing of the soft reset status register bits in RESET_SR0 and RESET_SR1. The implication is that any soft reset occurring in conjunction with the LVI interrupt event will not be properly reflected in the RESET_SR0, RESET_SR1 status registers. However, there will be no impact on any other device operation; the device will undergo the normal sequence after the reset occurs. This behavior is applicable for the HVI interrupt as well. 15.3.3.4 Reset on a Voltage Monitoring Interrupt The ALVI and DLVI can be configured to directly reset the device by setting the corresponding bits in RESET_CR3[7:6]. When this bit is set to ‘1’ along with the RESET_CR1[0/1] set to ‘1’, the corresponding LVI becomes an additional reset source through the PRES reset path. When this bit is cleared to ‘0’ along with the RESET_CR1[0/ 1] set to ‘1’, the corresponding LVI is only used as an interrupt source. If the RESET_CR1[0/1] is cleared to ‘0’, the bit state (either a zero or a one) has no impact on the reset or interrupt functionality. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Power Supply and Monitoring The LVI glitch mentioned in 15.3.3.3 Processing a Low/High Voltage Detect Interrupt triggers a system reset if the LVI monitor is enabled after enabling the LVI reset (the LVI reset is enabled by default). To avoid this, disable the LVI reset by clearing the corresponding bits in RESET_CR3[7:6] before enabling the LVI monitor. 15.4 Note that the LVI reset will not hold the device in reset until the voltage goes above the set value. When the LVI circuit detects a low voltage, reset is asserted. The reset is then released even if the voltage is still below the LVI set value. Register Summary Table 15-2. Power Supply Register Summary Register PWRSYS_CR0 Function Regulator control PWRSYS_CR1 Analog regulator control BOOST_CR0 Boost Thump, voltage selection and mode select BOOST_CR1 Boost enable and control BOOST_CR2 Boost control BOOST_CR3 Boost PWM duty cycle BOOST_SR Boost status RESET_CR0 LVI trip value setting RESET_CR1 Voltage monitoring control RESET_SR0 voltage monitoring status RESET_SR2 Real-time voltage monitoring status PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 133 Power Supply and Monitoring 134 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 16. Low-Power Modes The PSoC® 5LP devices feature a set of four power modes with a goal of reducing the average power consumption of the device. 16.1 Features The PSoC 5LP power mode features, in order of decreasing power consumption, are: Internal Low Speed Oscillator (ILO) and the Internal Main Oscillator (IMO) to schedule the alternation. ■ Active ■ Low Power Active ■ Alternative Active ■ Sleep ■ Hibernate Sleep and hibernate modes are used when processing is not necessary for an extended time. All subsystems are automatically disabled in these two modes, regardless of the settings in the active template register. Some subsystems have an additional available bit [PM_Avail_CRx] that can mark a subsystem as unused and prevent it from waking back up. This reduces the power overhead of waking up the part, in that not all subsystems are repowered. Active and alternative active are the main processing modes, and the list of enabled peripherals is programmable for each mode. The user can enter Active or Alternative Active manually or automatically alternate between them using Low Power Active (LPA) mode. LPA mode uses the The allowable transitions between power modes are illustrated in Figure 16-1. Figure 16-1. State Diagram of Allowable Power Mode Transitions Active Manual Sleep Hibernate Alternate Active PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 135 Low-Power Modes The various power modes reduce power by affecting the following resources: ■ Regulators for the digital and analog supply in the device ■ Clocks such as the IMO, ILO, and external crystal oscillator (ECO32K, ECOM) ■ Central processing unit (CPU) and all other peripherals Power savings, resume time, and supported wakeup sources depend on the particular mode. The four global power-reducing modes are described in Table 16-1 and are listed in decreasing order of power consumption. Table 16-1. Power Consumption-Reducing Modes Power Modes Description Active Primary mode of operation, all peripherals available (programmable) Alternate Active Similar to Active mode, and is typically configured to have fewer peripherals active to reduce power. One possible configuration is to use the UDBs for processing, with the CPU turned off Sleep Hibernate Entry Condition Wakeup Source Wakeup, reset, manual register entry Manual register entry All subsystems automati- Manual register entry cally disabled All subsystems automatically disabled Lowest power consuming mode with all periphManual register entry erals and internal regulators disabled, except hibernate regulator is enabled Configuration and memory contents retained Active Clocks Any (programmable) Any interrupt Any (programmable) Any interrupt Comparator, PICU, I2C, RTC, CTWa Comparator, PICUa ILO/kHzECO Regulator All regulators available. Digital and analog regulators can be disabled if external regulation used. All regulators available. Digital and analog regulators can be disabled if external regulation used. Both digital and analog regulators buzzed. Digital and analog regulators can be disabled if external regulation used. Only hibernate regulator active. a. On PSoC 5LP, an interrupt signal coming from a wakeup source should not be passed through the "Edge Detect" logic shown in Figure 7-2 on page 74. The interrupt signal should be passed directly to the interrupt controller. This is a requirement only for sleep and hibernate power mode wakeup sources. Alternate active mode wakeup sources can have their interrupt signals either passing directly to the interrupt controller, or through the edge detect logic. 16.2 Active Mode Active mode is the primary power mode of the PSoC device. This mode provides the option to use every possible subsystem/peripheral in the device. All of the clocks in the device are available for use in this mode. Each power-controllable subsystem is enabled or disabled in active mode, using the active power configuration template bits [PM_ACT_CFGx registers]. This is a set of 14 registers in which each bit is allocated to enable/disable a distinct power controllable subsystem. When a subsystem is disabled, the clocks are gated and/or analog bias currents are reduced. Firmware may be used to dynamically enable or disable subsystems by setting or clearing bits in the active configuration template. It is possible for the CPU to disable itself, while the rest of the system remains in active mode. The CPU active mode bit is not sticky; therefore the CPU is 136 always awakened whenever the system returns to active mode. 16.2.1 Entering Active Mode Any wakeup event, any reset, or writing 0 into PM_MODE_CSR[2:0] register while in alternate active mode transitions the device into active mode. When a wakeup event occurs in alternate active/sleep/hibernate mode, the global mode always returns to active and the CPU is automatically enabled, regardless of its template settings. Active mode is the default global power mode upon boot. 16.2.2 Exiting Active Mode A register write into PM_MODE_CSR[2:0] can transition to another mode. Firmware must ensure the SPC Idle bit in the SPC_SR[1] register is '1' prior to writing to the PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Low-Power Modes PM_MODE_CSR[2:0] register to ensure any SPC commands have completed. Any pending wakeup source prevents the device from exiting active mode. 16.3 Alternative Active Mode Alternative active mode is similar to active mode in most of its functionality. Alternative active mode also has its own additional set of subsystem template bits [PM_STBY_CFGx], which determine whether a subsystem is enabled or disabled. This mode is made available for quick transitions between active and an alternate low-power mode. For example, you can write to the template bits to disable CPU and enable certain peripherals to operate in alternate active mode. While in alternate active mode, if any interrupt is generated, the device automatically transitions to active mode and begins executing the firmware in active mode. and PPOR depends on the buzz rate. Selecting a higher buzz rate improves the response time at the cost of increased average power consumption. If the keepers are disabled, then the buzz rate should be set high enough to ensure that the external capacitors remain within a tightly bound voltage range. It is not fatal if the capacitor discharges below the desired voltage boundary. However, this increases wake time because the LDOs must charge the capacitor before it can enter active mode. 16.4.1 Sleep mode is entered by writing the appropriate code into PM_MODE_CSR[2:0]. Firmware must ensure the SPC Idle bit in the SPC_SR[1] register is '1' before writing to the PM_MODE_CSR[2:0] register. Entry must be from a state where the CPU is available (active). The system ignores any request to enter sleep mode for the first 1 ms after POR. 16.4.2 16.3.1 Entering Alternative Active Mode To enter alternative active mode, write into [PM_MODE_CSR]. Firmware must ensure the SPC Idle bit in the SPC_SR[1] register is '1' before writing to the PM_MODE_CSR[2:0] register. The essential difference between active and alternative active mode is that the device cannot wake up from sleep/ hibernate mode into the alternative active mode. 16.3.2 Exiting Alternative Active Mode Any interrupt or write to the [PM_MODE_CSR] register can return the system to active mode. 16.4 Sleep Mode Sleep mode powers down the CPU and other internal circuitry to reduce power consumption. System supervisory services, such as the central timewheel, RTC, and WDT remain active. When a wakeup event occurs, the system reactivates in a single phase and returns to active mode. The analog and digital LDO regulators are disabled during sleep mode. If the core supplies are configured for internal regulation, a weak keeper is used to hold the external capacitors at 1.8 V (nominal). Both regulators can be periodically activated (buzzed) to provide supervisory features for voltage monitoring and brownout detect (LVI, HVI, and PPOR). Buzzing is not required if these supervisory services are not used. The buzz rate is programmable using the {PWRSYS.BUZZ_TR} register. The response time of the LVI, HVI, PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Entering Sleep Mode Exiting Sleep Mode Only PICU interrupts, comparator wakeup, supervisory interrupts, or resets wake up the system. At wakeup, the system activates all previously available domains from active mode template and begins executing the firmware in active mode. 16.5 Hibernate Mode Hibernate mode consumes/dissipates the lowest power, and nearly all internal functions are disabled. There is no buzzing, and the external capacitors are permitted to discharge. The hibernate-regulator is always active to generate the keep-alive voltage (Vpwrka) used to retain the system state. See 15.3.3 Voltage Monitoring on page 131. Configuration state and all memory contents are preserved in hibernate mode. GPIOs configured as digital outputs maintain their previous values, and pin interrupt settings are preserved. The voltage used to retain state is lower than the nominal core voltage. In hibernate mode, voltage is monitored with a lower degree of precision than in the other power modes. The hibernate mode has a higher probability of having soft errors. Hence for safety critical applications the MFGCFG.PWRSYS.HIB.TR1[7] can be programmed to prevent hibernate mode. When this bit is asserted, the command to hibernate will put the system into sleep mode. This is important when there are chances of an accidental entry into hibernate mode and the watchdog is disabled. To achieve an extremely low current, a hibernate regulator with limited capacity is used. This limits the frequency of any signal present on the input pins - no GPIO should toggle at a 137 Low-Power Modes rate greater than 10 kHz while in hibernate mode. Because hibernate mode is intended to implement a dormant state in the application, this is not a practical limitation. Any system that has signals toggling at high rates in low-power modes can use the sleep mode without seeing a significant difference in total power consumption. 16.6.1 16.5.1 ■ Buzzing during sleep mode ■ Waking up the device from a low-power mode ■ Watchdog timer (WDT) ■ General timing purposes Entering Hibernate Mode Hibernate mode is entered by a write into PM_MODE_CSR[2:0]. Firmware must ensure the SPC Idle bit in the SPC_SR[1] register is '1' before writing to the PM_MODE_CSR[2:0] register. The extremely low current hibernate regulator requires at least 1 ms to start up after a reset. During this time, the system ignores requests to enter hibernate mode. 16.5.2 Exiting Hibernate Mode Return from hibernate mode can occur only in response to a PICU, comparator, or reset event. The digital, analog, and sleep regulators are disabled in hibernate mode. Upon wakeup, the system activates all previously available domains, unless the {PM_MODE_CFG1[2]} field is set. 16.6 Timers The PSoC power manager provides timers to facilitate advanced power management. Each of these timers are clocked off the low-speed clocking resources available on the chip: the ILO and 32-kHz crystal oscillator. Both of these oscillators have startup times associated with them. Ensure that firmware driven timer initialization occurs quickly after reset. Central Timewheel (CTW) The central timewheel (CTW) is a 1-kHz, free-running, 13-bit counter clocked by the ILO. The CTW is always available, except in hibernate mode and when the CPU is stopped during debug on-chip (DoC) mode. The main functions of the CTW are: CTW settings PM_TW_CFG1[3:0]. are programmable, using Although the CTW is free-running, separate settings are used for the wakeup and watchdog timeouts. The CTW can be programmed, using the {PM_TW_CFG2[2]} registers, to wake the system periodically and optionally issue an interrupt by programming the bit {PM_TW_CFG2[3]}. 16.6.2 Fast Timewheel (FTW) The fast timewheel (FTW) is a 100-kHz, 5-bit counter clocked by the ILO, which can also be used to wake the system from alternative active mode only. The FTW settings are programmable, using PM_TW_CFG0[4:0]; the counter automatically resets when the terminal count is reached. The FTW enables flexible, periodic wakeups of the CPU at a higher rate than the rate allowed using the CTW. To wake up on the FTW, write into register PM_TW_CFG2[0]. If the associated FTW interrupt is enabled using PM_TW_CFG2[1], an interrupt is generated each time the terminal count is reached. The output of these oscillators is considered unreliable during the specified startup time. User code should be authored to ensure that this unreliable clocking is accounted for. The best way to achieve this is to keep the timer disabled during the startup period. For the central timewheel, fast timewheel, and one pulse-per-Second timers, this is achieved with the associated enable bits in the PM.TW_CFG2 register. For the Watchdog timer, this is achieved with a watchdog clear followed by a watchdog reset enable (see the Watchdog Timer chapter on page 141). The ILO is automatically started upon power up and wake from hibernate. The ILO and 32-kHz crystal oscillator may also be stopped and restarted by user firmware. In this case, the user code must also account for oscillator startup time. Timers and timewheels schedule events. They can be programmed to generate periodic interrupts for timing or to wake the system from a low-power mode. 138 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Low-Power Modes 16.7 Register List Table 16-2. Low-Power Modes Register List Register Name Description General Registers PM_ACT_CFGx Active mode template PM_STBY_CFGx Alternate Active mode template PM_AVAIL_CRx Available settings for limited Active mode transition PM_AVAIL_SRx Availability Status register PM_MODE_CFG0 Not used PM_MODE_CFG1 Interrupt and settings for low-power modes PM_MODE_CSR Power Mode Control and Status register PM_INT_SR Power Mode Interrupt Status register PM_TW_CFG0 Fast Timewheel (FTW) Configuration register PM_TW_CFG1 Central Timewheel (CTW) Configuration register PM_TW_CFG2 Configuration settings for CTW and FTW PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 139 Low-Power Modes 140 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 17. Watchdog Timer The watchdog timer (WDT) circuit automatically reboots the system in the event of an unexpected execution path. This timer must be serviced periodically. If not, the CPU resets after a specified period of time. After the WDT is enabled, it cannot be disabled except during a reset event. This is done to prevent any errant code from disabling the WDT reset function. To use the WDT function, enable the WDT function during the startup code. 17.1 Features The WDT has the following features: ■ Protection settings to prevent accidental corruption of the WDT ■ Optionally-protected servicing (feeding) of the WDT ■ A configurable low-power mode to reduce servicing requirements during sleep mode ■ A status bit for the watchdog event that shows the status even after a watchdog reset 17.2 Block Diagram Figure 17-1 is a block diagram of the WDT circuit. Figure 17-1. Watchdog Timer Circuit 2.048 sec - 3.072 sec 1024 Ticks 256 ms – 384 ms 128 Ticks Watchdog Counter (3 Counts) 32 ms – 48 ms Watchdog Reset 16 Ticks 4 ms – 6 ms ILO PM_WDT_CFG[1:0] 2 Ticks Clear Enable PM_WDT_CR PM_WDT_CFG[4] 1 kHz Central Timewheel PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 141 Watchdog Timer 17.3 How It Works The WDT circuit asserts a hardware reset to the device after a preprogrammed interval, unless it is periodically serviced in firmware. The system restarts if an unexpected execution path is taken through the code and the preprogrammed interval times out. It can also restart the system from the CPU halt state. The WDT timeout is between two and three programmable tap periods, based on the free-running Central Timewheel. See the PSoC® 5LP Registers TRM (Technical Reference Manual). Each time the central timewheel crosses the programmed tap point, the Watchdog counter increments. When the counter reaches three, a Watchdog reset is asserted, and the counter is reset. When the WDT is serviced in software, the counter is reset to zero. The time between servicing and the first tap crossing is usually less than the complete tap period; therefore, program the software to service the WDT within two tap periods. Actual WDT timeouts may differ slightly from nominal, caused by inaccuracy of the ILO frequency. 17.3.1 Enabling and Disabling the WDT The WDT is enabled by setting the PM_WDT_CFG[4] register bit. After this bit is set, it cannot be cleared again except by a reset event. This is done so that errant code cannot accidentally disable the watchdog. You must either re-enable the Watchdog function at startup after a reset occurs or include code to re-enable the function should a reset occur, allowing a dynamic choice whether to enable the Watchdog. A status bit (RESET_SR0[3]) becomes set on the occurrence of a Watchdog reset. This bit remains set until cleared by the user, by reading or writing to the register, or until a POR reset. All other resets leave this bit untouched. 17.3.2 Setting the WDT Time Period and Clearing the WDT Select a tap from the central timewheel using the register PM_WDT_CFG[1:0]. Based on the tap selected, the WDT is timed at various periods, shown in Figure 17-1 on page 141. The WDT counts until reaching three, based on the tap from the central timewheel. If the firmware does not clear the WDT before this time, a Watchdog reset is initiated. To prevent an automatic reset, the WDT must be periodically serviced by firmware. In the default mode, this is accomplished by writing any value to the PM_WDT_CR field. It is a good idea to service the WDT in a firmware main 142 loop, that is, not in an interrupt handler. If the WDT is serviced in an interrupt handler, and the main loop code goes astray, the WDT may never generate a reset because the interrupt may still be active, causing the interrupt handler to continue to service the WDT. 17.3.3 Operation in Low-Power Modes A configurable low-power mode of the WDT reduces servicing requirements during sleep mode. The register PM_WDT_CFG[6:5] governs the low-power mode for the WDT. If the WDT is enabled, two bits define how the WDT behaves when the part enters Sleep/Idle/Hibtimers (lowpower) mode. The default is 01; the system will automatically use the longest WDT interval when Sleep/Idle/Hibtimers mode is entered, so software is not burdened with waking just to feed the WDT. This is true regardless of the value programmed in the wdt_interval register. Upon wakeup, the interval will remain at the highest setting until the WDT is fed the first time. A feeding at this point will cause the interval to automatically return to the normal setting (value in wdt_interval). If this field is set to NOCHANGE ('00'), the system does not change the interval and does not feed the WDT when entering Sleep/Idle/Hibtimers mode. If DISABLED (wdt_lpmode=11), the WDT is turned off when Sleep/Idle/Hibtimers mode is entered and remains disabled until the first feeding by the user after active mode is reentered. 17.3.4 Watchdog Protection Settings Using the MLOGIC_SEG_CR and MLOGIC_SEG_CFG0 registers, the WDT registers are protected from accidental corruption as follows: ■ Clear, low-power enable, and Watchdog enable registers are protected as segment 0 as one-time system settings. ■ The servicing of WDT clear is protected in segment 1 as a reconfigurable system setting. See 20.3 Configuration Segment Protection on page 168. 17.4 Register List Table 17-1. Reset Register List Register Name Comments PM_WDT_CFG Configuration register for Watchdog PM_WDT_CR Watchdog clear RESET_SRO Persistent Status register for Watchdog reset PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 18. Reset PSoC® 5LP architecture supports several types of resets, enable power-on reset, user-supplied external or software resets, and recovery from errant code operation. 18.1 Reset Sources PSoC 5LP supports a number of hard and soft resets. Hard resets such as power-on reset completely reset the device. Soft resets are similar but do not reset certain register bits; see the Register TRM for details. Resets sources include: ■ IPOR (hard): power-on reset ■ PRES (hard): internal regulators low-voltage monitor reset ■ HRES (hard): supply low-voltage monitor reset for sleep and hibernate modes ■ XRES (hard): external source activates reset pin ■ LVI (soft): supply low-voltage monitor for active mode, can generate interrupt or reset ■ SRES (soft): software reset, by setting a register bit ■ WRES (soft): watchdog reset 18.1.1 Voltage Monitor Resets PSoC 5LP provides a number of power supply voltage monitor reset sources. 18.1.1.1 Initial Power-On Reset (IPOR) At initial power on, IPOR monitors the supply voltages VDDD, VDDA, VCCD, and VCCA. The trip level is not precise. It is set to approximately 1 volt, which is below the lowest specified operating voltage but high enough for the internal circuits to be reset and to hold their reset state. The monitor generates a reset pulse that is at least 150 ns wide. It may be much wider if one or more of the voltages ramps up slowly. power up until VDDA, VDDD, VCCA, and VCCD are above the IPOR threshold. It also asserts for several hundred nanoseconds after the asserting edge of a PRES or XRES event. A small excursion on VCCA or VCCD may be seen due to this IPOR pulse. 18.1.1.2 Precise Low-Voltage Reset (PRES) This circuit monitors the VCCX outputs of the analog and digital regulators. The regulator outputs are compared to a precise reference voltage. Software can disable the analog regulator, which also disables the analog portion of PRES. The digital portion of PRES cannot be disabled. PRES disabled automatically in sleep and hibernate modes, with one exception: During sleep mode the regulators are periodically activated (buzzed) to provide supervisory services and reduce wakeup time. At these times PRES is also buzzed to allow periodic regulator monitoring. 18.1.1.3 Low-Voltage Interrupt/Reset (LVI) These circuits detect when VDDA and VDDD go below a programmable voltage, and generate either an interrupt or a reset. See section 15.3.3 Voltage Monitoring on page 131 for details. 18.1.1.4 Hibernate Reset (HRES) During hibernate and sleep, an ultra-low-power supply monitor circuit is used to issue a reset if the voltage on the external digital supply drops to the point that SRAM and register state information may be lost. See the Power Supply and Monitoring chapter on page 125 for details on the configuration of the LVI and HVI reset sources. In addition to power-on, all hard resets start an IPOR sequence. 18.1.2 The analog and digital regulator trim is forced when the IPOR circuit is in reset to provide the regulator a known trim value while the NVLs are loading. The IPOR asserts during External reset is available on a dedicated XRES pin on some devices, as well as a shared GPIO pin P1[2] on all devices. The shared pin supports low pin count parts that do PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F External Reset (XRES) 143 Reset not have a dedicated XRES pin. The shared pin is available through a programmable NV latch setting, and is configured during the boot phase immediately after power-up. See the Nonvolatile Latch chapter on page 85 for details. The external reset is active low; an internal pull-up resistor is provided for both dedicated and shared pins. Either pin (if P1[2] is configured) holds the part in reset while the pin is held low. XRES is active during sleep and hibernate modes. 18.1.3 Software Initiated Reset (SRES) Software initiated reset (SRES) is a mechanism that allows a software-driven reset. The RESET_CR2 register forces a device reset when a ‘1’ is written into bit 0. This setting can be made by firmware or with a DMA. The RESET_SR0[5] status bit is set by a software reset. This bit remains set until cleared by the user or until a hard reset. 18.1.4 Watchdog Reset (WRES) The watchdog timer (WDT) detects errant code by causing a reset if the watchdog timer is not cleared within the userspecified time limit. See the Watchdog Timer chapter on page 141. Enable the WDT by setting the PM_WDT_CFG[4] register bit. After this bit is set, it cannot be cleared again except by a reset event. The RESET_SR0[3] status bit is set by a watchdog reset. This bit remains set until cleared by the user or until a hard reset. 18.1.5 Identifying Reset Sources These two registers have specific status bits allocated for the various reset sources, except POR and XRES. The bits are set on the occurrence of the corresponding reset, and remain set after the reset, until cleared by the user or a device reset occurs due to one of the below mentioned sources. The RESET_SR0 register contains status bits for all soft reset sources. The register also has two general-purpose bits. These bits are persistent through soft resets. They can be cleared by firmware, or by a hard reset. 18.1.5.1 Preservation of Reset Status The device reset caused due to XRES, IPOR, PRES, LVI, and HVI sources clear the contents of the RESET_SR0 and RESET_SR1 registers. These sources are referred to as hard reset sources because they reset all the registers. The remaining reset sources, which include software reset, watchdog reset, and segment reset preserve the status of the RESET_SR0 and RESET_SR1 registers. For example, if an LVI reset and a software reset occur simultaneously, the LVI reset will clear the status bit corresponding to the software reset, making it impossible to detect a software reset condition. Also, the status bits corresponding to PRES, LVI, and HVI in RESET_SR0 and RESET_SR1 registers are meaningless because the respective reset conditions are hard resets, which clear all the register bits. The status bits corresponding to LVI and HVI will however be required when they are configured as interrupt sources instead of reset sources. Note PSoC Creator reads and preserves the RESET_SR0 register during boot before executing the code in main(). Refer to the “Preservation of Reset Status” section of the PSoC 3/PSoC 5LP System Reference Guide for more information. When the device comes out of reset, it is beneficial to know the cause of the reset. This is achieved in the device through the registers RESET_SR0 and RESET_SR1. 144 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Reset 18.2 Reset Diagram Figure 18-1 is a simplified logic diagram of the RESET module. Any active source of reset will make the system reset. Figure 18-1. Logic Diagram of the RESET Module IPOR, POR Hibernate System RESET WRES WRES_ENA Hibernate SRES Hibernate XRES Figure 18-2 shows the operation of various RESETs with the change in Vdd/Vcc. The diagram also shows the functioning of RESETs in a normal power-up. Figure 18-2. Resets Resulting from Various Reset Sources Reset held until XRES is released CPU State Vddd/ Vdda Pin Core Vccd/Vcca Trip Level POR XRES HRES WRES SRES Legend Reset Boot User Code Runs PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 145 Reset 18.3 Reset Summary All reset sources and their triggers/effects are described in Table 18-1 and Table 18-2. Table 18-1. Reset Sources and Triggers POR Trigger Vccd <1.6 V Enable by Default? Block Power Sleep Mode Operation WRES SRES XRES WDT not written in time window RESET_CR2[0] set External XRES pin active Yes No No Yes (nonvolatile latch setting) 50 µA <1 µA 0 0 Buzzed Not in Hibernate No Yes TC RESET IPOR KEY CLEARED GPSW[1:0] CLEARED HV & ANA DISABLED OPENS BOOTWINDOW DEVICE CONFIG RESET TYPE NVL RELOAD SYSTEM RESET Table 18-2. Reset Effects IPOR (Digital and Analog, Internal and External) PRES (Digital and Analog) reset_hard (hard reset) XRES (JTAG disabled or hibernate mode) X X X X X X X X X X X X Hibernate Reset (HBR) Software Reset (SWR) Watchdog Reset (WDR) reset_all (soft reset) XRES (JTAG enabled) checksum reset (CKSUMR) Segment Reset (SEGR) Power Domain Reset Notes ■ PRES, LPCOMP, and HBR can be disabled through RESET_CR4 and RESET_CR5 register settings, but generate resets by default. ■ JTAG is enabled if customer NVL setting in CNVL_DPS[1:0] is set or if Port 1 pins are actively configured to JTAG. 146 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Reset 18.4 Boot Process and Timing The boot process trims and configures the silicon to its ideal state before the first line of the user code is executed. The PSoC® 5LPlife cycle consists of reset, boot, and user phases. Figure 18-3 gives a brief view of these phases. Figure 18-3. Boot Process Reset Holds the part in reset until the operating conditions are stable. NV Latch configuration happens here Boot Configuration loaded from reserved area in Flash. Debug port acquire happens here The process from supply voltage stabilization to user code entry is shown in Figure 18-4. After the voltage is high enough, the NVL data load is initiated. The NVL load takes care of loading configuration data stored in the NV latches. The maximum time for this NVL load is 10 s from the time of initiation. This resets the I/Os to the NVL drive mode settings as well as setting the other Configuration data for the device. At this point, the device enters the reset state. The two types of NVL loads are explained in 18.4.1 Manufacturing Configuration NV Latch. User Mode CPU active. Start running code from Address0 . Loads configuration based on PSoC creator generated code . gling P1_0 and P1_1 implies a debug port acquire is being attempted which must trigger a debug port entry. If the external reset pin (XRES) is asserted low, the device stays in the reset state. If the external reset pin (XRES) is not asserted and all the voltages are at their correct operating values, it triggers the reset hold off circuitry to bring the device out of the reset state. The IMO clock is then started in a fast IMO (FIMO) mode, which is a faster startup version of the IMO. The reset holdoff counter continues to hold the device in reset until other systems, such as band-gap and precision resets stabilize. The length of the hold off is approximately 20 s to allow enough time for these circuits to stabilize. If the band-gap or precision reset blocks are not ready or there is a problem with any of these devices stabilizing by the end of the holdoff counter, a fresh reset cycle is initiated and the hold-off counter is restarted. If there are no problems, the hold-off counter completes and the device is released from reset. After releasing from reset, the IMO is switched to either 12 MHz or 48 MHz, the system bus clock is started and the boot cycle begins. Until now, the bus clock is fed from the FIMO, which has lesser accuracy compared to the IMO. After the reset is released, it moves into the IMO, which is more precise. The boot phase is explained in section 18.4.3 User Mode. During this boot configuration time, if there is no toggling of the external pins P1_0 and P1_1 and the configuration finishes, the system moves into the user mode. Tog- PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 147 Reset Figure 18-4. Power Up Reset Boot User Mode Cycle Reset Internal Boot User Mode Vddd / Vdda Pins Core Vccd/ Vcca XRES IPORXA NVL Data reset_ holdoff_ counter NV Manufacturing Configuration Valid counting count = 0 FIMO Counter expired IMO, Either12 MHz or48MHz IMO ~20us BG + Precision RESET READY System_ resets clk_bus Boot Configuration Data Configuration reads Configuration writes Checksum Done Boot Window Open In this phase, two types of NV latches are loaded to set reset states and trims in the device. The two types of the configuration, explained in sections 18.4.1 Manufacturing Configuration NV Latch and 18.4.1.1 Device Configuration NV Latch, occur simultaneously in the reset phase. 18.4.1 Manufacturing Configuration NV Latch There are some circuits that must receive part specific trim values before the device comes out of reset. Manufacturing NV latches provide these trim values. An example of this circuit is the power-on-reset. This circuit is responsible for holding the device in reset until a safe supply voltage is reached. The POR circuit requires a trim value, which is stored in an NV latch. The NV latch's output is stable at approximately 1 V while the lowest operating voltage in the PSoC® 5LPplatform is 1.71 V. 18.4.1.1 Device Configuration NV Latch Device configuration is similar to manufacturing configuration NV in that it occurs while the device is in reset; however, it differs in that customers select optional configuration settings not trim values for circuits. Manufacturing configuration and device configuration occur in parallel. One example of a device configuration is the NV latches that determine the I/O 148 drive modes during reset, which determine the reset state of the drive mode registers. 18.4.2 Boot Phase Though many device settings are done using NV latch during the preboot process, there are other trim values that require to be written during the boot process. These values are stored in reserved space in the flash memory (I/O System chapter on page 151) and the boot process takes care of moving this data to the corresponding blocks. This loading of the configuration happens using the DMA and PHUB. A DMA channel fetches the configuration bytes from the flash and places them in the SRAM. This move to SRAM causes indeterminate data to be present in SRAM addresses 0 through 127 after boot is complete. The checksum block does a checksum to determine integrity. After the data is verified, it is then transferred using the DMA to the corresponding configuration register. If the checksum fails, it triggers a system reset. Note that some circuits have mode dependent trim values, for example the IMO's trim value depends on the speed setting of the IMO. For circuits with mode dependent trim values, the boot process loads the trim value that matches the default mode. When the user's firmware or configuration changes the mode, the firmware also retrieves the correct PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Reset trim value corresponding to the modes from the tables stored in flash and writes them to the appropriate register. ■ The CPU halts until boot completes, therefore, you cannot use the CPU to complete the boot process. The PHUB, DMA, and a special checksum block are used to move the manufacturing configuration data from the flash to the appropriate registers. These three blocks work together to accomplish these objectives: When the boot process is complete, the device enters the user mode where the user code starts executing. ■ Minimize boot time, giving you the quickest path to firmware execution ■ Provide a data integrity check on the manufacturing configuration data 18.5 Provide flexibility in the order and addresses to which manufacturing configuration data is written 18.4.3 User Mode When the boot phase is complete, the device enters the user mode to enable firmware code execution. This is where code execution starts for the startup/configuration code developed by PSoC Creator. Only after executing this part of the PSoC Creator generated code does the code execution reach the main(). Register List Table 18-3. Reset Register List Register Name Comments RESET_CR2 RESET_SR0 Persistent status bits for WRES, SRES, XRES, and so on RESET_SR1 Persistent status bits for Segment reset, PRES RESET_SR2 Real-time Reset Status PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 149 Reset 150 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 19. I/O System The I/O system provides the interface between the CPU core and peripheral components to the outside world. The flexibility of PSoC® devices and the capability of its I/O to route any signal to any pin greatly simplifies circuit design and board layout. There are two types of I/O pins on every device, general purpose I/O (GPIO) and special I/O (SIO); those with USB provide a third type. Both GPIO and SIO provide similar digital functionality. The primary differences are their analog capability and drive strength. Devices that include USB also provide two USBIO pins that support specific USB functionality as well as specialized general purpose capability. All I/O pins are available for use as digital inputs and outputs for both the CPU and digital peripherals. In addition, all I/O pins can generate an interrupt. All GPIO pins can be used for analog input, CapSense®, and LCD segment drive, while SIO pins are used for voltages in excess of Vdda and for programmable output voltages and input thresholds. 19.1 Features The PSoC I/O system has these features, depending on the pin type. Supported by both GPIO and SIO pins: ■ User programmable I/O state and drive mode on device reset ■ Flexible drive modes ■ Support level and edge interrupts on pin basis ■ Slew rate control ■ Supports CMOS and low voltage TTL input thresholds ■ Separate port read and write registers ■ Separate I/O supplies and voltages for up to four groups of I/O Provided only on the GPIO pins: ■ Supports LCD drive ■ Supports CapSense ■ Supports JTAG interface ■ Analog input and output capability ■ 8 mA sink and 4 mA source current ■ Ports can be configured to support EMIF address and data Provided only on SIO pins: ■ Hot swap capability (5 V tolerance at any operating Vdd) ■ Single enable and differential input with programmable threshold ■ Regulated output voltage level option ■ Overvoltage tolerance up to 5.5 V ■ Higher drive strength than GPIO ■ 25 mA sink and 4 mA source current PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 151 I/O System USBIO features: ■ USB 2.0 compliant I/O ■ 25 mA source/24 mA sink current 19.2 Block Diagrams Figure 19-1, Figure 19-2 on page 153, and Figure 19-3 on page 153 are block diagrams of three main categories of I/Os: GPIO, SIO, and USBIO, respectively. Each diagram emphasizes the main blocks that drive the system, as well as the signals and register settings that control the main blocks. Figure 19-1. GPIO Block Diagram Digital Input Path Naming Convention ‘x’ = Port Number ‘y’ = Pin Number PRT[x]CTL PRT[x]DBL_SYNC_IN PRT[x]PS Digital System Input PICU[x]INTTYPE[y] Input Buffer Disable PICU[x]INTSTAT Interrupt Logic Pin Interrupt Signal PICU[x]INTSTAT Digital Output Path PRT[x]SLW PRT[x]SYNC_OUT Vddio Vddio PRT[x]DR 0 In Digital System Output 1 Vddio PRT[x]BYP Drive Logic PRT[x]DM2 PRT[x]DM1 PRT[x]DM0 Bidirectional Control PRT[x]BIE Analog Slew Cntl PIN OE 1 0 1 Capsense Global Control 0 1 CAPS[x]CFG1 Switches PRT[x]AG Analog Global PRT[x]AMUX Analog Mux LCD Display Data PRT[x]LCD_COM_SEG Logic & MUX PRT[x]LCD_EN LCD Bias Bus 152 5 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I/O System Figure 19-2. SIO Block Diagram Digital Input Path PRT[x]SIO_HYST_EN PRT[x]SIO_DIFF Reference Level PRT[x]DBL_SYNC_IN Naming Convention ‘x’ = Port Number ‘y’ = Pin Number Buffer Thresholds PRT[x]PS Digital System Input PICU[x]INTTYPE[y] Input Buffer Disable PICU[x]INTSTAT Interrupt Logic Pin Interrupt Signal PICU[x]INTSTAT Digital Output Path Reference Level PRT[x]SIO_CFG PRT[x]SLW PRT[x]SYNC_OUT Driver Vhigh PRT[x]DR 0 Digital System Output In 1 PRT[x]BYP Drive Logic PRT[x]DM2 PRT[x]DM1 PRT[x]DM0 Bidirectional Control PRT[x]BIE Slew Cntl PIN OE Figure 19-3. USBIO Block Diagram D ig ita l In p u t P a th N a m in g C o n v e n tio n ‘y ’ = P in N u m b e r U S B R e c e iv e r C irc u itry P R T [1 5 ]D B L _ S Y N C _ IN P R T [1 5 ]P S [6 ,7 ] U S B IO _ C R 1 [0 ,1 ] D ig ita l S y s te m In p u t P IC U [1 5 ]IN T T Y P E [y ] P IC U [1 5 ]IN T S T A T In te rru p t L o g ic P in In te rru p t S ig n a l P IC U [1 5 ]IN T S T A T D ig ita l O u tp u t P a th P R T [1 5 ]S Y N C _ O U T U S B IO _ C R 1 [5 ] U S B o r I/O U S B IO _ C R 1 [2 ] V ddd U S B S IE C o n tro l fo r U S B M o d e P R T [1 5 ]D R 1 [7 ,6 ] D ig ita l S y s te m O u tp u t P R T [1 5 ]B Y P In 1 D riv e L o g ic D+ Open D ra in P R T [1 5 ]D M 0 [7 ] D- Open D ra in P R T [1 5 ]D M 1 [7 ] V ddd V ddd V ddd 0 P R T [1 5 ]D M 0 [6 ] P R T [1 5 ]D M 1 [6 ] D + p in o n ly D + 1 .5 k 5 k 1 .5 k P IN D+ 5 k D- 5 k PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 153 I/O System 19.3 How It Works 19.3.1 PSoC I/Os provide: ■ Digital input sensing ■ Digital output drive ■ Pin interrupts ■ Connectivity for analog inputs and outputs ■ Connectivity for LCD segment drive and EMIF ■ Access to internal peripherals: Usage Modes and Configuration Because of the variety of I/O capabilities, it is necessary to understand the modes thoroughly and the configuration for each function. ❐ Directly for defined ports ❐ Through the universal digital blocks (UDB) via the Digital System Interconnect (DSI) 19.3.2 I/O Drive Modes Each GPIO and SIO pin is individually configurable into one of the eight drive modes listed in Table 19-1 and shown in Figure 19-4, which depicts a simplified pin view based on each of the eight drive modes. The I/O pin drive state is based on the port data register value (DR) or on a DSI signal, if bypass mode is selected. The actual I/O pin voltage is determined by a combination of the DR value, the selected drive mode, and the load at the pin. The state of the pin can be read from the Port Status register (PS) or routed to a DSI signal, or both. Three configuration bits are used for each pin (DM[2:0]) and set in the PRTxDM[2:0] registers. When the drive mode of a pin is changed, it is possible that the input buffer may be turned-off for a short period during the drive mode transition. Therefore, pin interrupts should be disabled while changing pin configuration. The I/Os are arranged into ports, with up to eight pins per port. Some of the I/O pins are multiplexed with special functions (USB, debug port, crystal oscillator). Special functions are enabled using control registers associated with the specific functions. For example, the Crystal Oscillator control register enables the crystal oscillator function for the I/O pin multiplexed with the crystal oscillator function. Table 19-1. I/O Drive Modes Mode Drive Mode Number 154 PRTxDM2 PRTxDM1 PRTxDM0 DM2 DM1 DM0 Data = 1 Data = 0 High Z 0 High Impedance Analog 0 0 0 High Z 1 High Impedance Digital 0 0 1 High Z High Z 2 Resistive Pull Up 0 1 0 Res 1 (5k) Strong 0 3 Resistive Pull Down 0 1 1 Strong 1 Res 0 (5k) 4 Open Drain, Drives Low 1 0 0 High Z Strong 0 5 Open Drain, Drives High 1 0 1 Strong 1 High Z 6 Strong Drive 1 1 0 Strong 1 Strong 0 7 Resistive Pull Up and Down 1 1 1 Res 1 (5k) Res 0 (5k) PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I/O System Figure 19-4. I/O Drive Mode Diagram Vdd Out In Pin Out In Pin Out In Vdd Pin Out In Pin An An An An 0. High Impedance Analog 1. High Impedance Digital 2. Resistive Pull Up 3. Resistive Pull Down Vdd Out In Pin Out In Vdd Pin Out In Vdd Pin Out In Pin An An An An 4. Open Drain, Drives Low 5. Open Drain, Drives High 6. Strong Drive 7. Resistive Pull Up and Pull Down The ‘Out’ connection is driven from either the Digital System (when the Digital Output terminal is connected) or the Data Register (when HW connection is disabled). The ‘In’ connection drives the Pin State register, and the Digital System if the Digital Input terminal is enabled and connected. The ‘An’ connection connects to the Analog System. 19.3.2.1 Drive Mode on Reset The factory drive mode default is high impedance analog mode, which is appropriate for most designs. The Drive Mode on Reset feature allows the user to change the factory default to any of the four listed drive modes if the application requires faster configuration to low or high logic levels. The Reset drive mode is set at POR release. The Drive Mode on Reset setting is a port wide setting and is not set per pin. Each pin is individually configured during the device configuration step after POR release; this setting overwrites the reset drive mode. The Resistive Pull Up Drive Mode on Reset also sets the Port Data Register to 0xFF to ensure the port is pulled up; all other modes leave the Data Register 0x00. ■ High impedance analog ■ High impedance digital ■ Resistive pull up ■ Resistive pull down See the Nonvolatile Latch chapter on page 85 for details. 19.3.2.2 High Impedance Analog High Impedance Analog mode is the default reset state; both output driver and digital input buffer are turned off. This state prevents a floating voltage from causing a current to PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F flow into the I/O digital input buffer. This drive mode is recommended for pins that are floating or that support an analog voltage. High impedance analog pins cannot be used for digital inputs. Reading the pin state register returns a 0x00 regardless of the data register value. To achieve the lowest device current in sleep modes, all I/ Os must either be configured to the high impedance analog mode, or they must have their pins driven to a power supply rail (ground) by the PSoC device or by external circuitry. 19.3.2.3 High Impedance Digital High Impedance Digital mode is the standard high impedance (High Z) state recommended for digital inputs. In this state, the input buffer is enabled for digital signal input. 19.3.2.4 Resistive Pull Up or Resistive Pull Down Resistive modes provide a series resistance in one of the data states and strong drive in the other. Pins can be used for digital input and output in these modes. Interfacing to mechanical switches is a common application for these modes. If a pull up is needed with the Resistive Pull Up Drive mode, a 1 must be written to that pin’s Data Register bit. If a pull down is required with the Resistive Pull Down 155 I/O System Drive mode, a 0 must be written to that pin’s Data Register bit. 19.3.2.5 Open Drain, Drives High and Drives Low Open Drain modes provide high impedance in one of the data states and strong drive in the other. Pins are used for digital input and output in these modes. A common application for these modes is driving I2C bus signal lines. Outputs are driven from the CPU by writing to the port data registers (PRTx_DR) Digital inputs are read by the CPU through the pin state registers (PRTx_PS}). 19.3.4.1 Port Configuration Registers Table 19-2 lists port control registers. Table 19-2. Functional Registers Accessed through Pin and Port Configuration Registers Address 19.3.2.6 Description Strong Drive The Strong Drive mode is the standard digital output mode for pins; it provides a strong CMOS output drive in both high and low states. Strong drive mode pins must not be used as inputs under normal circumstances. This mode is often used to drive digital output signals or external FETs. PRT[0..11]_BYP A bit set in this register connects the corresponding port pin to the Digital System Interconnect (DSI), and disconnects it from the DR register. PRT[0..11]_SLW Each bit controls the output edge rate of the corresponding port pin – fast edge rate mode (Slew=0) or slow edge rate mode (Slew=1) Each bit set controls the bidirectional mode of the corresponding port pin. PRT[0..11]_BIE 19.3.2.7 The Resistive Pull Up and Pull Down mode is a single mode and is similar to the Resistive Pull Up and Resistive Pull Down modes, except that, in the single mode, the pin is always in series with a resistor. The high data state is pull up while the low data state is pull down. This mode is used when the bus is driven by other signals that may cause shorts. 19.3.3 Slew Rate Control GPIO and SIO pins have fast and slow output slew rate options for strong drive modes – not resistive drive modes. The fast slew rate is for signals between 1 MHz and 33 MHz. Because it results in reduced EMI, the slow option is recommended for signals that are not speed critical – generally less than 1 MHz. Slew rate is individually configurable for each pin and is set by the PRTxSLW registers. 19.3.4 0 = Output always enabled Resistive Pull Up and Pull Down 1 = Output Enable controlled by DSI input PRT[0..11]_PS This register reads the logical pin state for the corresponding GPIO port. PRT[0..11]_DM[0..2] The combined value of these registers – PRTx_DM2, PRTx_DM1, and PRTx_DM0 – determines the unique drive mode of each pin in a GPIO port. PRT[0..11]_DR Data written to this register specifies the high (Data=1) or low (Data=0) state for the GPIO pin at each bit location of the selected port. 19.3.4.2 Pin Wise Configuration Register Alias The port pin configuration registers (PRTxPC0 through PRTxPC7) access several configuration or status bits of a single I/O port pin at once, as shown in Figure 19-5 on page 157. Figure 19-5 shows an example of a read from {PRT*_PC[4]}. Bit four of the port control registers associated with the port configuration register is read and driven onto the data bus. Digital I/O Controlled by Port Register The Port Control registers (see Table 19-2 on page 156) have separate configuration bit for each port pins. In addition to port control registers, the device also provides register for port-wide and pin wise configuration. The port wide configuration register writes the same configuration for all the port pins in a single write. This is useful to configure all the port pins to a specific configuration. The pin wise configuration register writes to all configuration bits for a specific I/O pin in a single write. This is useful to configure individual port pins to a specific configuration. 156 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I/O System Figure 19-5. Effect of a Read of the Pin Configuration Register {PRT*_PC[4]} Bit7 Bit6 Bit5 Bit4 Bit3 Bit2 Bit1 Bit0 Data Register Bypass – (Port 3 BYP) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Slow Slew Rate – (Port 3 SLW) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Bidirectional Enable – (Port 3 BIE) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Pin Input State – (Port 3 PS) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Drive Mode 2 – (Port 3 DM2) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Drive Mode 1 – (Port 3 DM1) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Drive Mode 0 – (Port 3 DM0) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Data Output – (Port 3 DR) Pin 7 Pin 6 Pin 5 Pin 4 Pin 3 Pin 2 Pin 1 Pin 0 Port Pin Configuration – Port 3, Pin 2 19.3.4.3 BYP SLW Port Wide Configuration Register Alias BIE PS DM2 DM1 DM0 DR Figure 19-6. {PRT*.PRT} Write Example Write Data bus The Port Configuration Register accesses several available configuration registers on a port-wide basis with a single bit write. 8 write PRT[x].PRT bit [1] This register PRT*_PRT aliases a subset of the configuration registers, allowing the user to configure a complete port in a single write. 1 PRT[x].DM0 7 0 bit [2] 1 PRT[x].DM1 7 0 bit [3] 1 PRT[x].DM2 7 0 bit [5] 1 PRT[x].BIE 7 0 bit [6] 1 PRT[x].SLW 7 0 bit [7] 1 PRT[x].BYP PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 7 0 157 I/O System 19.3.5 Digital I/O Controlled Through DSI GPIO, USBIO, and SIO pins are connected to the internal peripheral blocks through the UDB via the digital system interconnect (DSI). Any peripheral connected to the UDB can be connected to any I/O pin through the DSI. Each port has 20 unique connections to the UDB through DSI: eight inputs, eight outputs, and four output control signals. 19.3.5.1 DSI Output The bypass register {PRTx_BYP} selects either the selected DSI output signal or the data register (PRTx_DR) to drive the port pin. Mapping of the DSI signal to the output pin is illustrated in Figure 19-7 on page 158. Together, output select registers PRTx_OUT_SEL1 and PRTx_OUT_SEL0 select the DSI output signal to drive the corresponding output port pin. Figure 19-7. Digital System Input to Pad Selection PRT[x].OUT_SEL1[4],PRT[x].OUT_SEL0[4] PRT[x]_OUT_SEL1[5], PRT[x]_OUT_SEL0[5] PRT[x]_OUT_SEL1[6], PRT[x]_OUT_SEL0[6] DSI[7] DSI[6] DSI[5] DSI[4] PRT[x]_OUT_SEL1[7], PRT[x]_OUT_SEL0[7] Upper Nibble DSI IN DSI[3] DSI[2] DSI[1] DSI[0] PRT[x]_OUT_SEL1[1], PRT[x]_OUT_SEL0[1] PRT[x]_OUT_SEL1[0], PRT[x]_OUT_SEL0[0] PRT[x]_OUT_SEL1[3], PRT[x]_OUT_SEL0[3] PRT[x]_OUT_SEL1[2], PRT[x]_OUT_SEL0[2] Lower Nibble DSI IN PRT[x]_DR[7:0] Port Logic Control PRT[x]_BYP[7:0] 19.3.5.2 in in in in in in in in GPIO GPIO GPIO GPIO GPIO GPIO GPIO GPIO Px[0] Px[1] Px[2] Px[3] Px[4] Px[5] Px[6] Px[7] DSI Input The port pin input is directly connected to the UDB array through DSI for routing the input to various internal peripheral blocks. The control for these port inputs are at the DSI inputs. See the Universal Digital Blocks (UDBs) chapter on page 175 for port-to-DSI connections. 19.3.5.3 DSI for Output Enable Control High-speed bidirectional capability is provided through the {PRT*_BIE} register. When this mode is enabled and the auxiliary control signal is high, the I/O pin immediately goes 158 into a High Z output drive state with input buffer enabled. When this signal is low (or returns low), the I/O pin assumes the pin state configured through the {PRT*_DM[2]}, {PRT*_DM[1]}, and {PRT*_DM[0]} registers. This allows fast turnaround of the I/O pin. Four DSI control signals are available for dynamic drive control of the pins. Mapping of the DSI control signal to port pin output enable is shown in Figure 19-8 on page 159. Together, dynamic output enable select registers PRTx_OE_SEL1 and PRTx_OE_SEL0 select the DSI control signal for each port pin. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I/O System Figure 19-8. Mapping of DSI Control Signal to Port Pin Output Enable from UDB dsi_oe[3] dsi_oe[2] dsi_oe[1] dsi_oe[0] PRT[x]_OE_SEL1[1], PRT[x]_OE_SEL0[1] PRT[x]_OE_SEL1[3], PRT[x]_OE_SEL0[3] PRT[x]_OE_SEL1[2], PRT[x]_OE_SEL0[2] PRT[x]_OE_SEL1[5], PRT[x]_OE_SEL0[5] PRT[x]_OE_SEL1[4], PRT[x]_OE_SEL0[4] PRT[x]_OE_SEL1[7], PRT[x]_OE_SEL0[7] PRT[x]_OE_SEL1[6], PRT[x]_OE_SEL0[6] PRT[x]_OE_SEL1[0], PRT[x]_OE_SEL0[0] Dynamic Output Control PORT LOGIC CONTROL PRT[x]_BIE[7:0] 19.3.6 oe oe oe oe oe oe oe oe GPIO GPIO GPIO GPIO GPIO GPIO GPIO GPIO Px[0] Px[1] Px[2] Px[3] Px[4] Px[5] Px[6] Px[7] Analog I/O The only way that analog signals can pass to and from the PSoC core is through GPIO. To connect a pin to an internal analog resource through analog global bus or analog mux line, each GPIO connects to one of the analog global lines and to one of the analog mux lines. The switches that connect the I/O pin to Analog global lines and analog mux line are configured by the {PRT*_AG} and {PRT*_AMUX} registers. See the Analog Routing chapter on page 319 for a description of the analog global network configuration. Selected pins provide direct connections to specific analog features, such as DACs or uncommitted opamps. For analog I/O pins, the drive mode should be configured to High Z Analog in most situations, which disables the input buffer. The input buffer can also be disabled using the port input disable (PRTx_INP_DIS) register. The buffer should remain enabled to allow simultaneous use of the pin as a digital input and analog input or output. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 19.3.7 LCD Drive All GPIO pins can be configured for LCD drive capabilities. {PRT*_LCD_EN} registers are used to enable individual pins for LCD drive. {PRT*_LCD_COM_SEG} registers are used to select whether a pin is set as a common or segment drive pin. In LCD mode, the GPIO pins are configured into a High Z output mode, allowing the LCD drivers to control the pin state. 19.3.8 CapSense All GPIO pins can be used to create CapSense buttons and sliders. The primary analog bus for CapSense is the AMUXBUS, which has two nets (AMUXBUSL and AMUXBUSR) for two simultaneous sensing operations. These can also be shorted to form a single net that connects to all GPIOs. See the CapSense chapter on page 359 for more information. 159 I/O System 19.3.9 External Memory Interface (EMIF) Figure 19-9. SIO Configuration Diagram SIO Pair The EMIF uses the port interface and the UDB to connect to external memory. When in EMIF mode, the ports directly pass to the pads the address and data out from the PHUB. Data reads from the EMIF pass through the port to the PHUB. See the EMIF chapter on page 97 for more information. 19.3.10 SIO PRT[x]SIO_DIFF ANALOG Global SIO Functions and Features PRT[x]SIO_CFG GPIO and SIO provide similar digital functionality. The primary differences are in their analog capability and drive strength. This section describes adjustable input and output level and hot swap features that are available only with SIO. 19.3.10.1 SIO Regulated Output Level 19.3.10.2 SIO port pins support the ability to provide a regulated high output level. This can be useful for interfacing to external signals with voltages lower than the SIO Vddio. This regulated output sets the Voh for the SIO pair. The SIO are grouped into pairs. Each pair shares the same reference generator, thus the regulated output level applies for both pins. Table 19-4. SIO Differential Input Buffer Reference Voltage Selection vref_sel[y] vtrip_sel[y] Mode Description 0 0 0.5 × vddio 0 1 0.4 × vddio 1 0 0.5 × vohref 1 1 vohref Table 19-3. SIO Input and Output Configuration ibuf_sel[y] 0 0 Single Ended Input Buffer Non-Regulated Output Buffer 0 1 Differential Input Buffer Non-Regulated Output Buffer 1 0 Single Ended Input Buffer Regulated Output Buffer 1 1 Differential Input Buffer Regulated Output Buffer Adjustable Input Level SIO pins support a differential input mode with programmable thresholds. The SIO pair input buffer voltage levels are set by the vref_sel and vtrip_sel bits of the {PRT*_SIO_DIFF} register. See the following table. Configuration is provided for each SIO pair through the {PRT*_SIO_CFG} registers, as shown in the following table. vreg_en[y] Reference Generator Mode Description Figure 19-10. SIO Reference Voltage vddio vddio ibuf_sel *(vohref | vref_sel) vreg_en | (ibuf_sel*vref_sel) vohref (From Analog Global) 5 R R voutref (To Output Buffer) 0 R 1 4 R vtrip_sel R vgnd 0 1 0 vinref (To Input Buffer) vref_sel 1 vtrip_sel 160 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I/O System 19.3.10.3 ■ Hot Swap USB SIO pins support hot swap capability. It is possible to connect to another system without loading the signals connected to the SIO pins and without applying power to the PSoC device. Special functions and peripherals such as I2C, crystal oscillators, USB, XRES, JTAG TAP, SWD, high-current DAC outputs, VREF inputs, and high drive analog output buffers have fixed pin assignments. The unpowered PSoC device can maintain a high impedance load to the external device while preventing the PSoC device from being powered through a GPIO pin’s protection diode. The I2C block supports three pin assignment options: SIO pin pair P12[0:1], SIO pin pair P12[4:5], or any GPIO / SIO pin pair routed via the DSI. 19.3.11 System reset (XRES, active low, resistive pull up) functionality is supported on either the dedicated XRES pin or the P1[2] GPIO. The IEEE 1149.1 JTAG TAP five pin interface may be enabled on the P1[0:1,3:5] pins. Special Functionality Special purpose capability may uniquely exist on some pins such as: ■ 4 to 25 MHz crystal input and output ■ 32 kHz crystal input and output ■ Test modes ■ I2C ■ SPI ■ CAN Serial wire debug is supported over the USBIO pins (P15[6:7]) or the same pins as TMS / TCK (P1[0:1]). Analog function fixed pin assignments include two pairs of VIDAC outputs to support high-current mode, two VREF inputs, and four sets of analog output buffer pins. The “left side” VIDAC and analog buffer pins are assigned to port 0 and are available on all package options. The “right side” VIDAC and linear buffer pins are assigned to port 3 and are available on all package options. Table 19-5. Fixed Pin Assignments Function I2C MHz ECO 32 kHz ECO FS USB XRES IEEE 1149.1 JTAG TAP Signal Name Pad # Pad Name Pad Type SCL 4 P12[4] SIO 4 3 H6 SDA 5 P12[5] SIO 5 4 K7 99 CSP SCL 61 P12[0] SIO 53 38 C3 SDA 62 P12[1] SIO 54 39 C4 Xo 49 P15[0] GPIO / Xtal 42 27 D1 Xi 50 P15[1] GPIO / Xtal 43 28 D2 Xo 62 P15[2] GPIO / Xtal 55 40 E3 Xi 63 P15[3] GPIO / Xtal 56 41 E4 D+ 39 P15[6] USBIO 35 22 G2 F2 D- 40 P15[7] USBIO 36 23 19 XRES XRES 15 10 – 26 P1[2] GPIO 22 13 H3 TMS 24 P1[0] GPIO 20 11 H4 J3 XRES TCK 25 P1[1] GPIO 21 12 TDO 26 P1[3] GPIO 23 14 J2 TDI 28 P1[4] GPIO 24 15 G4 nTRST 29 P1[5] GPIO 25 16 G3 SWDIO Serial Wire Debug TQPF 100 QFN 68 SWDCK SWO Comment SIO pair on Vio2 SIO pair on Vio3 Fixed function XRES/TSTRST pin XRES/TSTRST 24 P1[0] GPIO 20 11 H4 SWD on GPIO pins option 39 P15[6] USBIO 35 22 G2 SWD on USB pins option 25 P1[1] GPIO 21 12 J3 SWD on GPIO pins option 40 P15[7] USBIO 36 23 F2 SWD on USB pins option 27 P1[3] GPIO 23 14 J2 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 161 I/O System Table 19-5. Fixed Pin Assignments (continued) Function VIDAC High Current Output External Vref Analog Linear Output Buffer 19.3.12 Signal Name Pad # Pad Name Pad Type Abuffer0L 82 P0[6] GPIO 83 P0[7] GPIO 79 56 D8 P3[0] GPIO 44 29 C1 Abuffer1R 52 P3[1] GPIO 45 30 C2 Extref0 78 P0[3] GPIO 74 51 D6 D3 Extref1 53 P3[2] GPIO 46 31 Abuf0+ 77 P0[2] GPIO 73 50 A8 Abuf0- 78 P0[3] GPIO 74 51 D6 Abuf0out 76 P0[1] GPIO 72 49 C6 B4 Abuf1- 55 P3[4] GPIO 48 33 Abuf1+ 56 P3[5] GPIO 49 34 A2 Abuf1out 58 P3[6] GPIO 51 36 B2 Abuf2+ 80 P0[4] GPIO 76 53 D7 E7 Abuf2- 81 P0[5] GPIO 77 54 Abuf2out 75 P0[0] GPIO 71 48 B6 Abuf3- 53 P3[2] GPIO 46 31 D3 Abuf3+ 54 P3[3] GPIO 47 32 D4 Abuf3out 59 P3[7] GPIO 52 37 B3 I/O Port Reconfiguration 2. The software writes this value into the data registers, {PRT*_DR}. I/O ports driven by the DSI must be driven by the data register by de-asserting the bypass register value, {PRT*_BYP}. At this point, it is safe to reconfigure the device. When reconfiguration is complete, the I/O sources can be driven by the DSI by setting the {PRT*_BYP} register value. Power Up I/O Configuration By default, all I/Os power up in a known state, either driving a 0, driving a 1, or set to High Z. Input buffers are disabled during power up. The value set in the nonvolatile (NV) latches determines the value driving each port. Comment B9 51 1. The software reads the GPIO / SIO pin state, {PRT*_PS}. 162 55 Abuffer1L However, if the ports are bypassed and driven by the DSI, the current value must be read and written to the data register ({PRT*_DR}) before initiating reconfiguration. Saving of the current configuration occurs as follows: 19.3.13 78 99 CSP Abuffer0R Care must be taken not to lose the current configuration during reconfiguration of pins when the device is connected directly to a digital peripheral. The I/O pins should hold their current configurations during a reconfiguration. If the ports are driven by the data registers, configuration maintenance is automatic. 3. TQPF 100 QFN 68 A pair of NV latches is associated with each I/O port; these latches serve two functions: ■ Latch values configure the pins on a port-wide basis during power up. ■ Latch values load reset values for the drive mode and data registers to correctly configure the port, when IPOR_disabled is deasserted. See the Nonvolatile Latch chapter on page 85 for more information. If the NVLs are set to 0x00 for the port, by default all I/Os reset to the High Impedance Analog state but are reprogrammable on a port-by-port basis. They can be reset as High Impedance Analog, Pull Down, or Pull Up, based on the requirements of the application. 19.3.14 Overvoltage Tolerance All I/O pins provide an overvoltage (Vddio < Vin < Vdda) tolerance feature at any operating voltage. Limitations include the following: ■ No current limitations for the SIO pins, because they present a high impedance load to the external circuit. ■ GPIO pins must be limited to 100 µA, using a current limiting resistor. Outside the current limitation, GPIO pins clamp the pin voltage to approximately one diode above the Vddio supply. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I/O System A common application for this feature is connection to a bus such as I2C, where different devices are running from different supply voltages. In the I2C case, the PSoC device is configured into the Open Drain, Drives Low mode using an SIO pin. This allows an external pull up to pull the I2C bus voltage above the pin’s Vddio supply. For example, the PSoC device can operate at 1.8 V, and an external device can run from 5 V. The SIO pin’s VIH and VIL levels are determined by the associated Vddio supply pin. The I/O pin must be configured into a High Impedance drive mode, Open Drain Low mode, or Resistive Pull Down mode, for overvoltage tolerance to work properly. Absolute maximum ratings for the device must be observed for all I/O pins. 19.3.15 SIO port pins support an additional regulated high output capability, as discussed in 19.3.10.2 Adjustable Input Level. Sleep Mode Behavior Port Interrupt Controller Unit This section describes the functions of the port interrupt controller unit (PICU) for PSoC I/O. 19.4.1 Features The features of the PICU are as follows: ■ All eight pins in each port interface with their own PICU and associated interrupt vector ■ Pin status bits provide easy determination of interrupt source down to the pin level ■ Rising/falling/either edge interrupts are handled ■ Pin interrupts can be individually enabled or disabled ■ Interfaces to the PHUB for read and write into its registers ■ Sends out a single interrupt request (PIRQ) signal to the interrupt controller I/O Power Supply The Vddio supply must be less than or equal to the voltage on the device’s Vdda pin. This feature allows users to provide different I/O interface levels for different pins on the device. See the datasheet to determine Vddio capability for a given device and pin. 19.3.16 19.4 19.4.2 Interrupt Controller Block Diagram Figure 19-11 is a block diagram of the PICU showing the function of control signal generation and data manipulation blocks. These blocks send appropriate control signals to interrupt-generating pin logic blocks, simultaneously recording these signals in status and snap registers. The GPIO/SIO pad will maintain the current pin state during sleep modes. Port pin interrupts remain active in all sleep modes, allowing the PSoC device to wake from an externally generated interrupt. 19.3.17 Low-power Behavior In all low-power modes, I/O pins retain their states until the part is awakened and changed or reset. To awaken the part, use a pin interrupt, because the port interrupt logic continues to function in all low-power modes. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 163 I/O System Figure 19-11. PICU Block Diagram From GPIO Pin From GPIO Pin From GPIO Pin From GPIO Pin From GPIO Pin From GPIO Pin From GPIO Pin From GPIO Pin Pin 0 Logic Pin1 Logic wakeup_in wakeup_out Pin 2 Logic Pin 3 Logic Pin 4 Logic Logical OR PIRQ To Interrupt Controller Input Pin 5 Logic Pin 6 Logic Pin 7 Logic Status Register PHUB PHUB Interface Snap Shot Register 19.4.3 Function and Configuration Each pin of the port can be configured independently to generate interrupt on rising edge, falling edge, or either edge. Level sensitive interrupts are not directly supported. UDB provides this functionality to the system when needed. This configuration is done by writing into the interrupt type register corresponding to each pin. The sequence is as follows: together with other pin interrupts to generate a wakeup_out signal, as shown in Figure 19-11. 5. All of the PICUs are daisy chained together to generate a final wakeup signal that goes to the power manager. 1. Depending on the configured mode for each pin, whenever the selected edge occurs on a pin, its corresponding status bit in the status register is set to ‘1’, and an interrupt request is sent to the interrupt controller. 2. Status bits that have ‘1’ are cleared upon a read of the status register. Other bits of the status register can still respond to incoming interrupt sources. 3. If an interrupt is pending, and the status register is being read, all of the incoming events on the same interrupt source (GPIO) are blocked until the read is complete. However, all of the other interrupt sources that were not pending an interrupt in status register are not blocked. 4. Each PICU has a wakeup_in input and a wakeup_out output signal. The wakeup_in signal in a PICU is ORed 164 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I/O System 19.5 Register Summary Registers shown in Table 19-6 are associated with a single I/O port and are specific to both the GPIO and SIO ports. Table 19-6. GPIO and SIO Port Registers Address {PRT*_DR} Table 19-7 shows registers specific to a GPIO port. Table 19-7. GPIO Registers Address {PRT*_CTL} Port-wide configuration register. This contains the portEmifCfg[2:0] and port-wide vtrip_sel for the corresponding GPIO register. {PRT*_LCD_COM_SEG} LCD com_seg setting. This selects common or segment mode when the LCD is enabled. {PRT*_LCD_EN} LCD enable, allows port pins not connected to LCD to be used for other functions. Description The Port Data Output register sets the data output state for the corresponding GPIO port. It is aliased to continuous address space in the PRT*_DR_ALIASED registers. Description {PRT*_PS} The Port Pin State register reads the logical pin state for the corresponding GPIO port. It is aliased to continuous address space in the PRT*_PS_ALIASED registers. {PRT*_DM*} The Port Drive Mode registers ({PRT*_DM[0]}, {PRT*_DM[1]}, and (PRT*_DM[2]}) specify the drive mode for I/O pins. {PRT*_SLW} The Port Slew Control register sets the slew rate for pin outputs. {PRT*_SIO_DIFF} Differential input buffer reference voltage select, 2 bits per SIO pair. {PRT*_BYP} The Port Bypass register selects port output data from either the data output register or digital global input. {PRT*_SIO_CFG} Input buffer enable and Output buffer Configuration, 2 bits per SIO pair. The Port Bidirectional Enable register enables dynamic bidirectional mode at any pin. {PRT*_SIO_HYST_EN} Differential hysteresis enable. (PRT*_BIE} {PRT*_INP_DIS} The Port Input buffer disable allows the user to override the input buffer default drive mode settings. {PRT*_BIT_MSK} Mask of which bits within the {PRT*_DR} and {PRT*_PS} are accessible via read / writes to {PRT*_DR_ALIAS} and reads of {PRT*_PS_ALIAS}. {PRT*_AG} The Analog global control enable register selects on a pin-by-pin basis whether to connect the pin to the analog global bus. {PRT*_AMUX} The Analog Global Multiplexer Register selects on a pin-by-pin basis whether to connect the pin to the analog mux bus. {PRT*_PRT} The Port Configuration Register allows configuration of several configuration bits of the entire I/O port simultaneously. This register aliases the port functional registers on a port-wide basis. {PRT*_PC*} The Port Pin Configuration Registers ({PRT*_PC[0] through {PRT*_PC[7]}) access several configuration or status bits of a single I/O port pin simultaneously. These registers alias the functional registers on a pinby-pin basis. {PRT*_DR_ALIAS} Aliased port data. Allows read / write access to {PRT*_DR} if {PRT*_BIT_MSK} is set. Allows access to all port data registers as a contiguous block simplifying DMA access. {PRT*_PS_ALIAS} Aliased port data. Allows read access to {PRT*_PS} if {PRT*_BIT_MSK} is set. Allows access to all port state registers as a contiguous block simplifying DMA access. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Table 19-8 shows registers specific to an SIO port. Table 19-8. SIO Port Registers Address Description Registers shown in Table 19-9 involve DSI bit selection. These registers are associated with all I/O ports and are located within the port logic. Table 19-9. DSI Selection Registers Address Description {PRT*_OUT_SEL*} Data output from UDB to Digital System Array Input Select registers. There are two select lines per port pin. {PRT*_OE_SEL*} UDB set dynamic Output Enable control select. There are two select lines per port pin. {PRT*_DBL_SYNC_IN} The Port Double Sync In register enables synchronization of the data in from the port before driving the digital system interconnect (DSI) signals to the UDB. {PRT*_SYNC_OUT} The Port Sync Out register enables synchronization of the data in from the UDB digital system interconnect (DSI) using the existing {PRT*_DR} register. Table 19-10 shows the register associated with the PICU. Table 19-10. PICU-Associated Registers Address Description {PICU*_INTTYPE*} This register defines the interrupt type to configure the pin interrupt – 1 register for each pin {PICU*_INTSTAT} Status register provides information on currently posted interrupts – 1 register for each PICU {PICU*_SNAP} The Port Snapshot register provides information on the state of the input pins at the most recent read to the status (INTSTAT) register – 1 register for each PICU 165 I/O System 166 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 20. Flash, Configuration Protection PSoC® 5LP devices offer a host of flash and configuration protection options and device security features that can be leveraged to meet the security and protection requirements of an application. These requirements range from protecting configuration settings or flash data to locking the entire device from external access. The following section discusses in detail these features together with their usage cases. 20.1 Flash Protection The objective of flash protection is to prevent access or modification to the flash contents. The only nonvolatile (NV) storage on a PSoC 5LP device that has protection options is the flash; there are no EEPROM and NV latch protection options. Flash memory in PSoC 5LP architecture is organized as flash arrays. Depending on the flash memory size, there can be one or more than one flash array. Each flash array can have a maximum of 256 rows. Each flash array row has 256 bytes of data. PSoC 5LP architecture offer customers the ability to assign one of four protection levels to each row of flash in a device. For each flash array, flash protection bits are stored in a hidden row in that array. In the hidden row, two protection bits per row are packed into a byte, so each byte in the hidden row has protection settings for four flash rows. The flash rows are ordered so that the first two bits in the hidden row correspond to the protection settings of flash row 0 (see Figure 20-1). See the Flash Program Memory chapter on page 93 to learn more about flash memory organization in PSoC 5LP devices. Figure 20-1. Flash Protection Bit Structure Row 0 Bits [0:1] Row 1 Bits [2:3] Row 2 Bits [4:5] Row 3 Bits [6:7] Byte 0 in Flash Hidden Row 0: Contains protection bits for Flash rows 0 through 3 Row 4 Bits [0:1] Row 5 Bits [2:3] Row 6 Bits [4:5] Row 7 Bits [6:7] Byte 1 in Flash Hidden Row 0: Contains protection bits for Flash rows 4 through 7 Protection is cumulative in that modes have successively higher protection levels and include the lower protection modes. Flash protection can only be set once. To change flash protection settings after they are set, the flash contents must be completely erased and reprogrammed, then the protection levels can be set again. See the Nonvolatile Memory Programming chapter on page 423 for erasing and programming flash. Table 20-1 shows the protection modes. Table 20-1. Flash Protection Modes Mode Description In PSoC Creator U - Unprotected Reada External Writeb Internal Writec Yes Yes Yes 00 Unprotected 01 Read Protect F - Factory Upgrade No Yes Yes 10 Disable External Write R - Field Upgrade No No Yes 11 Disable Internal Write W - Full Protection No No No a. Applies to Test Controller and Read commands, and cache data fetches. Cache code fetches are always allowed. b. Test controller/3rd party programmers. c. Boot loading or writes due to firmware execution. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 167 Flash, Configuration Protection When a read/write/erase operation is done for a row, the corresponding protection bits are checked. The command is executed only if allowed under the current protection mode. If the command is not allowed, then the command fails. As shown in Table 20-1, four flash protection levels are available for every row of flash in a device. A customer may choose any one of these protection levels independent of the protection choice for all other rows in the flash. The following list provides a few additional details on the features and use cases for each of these protection levels. ■ 00 – No Protection ■ 01 – Read Protect No external device can read a flash block that is read protected. The SPC Read commands cannot be used to read a block that is read protected. Only the processor and the PHUB can access a block of flash that is read protected. and Test modes. This precaution prevents anyone from erasing or altering the content of the internal memory. If the device is protected with a WO latch setting, Cypress cannot perform failure analysis and, therefore, cannot accept an RMA from customers. The WO latch is read out via serial wire debug (SWD) to electrically identify protected parts. The user writes the key in the WO latch to lock out external access only if no flash protection is set. However, after setting the values in the WO latch, a user still has access to the device until it is reset. The output of the WOL is only sampled upon reset. Therefore, you can write the key into the WO latch, program the flash protection data, and then reset the part to lock it. See the Nonvolatile Memory Programming chapter on page 423 for information about writing to the Write Once (WO) nonvolatile latch. 20.3 Offers only read protection. ■ 10 – External Write Protection No external device can erase or write a row of flash that is external write protected. Includes all Read Protect restrictions. Boot loaders work at this protection level. ■ 11 – Fully Protected The processor cannot erase or write a block of flash that is fully protected. Includes all protections from lower levels of flash data protection. This level is used when a block of flash should never be modified by an internal process or external device. Note that when the debug controller is enabled, it can read the entire flash memory regardless of the flash protection setting. Therefore, if flash protection is required, the debug controller also needs to be disabled. 20.2 Device Security The objective of device security is to prevent the PSoC 5LP device in an application from being used as a host to compromise the application. The device security feature is enabled by writing to the Write Once (WO) latch. The WO latch is a type of nonvolatile latch. When the output is ‘1’, the Write Once NVL locks the part out of Debug and Test modes; it also permanently gates off the ability to erase or alter the contents of the latch. The user can write a correct 32-bit key (0x50536F43) into the WO latch to disable the part from entering into Debug 168 Configuration Segment Protection Part of the PSoC platform’s value to customers is its ability to change the functionality of the device in real time. Changing the functionality can be as simple as enabling an external crystal or as dramatic as changing the functionality of UDBs from timers to CRC generators. Based on the application needs, the customer may also want to protect certain Configuration registers. Not all configuration registers need the same level of security and protection. Hence, the configuration registers are grouped into four segments, with registers assigned to a segment based on the presumed application use cases. The registers under each of the four segments are listed in Table 20-3 to Table 20-6. The device registers that are not listed in these tables do not have any segment protection. This is to ensure that the protection logic is supported only on important registers, thereby saving chip area where the protection logic is not required. Segment 0. One time system settings. This segment has system registers that are configured only once during program execution. The registers in this segment come under the following broad categories: ■ Power System ■ Reset ■ Watchdog ■ Internal low speed oscillator (ILO) Segment 1. Reconfigurable system settings. This segment has registers that can be reconfigured during program exe- PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Flash, Configuration Protection cution. The registers in this segment come under the following broad categories: ■ LVI Detect ■ Voltage regulators ■ Power Manager ■ Wakeup Sources ■ Boost Converter Segment 2. UDB array configuration registers. ■ All UDB array configuration registers, such as the clock selection and datapath input/output multiplexer selection, come under this segment. Segment 3. Analog interface (Registers related to analog interface configuration). It must be noted that Segment 0 registers can be configured either as the one time configurable or reconfigurable type. The same applies to Segment 1 and Segment 2 registers as well. But as a best practice, it is advisable to set Segment 0 registers as one time configurable. The settings for the rest of the segments depend on application requirements. To find out the segment to which a register is allocated, see the segment field for the register in the PSoC® 5LP Registers TRM (Technical Reference Manual). Write access to the Configuration registers in various segments is enabled using the Segment Configuration register (MLOGIC_SEG_CFG0). Write access to the Segment Configuration register (MLOGIC_SEG_CFG0) is enabled using the Segment Control register (MLOGIC_SEG_CR). 20.3.2 Locking and Protecting Segments The 8-bit Segment Configuration register (MLOGIC_SEG_CFG0) holds a pair of bits for each segment (Segment 0 to Segment 3) that are used to regulate access to the Configuration registers in that segment. The pair consists of one protect bit and one lock bit; these bits operate independently of each other. Protect Bit. The segment protect (LOCK_PROTECT_x) bit controls the ability to write the segment’s lock bit. If the segment protect bit is ‘0’, the segment’s lock bit can be written as a ‘0’ or ‘1’ at anytime. If the protect bit is ‘1’, the segment’s lock bit cannot be modified. The segment protect (LOCK_PROTECT_x) bit is a write-to1 once bit. It cannot change from a ‘1’ to a ‘0’ except as a result of a hardware reset, such as a POR or XRES. For one time configuration of a segment, it must be locked and protected after configuration. Lock Bit. The segment lock (LOCK_x) bit controls the write access to the Configuration registers in the segment. Setting the LOCK_x bit prevents write access to the Control registers; clearing the lock bit allows a write. For dynamic configuration of a segment, it must not be protected and can be locked after every configuration. Table 20-2 describes the behavior for different protect and lock bit settings. Table 20-2. Protect and Lock Bit Settings Protect/Lock Bits 20.3.1 Locking/Unlocking Segment Configuration Register The 8-bit Segment Control register (MLOGIC_SEG_CR) is used to control write access to the Segment Configuration register (MLOGIC_SEG_CFG0) bits. By default, write access to the Segment Configuration register is disabled. Attempted writes will appear to execute normally, but the contents of the register will remain unchanged. Description 00b The Configuration registers are not protected and not locked. They can be written at anytime. 01b The Configuration registers are not protected but locked. This is used to temporarily lock the configuration and is used in the case of dynamic reconfiguration. 10b The Configuration register are protected and not locked. They can be written at anytime. 11b The Configuration registers are protected and locked. This is used for one time configuration. Segment configuration write access is enabled by writing 0xB5 to the Segment Control register and is disabled by writing 0xB4 to the Segment Control register. Upon device reset, the Segment Control register resets to the locked state and disables write to the Segment Configuration register. When illegal values (values other then 0xB4 and 0xB5) are written to the Segment Control register, it causes a device reset and is indicated by the Segment reset (SEGRS) bit in Reset Status (RESET_SR1) register. The segment reset bit remains set until cleared by the user or POR. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 169 Flash, Configuration Protection Table 20-3. Segment 0: One Time System Settings Category Reset Power System ILO Watchdog Register Names Table 20-4. Segment 1: Reconfigurable System Settings PHUB Address RESET_CR3 0x46F7 RESET_CR4 0x46F8 RESET_CR5 0x47F9 RESET_TR 0x46FB RESET_IPOR_CR0 0x46F0 RESET_IPOR_CR1 0x46F1 RESET_IPOR_CR2 RESET_IPOR_CR3 PWRSYS_HIB_TR0 0x4680 PWRSYS_HIB_TR1 Category Register Names PHUB Address RESET_CR0 0x46F4 RESET_CR1 0x46F5 RESET_CR2 0x46F6 PWRSYS_CR1 0x4331 0x46F2 PM_TW_CFG0 0x4380 0x46F3 PM_TW_CFG1 0x4381 PM_TW_CFG2 0x4382 LVI Detect Volt Regulators Power Manager PM_WDT_CR 0x4384 0x4681 PM_MODE_CFG0 0x4391 PWRSYS_I2C_TR 0x4682 PM_MODE_CFG1 0x4392 PWRSYS_SLP_TR 0x4683 PM_MODE_CSR 0x4393 PWRSYS_BUZZ_TR 0x4684 PWRSYS_WAKE_TR0 0x4685 PM_WAKEUP_CFG0 0x4398 PWRSYS_WAKE_TR1 0x4686 PM_WAKEUP_CFG1 0x4399 Wakeup Sources PWRSYS_BREF_TR 0x4687 PWRSYS_BG_TR 0x4688 BOOST_CR0 0x4320 PWRSYS_WAKE_TR2 0x4689 BOOST_CR1 0x4321 PWRSYS_WAKE_TR3 0x468a BOOST_CR2 0x4322 PWRSYS_CR0 0x4330 BOOST_CR3 0x4323 ILO_TR0 0x4690 FASTCLK_* 0x4200-0x42FF ILO_TR1 0x4691 IMO_* 0x46A0-0x46A7 SLOWCLK_ILO_CR0 0x4300 XMHZ_TR 0x46A8 PM_WDT_CFG 0x4383 CACHE_CR1 0x4801 Boost Fast Clock FLASH LPM Table 20-5. Segment 2: UDB Array Category UDB Config Register Names UCFG_* PHUB Address 0x100000x150FF Table 20-6. Segment 3: Analog Interface Category 170 Register Names PHUB Address Analog Interface Routing and Configuration Registers 0x5800-0x5FFF Analog Interface Trim Registers 0x4600-0x467F PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Flash, Configuration Protection 20.3.3 Example The device peripherals are enabled/disabled by the PM_ACT_CFG* registers in Active mode. These registers are mapped in Segment1. The following steps explain the procedure to configure these registers and then lock the configuration information so that runaway code does not overwrite the values. 1. Write 0xB5 to the Segment Control register (MLOGIC_SEG_CR) to enable the write access to the Segment Configuration register. 2. Clear the lock bit for Segment 1 to get write access to the Configuration registers in Segment 1. This is done by clearing the lock bit corresponding to Segment 1, which is MLOGIC_SEG_CFG0[2]. Here, it is assumed that the 20.4 protect bit for this segment, MLOGIC_SEG_CFG0[3], is not set. If the protect bit has been set by the user, the lock bit cannot be modified, other than by a device reset. 3. Write to the Active Power Mode Template registers (PM_ACT_CFG*) to enable/disable the required peripherals. 4. Set the lock bit (MLOGIC_SEG_CFG0[2]) and clear the protect bit (MLOGIC_SEG_CFG0[3]) for Segment 1 in the Segment Configuration register (MLOGIC_SEG_CFG0). 5. Write 0xB4 to the Segment Control register to disable the write access to the Segment Configuration register. Frequently Asked Questions About Flash Protection and Device Security Question 1. How do I decide on the flash protection level needed for the application? The protection settings for flash memory must be set based on the following criteria: ■ If the application warrants the need for a field upgrade, then set the Disable External Write mode for the flash rows that are going to be updated in the field. This allows you to use the bootloader application to update the flash using communication interfaces such as I2C and USB. ■ If the application code must be protected from being copied or modified to protect IP, the flash security level for the rows containing the IP code must be set to Full Protection mode. Question 2. Is it possible to modify the flash protection settings that have already been set? It is not possible to directly alter the flash protection setting. The only way to change the flash protection settings is to completely erase the entire flash memory using the Erase All command, reprogram the flash memory, and then set the new protection settings. See the Nonvolatile Memory Programming chapter on page 423 to learn more about flash erase/program commands. Question 3. Is it possible to reprogram a flash memory that is configured with Full Protection? The only way to reprogram the fully protected rows is to erase the entire flash memory using the Erase All command, reprogram the flash memory, and then set the new protection settings as described in Question 2 above. Question 4. Is it necessary to enable protection for the entire flash memory, or only the for the region of flash memory that the application uses? It is sufficient to configure flash security for memory regions that are used by the application, leaving the unused locations unprotected, provided that there is no possibility of the program execution going to the unprotected region. If there is a possibility of code executing from the unprotected region (due to, for instance, function calls), malicious code can be written in the unprotected region to read the flash data in the fully protected region. Remember that internal read is permitted in all protection modes; therefore, it is always a good practice to set protection for the entire flash memory. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 171 Flash, Configuration Protection Question 5. Is it ever necessary to configure different protection settings for different memory regions? Yes, depending on the application requirements. Different flash rows may need different protection settings. A typical example is the case of field upgrade using the bootloader component. The portion of flash that needs to be upgraded in the field with bootloadable code must be configured in External Write Protect mode. The remaining flash memory (base code or bootloader code, unused flash memory) can be set to Full Protection. Question 6. Are flash protection settings obeyed in Debug mode? The Read Protection setting is not obeyed in Debug mode, which means the flash memory can be read regardless of flash protection setting. The Write Protection setting is still intact. Setting Full Protection makes it impossible to write to the flash memory in Debug mode. Because the Debug mode is used during the application development phase, there is no need to protect the flash. After the application development phase is over, and code has been finalized, the user can disable the debug feature. Question 7. What is device security? Device security is the feature in PSoC 5LP architecture that prevents the device from entering Debug and Test modes. To enable device security, write a 32-bit key (0x50536F43) into the Write Once (WO) latch. After writing this key, the device cannot be reprogrammed by entering test mode. Entering debug mode while using JTAG boundary scan is also not possible. This prevents external access to registers and nonvolatile memory. See Device Security on page 168 of this chapter to learn more about device security. Question 8. What are the risks associated with enabling device security? If the device is protected with a WO latch setting, Cypress cannot perform failure analysis and, therefore, cannot accept RMAs from customers. The WO latch can be read via the SWD to electrically identify protected parts. Question 9. Are device security and flash protection interrelated or independent? The answer is both. While flash protection settings and device security are configured independently, enabling device security does not allow external read or write of flash memory, regardless of the flash protection settings. There is one important exception. Even with device security enabled, it is still possible to update the flash memory using a bootloader application, provided the flash memory is not fully protected. Question 10. Is it possible to implement OTP (one time programmable) functionality such that flash content can never be altered after it is programmed? The Full Protection setting for flash memory, along with the device security feature can prevent the flash from ever being modified. This combination is the highest level of security setting available in PSoC 5LP devices. The steps to do this are given below 1. Erase the entire flash memory using the Erase All command 2. Reprogram the flash content. 3. Write a 32-bit key (0x50536F43) into the WO latch to enable device security. 4. Set flash Protection setting to Full Protection. 5. Reset the part to lock it. 172 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Section E: Digital System The digital subsystems of PSoC® 5LP architecture provides these devices their first half of unique configurability. The subsystem connects a digital signal from any peripheral to any pin through the Digital System Interconnect (DSI). It also provides functional flexibility through an array of small, fast, low-power universal digital blocks (UDBs). PSoC Creator™ provides a library of pre-built and tested standard peripherals that are mapped onto the UDB array by the tool (UART, SPI, LIN, PRS, CRC, timer, counter, PWM, AND, OR, and so on). Nonstandard peripherals are easily implemented using a Hardware Description Language (HDL) such as Verilog. Each UDB contains Programmable Array Logic (PAL) and Programmable Logic Device (PLD) functionality, together with a small state machine engine to support a wide variety of peripherals. In addition to the flexibility of the UDB array, PSoC devices provide configurable digital blocks targeted at specific functions. These blocks can include 16-bit timer/counter/PWM blocks, I2C slave/master/multi-master, Full Speed USB, and CAN 2.0b. See the device datasheet for a list of available specific function digital blocks. This section encompasses the following chapters: ■ Universal Digital Blocks (UDBs) chapter on page 175 ■ UDB Array and Digital System Interconnect chapter on page 217 ■ Controller Area Network (CAN) chapter on page 225 ■ USB chapter on page 241 ■ Timer, Counter, and PWM chapter on page 257 ■ I2C chapter on page 273 ■ Digital Filter Block (DFB) chapter on page 287 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 173 Section E: Digital System Top Level Architecture Digital System Block Diagram System Bus DIGITAL SYSTEM Universal Digital Block Array (N x UDB) Quadrature Decoder UDB UDB UDB UDB UDB 8-Bit SPI I2C Slave 8-Bit Timer UDB UDB Logic 12-Bit SPI UDB UDB UDB UDB UDB UDB UDB UDB Logic UART 174 Usage Example for UDB UDB 16-Bit PRS 16-Bit PWM Sequencer 8-Bit Timer CAN 2.0 Nx Timer Counter PWM I2C Master/Slave FS USB 2.0 USB PHY D+ D- 12-Bit PWM PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 21. Universal Digital Blocks (UDBs) This chapter shows how the PSoC® 5LP universal digital blocks (UDBs) enable the development of programmable digital peripheral functions. The UDB architecture provides balance between configuration granularity and efficient implementation; UDBs consist of a combination of uncommitted logic similar to programmable logic devices (PLDs), structured logic (datapaths), and a flexible routing scheme. 21.1 ■ Features For optimal flexibility, each UDB contains several components: ❐ ALU-based 8-bit datapath (DP) with an 8-word instruction store and multiple registers and FIFOs ❐ Two PLDs, each with 12 inputs, eight product terms and four macrocell outputs ❐ Control and status modules ❐ Clock and reset modules ■ PSoC 5LP contains an array of up to 24 UDBs ■ Flexible routing through the UDB array ■ Portions of UDBs can be shared or chained to enable larger functions ■ Flexible implementations of multiple digital functions, including timers, counters, PWM (with dead band generator), UART, I2C, SPI, and CRC generation/checking 21.2 Block Diagram Figure 21-1 on page 176 illustrates the UDB as a construct containing a pair of basic PLD logic blocks, a datapath, and control, status, clock and reset functions. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 175 Universal Digital Blocks (UDBs) Figure 21-1. UDB Block Diagram PLD Chaining Clock and Reset Control PLD 12C4 (8 PTs) PLD 12C4 (8 PTs) Status and Control Datapath Datapath Chaining Routing Channel 21.3 How It Works The major components of a UDB are: ■ PLDs (2) – These blocks take inputs from the routing channel and form registered or combinational sum-ofproducts logic to implement state machines, control for datapath operations, conditioning inputs, and driving outputs. ■ Datapath – This block contains a dynamically programmable ALU, four registers, two FIFOs, comparators, and condition generation. ■ Control and Status – These modules provide a way for CPU firmware to interact and synchronize with UDB operation. Control registers drive internal routing, and status registers read internal routing. ■ Reset and Clock Control – These modules provide clock selection and enabling, and reset selection, for the other blocks in the UDB. ■ Chaining Signals – The PLDs and datapath have chaining signals that enable neighboring blocks to be linked, to create higher precision functions. ■ Routing Channel – UDBs are connected to the routing channel through a programmable switch matrix for connections between blocks in one UDB, and to all other UDBs in the array. Routing is covered in detail in the UDB Array and Digital System Interconnect chapter on page 217. ■ System Bus Interface – All registers and RAM in each UDB are mapped into the system address space and are accessible by the CPU and DMA as both 8-bit and 16-bit data. 176 21.3.1 PLDs There are two “12C4” PLDs in each UDB. PLD blocks, shown in Figure 21-2 on page 177, can be used to implement state machines, perform input or output data conditioning, and to create lookup tables (LUTs). PLDs may also be configured to perform arithmetic functions, sequence the datapath, and generate status. General purpose RTL can be synthesized and mapped to the PLD blocks. This section presents an overview of the PLD design. A PLD has 12 inputs which feed across eight product terms (PT) in the AND array. In a given product term, the true (T) or complement (C) of the input can be selected. The output of the PTs are inputs into the OR array. The 'C' in 12C4 indicates that the OR terms are constant across all inputs, and each OR input can programmatically access any or all of the PTs. This structure gives maximum flexibility and ensures that all inputs and outputs are permutable. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-2. PLD 12C4 Structure PT0 PT1 PT2 PT3 PT4 PT5 PT6 PT7 IN0 TC TC TC TC TC TC TC TC IN1 TC TC TC TC TC TC TC TC IN2 TC TC TC TC TC TC TC TC IN3 TC TC TC TC TC TC TC TC IN4 TC TC TC TC TC TC TC TC IN5 TC TC TC TC TC TC TC TC IN6 TC TC TC TC TC TC TC TC IN7 TC TC TC TC TC TC TC TC IN8 TC TC TC TC TC TC TC TC IN9 TC TC TC TC TC TC TC TC IN10 TC TC TC TC TC TC TC TC IN11 TC TC TC TC TC TC TC TC AND Array Carry In T T T T T T T T MC0 OUT0 T T T T T T T T MC1 OUT1 T T T T T T T T MC2 OUT2 T T T T T T T T MC3 OUT3 OR Array 21.3.1.1 Carry Out PLD Macrocells The macrocell architecture is shown in Figure 21-3 on page 178. The output drives the routing array, and can be registered or combinational. The registered modes are D Flip-Flop with true or inverted input, and Toggle Flip-Flop on input high or low. The output register can be set or reset for purposes of initialization, or asynchronously during operation under control of a routed signal. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 177 Universal Digital Blocks (UDBs) Figure 21-3. Macrocell Architecture XOR Feedback (XORFB) 00: D FF 01: Arithmetic (Carry) 10: T FF on high 11: T FF on low (from prev MC) XORFB[1:0] SSEL selin cpt1 cpt0 3 2 1 0 CONST 1 1 0 To macrocell read-only register Constant (CONST) 0: D FF true in 1: D FF inverted in 0 Set Select (SSEL) 0: Set not used 1: Set from input 1 out set D Q From OR gate clk 0 QB res pld_en reset 1 Carry Out Enable (COEN) 0:Carry Out disabled 1: Carry Out enabled RSEL selout Output Bypass (BYP) 0: Registered 1: Combinational BYP 0 COEN Reset Select (RSEL) 0: Set not used 1: Set from input (to next MC) PLD Macrocell Read Only Register In addition to driving the routing array, the outputs of the macrocells from both PLDs are mapped into the address space as an 8-bit read only register, which can be accessed by the CPU or DMA. Figure 21-4. PLD Macrocell Read Only Register PLD1 MC3 MC2 7 PLD0 MC1 6 MC0 5 MC3 4 MC2 3 MC1 2 MC0 1 0 RD MC (Read Only) System Bus 178 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) 21.3.1.2 PLD Carry Chain of the PLDs, and then to the next UDB as the carry chain out “selout”. To support the efficient mapping of arithmetic functions, special product terms are generated and used in the macrocell in conjunction with the carry chain. PLDs are chained together in UDB address order. As shown in Figure 21-5 the carry chain input “selin” is routed from the previous UDB in the chain, through each macrocell in both Figure 21-5. PLD Carry Chain and Special Product Term Inputs PLD1 {PT7,PT6} {PT5,PT4} {PT3,PT2} {PT1,PT0} {PT7,PT6} {PT5,PT4} {PT3,PT2} {PT1,PT0} selout PLD0 cpt1,cpt0 cpt1,cpt0 cpt1,cpt0 cpt1,cpt0 cpt1,cpt0 cpt1,cpt0 cpt1,cpt0 cpt1,cpt0 MC3 MC2 MC1 MC0 MC3 MC2 MC1 MC0 To the next PLD block in the chain 21.3.1.3 PLD Configuration Each PLD appears to the CPU or DMA as a 16-bit wide RAM. The AND array has 12 X 8 X 2 bits, or 24 bytes, for programming, and the OR array has 4 x 8 bits, or 4 bytes, for programming. In addition, each macrocell has one configuration byte, resulting in 32 total configuration bytes per PLD. Because each UDB contains two PLDs, there are 64 total PLD configuration bytes per UDB. See UDB Configuration Address Space on page 214 for more information. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F selin From previous PLD block in the chain 21.3.2 Datapath The datapath, shown in Figure 21-6 below, contains an 8-bit single-cycle ALU, with associated compare and condition generation circuits. A datapath may be chained with datapaths in neighboring UDBs to achieve higher precision functions. The datapath includes a small dynamic configuration RAM, which can dynamically select the operation to perform in a given cycle. The datapath is optimized to implement typical embedded functions such as timers, counters, PWMs, PRS, CRC, shifters and dead band generators. The addition of add and subtract functions allow support for digital delta-sigma operations. 179 Universal Digital Blocks (UDBs) Figure 21-6. Datapath Top Level 8-bit Datapath 8-bits parallel I/O In/Out FIFOs f1_blk_stat f1_bus_stat f0_blk_stat f0_bus_stat F1 PI 8 PO 8 F0 Output Mux co_msb srcb cmsbo cfbi Alu ALU sor sol_msb ci f1_bus_stat f1_blk_stat cmsb f0_bus_stat f0_blk_stat sol_msb sol_msb_reg sil sir sor_reg sor Shift 8 Mask 1 3 ov_msb 14 13 12 11 10 9 cmsbi cfbo co_msb-1 co_msb co_msb_reg z1 ff1 cfb_en srca cl1 5 A0 ce1 6 A1 7 16 ff0 4 z0 Accumulators 2 cl0 8 D0 15 f0_ld f1_ld d0_ld d1_ld ci si ce0 Sink Source Output Mux (6 - 16 to 1) Input Mux (9 6-to-1) dp_in[5:0] 8x16 Dynamic Configuration RAM rad0 rad1 rad2 D1 0 Data Registers Configuration Input Mux RAM 6 dp_out[5:0] amask out cmsb A0 cmask0 A0 z0 z1 ZDET ZDET A0 ff0 ff1 FFDET 180 co_msb-1 D0 Conditions z0i ce0 CMP == ce0i z1i cl0 CMP < cl0i A1 A0 A1 FFDET co_msb ov_msb A1 ff0i D1 A0 cmask1 ff1i ce1 CMP == ce1i cl1 CMP < cl1i OVDET PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) 21.3.2.1 Overview The following sections present an overview description of key datapath features: Dynamic Configuration Dynamic configuration is the ability to change the datapath function and interconnect on a cycle-by-cycle basis, under sequencer control. This is implemented using the configuration RAM, which stores eight unique configurations. The address input to this RAM can be routed from any block connected to the routing fabric, most typically PLD logic, I/O pins, or other datapaths. ALU The ALU can perform eight general-purpose functions: increment, decrement, add, subtract, AND, OR, XOR, and PASS. Function selection is controlled by the configuration RAM on a cycle-by-cycle basis. Independent shift (left, right, nibble swap) and masking operations are available at the output of the ALU. Conditionals Each datapath has two comparators, with bit masking options, which can be configured to select a variety of datapath register inputs for comparison. Other detectable conditions include all zeros, all ones, and overflow. These conditions form the primary datapath output selects to be routed to the digital routing fabric or inputs to other functions. Built in CRC/PRS The datapath has built-in support for single-cycle Cyclic Redundancy Check (CRC) computation and Pseudo Random Sequence (PRS) generation of arbitrary width and arbitrary polynomial specification. To achieve longer than 8-bit CRC/PRS widths, signals may be chained between datapaths. This feature is controlled dynamically, and therefore can be interleaved with other functions. FIFO), or an output buffer (datapath internals write to the FIFO, the CPU or DMA reads from the FIFO). These FIFOs generate status that can be routed to interact with sequencers, interrupt, or DMA requests. Chaining The datapath can be configured to chain conditions and signals with neighboring datapaths. Shift, carry, capture, and other conditional signals can be chained to form higher precision arithmetic, shift, and CRC/PRS functions. Time Multiplexing In applications that are oversampled, or do not need the highest clock rates, the single ALU block in the datapath can be efficiently shared between two sets of registers and condition generators. ALU and shift outputs are registered and can be used as inputs in subsequent cycles. Usage examples include support for 16-bit functions in one (8-bit) datapath, or interleaving a CRC generation operation with a data shift operation. Datapath Inputs The datapath has three types of inputs: configuration, control, and serial and parallel data. The configuration inputs select the dynamic configuration RAM address. The control inputs load the data registers from the FIFOs and capture accumulator outputs into the FIFOs. Serial data inputs include shift in and carry in. A parallel data input port allows up to eight bits of data to be brought in from routing. Datapath Outputs There are a total of 16 signals generated in the datapath. Some of these signals are conditional signals (for example, compares), some are status signals (for example, FIFO status), and the rest are data signals (for example, shift out). These 16 signals are multiplexed into the six datapath outputs and then driven to the routing matrix. By default the outputs are single synchronized (pipelined). A combinational output option is also available for these outputs. Variable MSB The most significant bit of an arithmetic and shift function can be programmatically specified. This supports variable width CRC/PRS functions and, in conjunction with ALU output masking, can implement arbitrary width timers, counters, and shift blocks. Input/Output FIFOs Each datapath contains two 4-byte FIFOs, which can be individually configured for direction as an input buffer (CPU or DMA writes to the FIFO, datapath internals read the PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 181 Universal Digital Blocks (UDBs) Datapath Working Registers Table 21-2. FIFO Modes and Configurations Each datapath module has six 8-bit working registers. All registers are readable and writable by CPU or DMA: Normal/Fast The control to load the FIFO from the datapath source is sampled on the currently selected datapath clock (normal) or the bus clock (fast). This allows captures to occur at the highest rate in the system (bus clock), independent of the datapath clock. Software Capture When this mode is enabled, and the FIFO is in output mode, a read by the CPU or DMA of the associated accumulator (A0 for F0, A1 for F1) initiates a synchronous transfer of the accumulator value into the FIFO. The captured value may then be immediately read from the FIFO. If chaining is enabled, the operation follows the chain to the MS block for atomic reads by datapaths of multi-byte values. Asynch When the datapath is being clocked asynchronously to the bus clock, the FIFO status signals can be routed to the rest of the datapath either directly, single sampled to the DP clock, or double sampled in the case of an asynchronous DP clock Table 21-1. Datapath Working Registers Type Accumulator Data FIFOs 21.3.2.2 Name Description A0, A1 The accumulators may be both a source and a destination for the ALU. They may also be loaded from a Data register or a FIFO. The accumulators typically contain the current value of a function, such as a count, CRC, or shift. These registers are nonretention; they lose their values in sleep and are reset to 0x00 on wakeup. D0, D1 The Data registers typically contain constant data for a function, such as a PWM compare value, timer period, or CRC polynomial. These registers retain their values across sleep intervals. F0, F1 The two 4-byte FIFOs provide both a source and a destination for buffered data. The FIFOs can be configured as both input buffers, both output buffers, or as one input buffer and one output buffer. Status signals indicate the read and write status of these registers. Usage examples include buffered TX and RX data in the SPI or UART and buffered PWM compare and buffered timer period data. These registers are nonretention; they lose their values in sleep and are reset to 0x00 on wakeup. Datapath FIFOs FIFO Modes and Configurations Independent Each FIFO has a control bit to invert polarity of the FIFO Clock Polarity clock with respect to the datapath clock. Figure 21-7 shows the possible FIFO configurations controlled by the input/output modes. The TX/RX mode has one FIFO in input mode and the other in output mode. The primary usage example of this configuration is SPI. The dual capture configuration provides independent capture of A0 and A1, or two separately controlled captures of either A0 or A1. Finally, the dual buffer mode can provide buffered periods and compares, or two independent periods/compares. Each FIFO has a variety of operation modes and configurations available: Table 21-2. FIFO Modes and Configurations Mode Description Input/Output In input mode the CPU or DMA writes to the FIFO and the data is read and consumed by the datapath internals. In output mode the FIFO is written to by the datapath internals and is read and consumed by the CPU or DMA Single Buffer The FIFO operates as a single buffer with no status. Data written to the FIFO is immediately available for reading, and can be overwritten at anytime. Level/Edge The control to load the FIFO from the datapath internals can be either level or edge triggered. 182 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-7. FIFO Configurations System Bus System Bus F0 F0 F1 D0/D1 D0 D1 A0 A1 A0/A1/ALU A0/A1/ALU A0/A1/ALU F1 F0 F1 System Bus System Bus TX/RX Dual Capture Dual Buffer Figure 21-8 shows a detailed view of the FIFO sources and sinks. UDB Local Data Bus FIFO F0 FIFO F1 D0 D1 A0 A1 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F ALU A1 A0 ALU A1 A0 Figure 21-8. FIFO Sources and Sinks 183 Universal Digital Blocks (UDBs) When the FIFO is in input mode, the source is the system bus and the sinks are the Dx and Ax registers. When in output mode, the sources include the Ax registers and the ALU, and the sink is the system bus. The multiplexer selection is statically set in UDB configuration register CFG15 as shown in the following table for the F0_INSEL[1:0] or F1_INSEL[1:0]: Table 21-3. FIFO Multiplexer Set in UDB Configuration Register Fx_INSEL[1:0] 00 Description Input Mode - System bus writes the FIFO, FIFO output destination is Ax or Dx. 01 Output Mode - FIFO input source is A0, FIFO output destination is the system bus. 10 Output Mode - FIFO input source is A1, FIFO output destination is the system bus. 11 Output Mode - FIFO input source is the ALU output, FIFO output destination is the system bus. FIFO Status Each FIFO generates two status signals, “bus” and “block,” which are sent to the UDB routing through the datapath output multiplexer. The “bus” status can be used to assert an interrupt or DMA request to read/write the FIFO. The “block” status is primarily intended to provide the FIFO state to the UDB internals. The meanings of the status bits depend on the configured direction (Fx_INSEL[1:0]) and the FIFO level bits. The FIFO level bits (Fx_LVL) are set in the Auxiliary Control Working register in working register space. Options are shown in the following table: Table 21-4. FIFO Status Options Fx_INSEL[1:0] Fx_LVL Status Signal Description Input 0 Not Full Bus Status Asserted when there is room for at least 1 byte in the FIFO. Input 1 At Least Half Empty Bus Status Asserted when there is room for at least 2 bytes in the FIFO. Input NA Empty Block Status Asserted when there are no bytes left in the FIFO. When not empty, the datapath internals may consume bytes. When empty the datapath may idle or generate an underrun condition. Output 0 Not Empty Bus Status Asserted when there is at least 1 byte available to be read from the FIFO. Output 1 At Least Half Full Bus Status Asserted when there are at least 2 bytes available to be read from the FIFO. Output NA Full Block Status Asserted when the FIFO is full. When not full, the datapath internals may write bytes to the FIFO. When full, the datapath may idle or generate an overrun condition. FIFO Illustrated Operation Figure 21-9 on page 185 illustrates a typical sequence of reads and writes and the associated status generation. Although the figure shows reads and writes occurring at different times, a read and write can also occur simultaneously. 184 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-9. Detailed FIFO Operation Sinks Reset Write 2 bytes Write 2 more bytes Empty = 1 Empty = 0 Empty = 0 Empty = 0 At Least Half Empty = 1 At Least Half Empty = 1 At Least Half Empty = 0 At Least Half Empty = 1 Full = 0 Full = 0 Full = 1 Full = 0 At Least Half Full = 0 At Least Half Full = 1 At Least Half Full = 1 At Least Half Full = 0 WR_PTR WR_PTR D0 RD_PTR RD_PTR RD_PTR D1 D0 WR_PTR D1 WR_PTR Write 2 bytes Read 2 bytes Empty = 0 Empty = 0 Empty = 1 At Least Half Empty = 0 At Least Half Empty = 1 At Least Half Empty = 1 Full = 0 Full = 0 Full = 0 At Least Half Full = 0 At Least Half Full = 0 D4 X X D5 X RD_PTR X RD_PTR WR_PTR X D3 X RD_PTR D3 Read 1 bytes At Least Half Full = 1 D5 X X D2 D3 WR_PTR Read 3 bytes WR_PTR RD_PTR X X X FIFO Fast Mode (FIFO FAST) When the FIFO is configured for output, the FIFO load operation normally uses the currently selected datapath clock for sampling the write signal. As shown in Figure 21-10, with the FIFO fast mode set, the bus clock can be optionally selected for this operation. Used in conjunction with edge sensitive mode, this operation reduces the latency of accumulator-to-FIFO transfer from the resolution of the DP clock to the resolution of the bus clock, which can be much higher. This allows the CPU or DMA to read the captured result in the FIFO with minimal latency. As shown in Figure 21-10, the fast load operation is independent of the currently selected datapath clock, however, using the bus clock may cause higher power consumption. Figure 21-10. FIFO Fast Configuration Sinks UDB DP Clock Mux digital clocks DP clk DP Operation bus clk Write fx_ld 0 bus clk FIFO (In Output Mode) 1 FIFO Fast PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 185 Universal Digital Blocks (UDBs) FIFO Edge/Level Write Mode There are two modes for writing the FIFO from the datapath. In the first mode, data is synchronously transferred from the accumulators to the FIFOs. The control for that write (FX_LD) is typically generated from a state machine or condition that is synchronous to the datapath clock. The FIFO will be written in any cycle where the input load control is a '1'. In the second mode, the FIFO is used to capture the value of the accumulator in response to a positive edge of the FX_LD signal. In this mode the duty cycle of the wave- form is arbitrary (however, it must be at least one datapath clock cycle in width). An example of this mode is capturing the value of the accumulator using an external pin input as a trigger. The limitation of this mode is that the input control must revert to '0' for at least one cycle before another positive edge is detected. Figure 21-11 shows the edge detect option on the FX_LD control input. One bit for this option sets the mode for both FIFOs in a UDB. Note that edge detection is sampled at the rate of the selected FIFO clock. Figure 21-11. Edge Detect Option for Internal FIFO Write Sinks 0 fx_ld (from Routing) fx_write 1 FF dp_clk 0 bus_clk 1 FIFO Edge FIFO Fast FIFO Software Capture Mode A common and important requirement is to allow the CPU or DMA the ability to reliably read the contents of an accumulator during normal operation. This is done with software capture and is enabled by setting the FIFO Cap configuration bit. This bit applies to both FIFOs in a UDB, but is only operational when a FIFO is in output mode. When using software capture, F0 should be set to load from A0 and F1 from A1. As shown in Figure 21-12, reading the accumulator triggers a write to the FIFO from that accumulator. This signal is chained so that a read of a given byte simultaneously captures accumulators in all chained UDBs. This allows an 8-bit processor to reliably read 16 bits or more simultaneously. The data returned in the read of the accumulator should be ignored; the captured value may be read from the FIFOs immediately. The routed FX_LD signal, which generates a FIFO load, is ORed with the software capture signal; the results can be unpredictable when both hardware and software capture are used at the same time. As a general rule these functions should be mutually exclusive, however, hardware and software capture can be used simultaneously with the following settings: ■ FIFO capture clocking mode is set to FIFO FAST ■ FIFO write mode is set to FIFO EDGE With these settings, hardware and software capture work essentially the same and in any given bus clock cycle, either signal asserted initiates a capture. It is also recommended to clear the target FIFO in firmware (ACTL register) before initiating a software capture. This initializes the FIFO read and write pointers to a known state. 186 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-12. Software Capture Configuration Chain X capxi (chaining in) capx (chaining out) read ax FIFO Cap 0 fx_write 1 fx_ld FIFO EDGE bus clk (FIFO FAST) FIFO Control Bits There are four bits in the Auxiliary Control register that may be used to control the FIFO during normal operation. The FIFO0 CLR and FIFO1 CLR bits are used to reset or flush the FIFO. When a '1' is written to one of these bits, the associated FIFO is reset. The bit must be written back to '0' for FIFO operation to continue. If the bit is left asserted, the given FIFO is disabled and operates as a one byte buffer without status. Data can be written to the FIFO; the data is immediately available for reading and can be overwritten at anytime. Data direction using the Fx INSEL[1:0] configuration bits is still valid. The FIFO0 LVL and FIFO1 LVL bits control the level at which the 4-byte FIFO asserts bus status (when the bus is either reading or writing to the FIFO) to be asserted. The meaning of FIFO bus status depends on the configured direction, as shown in the table below. and F1 is set for output mode, which is a typical configuration for TX and RX registers. On the TX side, the datapath state machine uses "empty" to determine if there are any bytes available to consume. Empty is set synchronously to the DP state machine, but is cleared asynchronously due to a bus write. When cleared, the status is synchronized back to the DP state machine. On the RX side, the datapath state machine uses “full” to determine whether there is a space left to write to the FIFO. Full is set synchronously to the DP state machine, but is cleared asynchronously due to a bus read. When cleared, the status is synchronized back to the DP state machine. A single FIFO ASYNCH bit is used to enable this synchronization method; when set it applies to both FIFOs. It is only Table 21-5. FIFO Level Control Bits 0 1 FIFOx Input Mode Output Mode LVL (Bus is Writing FIFO) (Bus is Reading FIFO) Not Full Not Empty At least 1 byte can be written At least 1 byte can be read At Least Half Empty At Least Half Full At least 2 bytes can be written At least 2 bytes can be read FIFO Asynchronous Operation Figure 21-13 illustrates the concept of asynchronous FIFO operation. As an example, assume F0 is set for input mode PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 187 Universal Digital Blocks (UDBs) FIFO Overflow Operation applied to the block status, as it is assumed that bus status is naturally synchronized by the interrupt process. Use FIFO status signaling to safely implement both internal (datapath) and external (CPU or DMA) reads and writes. There is no built-in protection from underflow and overflow conditions. If the FIFO is full, and subsequent writes occur (overflow), the new data overwrites the front of the FIFO (the data currently being output, the next data to read). If the FIFO is empty, and subsequent reads occur (underflow), the read value is undefined. FIFO pointers remain accurate regardless of underflow and overflow. Table 21-6. FIFO Block Status Synchronization Options ASYNC ADD SYNC Operation Usage Model 0 CPU read/write status changes Synchronous to occur at bus clock resolution. Can bus clock be used for minimum latency if bus clock timing can be met. 0 1 Re-sampled from bus clock to DP clock This should be the default synchronous operating mode. When the CPU read/write status changes are synchronously re-sampled with the currently selected DP clock. Gives a full cycle of DP clock setup time to the UDB logic. 1 0 Reserved – 1 When a free running asynchronous Double synDP clock is in use, this setting can chronized from be used to double synchronize the bus clock to DP CPU read and write actions to the clock DP clock. 0 1 Figure 21-13. FIFO Asynchronous Operation System Bus async F0 (TX) blk_stat empty Synch to DP empty Asynchronously cleared by bus write, sycnhyronously set by DP read 0 d set 1 q DP clk Datapath Process (Asynch) async Synch to DP full blk_stat F1 (RX) Empty to DP state machine full Asynchronously cleared by bus read, sycnhyronously set by DP write System Bus FIFO Clock Inversion Option Each FIFO has a control bit called Fx CK INV that controls the polarity of the FIFO clock, with respect to the polarity of the DP clock. By default the FIFO operates at the same 0 d set 1 Full to DP state machine q DP clk polarity as the DP clock. When this bit is set, the FIFO operates at the opposite polarity as the DP clock. This provides support for “both clock edge” communication protocols, such as SPI. FIFO Dynamic Control Normally, the FIFOs are configured statically in either input or output mode. As an alternative, each FIFO can be configured into a mode where the direction is controlled dynamically, that is, by routed signals. One configuration bit per 188 FIFO (Fx DYN) enables the mode. Figure 21-14 on page 189 shows the configurations available in dynamic FIFO mode. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) ALU A1 A0 Figure 21-14. FIFO Dynamic Mode UDB Local Data Bus FIFO Fx FIFO Fx UDB Local Data Bus Ax Internal Access In internal access mode, the datapath can read and write the FIFO. In this configuration, the Fx INSEL bits must be configured to select the source for the FIFO writes. Fx INSEL = 0 (CPU bus source) is invalid in this mode; they can only be 1, 2 or 3 (A0, A1, or ALU). Note that the only read access is to the associated accumulator; the data register destination is not available in this mode. In external access mode, the CPU or DMA can both read and write the FIFO. The configuration between internal and external access is dynamically switchable using datapath routing signals. The datapath input signals d0_load and d1_load are used for this control. Note that in the dynamic control mode, d0_load and d1_load are not available for their normal use in loading the D0/D1 registers from F0/F1. The dx_load signals can be driven by any routed signal, including constants. In one usage example, starting with external access (dx_load == 1), the CPU or DMA can write one or more bytes of data to the FIFO. Then toggling to internal access (dx_load == 0), the datapath can perform operations on the data. Then toggling back to external access, the CPU or DMA can read the result of the computation. Because the Fx INSEL must always be set to 01, 10 or 11 (A0, A1, or ALU), which is “output mode” in normal opera- PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F External Access tion, the FIFO status signals have the following definitions (also dependent on Fx LVL control): Table 21-7. FIFO Status Status Signal Meaning Fx LVL = 0 Fx LVL = 1 fx_blk_stat Write Status FIFO full FIFO full fx_bus_stat Read Status FIFO not empty At least ½ full Because the datapath and CPU may both write and read the FIFO, these signals are no longer considered “block” and “bus” status. The blk_stat signal is used for write status, and the bus_stat signal is used for read status. 21.3.2.3 FIFO Status There are four FIFO status signals, two for each FIFO: fifo0_bus_stat, fifo0_blk_stat, fifo1_bus_stat and fifo1_blk_stat. The meaning of these signals depends on the direction of the given FIFO, which is determined by static configuration. FIFO status is covered in detail in section 21.3.2.2 Datapath FIFOs on page 182. 21.3.2.4 Datapath ALU The ALU core consists of three independent 8-bit programmable functions, which include an arithmetic/logic unit, a shifter unit, and a mask unit. 189 Universal Digital Blocks (UDBs) Arithmetic and Logic Operation The ALU functions, which are configured dynamically by the dynamic configuration RAM, are shown in the following table: Table 21-11. Routed Carry In Functions Table 21-8. ALU Functions Func[2:0] When a routed carry is used, the meaning with respect to each arithmetic function is shown in Table 21-11. Note that in the case of the decrement and subtract functions, the carry is active low (inverted). Function Operation 000 PASS srca 001 INC ++srca 010 DEC --srca 011 ADD srca + srcb 100 SUB srca - srcb 101 XOR srca ^ srcb 110 AND srca & srcb 111 OR srca | srcb Carry In The carry in is used in arithmetic operations. There is a default carry in value for certain functions as shown in Table 21-9. Function INC Carry In Carry In Polarity Active Carry In Inactive True ++srca srca DEC Inverted --srca srca ADD True (srca + srcb) + 1 srca + srcb SUB Inverted (srca - srcb) - 1 (srca - srcb) Carry Out The carry out is a selectable datapath output and is derived from the currently defined MSB position, which is statically programmable. This value is also chained to the next most significant block as an optional carry in. Note that in the case of decrement and subtract functions, the carry out is inverted. Table 21-12. Carry Out Functions Table 21-9. Carry In Functions Function Operation INC ++srca DEC ADD SUB Default Carry In Implementation Function srca + 00h + ci, where ci is forced to 1 INC --srca srca + ffh + ci, where ci is forced to 0 srca + srcb srca + srcb + ci, where ci is forced to 0 srca - srcb srca + ~srcb + ci, where ci is forced to 1 In addition to this default arithmetic mode for carry operation, there are three additional carry options. The CI SELA and CI SELB configuration bits determine the carry in for a given cycle. Dynamic configuration RAM selects either the A or B configuration on a cycle-by-cycle basis. The options are defined in Table 21-10. Carry Out Carry Out Polarity Active Carry Out Inactive True ++srca == 0 srca DEC Inverted --srca == -1 srca ADD True srca + srcb > 255 srca + srcb SUB Inverted srca - srcb < 0 (srca - srcb) Carry Structure Options for carry in, and for MSB selection for carry out generation, are shown in Figure 21-15 on page 191. The registered carry out value may be selected as the carry in for a subsequent arithmetic operation. This feature can be used to implement higher precision functions in multiple cycles. Table 21-10. Additional Carry In Functions CI SEL A CI SEL B Carry Mode Description Default Default arithmetic mode as described in Table 21-9. 01 Registered Carry Flag, result of the carry from the previous cycle. This mode is used to implement add with carry and subtract with borrow operations. It can be used in successive cycles to emulate a double precision operation. 10 Routed Carry is generated elsewhere and routed to this input. This mode can be used to implement controllable counters. Chained Carry is chained from the previous datapath. This mode can be used to implement single cycle operations of higher precision involving two or more datapaths. 00 11 190 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-15. Carry Operation Selected MSB Arithmetic ALU Function (inc, dec, add, sub) Default function value ALU Bit 7 ALU Bit 6 ALU Bit 5 ALU Bit 4 ALU Bit 3 ALU Bit 2 ALU Bit 1 ci ALU Bit 0 Chained (from prev datapath) Registered (from co_msb_reg) Routed (from interconnect) co_msb (to DP output mux) co_msb_reg Shift Operation Table 21-14. Shift In Functions SI SEL A The shift operation occurs independently of the ALU operation, according to Table 21-13 Table 21-13. Shift Operation Functions Shift[1:0] Shift In Source SI SEL B Default/Arithmetic The default input is the value of the DEF SI configuration bit (fixed 1 or 0). However, if the MSB SI bit is set, then the default input is the currently defined MSB (for right shift only). 01 Registered The shift in value is driven by the current registered shift out value (from the previous cycle). The shift left operation uses the last shift out left value. The shift right operation uses the last shift out right value. 10 Routed Shift in is selected from the routing channel (the SI input). 11 Chained Shift in left is routed from the right datapath neighbor and shift in right is routed from the left datapath neighbor. 00 Function 00 Pass 01 Shift Left 10 Shift Right 11 Nibble Swap A shift out value is available as a datapath output. Both shift out right (sor) and shift out left (sol_msb) share that output selection. A static configuration bit (SHIFT SEL in register CFG15) determines which shift output is used as a datapath output. When no shift is occurring, the sor and sol_msb signal is defined as the LSB or MSB of the ALU function, respectively. The SI SELA and SI SELB configuration bits determine the shift in data for a given operation. Dynamic configuration RAM selects the A or B configuration on a cycle-by-cycle basis. Shift in data is only valid for left and right shift; it is not used for pass and nibble swap. The selections and usage apply to both left and right shift directions and are shown in Table 21-14. Description The shift out left data comes from the currently defined MSB position, and the data that is shifted in from the left (in a shift right operation) goes into the currently defined MSB position. Both shift out data (left or right) are registered and can be used in a subsequent cycle. This feature can be used to implement a higher precision shift in multiple cycles. Figure 21-16. Shift Operation S e le c t d e fa u lt v a lu e o r a r ith m e tic s h ift D e fa u lt ( tie v a lu e ) s o r_ r e g s h ift in le ft ( s il) R e g is te r e d ( s o r_ r e g ) S e le c te d M S B R o u te d ( fr o m in te r c o n n e c t) C h a in e d ( fr o m n e x t D a ta p a th ) s il 7 6 s h ift o u t r ig h t (s o r ) ( to D P o u tp u t m u x ) S h ift r ig h t o r s h ift le ft 5 4 3 2 1 0 D e fa u lt ( tie v a lu e ) s h ift o u t le ft (s o l_ m s b ) ( to D P o u tp u t m u x ) s o l_ m s b _ r e g PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F R e g is te r e d ( fr o m s o l_ m s b _ r e g ) s h ift in r ig h t (s ir ) R o u te d ( fr o m in te r c o n n e c t) C h a in e d ( fr o m p r e v D a ta p a th ) 191 Universal Digital Blocks (UDBs) Note that the bits that are isolated by the MSB selection are still shifted. In the example shown, bit 7 still shifts in the sil value on a right shift and bit 5 shifts in bit 4 on a left shift. The shift out either right or left from the isolated bits is lost. ALU Masking Operation An 8-bit mask register in the UDB static configuration register space defines the masking operation. In this operation, the output of the ALU is masked (ANDed) with the value in the mask register. A typical use for the ALU mask function is to implement free-running timers and counters in power of two resolutions. 21.3.2.5 Datapath Inputs and Multiplexing The datapath has a total of nine inputs as shown in Table 24-16, including six inputs from the channel routing. These consist of the configuration RAM address, FIFO and data register load control signals, and the data inputs shift in and carry in. Table 21-15. Datapath Inputs Input Description RAD2 RAD1 RAD0 Asynchronous dynamic configuration RAM address. There are eight 16-bit words, which are user programmable. Each word contains the datapath control bits for the current cycle. Sequences of instructions can be controlled by these address inputs. F0 LD F1 LD When asserted in a given cycle, the selected FIFO is loaded with data from one of the A0 or A1 accumulators or from the output of the ALU. The source is selected by the Fx INSEL[1:0] configuration bits. This input is edge sensitive. It is sampled at the datapath clock; when a '0' to '1' transition is detected, a load occurs at the subsequent clock edge. D0 LD D1 LD When asserted in a given cycle, the Dx register is loaded from associated FIFO Fx. This input is edge sensitive. It is sampled at the datapath clock; when a '0' to '1' transition is detected, a load occurs at the subsequent clock edge. SI This is a data input value that can be used for either shift in left or shift in right. CI This is the carry in value used when the carry in select control is set to "routed carry." As shown in Figure 21-17, each input has a 6-to-1 multiplexer, therefore, all inputs are permutable. Inputs are handled in one of two ways, either level sensitive or edge sensitive. RAM address, shift in and data in values are level sensitive; FIFO and data register load signals are edge sensitive. Figure 21-17. Datapath Input Select rad0 (similar for rad1, rad2, si, ci) {0, dp_in[5:0], 0} CFGx RAD0 MUX[2:0] These inputs are edge sensitive f0_ld (similar for f1_ld, d0_ld, d1_ld) {0, dp_in[5:0], 0} CFGx F0 LD MUX[2:0] 21.3.2.6 CRC/PRS Support The datapath can support Cyclic Redundancy Checking (CRC) and Pseudo Random Sequence (PRS) generation. Chaining signals are routed between datapath blocks to support CRC/PRS bit lengths of longer than 8 bits. The most significant bit (MSB) of the most significant block in the CRC/PRS computation is selected and routed (and chained across blocks) to the least significant block. The MSB is then XORed with the data input (SI data) to provide the feedback (FB) signal. The FB signal is then routed (and chained across blocks) to the most significant block. This feedback value is used in all blocks to gate the XOR of the polynomial (from the Data0 or Data1 register) with the current accumulator value. 192 Figure 21-18 shows the structural configuration for the CRC operation. The PRS configuration is identical except that the shift in (SI) is tied to '0'. In the PRS configuration, D0 or D1 contain the polynomial value, while A0 or A1 contain the initial (seed) value and the CRC residual value at the end of the computation. To enable CRC operation, the CFB_EN bit in the dynamic configuration RAM must be set to '1'. This enables the AND of SRCB ALU input with the CRC feedback signal. When set to zero, the feedback signal is driven to '1', which allows for normal arithmetic operation. Dynamic control of this bit on a cycle-by-cycle basis gives the capability to interleave a CRC/PRS operation with other arithmetic operations. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-18. CRC Functional Structure D0/D1 (POLY) A0/A1 (CRC) MSB (most significant bit) SI (shift in) FB (feedback) srcb Tie input to zero for PRS operation srca ALU (XOR) SHIFTER (LEFT) CRC/PRS Chaining Figure 21-19 illustrates an example of CRC/PRS chaining across three UDBs. This scenario can support a 17- to 24-bit operation. The chaining control bits are set according to the position of the datapath in the chain as shown. Figure 21-19. CRC/PRS Chaining Configuration Set msb_sel CHAIN MSB = 1 CHAIN FB = 1 CHAIN FB = 1 cm sbi cmsbo cmsbi UDB 2 cfbo CHAIN M SB = 1 cmsbo cmsbi UDB 0 UDB 1 cfbi cfbo cfbi cfbo cmsbo sir CRC data in cfbi How the CRC/PRS feedback signal (cfbo, cfbi) is chained: CRC/PRS Polynomial Specification ■ If a given block is the least significant block, then the feedback signal is generated in that block from the builtin logic that takes the shift in from the right (sir) and XORs it with the MSB signal. (For PRS, the "sir" signal is tied to '0'.) As an example of how to configure the polynomial for programming into the associated D0/D1 register, consider the CCITT CRC-16 polynomial, which is defined as x16 + x12 +x5 + 1. The method for deriving the data format from the polynomial is shown in Figure 21-20. ■ If a given block is not the least significant block, the CHAIN FB configuration bit must be set and the feedback is chained from the previous block in the chain. The X0 term is inherently always '1' and therefore does not need to be programmed. For each of the remaining terms in the polynomial, a '1' is set in the appropriate position in the alignment shown. How the CRC/PRS MSB signal (cmsbo, cmsbi) is chained: ■ If a given block is the most significant block, the MSB bit (according to the polynomial selected) is configured using the MSB_SEL configuration bits. ■ If a given block is not the most significant block, the CHAIN MSB configuration bit must be set and the MSB signal is chained from the next block in the chain. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Note This polynomial format is slightly different from the format normally specified in HEX. For example, the CCITT CRC16 polynomial is typically denoted as 1021H. To convert to the format required for datapath operation, shift right by one and add a '1' in the MSB bit. In this case, the correct polynomial value to load into the D0 or D1 register is 8810H 193 Universal Digital Blocks (UDBs) Figure 21-20. CCITT CRC16 Polynomial Format X16 X15 X14 X16 1 X13 0 X11 X10 X9 X12 + 0 X12 0 1 X8 X7 X6 0 0 X4 X3 X5 + 0 X5 0 0 0 1 X2 X1 + 0 0 X0 1 0 0 CCITT 16-Bit Polynomial is 0x8810 Example CRC/PRS Configuration The following is a summary of CRC/PRS configuration requirements, assuming that D0 is the polynomial and the CRC/PRS is computed in A0: 1. Select a suitable polynomial (example above) and write it into D0. 2. Select a suitable seed value (for example, all zeros for CRC, all ones for PRS) and write it into A0. used, but this feature gives the capability for more elaborate configurations, such as up to a 16-bit CRC/PRS function in one UDB using time division multiplexing. In this mode, the dynamic configuration RAM bit CFB_EN still controls whether the CRC feedback signal is ANDed with the SRCB ALU input. Therefore, as with the built-in CRC/PRS operation, the function can be interleaved with other functions if desired. 3. Configure chaining if necessary as described above. 4. Select the MSB position as defined in the polynomial from the MSB_SEL static configuration register bits and set the MSB_EN register bit. 5. Configure the dynamic configuration RAM word fields: a. Select D0 as the ALU "SRCB" (ALU B Input Source) b. Select A0 as the ALU "SRCA" (ALU A Input Source) c. Select "XOR" for the ALU function d. Select "SHIFT LEFT" for the SHIFT function e. Select "CFB_EN" to enable the support for CRC/ PRS f. Select ALU as the A0 write source If a CRC operation, configure "shift in right" for input data from routing and supply input on each clock. If a PRS operation, tie "shift in right" to '0'. Clocking the UDB with this configuration generates the required CRC or outputs the MSB, which may be output to the routing for the PRS sequence. External CRC/PRS Mode A static configuration bit may be set (EXT CRCPRS) to enable support for external computation of a CRC or PRS. As shown in Figure 21-21, computation of the CRC feedback is done in a PLD block. When the bit is set, the CRC feedback signal is driven directly from the CI (Carry In) datapath input selection mux, bypassing the internal computation. The figure shows a simple configuration that supports up to an 8-bit CRC or PRS. Normally the built-in circuitry is 194 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-21. External CRC/PRS Mode PLD Tie shift in to zero for PRS operation SI (shift in) Routing Routing D0/D1 (POLY) When the EXT_CRCPRS bit is set, the CI selection drives the CRC feedback line. A0/A1 (CRC) FB (feedback) srcb srca CI Mux ALU (XOR) SHIFTER (LEFT) Datapath Outputs and Multiplexing Conditions are generated from the registered accumulator values, ALU outputs, and FIFO status. These conditions can be driven to the digital routing for use in other UDB blocks, for use as interrupts or DMA requests, or to I/O pins. The 16 possible conditions are shown in the table below: There are a total of six datapath outputs. As shown in Figure 21-22, each output has a 16-1 multiplexer that allows any of these 16 signals to be routed to any of the datapath outputs. Figure 21-22. Output Mux Connections Output Mux Table 21-16. Datapath Condition Generation Description ce0 Y A0 == D0 cl0 Compare Less Than Y A0 < D0 cl0 z0 Compare Equal Y cl1 Compare Less Than Y A1 or A0 < D1 or A0 (dynamic selection) z1 Zero Detect Y A1 == 00h ff1 Ones Detect Y A1 == FFh ov_msb Overflow N Carry(msb) ^ Carry(msb-1) co_msb Carry Out Y Carry out of MSB defined bit cmsb CRC MSB Y MSB of CRC/PRS function so Shift Out Y Selection of shift output N Definition depends on FIFO configuration f0_blk_stat FIFO0 Block Status f1_blk_stat FIFO1 Block Status N Definition depends on FIFO configuration f0_bus_stat FIFO0 Bus Status N Definition depends on FIFO configuration f1_bus_stat FIFO1 Bus Status N Definition depends on FIFO configuration PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F ff0 ce1 cl1 z1 ff1 ov_msb co_msb sor sol_ msb f0_bus_stat f1_bus_stat cmsb f0_blk_stat f1_blk_stat Output Mux (6 - 16 to 1) ce1 A1 or A0 == D1 or A0 (dynamic selection) 5 A0 = FFh 6 A0 == 00h Y 7 Y Ones Detect 14 13 12 11 10 9 Zero Detect ff0 15 z0 3 Compare Equal 4 ce0 0 Chain? 1 Condition 2 Name SI Mux 8 21.3.2.7 DP Inputs MSB (Most Significant Bit) 6 dp_out[5:0] 195 Universal Digital Blocks (UDBs) Compares There are two compares, one of which has fixed sources (Compare 0) and the other has dynamically selectable sources (Compare 1). Each compare has an 8-bit statically programmed mask register, which enables the compare to occur in a specified bit field. By default, the masking is off (all bits are compared) and must be enabled. Comparator 1 inputs are dynamically configurable. As shown in the table below, there are four options for Comparator 1, which applies to both the "less than" and the "equal" conditions. The CMP SELA and CMP SELB configuration bits determine the possible compare configurations. A dynamic RAM bit selects one of the A or B configurations on a cycle-by-cycle basis. Table 21-17. Compare Configuration CMP SEL A Comparator 1 Compare Configuration CMP SEL B 00 A1 Compare to D1 01 A1 Compare to A0 10 A0 Compare to D1 11 A0 Compare to A0 Compare 0 and Compare 1 are independently chainable to the conditions generated in the previous datapath (in addressing order). Whether to chain compares or not is statically specified in UDB configuration registers. Figure 21-23 illustrates compare equal chaining, which is just an ANDing of the compare equal in this block with the chained input from the previous block. Figure 21-23. Compare Equal Chaining CFGx CCHAIN0 ce0 (to routing and chaining) ce0i (from chaining) Compare Equal Figure 21-24 illustrates compare less than chaining. In this case, the “less than” is formed by the compare less than output in this block, which is unconditional. This is ORed with the condition where this block is equal, and the chained input from the previous block is asserted as less than. Figure 21-24. Compare Less Than Chaining CFGx CCHAIN0 cl0i (from chaining) cl0 (to routing and chaining) Compare Less Than 196 Compare Equal PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) All Zeros and All Ones Detect the currently defined MSB as specified by the MSB_SEL bits. This condition is not chainable, however the computation is valid when done in the most significant datapath of a multi-precision function as long as the carry is chained between blocks. Each accumulator has dedicated all zeros detect and all ones detect. These conditions are statically chainable as specified in UDB configuration registers. Whether to chain these conditions is statically specified in UDB configuration registers. Chaining of zero detect is the same concept as the compare equal. Successive chained data is ANDed if the chaining is enabled. 21.3.2.8 Datapath Parallel Inputs and Outputs As shown in Figure 21-25, the datapath Parallel In (PI) and Parallel Out (PO) signals give limited capability to bring routed data into and out of the Datapath. Parallel Out signals are always available for routing as the ALU asrc selection between A0 and A1. Overflow Overflow is defined as the XOR of the carry into the MSB and the carry out of the MSB. The computation is done on Figure 21-25. Datapath Parallel In/Out PI[7:0] A0[7:0] A1[7:0] CFB_EN PI DYN (static config bit) 1 0 ASRC[7:0] PI SEL (static config bit) Alu PO[7:0] Parallel In needs to be selected for input to the ALU. There are two options, static operation or dynamic operation. For static operation, the PI SEL bit forces the ALU asrc to be PI. The PI DYN bit is used to enable the PI dynamic operation. When it is enabled, and assuming the PI SEL is 0, the PI multiplexer may then be controlled by the CFB_EN dynamic control bit. The primary function of the CFB_EN bit is to enable PRS/CRC functionality. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 21.3.2.9 Datapath Chaining Each datapath block contains an 8-bit ALU, which is designed to chain carries, shifted data, capture triggers, and conditional signals to the nearest neighbor datapaths, to create higher precision arithmetic functions and shifters. These chaining signals, which are dedicated signals, allow single-cycle 16-, 24- and 32-bit functions to be efficiently implemented without the timing uncertainty of channel routing resources. In addition, the capture chaining supports the ability to perform an atomic read of the accumulators in chained blocks. As shown in Figure 21-21, all generated conditional and capture signals chain in the direction of least significant to most significant blocks. Shift left also chains from least to most significant. Shift right chains from most to least significant. The CRC/PRS chaining signal for feedback chains least to most significant; the MSB output chains from most to least significant. 197 Universal Digital Blocks (UDBs) Figure 21-26. Datapath Chaining Flow CE0i CL0i CE1i CL1i CE0 CL0 CE1 CL1 Z0 Z0i Z1 FF0 Z1i FF0i FF1 0 0 21.3.2.10 Z0 CAP0i CAP1 CAP1i Z0 Z1i FF0i FF1 Z1 FF0 FF1i UDB1 CAP0 CAP0i CAP1 CAP1i 0 CE0i CL0i CE1i CL1i CE0 CL0 CE1 CL1 Z0i Z1 FF0 FF1i UDB2 CAP0 CE0i CL0i CE1i CL1i CE0 CL0 CE1 CL1 FF1 0 0 Z0i 0 0 Z1i FF0i 0 0 FF1i CAP0 CAP0i 0 0 CAP1 UDB0 CAP1i 0 CO_MSB CI CO_MSB CI CO_MSB CI SOL_MSB SIR SOL_MSB SIR SOL_MSB SIR 0 0 0 CFBO CFBI CFBO CFBI CFBO CFBI SIL SOR SIL SOR SIL SOR CMSBI CMSBO CMSBI CMSBO Dynamic Configuration RAM CMSBI CMSBO An additional asynchronous read port is provided as a fast path to output these 16-bit words as control bits to the datapath. The asynchronous address inputs are selected from datapath inputs and can be generated from any of the possible signals on the channel routing, including I/O pins, PLD outputs, control block outputs, or other datapath outputs. The primary purpose of the asynchronous read path is to provide a fast single-cycle decode of datapath control bits. Each datapath contains a 16 bit-by-8 word dynamic configuration RAM, which is shown in Figure 21-27. The purpose of this RAM is to control the datapath configuration bits on a cycle-by-cycle basis, based on the clock selected for that datapath. This RAM has synchronous read and write ports for purposes of loading the configuration via the system bus. Figure 21-27. Configuration RAM I/O Read Only UDBLocal Bus wrl bus_data[15:0] R/W Read 16 bus_addr [2:0] Wr Ctrl RO Read Address 16 Bit-by-8 Word RAM Array Read/Write rad[2:0] Decoder Datapath Control Inputs Address Decoder 16 wrh 16 Config RAM dyn_cfg_ram [15:0] rd dpram The fields of this dynamic configuration RAM word are shown in the following tables. A description of the usage of each field follows. Register Address CFGRAM 61h - 6Fh (Odd) 15 14 FUNC[2:0] Register Address 7 CFGRAM 60h - 6Eh (Even) A0 WRSRC[1:0] 198 13 6 12 11 SRCA 5 4 A1 WRSRC[1:0] 10 9 SRCB[1:0] 8 SHIFT[1:0] 3 2 1 0 CFB EN CI SEL SI SEL CMPSEL PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Table 21-18. Dynamic Configuration Quick Reference Field Bits Parameter Values FUNC[2:0] 3 ALU Function 000 PASS 001 INC SRCA 010 DEC SRCA 011 ADD 100 SUB 101 XOR 110 AND 111 OR SRCA 1 ALU A Input Source 0 A0 1 A1 SRCB 2 ALU B Input Source 00 D0 01 D1 10 A0 11 A1 SHIFT[1:0] 2 SHIFT Function 00 PASS 01 Left Shift 10 Right Shift 11 Nibble Swap A0 WR SRC[1:0] 2 A0 Write Source 00 None 01 ALU 10 D0 11 F0 A1 WR SRC[1:0] 2 A1 Write Source 00 None 01 ALU 10 D1 11 F1 CFB EN 1 CRC Feedback Enable 0 Enable 1 Disable CI SEL 1 Carry In Configuration Select 0 ConfigA 1 ConfigBa SI SEL 1 Shift In Configuration Select 0 ConfigA 1 ConfigBa CMP SEL 1 Compare Configuration Select 0 ConfigA 1 ConfigBa a. For CI, SI, and CMP, the RAM fields select between two predefined static settings. See Static Register Configuration. 21.3.3 Status and Control Module A high level view of the Status and Control module is shown in Figure 21-28. The Control register drives into the routing to provide firmware control inputs to UDB operation. The Status register read from routing provides firmware a method of monitoring the state of UDB operation. Figure 21-28. Status and Control Registers System Bus 8-Bit Status Register (Read Only) 8-Bit Control Register (Write/Read) Routing Channel PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 199 Universal Digital Blocks (UDBs) A more detailed view of the Status and Control module is shown in Figure 21-29. The primary purpose of this block is to coordinate CPU firmware interaction with internal UDB operation. However, due to its rich connectivity to the routing matrix, this block may be configured to perform other functions. Figure 21-29. Status and Control Module Status and Control Module 7-Bit Period Register (same as Mask) 8-Bit Control Register From Datapath Parallel Output (po[7:0]) TC 8 Interrupt Gen EN/LD CTL 7-Bit Down Count CNT 7 8 8-Bit Status Register INT 8 sc_in[3:0] 7-Bit Mask Register (same as Period) 8 8 To Datapath Parallel Input (pi[7:0]) 4 4-Bit Sync 8 CFGx SC OUT CTL[1:0] 3 8 CFGx INT MD CFGx SYNC MD sc_out[7:0] 8 sc_io_out[3] sc_io_out[2:0] {sc_io_in[3:0],sc_in[3:0]} Horizontal Channel Routing Modes of operation include: ■ Status Input – The state of routing signals can be input and captured as status and read by the CPU or DMA. ■ Control Output – The CPU or DMA can write to the control register to drive the state of the routing. ■ Parallel Input – To datapath parallel input. ■ Parallel Output – From datapath parallel output. ■ Counter Mode – In this mode, the control register operates as a 7-bit down counter with programmable period and automatic reload. Routing inputs can be configured to control both the enable and reload of the counter. When this mode is enabled, control register operation is not available. ■ Sync Mode – In this mode, the status register operates as a 4-bit double synchronizer. When this mode is enabled, status register operation is not available. 21.3.3.1 Status and Control Mode When operating in status and control mode, this module functions as a status register, interrupt mask register, and control register in the configuration shown in Figure 21-30 on page 201. 200 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-30. Status and Control Operation 00: Read Transparently 01: Sticky, Clear on Read CFGx STAT MD[7:0] System Bus Read Write Reset (Routed Reset Read Only 8-Bit Control Register Read Write 8-Bit Status Register 7 7 7-Bit Mask Register from Reset and Clock Control Block 7 ACTL INT EN CFGx SC OUT CTL[1:0] INT SC OUT CTL bits must be set to select Control register bits for output 8 sc_out[7:0] CFGx INT MD 8 {sc_io_in[3:0],sc_in[3:0] sc_io_out[3] Status Register Operation Sticky Status, with Clear on Read One 8-bit, read only status register is available for each UDB. Inputs to this register come from any signal in the digital routing fabric. The Status register is nonretention; it loses its state across sleep intervals and is reset to 0x00 on wakeup. Each bit can be independently programmed to operate in one of two ways, as shown below: In this mode, the status register inputs are sampled on each cycle of the status and control clock. If the signal is high in a given sample, it is captured in the status bit and remains high, regardless of the subsequent state of the input. When the CPU or DMA reads the status register the bit is cleared. The status register clearing is independent of mode and occurs even if the UDB clock is disabled; it is based on the bus clock and occurs as part of the read operation. Table 21-19. Status Register STAT MD Description 0 Transparent read. A read returns the current value of the routed signal. 1 Sticky, clear on read. A high on the input is sampled and captured. It is cleared when the register is read. An important feature of the status register clearing operation is to note that the clear of status is only applied to the bits that are set. This allows other bits that are not set to continue to capture status, so that a coherent view of the process can be maintained. Status Latching During Read Figure 21-31 on page 202 shows the structure of the status read logic. The sticky status register is followed by a latch, which latches the status register data and holds it stable during the duration of the read cycle, regardless of the number of wait states in a given read. Transparent Status Read By default, a CPU read of this register transparently reads the state of the associated routing net. This mode can be used for a transient state that is computed and registered internally in the UDB. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 201 Universal Digital Blocks (UDBs) Figure 21-31. Status Read Logic S ticky /!T ra n sp a re n t R e a d L a tch 0 D D fro m R o u tin g Q AR S ta tu s a n d C o n tro l C lo ck U D B L o ca l B u s Q EN 1 S ticky S ta tu s R e g iste r S ta tu s R e g iste r Read E n d o f S ta tu s R e g iste r R e a d Interrupt Generation In most functions, interrupt generation is tied to the setting of status bits. As shown in Figure 21-31, this feature is built into the status register logic as the masking and OR reduction of status. Only the lower 7 bits of status input can be used with the built-in interrupt generation circuitry. The most significant bit is typically used as the interrupt output and may be routed to the interrupt controller through the digital routing. In this configuration, the MSB of the status register is read as the state of the interrupt bit. 21.3.3.2 output of the control register is driven directly to the routing on that write cycle. Figure 21-32. Control Register Direct Mode To Routing Data Bus Bus Write Clock Control Register Operation One 8-bit control register is available for each UDB. This operates as a standard read/write register on the system bus, where the output of these register bits are selectable as drivers into the digital routing fabric. Control Register Sync Mode The Control register is nonretention; it loses its contents across sleep intervals and is reset to 0x00 on wakeup. In Sync mode, as shown in Figure 21-33, the control register output is driven by a re-sampling register clocked by the currently selected Status and Control (SC) clock. This allows the timing of the output to be controlled by the selected SC clock, rather than the bus clock. Control Register Operating Modes Figure 21-33. Control Register Sync Mode There are three available modes that may be configured on a bit-by-bit basis. The configuration is controlled by the concatenation of the bits of the two 8-bit registers CTL_MD1[7:0] and CTL_MD0[7:0]. For example {CTL_MD1[0],CTL_MD0[0]} controls the mode for Control Register bit 0, as shown in Figure 21-20. To Routing Data Bus Bus Write Clock SC CLK Table 21-20. Mode for Control Register Bit 0 CTL MD Description 00 Direct mode 01 Sync mode 10 (reserved) 11 Pulse mode Control Register Direct Mode The default mode is Direct mode. As shown in Figure 21-32, when the Control Register is written by the CPU or DMA the 202 Control Register Pulse Mode Pulse mode is similar to Sync mode in that the control bit is re-sampled by the SC clock; the pulse starts on the first SC clock cycle following the bus write cycle. The output of the control bit is asserted for one full SC clock cycle. At the end of this clock cycle, the control bit is automatically reset. With this mode of operation, firmware can write a 1 to a control register bit to generate a pulse. After it is written as a 1 it PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Control Register Reset will be read back by firmware as a 1 until the completion of the pulse, after which it will be read back as a 0. The firmware can then write another 1 to start another pulse. A new pulse cannot be generated until the previous one is completed. Therefore the maximum frequency of pulse generation is every other SC clock cycle. The control register has two reset modes, controlled by the EXT RES configuration bit, as shown in Figure 21-34. When EXT RES is 0 (the default) then in sync or pulse mode the routed reset input resets the synced output but not the actual control bit. When EXT RES is 1 then the routed reset input resets both the control bit and the synced output. Figure 21-34. Control Register Reset Routed Reset 0 EXT RES 1 Static configuration bit res Data Bus To Routing res Bit by Bit CFG Bus Write Clock 21.3.3.3 Parallel Input/Output Mode In this mode, the status and control routing is connected to the datapath parallel in and parallel out signals. To enable this mode, the SC OUT configuration bits are set to select SC CLK datapath parallel out. The parallel input connection is always available, but these routing connections are shared with the status register inputs, counter control inputs, and the interrupt output. Figure 21-35. Parallel Input/Output Mode Datapath SC OUT CTL bits must be set to select datapath parallel out bits for output to routing. po[7:0] Datapath Parallel Out 8 sc_out[7:0] PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F pi[7:0] Datapath Parallel In 8 The INT MD and SYNC MD control bits should be cleared to enable SC_IO bits to input mode. {sc_io_in[3:0], sc_in[3:0]} 203 Universal Digital Blocks (UDBs) 21.3.3.4 Counter Mode As shown in Figure 21-36, when the block is in counter mode, a 7-bit down counter is exposed for use by UDB internal operation or firmware applications. This counter has the following features: ■ A 7-bit read/write period register. ■ A 7-bit read/write count register. It can be accessed only when the counter is disabled. ■ Automatic reload of the period to the count register on terminal count (0). ■ A firmware control bit in the Auxiliary Control Working register called CNT START, to start and stop the counter. (This is an overriding enable and must be set for optional routed enable to be operational.) ■ Selectable bits from the routing for optional dynamic control of the counter enable and load functions: ❐ EN, routed enable to start or stop counting. ❐ LD, routed load signal to force the reload of period. When this signal is asserted, it overrides a pending terminal count. It is level sensitive and continues to load the period while asserted. ■ The 7-bit count may be driven to the routing fabric as sc_out[6:0]. ■ The terminal count may be driven to the routing fabric as sc_out[7]. ■ In default mode the terminal count is registered. In alternate mode the terminal count is combinational. ■ In default mode, the routed enable, if used, must be asserted for routed load to operate. In alternate mode the routed enable and routed load signals operate independently. To enable the counter mode, the SC_OUT_CTl[1:0] bits must be set to counter output. In this mode the normal operation of the control register is not available. The status register can still be used for read operations, but should not be used to generate an interrupt because the mask register is reused as the counter period register. The Period register is retention and will maintain its state across sleep intervals. For a period of N clocks, the period value of N-1 should be loaded. N = 1 (period of 0) is not supported as a clock divide value, and will result in the terminal count output of a constant 1.The use of SYNC mode depends on whether or not the dynamic control inputs (LD/EN) are used. If they are not used, SYNC mode is unaffected. If they are used, SYNC mode is unavailable. Figure 21-36. Counter Mode P3B Bus System Bus Read Only* Read Write *Current count value is only readable when not enabled. 7-Bit Period Register 0: Reload is only controlled by terminal count 1: Reload is also controlled by routing Routed Reset from Reset and Clock Control Block RES CFGx ROUTE LD CFGx ROUTE EN LD 7-Bit Counter EN Zero Detect Terminal Count (TC) SC OUT CTL bits must be set to select the counter output as the selected output to routing. sc_out[7] 0: Enable is only controlled by firmware 1: Enable is also controlled by routing ACTL CNT START CFGx EN SEL[1:0] 7 sc_out[6:0] The INT MD and SYNC MD bits should be cleared to configure the SC_IO bits to input mode. CFGx LD SEL[1:0] 4 4 [7:4] [3:0] 8 {sc_io_in[3:0], sc_in[3:0]} 204 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) 21.3.3.5 Sync Mode 21.3.3.7 As shown in Figure 21-37, the status register can operate as a 4-bit double synchronizer, clocked by the current SC_CLK, when the SYNC MD bit is set. This mode may be used to implement local synchronization of asynchronous signals, such as GPIO inputs. When enabled, the signals to be synchronized are selected from SC_IN[3:0], the outputs are driven to the SC_IO_OUT[3:0] pins, and SYNC MD automatically puts the SC_IO pins into output mode. When in this mode, the normal operation of the status register is not available, and the status sticky bit mode is forced off, regardless of the control settings for this mode. The control register is not affected by the mode. The counter can still be used with limitations. No dynamic inputs (LD/EN) to the counter can be enabled in this mode. Figure 21-37. Sync Mode Sync Module (Status Register) Auxiliary Control Register The read-write Auxiliary Control register is a special register that controls fixed function hardware in the UDB. This register allows CPU or DMA to dynamically control the interrupt, FIFO, and counter operation. The register bits and descriptions are: Auxiliary Control Register 7 6 5 CNT START 4 3 2 1 0 INT EN FIFO1 LVL FIFO0 LVL FIFO1 CLR FIFO0 CLR FIFO0 Clear, FIFO1 Clear The FIFO0 CLR and FIFO1 CLR bits are used to reset the state of the associated FIFO. When a '1' is written to these bits, the state of the associated FIFO is cleared. These bits must be written back to '0' to allow FIFO operation to continue. When these bits are left asserted, the FIFOs operate as simple one-byte buffers, without status. FIFO0 Level, FIFO1 Level 7 6 5 4 3 4 2 1 0 4 The FIFO0 LVL and FIFO1 LVL bits control the level at which the 4-byte FIFO asserts bus status (when the bus is either reading or writing to the FIFO) to be asserted. The meaning of FIFO bus status depends on the configured direction, as shown in the table below. Table 21-21. FIFO Level Control Bits CFGx SYNC MD sc_in[3:0] sc_io_out[3:0] 0 Digital Routing 1 21.3.3.6 FIFOx Input Mode Output Mode LVL (Bus is Writing FIFO) (Bus is Reading FIFO) Not Full Not Empty At least 1 byte can be written At least 1 byte can be read At Least Half Empty At Least Half Full At least 2 bytes can be written At least 2 bytes can be read Status and Control Clocking The status and control registers require a clock selection for any of the following operating modes: ■ Status register with any bit set to sticky, clear on read mode. ■ Control register in counter mode. ■ Sync mode. The clock for this is allocated in the reset and clock control module. See 21.3.4 Reset and Clock Control Module on page 206. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Interrupt Enable When the status register’s generation logic is enabled, the INT EN bit gates the resulting interrupt signal. Count Start The CNT START bit may be used to enable and disable the counter (only valid when the SC_OUT_CTL[1:0] bits are configured for counter output mode). 205 Universal Digital Blocks (UDBs) 21.3.3.8 Status and Control Register Summary The table below summarizes the function of the status and control registers. Note that the control and mask registers are shared with the count and period registers and the meaning of these registers is mode dependent. Table 21-22. Status, Control Register Function Summary Mode Control/Count Control Control Out Count Count Out Status/SYNC Status Mask Status In or SYNC Status Count Perioda Status In Status Mask SYNC NAb Control Out or Count Out SYNC Mask/Period a. Note that in counter mode, the mask register is operating as a period register and cannot function as a mask register. Therefore, interrupt output is not available when counter mode is enabled. b. Note that in SYNC mode, the status register function is not available, and therefore, the mask register is unusable. However, it can be used as a period register for count mode. 21.3.4 Reset and Clock Control Module The primary function of the reset and clock block is to select a clock from the available global system clocks or bus clock for each of the PLDs, the datapath, and the status and control block. It also supplies dynamic and firmware-based resets to the UDB blocks. As shown in Figure 21-38, there are four clock control blocks, and one reset block. Four inputs are available for use from the routing matrix (RC_IN[3:0]). Each clock control block can select a clock enable source from these routing inputs, and there is also a multiplexer to select one of the routing inputs to be used as an external clock source. As shown, the external clock source selection can be optionally synchronized. There are a total of 10 clocks that can be selected for each UDB component: 8 global digital clocks, bus clock, and the selected external clock (ext clk). Any of the routed input signals (rc_in) can be used as either a level sensitive or edge sensitive enable. The reset function of this block provides a routed reset for the PLD blocks and SC counter, and a firmware reset capability to each block to support reconfiguration. The bus clock input to the reset and clock control is distinct from the system bus clock. This clock is called “bus_clk_app” because it is gated just like the other global digital clocks and used for UDB applications. The system bus clock is only used for I/O access and is automatically gated, per access. The datapath clock generator produces three clocks: one for the datapath in general, and one for each of the FIFOs. 206 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Figure 21-38. Reset and Clock Control global_enable From channel routing rc_in[3:0] bus_clk_app, gclks[7:0] rc_in_gated[3:0] PLD0 Clock Select/Enable pld0_clk (to PLD0) PLD1 Clock Select/Enable pld1_clk (to PLD1) ext_clk 2 CFGx EXT CLK SEL[1:0] bus_clk CFGx EXT SYNC dp_clk (to Datapath) DP Clock Select/Enable f0_clk (to FIFO0) f1_clk (to FIFO1) SC Clock Select/Enable sc_clk (to Status and Control) mf rc_in_gated[3:0] cnt_routed_ reset (to SC counter) sysreset Reset Select/Enable pld0_reset (firmware/system reset) pld1_reset (firmware/system reset) sc_reset (firmware/system reset) dp_reset (firmware/system reset) 21.3.4.1 Clock Control Figure 21-39 illustrates one instance of the clock selection and enable circuit. There are four of these circuits in each UDB: one for each of the PLD blocks, one for the datapath, and one for the status and control block. The main components of this circuit are a global clock selection multiplexer, clock inversion, clock enable selection multiplexer, clock enable inversion, and edge detect logic. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 207 Universal Digital Blocks (UDBs) Figure 21-39. Clock Select/Enable Control 3 1 rc_in_gated[3:0] 2 0 2 1 1 0 0 2 CFGx EN SEL[1:0] Enable Select 00: rc_in[0] 01: rc_in[1] 10: rc_in[2] 11: rc_in[3] FF clk Latch 2 CFGx EN INV CFGx EN MODE[1:0] Enable Mode 00: off 01: on 10: positive edge 11: level Enable Invert 0: true 1: inverted 1 {bus_clk_app,ext_clk, gclk[7:0]} Clock Select 0000: gclk[0] 0100: gclk[4] 0001: gclk[1] 0101: gclk[5] 0010: gclk[2] 0110: gclk[6] 0011: gclk[3] 0111: gclk[7] 1000: ext_clk 1001: bus_clk_app 0 4 CFGx CK SEL[3:0] 2 CFGx CK INV Clock Invert 0: true 1: inverted Clock Selection Clock Enable Inversion There are eight global digital clocks routed to all UDBs; any of these clocks may be selected. Global digital clocks are the output of user selectable clock dividers. See the Clocking System chapter on page 109. Another selection is bus clock, which is the highest frequency in the system. Called “bus_clk_app,” this signal is routed separately from the system bus clock. In addition, an external routing signal can be selected as a clock input to support direct-clocked functions such as SPI. Because application functions are mapped to arbitrary boundaries across UDBs, individual clock selection for each UDB subcomponent block supports a fine granularity of programming. The clock enable signal may be optionally inverted. This feature allows the clock enable to be generated in any polarity. Clock Inversion The selected clock may be optionally inverted. This limits the maximum frequency of operation due to the existence of one half cycle timing paths. Simultaneous bus writes and internal writes (for example writing a new count value while a counter is counting) are not supported when the internal clock is inverted and the same frequency as bus clock. This limitation affects A0, A1, D0, D1, and the Control register in counter mode. Clock Enable Selection The clock enable signal may be routed to any synchronous signal and can be selected from any of the four inputs from the routing matrix that are available to this block. 208 Clock Enable Mode By default, the clock enable is OFF. After configuring the target block operation, software can set the mode to one of the following using the CFGxEN MODE[1:0] register shown in Figure 21-39. Table 21-23. Clock Enable Mode Clock Enable Mode Description OFF Clock is OFF. ON Clock is ON. The selected global clock is free running. Positive Edge A gated clock is generated on each positive edge detect of the clock enable input. Maximum frequency of enable input is the selected global clock divided by two. Level Clocks are generated while the clock enable input is high ('1'). Clock Enable Usage There are two general usage scenarios for the clock enable. Firmware Enable – It is assumed that most functions require a firmware clock enable to start and stop the function. Because the boundary of a function mapped into the UDB array is arbitrary–it may span multiple UDBs and/or portions of UDBs–there must be a way to enable a given function atomically. This is typically implemented from a bit PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) in a control register routed to one or more clock enable inputs. This scenario also supports the case where applications require multiple, unrelated blocks to be enabled simultaneously. Emulated Local Clock Generation – This feature allows local clocks to be generated by UDBs, and distributed to other UDBs in the array by using a synchronous clock enable implementation scheme, rather than directly clocking from one UDB to another. Using the positive edge feature of the clock enable mode eliminates restrictions on the duty cycle of the clock enable waveform. Special FIFO Clocking The datapath FIFOs have special clocking considerations. By default, the FIFO clocks follow the same configuration as the datapath clock. However, the FIFOs have special control bits that alter the clock configuration: ■ Each FIFO clock can be inverted with respect to the selected datapath clock polarity. ■ When FIFO FAST mode is set, the bus clock overrides the datapath clock selection normally in use by the FIFO. 21.3.4.2 Reset Control There are two modes of reset control: legacy mode and standard mode. The modes are controlled by the ALT RES bit in each UDB configuration register CFG31. The default for this bit is 0 (legacy mode); it is recommended that it be set to 1 for standard mode. Standard mode has greater granularity - routed resets can be used by individual blocks within the UDB. Contact Cypress for information on legacy mode reset. PLD Reset Control Figure 21-40 shows the PLD reset system. Figure 21-40. PLD Reset Structure PLD0 pld_routed_reset rc_in[3:0] M C sysreset 2 CFGx PLD0 RES SEL[1:0] Reset Select 00: rc_in[0] 01: rc_in[1] 10: rc_in[2] 11: rc_in[3] CFGx PLD0 RES POL Reset Invert 0: true 1: inverted M C SSEL routed reset 1 M C 0 SSEL M C set D Q QB res 1 0 PLD1 M C System Reset RSEL PLD Macrocell M C M C M C Datapath Reset Control Figure 21-41 shows the datapath reset system. The routed reset is applied to all datapath registers and states except the data registers D0 and D1. The data registers are retention registers. The FIFO data is unknown after reset because it is RAM based. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 209 Universal Digital Blocks (UDBs) Figure 21-41. Datapath Reset Structure sysreset_ret RES Accumulator Data Registers RES Accumulator Accumulators Output Sync Registers RES Carry Out Register sysrese t RES Shift Out Left Register RES rc_in[3:0] RES Shift Out Right Register 2 CFGx DP RES SEL[1:0] Reset Select 00: rc_in[0] 01: rc_in[1] 10: rc_in[2] 11: rc_in[3] 210 CFGx DP RES POL Reset Invert 0: true 1: inverted CFGx EN RES DP ACTL F0 CLR RES FIFO0 Status ACTL F1 CLR RES FIFO1 Status PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) Status and Control Reset Control Figure 21-42 shows the status and control block reset system. The status and control/count registers share the routed reset, however they are individually enabled. The mask/period and auxiliary control registers are retention registers. Figure 21-42. Status and Control Reset Control sysreset_ret All elements of the Datapath are reset by the selected DP routed reset signal, EXCEPT the Data Registers RES Accumulator Data Registers RES Accumulator Accumulators RES Carry Out Register RES Shift Out Left Register RES Shift Out Right Register Output Sync Registers sysreset RES rc_in[3:0] 2 CFGx DP RES SEL[1:0] Reset Select 00: rc_in[0] 01: rc_in[1] 10: rc_in[2] 11: rc_in[3] 21.3.4.3 CFGx DP RES POL Reset Invert 0: true 1: inverted CFGx EN RES DP UDB POR Initialization ACTL F0 CLR RES FIFO0 Status ACTL F1 CLR RES FIFO1 Status As a result of this initialization, conflicting drive states on the routing are avoided and initial configuration occurs in an order-independent sequence. Register and State Initialization Table 21-24. UDB POR State Initialization State Element State Element POR State Configuration Latches CFG 0 - 31 0 Ax, Dx, CTL, ACTL, MSK Accumulators, data registers, auxiliary control register, mask register 0 ST, MC Status and macrocell read only registers 0 DP CFG RAM & Fx (FIFOs) Datapath configuration RAM and FIFO RAM Unknown PLD RAM PLD configuration RAM Unknown Routing Initialization On POR, the state of input and output routing is as follows: ■ All outputs from the UDB that drive into the routing matrix are held at '0'. ■ All drivers out of the routing and into UDB inputs are initially gated to '0'. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 211 Universal Digital Blocks (UDBs) 21.3.5 UDB Addressing There are three unique address spaces in a UDB pair: ■ ■ ■ due to the even address alignment. The upper 4 bits is still the register number. 8-Bit Working Registers – A bus master that can only access 8 bits of data per bus cycle can use this address space to read or write any UDB working register. These are the registers with which the CPU and DMA interact during normal operation. Figure 21-43. UDB Working Registers 8-Bit Addresses UDB Working Base + 0xh 16-Bit Working Registers – A bus master with 16-bit capability, such as the DMA or the PSoC 5LP CortexM3, can access 16 bits per bus cycle to facilitate the data transfer of functions that are inherently 16 bits or greater. Although this address space is mapped into a different area than the 8-bit space, the same registers are accessed, two registers at a time. 8- or 16-Bit Configuration Registers – These registers configure the UDB to perform a function. When configured, they are normally left in a static state during operation. These registers maintain their state through sleep. 21.3.5.1 16-Bit Addresses Working Register Address Space A0 0xh 1xh A1 2xh 2xh D0 4xh 3xh D1 6xh 4xh F0 8xh 5xh F1 Axh 6xh ST Cxh 7xh CTL/CNT Exh 8xh MSK/PER 10xh 9xh ACTL 12xh Axh MC 14xh Bxh Working registers are accessed during normal operation and include accumulators, data registers, FIFOs, status and control registers, mask register, and the auxiliary control register. 16xh 8-Bit Working Register Access In this mode, all UDB registers are accessed on bytealigned addresses. In 8-bit register access mode, as shown in Figure 21-44, all data bytes written to the UDBs are aligned with the low byte of the 16-bit UDB bus. Figure 21-43 shows the register map for one UDB. On the right in Figure 21-43 is the 16-bit address, which is always even aligned. The UDB number is 5 bits instead of 4, Only one byte at a time can be accessed in this mode. Figure 21-44. 8-Bit Working Register Access UDB 2 UDB 1 UDB 0 A0 A0 A0 A1 A1 A1 Low byte Low byte Low byte 16-Bit UDB Array Data Bus 212 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) 16-Bit Working Register Address Space The 16-bit address space is designed for efficient DMA access and to provide support for CPU firmware access in processors that can support it, such as the Cortex-M3 in PSoC 5LP. There are two modes of 16-bit register access, the “default” mode and the “concat” mode. As shown in Figure 21-45, the default mode accesses a given register in UDB 'i' in the lower byte and the same register in UDB 'i+1' in the upper byte. This makes 16-bit data handling efficient in neighboring UDBs (address order) that are configured as a 16-bit function. Figure 21-45. 16-Bit Working Register Default Access Mode UDB 2 UDB 1 UDB 0 A0 A0 A0 A1 A1 A1 High byte Low byte Low byte 16 bits at UDB 2 High byte Low byte 16 bits at UDB 0 16 bits at UDB 1 16-Bit UDB Array Data Bus In concat mode, the registers of a single UDB are concatenated to form 16-bit registers as shown in Figure 21-46. In this mode, the 16-bit UDB array data bus has access to pairs of registers in the UDB in the format shown in the figure. For example, an access at A0 accesses A0 in the low byte and A1 in the high byte. Figure 21-46. 16-Bit Working Register Concat Access Mode UDB i A1 A0 D1 D0 F1 F0 CTL/CNT ST ACTL MSK/PER 00h MC High byte 16 bits at UDB i Low byte There is a limitation in the use of DMA with respect to the 16-bit working register address space. It is inefficient for use when the function is greater than 16 bits. This is because the addressing overlaps, as shown in Table 21-25. Table 21-25. Optimized Address Space for 16-Bit UDB Function Address Upper Byte Goes Lower Byte Goes 0 UDB1 UDB0 2 UDB2 UDB1 4 UDB3 UDB2 When the DMA transfers 16 bits to address 0, the lower and upper bytes are written to UDB0 and UDB1, respectively. On the next 16 bit DMA transfer at address 2, you overwrite the value in UDB1 with the lower byte of that transfer. To avoid having to provide redundant data organization in memory buffers to support this addressing, it is recommended that 8-bit DMA transfers in the 8-bit working space be used for functions over 16 bits. 16-Bit UDB Array Data Bus PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 213 Universal Digital Blocks (UDBs) 21.3.5.2 Configuration Register Address Space Configuration is done at the UDB pair level. A UDB pair consists of two UDBs and an associated routing channel, as shown in Figure 21-47. Figure 21-47. UDB Pair Configuration Address Map UDB Pair k Base + 0 21.3.5.4 Routing Configuration Address Space UDB routing configuration consists of embedded RAM bits to control the state of transmission gate switches, segmentation, and input/output buffers. For more information, see the UDB Array and Digital System Interconnect chapter on page 217. UDB i 128 bytes 80h UDB i+1 128 bytes UDB Pair k 512 bytes 100h UDB Pair k Routing 256 bytes 200h 21.3.5.3 UDB Configuration Address Space Figure 21-48 shows the address map for configuration of a given UDB. As shown, this UDB configuration space is replicated for the two UDBs in the UDB pair. There are 128 bytes (7 bits of address) reserved for each UDB configuration, which is organized in 16-bit width. There are individual byte write enables for this address space to support both 16- and 8-bit access. Note that 16-bit access on odd boundaries is not supported. Reads always return 16 bits in configuration space, and the byte not required can be ignored. Figure 21-48. UDB Configuration Address Space Write High Byte Write Low Byte 00h PLD0/PLD1 64 bytes (32 words x 16 bits) 40h UDB Config Registers (32 bytes) (16 words x 16 bits) 60h 128 bytes Dynamic Configuration RAM (16 bytes) (8 words x 16 bits) 70h Reserved (16 bytes) 80h (MS Byte) (LS byte) Read Word (16 bits) 214 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Universal Digital Blocks (UDBs) 21.3.6 System Bus Access Coherency UDB registers have dual access modes: ■ System bus access, where the CPU or DMA is reading or writing a UDB register. ■ UDB internal access, where the UDB function is updating or using the contents of a register. 21.3.6.1 Simultaneous System Bus Access The following table lists the possible simultaneous access events and required behavior: Table 21-26. Simultaneous System Bus Access Register UDB Write Bus Write UDB Write Bus Read Ax UDB Read Bus Write Undefined result Not allowed directlya, b Not supported (UDB and bus must be opposite access) If FIFO status flags are used, no simultaneous read/write at the same location is possible ST NA, bus does not write Bus reads previous value CTL NA, UDB does not write CNT Undefined result Dx Fx UDB reads previous value Current value is read by both Not supported (UDB and bus must be opposite access) NA, UDB does not read Not allowed directlyc UDB reads previous value ACTL MSK UDB Read Bus Read NA, UDB does not write Current value is read by both PER MC (RO) NA, bus does not write Not allowed directlyd NA, bus does not write a. The Ax registers can be safely read by using software capture feature of the FIFOs. b. The Dx registers can only be written to dynamically by the FIFOs. When this mode is programmed, direct read of the Dx registers is not allowed. c. The CNT register can only be safely read when it is disabled. An alternative for dynamically reading the CNT value is to route the output to the SC register (in transparent mode). d. MC register bits can also be routed to the status register (in transparent mode) inputs for safe reading. 21.3.6.2 Coherent Accumulator Access (Atomic Reads and Writes) The UDB accumulators are the primary target of data computation. Therefore, reading these registers directly during normal operation gives an undefined result, as indicated in the table above). However, there is built-in support for atomic reads in the form of software capture, which is implemented across chained blocks. In this usage model, a read of the least significant accumulator transfers the data from all chained blocks to their associated FIFOs. This operation is explained in FIFO Software Capture Mode on page 186. Atomic writes to the accumulator can be implemented programmatically. Individual writes can be performed to the input FIFOs, and then the status signal of the last FIFO written can be routed to all associated blocks and simultaneously transfer the FIFO data into the Dx or Ax registers. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 215 Universal Digital Blocks (UDBs) 21.4 UDB Working Register Reference All registers except the FIFO are cleared upon any system reset. The FIFO status is cleared, but FIFO data is random. These registers are not retention registers so they must be reset upon wakeup from a power cut-off sequence. Register 8-Bit Address 16-Bit Address 7 6 5 4 3 2 1 0 FIFO1 CLR FIFO0 CLR Datapath Registers A0 0xh 00xh A1 1xh 02xh D0 2xh 04xh D1 3xh 06xh F0 4xh 08xh F1 5xh 0Axh A0[7:0] (Accumulator 0 Value) A1[7:0] (Accumulator 1 Value) D0[7:0] (Data Register 0) D1[7:0] (Data Register 1) F0[7:0] (FIFO 0) F1[7:0] (FIFO 1) Status and Control Registers ST 6xh 0Cxh CTL/CNT 7xh 0Exh MSK/PER 8xh 10xh ST[7:0] (Status Register) CTL[7:0] / CNT[6:0] (Control / Count Register) MSK[6:0] / PER[6:0] (Interrupt Mask / Period Register) Auxiliary Control Register ACTL 9xh 12xh CNT START INT EN FIFO1 LVL FIFO0 LVL PLD Macrocell Register MC 216 Axh 14xh PLD1 MC[3:0] PLD0 MC[3:0] PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 22. UDB Array and Digital System Interconnect This chapter describes the structure of the UDB Array and Digital System Interconnect (DSI). Universal digital blocks (UDBs) are organized in the form of a two-dimensional array with programmable interconnect provided by the DSI. In addition to connecting UDB components, the DSI routing also provides connection between other hardware resources on the device, such as I/O pins, interrupts, and fixed function blocks. 22.1 Features ■ Offers a homogeneous array of UDBs which provide flexible function mapping ■ Provides array level interconnect routing between the components of the UDB hardware ■ Provides device level interconnect routing between UDBs, device peripherals, and I/O pins 22.2 Block Diagram Figure 22-1 illustrates the programmable digital architecture for PSoC 5LP. Figure 22-1. Programmable Digital Architecture Digital Core System and Fixed Function Peripherals IO Port INT/DMA Controller IO Port INT/DMA Routing UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB UDB Array UDB Array DSI Routing Interface Digital Core System and Fixed Function Peripherals PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F IO Port IO Port DSI Routing Interface 217 UDB Array and Digital System Interconnect The main components of this system are: ■ ■ ■ system blocks that require connectivity are routed to this interface at the UDB array, which allows connections into the core of the array or directly between device peripherals. UDB Array:- UDB blocks are arrayed within a matrix of programmable interconnect. UDB pairs consisting of 2 UDBs are the basic building blocks of the UDB array. UDB pairs are tiled to create an array. UDB pairs can connect with neighboring UDB pairs in seamless fashion Signals in this category include: DSI- Routing interface tiled at top and bottom of UDB array core. Provides general purpose programmable routing between device peripherals, including UDBs, I/ Os and fixed function blocks. System Interface (not shown)- Built in 8/16-bit bus interface with parallel access to all registers to support fast configuration. Also provides clock distribution and clock gating functionality. The following section explain in detail the DSI routing and System Interface. 22.3 ■ Interrupt requests from all digital peripherals in the system ■ DMA requests from all digital peripherals in the system ■ Digital peripheral data signals that need flexible routing to I/Os ■ Digital peripheral data signals that need connections to UDBs ■ Connections to the interrupt and DMA controllers ■ Connection to I/O pins ■ Connection to analog system digital signals Figure 22-2 and Figure 22-4 show some examples of the device peripherals that are connected to this interface, including UDBs, I/Os, analog peripherals, interrupts, DMA, and fixed function peripherals. How It Works The purpose of the DSI is to provide general purpose programmable connectivity across the device. Peripherals and Figure 22-2. DSI Example Connections to the Interrupt and DMA Controller DMA Controller Interrupt Controller CAN Interrupt request CAN USB REQ REQ USB DMA request DSI Routing Interface UDB DMA Request UDB UDB UDB UDB Interrupt Request UDB 218 UDB UDB PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F UDB Array and Digital System Interconnect Figure 22-3. DSI Example Connections between Peripherals, I/O Pins, and UDBs I2C I/O Pin I/O Pin SDA IN Timer SDA SDA OUT IN SDA OUT I/O Pin EN TC DSI Routing Interface UDB UDB UDB UDB UDB UDB PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 219 UDB Array and Digital System Interconnect 22.4 UDB Array System Interface The system interface consists of infrastructure blocks that distribute and interface the device system bus to the UDB array bus and to the UDB blocks, the DSI channel routing, and the UDB pair channel routing. Depending on the configuration of the array, there is one or more AHB interfaces that connect to PHUB spokes providing an interface to the UDB array system bus. Both 8-bit and 16-bit bus access is supported. The system interface also provides support for clock distribution and gating for the digital global clocks and bus clock. A gated clock tree distribution is implemented to allow only those clocks that are in use to be activated. There are eight digital global clocks, plus the application bus clock, routed to each bank of UDBs. The UDB local interface blocks contain clock gating control registers, which must be set by configuration firmware to enable clock distribution. There are four registers in each block: ■ 8-Bit MDCLK_EN (Master Digital Clock Enable) – This register individually enables the digital global clocks at the input to the UDB array. ■ 1-Bit MBCLK_EN (Master Bus Clock Enable) – This register individually enables the application bus clock at the input to the UDB array. ■ 8-Bit DCLK_ENx (Quadrant Digital Clock Enable) – This register individually enables the digital global clocks to the associated quadrant (4 UDBs) of the UDB array. ■ 1-Bit BCLK_ENx (Quadrant Bus Clock Enable) – This register individually enables the bus clock to the associated quadrant (4 UDBs) of the UDB array. It also contains bits to put the associated routing channel RAM into global write mode. Following are the system interface components: ■ AHB Interface – Connects to a standard PHUB spoke and provides support for up to 1 bank of UDBs (16). Controls array wait states and translates AHB signaling into array register and routing configuration access control. ■ DSI Channel IF – Interfaces the UDB array bus to the DSI routing channel for writing and reading configuration. ■ UDB Local IF – Interfaces the UDB array bus to the UDB blocks for registers and RAM access, and provides local clock gating. ■ UDB Pair Channel – Interfaces the UDB array bus to the pair routing channel for writing and reading configuration. ■ Bank IF – Contains the master clock gating and bank wide configuration interface signals. ■ 8-Bit WAIT_CFG Register – Sets the read and write wait states for working and configuration registers. ■ 4-Bit BANK_CTL Register – Contains global bank control bits. ❐ One bit to globally enable all DSI inputs. On POR, all DSI inputs are gated off until the DSI channel is configured. This bit globally enables DSI inputs to drive the routing. ❐ One to disable all UDB status register clear-on-read function for debug support. ❐ One to put the embedded DP RAM into test mode for DFT support. ❐ One to put the bank into global write mode, also for DFT support. 220 22.4.1 UDB Array POR Initialization The key aspects of POR initialization are summarized as follows. ■ All UDB clocks are gated off. There are three levels of clock gating configuration: one at the UDB level for each individual block clock control and a set of registers at the array level that controls master and quadrant clock gating. ■ The state of all drivers into the routing matrix is gated to ‘0’ with a global routing enable control. This includes UDB block outputs, DSI inputs, and segmentation buffers. Because the routing is initialized to a random state, the state of routing nets will be either ‘0’ or ‘Z’. ■ The inputs of all routing output buffers, including segmentation buffers, are gated to ‘0’ with a global routing enable control. This prevents floating routes from causing high power states. This also drives the buffer outputs to ‘0’ and that is the state for all DSI outputs. ■ Configuration can occur in an order-independent way. When configuration is complete, each bank of UDBs has a global routing enable which is asserted to activate the connections (forced gating is disabled). ■ After routing is enabled, a global clock enable bit (bank enable) can be set (residing in the power manager) which then enables clocking in the array. The bank enable bit prevents any spurious operation until the array is completely configured. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F UDB Array and Digital System Interconnect 22.4.2 UDB POR Configuration Sequence The previous section documented the POR state for the UDB array. From this initial state, configuration will proceed in the order shown in Figure 22-4. Figure 22-4. POR Configuration Sequence Step 1 POR Configuration Engine Performs These Steps Step 2 Configure Array Step 3 Enable Clock Configuration to Input Routed Clock Enable Routing configuration is random, but due to the fact that routing and clocking is disabled, array is in a benign state. Configure routing, PLD RAM, Datapath RAM, and Datapath CFG registers in the UDB blocks. Also can configure working registers if desired. Configuration is order independent. Now that routing and UDB block configuration is done, you can enable clock configuration input to use the routed enable for the clock. The clock trees are still gated off. Step 4 Enable Quadrant Clock Enables and Then Global Clock Enables The set of clocks to enable is determined by how clocks are allocated in the array. Enabling only clocks that are used reduces the UDB array power. This does not actually enable the clocks. There is a global clock enable for the UDB array in the power manager block (Step 6). Step 5 Enable Routing Routing between blocks and initial DSI outputs are enabled. Step 6 Enable the UDB Bank Enable Bit in the Power Manager Step 7 - Done. User Can Write to Firmware Enable Bits in Control Register to Start Functions PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F There is one bit per bank (in one register) in the Power Manager register set. This allows for a global atomic clock enable for the entire UDB array. Need to start from the sources and work forward to avoid temporary high current states. Need to include a routed enable to implement an atomic firmware enable. 221 UDB Array and Digital System Interconnect 22.4.2.1 Quadrant Route Disable To support fast bring up of initial functionality, the Quadrant Bus Clock Enable register contains a bit called Route Disable to disable the routing for the associated UDB quadrant (2 UDB pairs). By default, this bit is cleared and is not disabling the routing. If this bit is set to ‘1’ during initial configuration, the associated channel routing RAM does not need to be configured. The global route enable bit can be set and this routing will remain in a benign state. Routing configuration for this quadrant can occur at a future time when this bit can be cleared to ‘0’ to enable the routing (assuming that the global route enable bit is set). 22.4.3 UDB Sleep and Power Control The UDB array has support for low-power operation in the form of a sleep control input and power switch control inputs. All static configurations are on the “keep-alive” domain which retains state during a sleep/power down period. However, all application level working registers, including the accumulators, the data registers, the FIFO, control and status registers, etc., lose their state and must be reinitialized on power up. Nonretention registers and FIFO state are reset after a sleep period to insure a good initial state. 22.4.4 UDB Register References and Address Mapping UDB registers are classified as shown in the Figure 22-5. There are five address spaces: one for 8-bit working registers (registers that are accessed during normal operation), one for 16-bit working registers, and three for configuration. Each bank of UDBs is on a separate spoke, so a total of 6 select lines are generated from the PHUB to support the UDB array. The working registers are on the main 64K page (Page 0). The configuration registers have their own page (Page 1). Details of these registers are located in the PSoC® 5LP Registers TRM (Technical Reference Manual). Figure 22-5. UDB Register Mapping UDB Registers Working Registers 8-Bit Working 222 Configuration Registers 16-Bit Working UDB Bank Array DSI Interface Configuration Concatinated UDB Pair Configuration Default Access UDB Channel Interface Configuration Bank Control PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F UDB Array and Digital System Interconnect Figure 22-6 shows the register mapping for working and configuration registers of UDB and DSI. Figure 22-6. UDB Array Base Addresses Address Page 1 Address Page 0 6400h 6500h 8-Bit Working Register Address Space Bank 0 8-Bit Working Registers Bank 1 8-Bit Working Registers 0000h Bank 0 Array Configuration 512 Bytes 6600h Reserved 6700h Reserved 8K Bytes 1000h UDB Array Configuration Space (maximum used: 13K) Bank 1 Array Configuration 2000h Reserved 6800h Bank 0 16-Bit Working Registers 3000h Reserved 1K Bytes 6A00h 16-Bit Working Register Address Space Bank 1 16-Bit Working Registers 4000h DSI Configuration Space (maximum used: 4K) 6C00h Reserved Bank Configuration Space DSI Configuration 3K Bytes 4800h Reserved 5000h Bank 0 Control 5010h Bank 1 Control 5020h Reserved 32 Bytes Reserved 6E00h Reserved Reserved Interrupt and DMA Configuration 5100h INT/DMA Configuration 256 Bytes 5200h 7000h PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 223 UDB Array and Digital System Interconnect 224 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 23. Controller Area Network (CAN) The CAN peripheral is a fully functional Controller Area Network (CAN) supporting communication baud rates up to 1 Mbps. The CAN controller is CAN2.0A and CAN2.0B compliant per the ISO-11898 specification. The CAN protocol was originally designed for automotive applications with a focus on a high level of fault detection and recovery. This ensures high communication reliability at a low cost. Because of its success in automotive applications, CAN is used as a standard communication protocol for motion oriented embedded control applications (CANOpen) and factory automation applications (DeviceNet). The CAN features allow the efficient implementation of higher level protocols without affecting the performance of the microcontroller CPU. Figure 23-1. CAN Bus System Implementation CAN Node 1 CAN Node 2 CAN Node n PSOC CAN Drivers CAN Controller RX TX EN CAN Transceiver CAN_H CAN_L CAN_H CAN_L CAN_H CAN_L CAN Bus 23.1 ■ ■ ■ Features Compliant with CAN2.0A/B protocol specification: ❐ Standard and extended frames ❐ Remote Transmission Request (RTR) support ❐ Programmable bit rate up to 1 Mbps Receive path: ❐ 16 receive message buffers ❐ 16 acceptance filters and acceptance masks ❐ DeviceNet addressing support ❐ Option to link multiple receive buffers to form a hardware FIFO Transmit path: ❐ Eight transmit message buffers ❐ Programmable priority for each transmit message buffer ■ CAN Transmit (Tx), Receive (Rx), and EN can be routed to any I/O ■ Listen Only mode for auto baud detection ■ Ability to wake up the device from Sleep mode on bus activity PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 225 Controller Area Network (CAN) 23.2 Block Diagram To transmit a message, the host controller stores a message in the transmit message buffer and informs the transmit message handler which transmits the message. When a message is received, it is stored in the memory buffer and the host controller can process it on demand. The transmission and reception are mainly governed by the status and configuration registers. The various interrupts of the CAN module are handled by the interrupt controller unit. Figure 23-2 illustrates this process. Figure 23-2. CAN Block Diagram Memory Buffer (SRAM) CAN Module Memory Arbiter Receive Message Handler TO CPU/PHUB Advanced Peripheral Bus (APB) Coupler CAN Transmit Message Handler Bus CAN Framer Interrupt Controller Status and Configuration Control and Command 23.3 CAN Message Frames In CAN the transmission and reception of messages are governed by four main frame types: ■ Data frames ■ Remote frame ■ Error frame ■ Overload frame 23.3.1 23.3.1.1 Standard Data Frame The standard data frame for CAN is illustrated in Figure 23-3 on page 227. Data Frames Data frames are mainly used to transfer data between transmitter and receiver. CAN supports mainly two types of data frames: Standard Data Frame and Extended Data Frame. For a CAN frame, '0' is referred to as the dominant bit and '1' as a recessive bit. 226 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Controller Area Network (CAN) Figure 23-3. Standard Data Frame Arbitration Field Interframe Space Start of Frame Identifier (11 Bits) RTR IDE R0 DLC (4 Bits) Data (Maximum 8 Bytes) CRC Field ACK Field End of Frame Interframe Space Control Field Start of frame. The beginning of a data frame is indicated by the start of frame bit. It is a single dominant bit. Data Length Code (DLC). These 4 bits indicate the number of data bytes in the data field. The IDE, R0, and DLC bits constitute the Control Field. Identifier. For a basic CAN data frame, the identifier is 11 bits long. It is mainly used to filter the data at the receiver side. Data Field. This field contains the message data. It is of variable length and can have a maximum of 8 bytes. Remote Transmission Request Bit (RTR). Set the RTR bit '0' (dominant) for a data frame and set to '1' (recessive) for a remote frame. The identifier and RTR bit are known as the Arbitration Field. Cyclic Redundancy Check (CRC). Frame checking is carried out by the method of cyclic redundancy check (CRC). The field consists of a 15-bit CRC code followed by a CRC delimiter. Extended Identifier Bit (IDE). This bit must be a ‘0’ (dominant for a standard data frame and a ‘1’ (recessive) for extended CAN data frame. Acknowledgement Field (ACK). The ACK field is two bits long and recessive by default. When a receiver receives a message correctly, it overwrites the ACK field with a dominant bit. R0. Reserved bit. End of Frame. The end of every frame is indicated by End of Frame field and it consists of seven recessive bits. 23.3.1.2 Extended Data Frame The extended CAN frame format is illustrated in Figure 23-4. The extended CAN has a 29-bit identifier. It is arranged as an 11-bit identifier field and an 18-bit identifier field separated by a Substitute Remote Request (SRR) bit and an IDE bit. The SRR bit is in the same position as the RTR bit in the standard frame, and is recessive. The IDE bit is set for extended frames. The Control Field of the extended data frame has an additional reserve bit ‘R1’ compared to the standard data frame. Figure 23-4. Extended Data Frame Arbitration Field Interframe Space Start of Frame Identifier (11 Bits) SRR IDE Identifier (18 Bits) RTR R1 R0 DLC (4 Bits) Data (Maximum 8 Bytes) CRC Field ACK Field End of Frame Interframe Space Control Field PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 227 Controller Area Network (CAN) 23.3.2 Remote Frame thereby forcing all other nodes to send out error flags resulting in a series of six to twelve dominant bits on the bus. The CAN bus allows a destination node to request data from the source by sending a Remote Frame. There are two differences between a Data Frame and a Remote Frame. First, the RTR bit is transmitted as a recessive bit in the remote Frame. Second, there is no Data Field in the Remote Frame. Error Passive Flag. An error passive flag consists of six recessive bits. When an error passive station detects an error it sends a passive error flag. A passive error does not affect any other nodes and the error is detected only if the transmitting node detects a bus error. The Error Delimiter consists of eight recessive bits. For extended remote frame, the SRR bit is also transmitted as a recessive bit. 23.3.4 Interframe Space. Interframe space separates the data frames and remote frames from the preceding frames. 23.3.3 The overload frame (EOF) consists of an overload flag and an overload delimiter. CAN supports reactive overload frame which is activated when the following conditions occur: Error Frame The Error frame is generated by a node when it detects any bus error. The error frame consists of an error flag and error delimiter. The error flag are classified into two types: error active flag and error passive flag. Error Active Flag. When an error active station detects an error it sends six dominant bits as an active error flag. The format of the error flag thus violates the rule of bit stuffing 23.4 Overload Frame ■ Detection of a dominant bit during first two bits of intermission ■ Detection of a dominant bit in the last bit of EOF by a receiver ■ Detection of a dominant bit by any node at the last bit of error delimiter or overload delimiter Transmitting Messages in CAN The CAN module supports eight transmit message holding buffers. An internal priority arbiter selects the message according to the chosen arbitration scheme. The arbitration scheme is either a round robin or fixed priority scheme. When a message is transmitted or when there is a message arbitration loss, the priority arbiter re-evaluates the message priority of the next message. The receive message buffers can also transmit remote transmit requests, which are explained later in this chapter. 228 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Controller Area Network (CAN) Figure 23-5. Transmit (Tx) Block Diagram CAN Module TxREQ TxMESSAGE0 TxREQ TxMESSAGE1 CAN Bus To CPU/PHUB TxREQ TxMESSAGE7 ABP Bus Coupler Priority Arbiter RxMESSAGE0 CAN Framer RTR REQ RTR REQ RxMESSAGE1 RTR REQ RxMESSAGE15 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 229 Controller Area Network (CAN) 23.4.1 Message Arbitration flag equal to zero. PSoC 5LP CAN block initialization must be complete before starting any message transmission on the CAN bus. If initialization is not complete before the first CAN message, the device may fail to comply with the CAN start up synchronization wait period. The priority arbiter supports a round robin and fixed priority arbitration. The arbitration mode is selected using the configuration register. Round Robin. In a round robin scheme, Buffer 0 is selected first, then Buffer 1 and so on till Buffer 7, and it continues again with Buffer 0 thus forming a cycle. A particular buffer is only selected if its TxREQ flag is set. This scheme guarantees that all buffers receive the same probability to send a message. a. For standard data frame, write '0' (dominant) to the RTR and IDE bit. Fixed Priority. Buffer 0 has the highest priority. Designate Buffer 0 as the buffer for critical messages to guarantee that message is sent first. Priority arbitration is selected using the CFG_ARBITER bit in the Configuration register (CAN_CSR_CFG[12]). c. The 11-bit message identifiers are written to the ID[31:21] bit field. b. Write the DLC bits appropriately to specify the number of data bytes to be transferred. The maximum number of data bytes is limited to eight. Data bytes with MSb (most significant bit) first in each byte are written in D0, D1…D7 locations. 2. Choose an appropriate priority arbitration scheme. The internal message priority arbiter selects the message according to the chosen arbitration scheme. 3. Request transmission by setting the respective TxREQ flag to ‘1’. Note RTR reply messages have higher priority than Tx messages. If there are multiple Tx messages in the transmission queue of the CAN block while an RTR message is received, it completes transmission of the next Tx message before sending the RTR reply. 23.4.2 4. The TxREQ flag remains set as long as the message transmit request is pending. The content of the message buffer must not be changed while the TxREQ flag is set. After the message is transmitted, the TxREQ flag is cleared and the TX_MSG interrupt status bit [CAN_CSR_INT_SR[11] in the interrupt status register CAN_CSR_INT_SR is asserted. The interrupt status bit is only asserted if the TxINT ENBL (CAN_TX[n]_CMD[2]) is set to ‘1’. Message Transmit Process Figure 23-6 shows the registers associated with a message that is transmitted. The main steps in transmitting a standard data frame are: 1. Write the message into an empty transmit message holding buffer. An empty buffer is indicated by TxREQ Figure 23-6. Transmit (Tx) Message Registers REGISTERS COMMAND REGISTER (CAN_Txn_CMD) Reserved [31:24] WPN2 Reserved 1 [23] [22] RTR [21] IDE [20] DLC [19:16] Reserved [15:4] ID [31:3] IDENTIFIER (CAN_Txn_ID) WPN1 [3] Tx INT ENBL [2] Tx ABORT [1] Tx REQ [0] Reserved [2:0] DATA REGISTER High (CAN_Txn_DH) D0 [63:56] D1 [55:48] D2 [47:40] D3 [39:32] DATA REGISTER Low (CAN_Txn_DL) D4 [31:24] D5 [23:16] D6 [15:8] D7 [7:0] n = 0,1,…,7 230 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Controller Area Network (CAN) 23.4.3 Message Abort Note 4. Using the WPN flags(wpn1 and wpn2) enables simple retransmission of the same message by only having to set the TxREQ flag without taking care of the special flags (RTR,IDE,DLC and TxINTENBL). A message is aborted by setting the TxABORT flag (CAN_TX[n]_CMD[1]) in the CAN_TX[n]_CMD register. This bit is automatically cleared by the hardware when the message is aborted. 23.4.4 Note 1. The CAN Buffer register (CAN_CSR_BUF_SR) is used to read whether any transmission requests are pending. Transmitting Extended Data Frames For transmitting an extended data frame certain register settings must change compared to that of a standard data frame. These changes are as follows. Note 2. If the write protect bit wpn2 (CAN_TX[n]_CMD[23]) is ‘0’, then the bits [21:16] of the Command register cannot be modified because they are protected and provides an undefined value on read back. ■ For extended date frame, write '1' (recessive) to the IDE bit. ■ The message identifiers are written to the ID[31:3] bit field. Note 3. If the write protect bit wpn1 (CAN_TX[n]_CMD[3]) is ‘0’, then the bit [2] of the Command register cannot be modified. This bit gives a ‘0’ upon read back. 23.5 Receiving Messages in CAN The CAN module has 16 receive message buffers as illustrated in Figure 23-7. Each message buffer has a dedicated acceptance filter. The CAN message is received by the CAN framer and then the received message is simultaneously compared with all the acceptance filters and the accepted message is stored in the respective receive message buffer. The message available (MSG AV) bit in the message buffer is set to indicate the availability of the new message. Message receipt must be acknowledged by clearing the MSG AV flag to allow receipt of another message. The acceptance filter is configured by the Acceptance Mask Register (AMR) and the Acceptance Code Register (ACR). Figure 23-7. Receive (Rx) Block Diagram CAN Module RxMESSAGE0 Acceptance Filter 0 RxMESSAGE1 Acceptance Filter 1 1 RxMESSAGE2 Acceptance Filter 2 2 3 RxMESSAGE Handler CAN Framer CAN Bus 16 RxMESSAGE15 Acceptance Filter 15 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 231 Controller Area Network (CAN) 23.5.1 Message Receive Process Figure 23-8 shows the registers associated with a received message. Figure 23-8. Receive (Rx) Message Registers REGISTERS COMMAND REGISTER (CAN_Rxn_CMD) Reserved WPN2 [31:24] [23] Reserved1 [22] RTR [21] IDE [20] DLC [19:16] Reserved WPNL [7] [15:8] LINK FLAG [6] Rx INT ENBL [5] RTR REPLY [4] BUFF ENBL [3] RTR ABORT [2] Reserved [2:0] ID [31:3] IDENTIFIER (CAN_Rxn_ID) RTR REPLY MSG AV PNDG [0] [1] DATA REGISTER High (CAN_Rxn_DH) D0 [63:56] D1 [55:48] D2 [47:40] D3 [39:32] DATA REGISTER Low (CAN_Rxn_DL) D4 [31:24] D5 [23:16] D6 [15:8] D7 [7:0] n = 0,1,…,15 The main steps in receiving a message are: Following message fields are covered: 1. After receipt of a new message, the RxMessageHandler hardware (as seen in Figure 23-7) searches all receive buffer starting from RxMessage0 until it finds a valid buffer. A valid buffer is indicated by: ■ Identifier ■ IDE ■ RTR a. Receive buffer is enabled indicated by BUFF ENBL = ‘1’ (CAN_RX[n]_CMD[3]). ■ Data byte 1 and data byte 2 For a standard CAN message when IDE=0, the 11 bit identifier are the bits [31:21] of AMR and ACR. b. Acceptance filter of the receive buffer matches incoming message. 2. If the RxMessageHandler finds a valid buffer that is empty, then the message is stored and the MSG AV flag of this buffer is set to ‘1’. 23.5.2.1 Example A message and the acceptance filter settings to accept that message are shown in Figure 23-9 on page 233. 3. If the Rx INT ENBL flag is set, then the RX_MSG flag (CAN_CSR_INT_SR[12]) of the interrupt controller is asserted. 4. If the receive buffer already contains a message indicated by MSG AV = ‘1’ and the Link Flag is not set, then the RX_MSG_LOSS interrupt flag (CAN_CSR_INT_SR[10]) is asserted. The existing message is overwritten with the new received message. Note The CAN Buffer register (CAN_CSR_BUF_SR) determines if any receive message buffer is available. 23.5.2 Acceptance Filter Each receive buffer has its own acceptance filter that is used to filter incoming messages. An acceptance filter is configured by the Acceptance Mask register (AMR) and the Acceptance Code register (ACR). AMR: ‘0’. The incoming bit is checked against the respective ACR bit. The message is not accepted when the incoming bit does not match respective ACR bit. AMR: ‘1’. The incoming bit is Do Not Care. 232 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Controller Area Network (CAN) Figure 23-9. Acceptance Filter M e s s a g e F ra m e Id e n tifie r S ta rt of F ra m e 0 X X 0 0 1 1 0 0 1 0 RTR ID E 0 0 DLC YES ACCEPT MESSAGE = NO REJECT M ESSAG E ACR 0 X 31 30 X 29 0 0 28 1 27 1 26 0 25 0 24 1 23 22 ID E Do Not C a re 0 21 20 RTR 0 3 2 ID E AMR 0 31 1 30 1 29 0 28 0 27 0 26 0 25 0 24 0 23 0 22 0 21 20 3 1 RTR 2 0 0 R S V D 0 0 A ll O n e s R S V D 0 1 0 M asked As seen in the Figure 23-9, the shaded areas are masked bits. When a bit is set to ‘1’ in the AMR register, the corresponding bit in the ACR register is not checked against the received message frame. In the example, bits 30, 29, and bits from 3 to 20 are set to ‘1’ and are masked. Because other bits in the AMR register are written as ‘0’, the respective bits in the ACR register are compared with message bits as shown in Figure 23-9. If the corresponding bits in ACR match with that of the message, the message is then stored in the receive message buffer. If the corresponding bits in ACR do not match with the message, the incoming message is rejected. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F AMR Settings: ID[28:21],ID[31] = 0 ID[30],ID[29]= 1 ID[20:3] = All Ones IDE = 0 RTR = 0 ACR Settings: ID[31:21] = 0x032 ID[20:3] = Do Not Care IDE = 0 RTR = 0 233 Controller Area Network (CAN) 23.5.3 DeviceNet Filtering For some CAN high level protocols such as DeviceNet, additional protocol related information is contained in the first and second data bytes. The acceptance filters provide additional coverage of these two bytes for a more efficient implementation of the protocol. The data bits of the first two bytes of the incoming message are compared with the ACRD register (CAN_RX[n]_ACRD) and the respective bits that are compared are specified using AMRD register (CAN_RX[n]_AMRD). Using the Example on page 232, DeviceNet filtering is illustrated in Figure 23-10. Figure 23-10. DeviceNet Filter Message Frame Start of Frame 0 Identifier X X 0 0 1 Data 1 0 0 1 0 RTR IDE 0 0 DLC 0 0 0 0 0 0 0 1 1 0 Do Not Care Yes Accept Message = No Reject Message ACR 0 X 31 30 AMR 0 1 31 30 X 29 1 29 0 28 0 28 0 27 0 27 1 26 0 26 1 25 0 25 0 24 0 24 0 23 0 23 1 22 0 22 Do Not Care 0 21 20 0 0 3 All Ones 21 20 2 0 1 IDE RTR 0 0 3 R S V D 0 2 0 15 14 13 12 11 10 9 8 7 6 5 AMRD 0 0 0 0 0 1 1 0 0 0 R S V D 1 Do Not ACRD 0 0 0 0 0 X X 1 1 0 Care 15 14 13 12 11 10 9 8 7 0 All Ones 6 5 0 0 Masked In Figure 23-10 the data field of the message frame is compared with those bits of the ACRD register, which are not masked by the AMRD register. To accept this message, the acceptance filter settings are as follows. AMR Settings: ID[28:21],ID[31] = 0 ID[30],ID[29] = 1 ID[20:3] = All Ones IDE = 0 RTR = 0 AMRD[15:11], AMRD[8:6] = 0 AMRD[10:9], AMRD[5:0] = All Ones ACR Settings: ID[31:21] = 0x182 ID[20:3] = Do Not Care IDE = 0 RTR = 0 ACRD[15:6] = 0x006 ACRD[5:0] = Do Not Care The example in Figure 23-10 shows the filtering using 10 data bits. Using AMRD, up to 16 data bits, can be used for filtering. 23.5.4 Filtering the extended data frame is very similar to the standard date frame with the following exception. ■ 234 Filtering of Extended Data Frames IDE bit in AMR and ACR registers must be set to check for extended data frame. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Controller Area Network (CAN) 23.5.5 Receiver Message Buffer Linking Several receive buffers can link together to form a receive buffer array that acts almost like a receive FIFO. To accomplish this, do the following: 23.6 Remote Frames ■ Set the Link flag CAN_RX[n]_CMD[6] for the buffers that need to be linked. ■ Make sure that all buffers of the same array have the same message filter setting (AMR and ACR are identical). Remote frames are used for initiating transmission between two nodes and the node acting as a receiver sends the remote frame. A remote frame can use either standard format or extended format. A remote frame is different from a data frame in that the RTR bit is always equal to ‘1’ and the data field is absent, independent of the value of DLC field. The flow of a remote transmit request is illustrated in Figure 23-11. ■ Do not set the Link flag of the last buffer of an array. As shown in Figure 23-11: When a receive buffer already contains a message (MSG AV=’1’) and a new message arrives for this buffer, then this message is discarded (RX_MSG_LOSS Interrupt). To avoid this situation, several receive buffers are linked together. When the CAN controller receives a new message, the RxMessageHandler searches for a valid receive buffer. If one is found that is already full (MSG AV = ‘1’) and the ‘Link Flag’ is set, the search for a valid receive buffer is continued. If found, the message is transferred to that buffer thereby forming an array. If no other buffer is found, then the RX_MSG_LOSS interrupt is set. ■ The message buffer0 of node1 transmits a remote frame into the CAN bus. ■ The RTR request is received by the RxMessageHandler of node 2 and sends it to the acceptance filters. ■ The acceptance filter settings of the receive message buffer 15 matches with that of the message and then the message is moved to the receive message buffer 15. ■ If the RTR Auto Reply feature is enabled, the receive message buffer 15 will transmit the message with the same identifier as it received (without CPU intervention). It is possible to build several message arrays. Each of these arrays must use the same AMR and ACR. ■ The acceptance filter of the receive message buffer 1 of node1 has the same identifier settings as that of the transmitted message node 1. Hence, the RTR message will be stored in the receive message buffer 1 of node 1. Figure 23-11. Remote Transmit Request Node 1 Node 2 Message Buffers Node 1 TxMESSAGE0 Acceptance Filters RTR REQ FILTER0 Priority Arbiter TxMESSAGE1 CAN Bus Rx Message Handler TxMESSAGE7 RxMESSAGE0 Receive Message Buffers Node 2 RxMESSAGE0 FILTER1 RxMESSAGE1 FILTER15 RxMESSAGE15 Acceptance Filters FILTER0 RxMESSAGE1 FILTER1 RxMESSAGE15 FILTER15 Rx Message Handler CAN Bus PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Priority Arbiter 235 Controller Area Network (CAN) 23.6.1 Transmitting a Remote Frame by the Requesting Node The process to transmit a remote frame by a requesting node (Node 1 as shown in Figure 23-11 on page 235) is as follows. 1. Write a message to an empty transmit buffer. An empty buffer is indicated by Tx_REQ = ‘0’ (CAN_TX[n]_CMD[0]). 2. Set the RTR bit (CAN_TX[n]_CMD[21]) to ‘1’. 3. Choose an appropriate priority arbitration scheme. 4. Set the transmit request flag to initiate transmission. 23.7 Bit Time Configuration The CAN module operates on a single clock input CLK_BUS. This section explains how to configure the programmable bit-rate divider to achieve the desired bit rate and its relationship with CLK_BUS. 23.7.1 Across the industry, most implementations of CAN-Bus use one of 10 bit rates: ■ 1 Mbps ■ 800 Kbps ■ 500 Kbps ■ 250 Kbps The process to receive a remote frame is as follows. ■ 125 Kbps 1. The acceptance filter must be configured to receive the desired message ID. ■ 100 Kbps ■ 50 Kbps ■ 20 Kbps ■ 10 Kbps ■ 5 Kbps 5. The Identifier transmitted in a message must be the same as the identifier of receiving message. 23.6.2 Receiving a Remote Frame 2. Enable the automatic RTR message handling by setting bit ‘RTR REPLY’ to ‘1’. a. If enabled, it will automatically transmit the remote frame with the same identifier. b. Else the remote frame must be transmitted following the standard routine as that of a data frame. 3. Set the requesting node that receives the replied RTR message to receive a normal message. Do not set the RTR Reply bit. 23.6.3 RTR Auto Reply The CAN module supports automatic answering of RTR message requests. All 16 receive buffers support this feature. If an RTR message is accepted in a receive buffer where the RTR REPLY FLAG is set, then this buffer automatically replies to this message with the content of the receive buffer. The ‘RTR REPLY PNDG FLAG’ is set when the RTR message request is received. It is reset when the message is sent or when the message buffer is disabled. To abort a pending RTRreply message, use the RTR ABORTcommand. 23.6.4 Allowable Bit Rates and System Clock (CLK_BUS) These bit rates are configurable if CLK_BUS is 8 MHz or a multiple. All except 800 Kb are configurable if CLK_BUS is 10 MHz or a multiple. With a very few exceptions, all 10 bit rates are not possible if CLK_BUS is not evenly divisible by 1,000,000 Hz. From a bit rate generation point of view, the accuracy for CLK_BUS must be at least 1.58% for 125 Kbps and slower bit rates, and 0.5% or better for bit rates faster than 125 Kbps. Figure 23-12 on page 237 shows a table of the 10 bit rates that are supported for any given fclk frequency from 8 MHz to 100 MHz. Note that maximum possible frequency for PSoC 5LP is 80 MHz. Remote Frames in Extended Format The transmission and reception of remote frames in extended format is similar to standard format except for the following. ■ The IDE bit (CAN_TX[n]_CMD[20]) is set to ‘1’ to make it an extended data frame. ■ The identifier is 29 bits long compared to the 11 bits of a standard data frame. 236 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Controller Area Network (CAN) Figure 23-12. Bit Rate Versus CLK_BUS clk_bus Freq (MHz) 1 Mb 800 Kb 500 Kb 250 Kb 125 Kb 100 Kb 50 Kb 20 Kb 10 Kb clk_bus Freq (MHz) 5 Kb 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 1 Mb 800 Kb 500 Kb 250 Kb 125 Kb 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 100 Kb 50 Kb 20 Kb 10 Kb 5 Kb clk_bus Freq (MHz) 1 Mb 800 Kb 500 Kb 250 Kb 125 Kb 100 Kb 50 Kb 20 Kb 10 Kb 5 Kb 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 Configurable Bit Rates Non Configurable Bit Rates 23.7.2 Setting Bit Rate TSEG1 and TSEG2 The bit rate is defined as the number of bits transmitted on a CAN bus per second. Bit time is the reciprocal of bit rate. Bit time is divided into three segments as shown in Figure 23-13. Each segment is represented in terms of fixed units of time called Time Quanta (TQ) which is derived from the oscillator clock. BRP + 1 TQ = --------------------clk_bus Equation 2 Note Bit rate pre scaler is a register that performs a pre scaling function on CLK_BUS to generate the clock for CAN module. See Figure 23-14. Synchronization Segment. This is the first segment with 1 TQ length and is mainly used for synchronization. An edge is expected to fall within this segment. Figure 23-13. Bit Time Nominal Bit Time = 8...25 TQ (Time Quanta) SJW: 1...4TQ Tseg1 prop_seg + phase_seg1 Synchronization Segment Tseg2 phase_seg2 Sample Point 1 or 3 Sample Mode BitTime = 1 + tseg1 + tseg2 TQ Equation 1 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Tseg1, Tseg2. These segments compensate for the edge phase shift errors. The tseg1 also takes in the propagation time which includes any delays in the network. The length of the segments is increased or decreased to compensate for the error due to phase shift of edges which is known as resynchronization. Sample Point. This is the point at which the state of the bus is read and the bit is interpreted. It is located at the end of tseg1. Synchronization Jump Width. By resynchronization, the tseg1 is lengthened or tseg2 is shortened. Synchronization jump width puts a limit to this resynchronization. The length 237 Controller Area Network (CAN) of tseg2 must be greater than the synchronization jump width. The Configuration register CAN_CSR_CFG is used for setting the bit rate prescaler (BRP), tseg1, tseg2, and the synchronization jump width. CAN peripheral clock (CAN_CLK) is generated by dividing the system clock (CLK_BUS) by (BRP+1). See the Clocking section for detailed information on available options to generate the system clock. For N time quanta in a bit time, the CAN peripheral clock frequency must be configured to N time the CAN bus bit rate. Note 1. Sampling_mode bit in the Configuration register (CAN_CSR_CFG) specifies whether or not one sampling point is used in the receiver path or three sampling points with majority decision are used. Note 2. Edge_mode bit in the Configuration register (CAN_CSR_CFG) specifies whether or not the high to low edge is used for synchronization or both edges are used. 23.8 Figure 23-14. Bit Timing Block Diagram clk_bus Divider can_clk BRP (CAN_CSR_CFG[14:0]) 23.7.2.1 Example An example to achieve 1 Mbps speed with 40 MHz is described as follows. 1. The speed is 1 MHz and the bit time is 1 µs. 2. Choosing a minimum value of 8 TQ in the bit time, 1TQ = 0.125 µs. 3. BRP = ((time quanta * clk_bus) – 1) = 4. 4. Therefore write a value of ‘4’ into the CFG_BITRATE bits in the configuration register. 5. Choose the sampling point to be 60% of the bit time, which is approximately equal to 5TQ. Because the sampling point is at the end of tseg1, this implies that (tseg2+1) = 3TQ or tseg2 = 2TQ. Error Handling and Interrupts in CAN According to the CAN protocol specification, there are five different types of errors. Each CAN node in the bus tries to detect an error, and when it does, it sends out an error frame. The different types of errors and the process of error handling are explained in the following sections. 23.8.1 23.8.1.1 Types of Errors BIT Error A CAN unit sending a bit on the bus also monitors the bus. When the bit value that is monitored is different from the bit value that is sent, a BIT error is detected. An exception is the sending of a ‘recessive’ bit during the stuffed bit stream of the Arbitration Field or during the ACK Slot. A Transmitter sending a Passive Error Flag and detecting a ‘dominant’ bit does not interpret this as a BIT error. 23.8.1.2 FORM Error 6. To fix the sampling point synchronization jump width, use a value ‘1’ by writing to the bits CFG_SJW = ‘1’. A FORM error is detected when there is an error in the CAN message format. The fixed format fields in the message frame such as End of Frame, Interframe Space, etc., contains illegal bits. 7. Write to the bits cfg_tseg2 a value of ‘2’ to set the value of tseg2 to 2TQ. 23.8.1.3 8. Now tseg1 is calculated using the following equation: tseg1 = ((BitTime - (1TQ + tseg2 + 1TQ)) - 1TQ) ...which is tseg1 = 3TQ. 9. Therefore, write a value of ‘3’ into the bits cfg_tseg1 in the configuration register. This procedure above is applied to achieve the standard bit rates using the clock frequencies as specified in the table in Figure 23-12 on page 237. Observe the following conditions for setting tseg1 and tseg2: ■ tseg1 = 0 or tseg1 = 1 are not allowed. ■ tseg2 = 0 is not allowed; tseg2 = 1 is only allowed in direct sampling mode. 238 ACKNOWLEDGE Error A transmitter sending a recessive bit during the ACK slot monitors the ACK slot for a dominant bit. If a receiver receives a message correctly, a dominant bit is written in the ACK slot. Therefore, if the transmitter does not find a dominant bit in the ACK slot after transmission, then an ACKNOWLEDGE error is detected. 23.8.1.4 CRC Error A transmitting node performs certain calculations to generate a CRC code and transmits it in the CRC field. A receiving node also performs the same calculations to generate a CRC code. If the code generated by the receiver does not match the code transmitted then a CRC error is detected. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Controller Area Network (CAN) 23.8.1.5 STUFF Error When there are six consecutive equal bit levels in a message field that is coded by the message of bit stuffing, a STUFF error is detected during the bit time of the sixth consecutive bit level. 23.8.2 Error States in CAN There are three main error states in CAN: Error Active. An error active node can take part in normal bus communication. When it detects an error it sends out an ERROR ACTIVE FLAG. Error Passive. An error passive node takes part in bus communication. When it detects an error it sends out an ERROR PASSIVE FLAG. After sending out the ERROR PASSIVE FLAG, it waits before proceeding with further transmission. An error passive node sends additional 8 recessive bits during the interframe space. This period is also known as suspend transmission because no transmission takes place. that indicate if the Transmit Error Counter and Receive Error Counter, respectively, are greater than or equal to 96 decimal. This is a feature. It serves as an error warning because an error count value greater than and around 96 indicates a heavily disturbed bus. 23.8.3 Interrupt Sources in CAN The interrupt controller governs the various interrupt sources in CAN. The interrupt controller contains an interrupt status and an interrupt enable register. The interrupt status register (CAN_CSR_INT_SR) stores internal interrupt events. When a bit is set, it remains set until it is cleared by writing a '1' to it. The interrupt enable register has no effect on the interrupt status register. The interrupt enable register (CAN_CSR_INT_EN) controls which particular bits from the interrupt status register are used to assert the interrupt output INT_N. INT_N is asserted if a particular interrupt status bit and the respective enable bit are set. The various interrupt sources in CAN are as follows: Bus Off. A node that is in Bus Off does not take part in any bus communication. It has no effect on the bus. The error status in CAN is indicated by the error status register CAN_CSR_ERR_SR. The bits ERR_STATE (CAN_CSR_ERR_SR[17:16]) indicate which error state the CAN node is in. The error states in CAN are determined according to the values of two counters: ■ Transmit Error Counter (CAN_CSR_ERR_SR[7:0]) ■ Receive Error Counter (CAN_CSR_ERR_SR[15:8]) rx_msg. Indicates a message received. tx_msg. Indicates a message sent. rx_msg_loss. Is set when a new message arrives but the RxMessage flag MSG AV is set. bus_off. The CAN has reached the bus off state. crc_err. A CAN CRC error detected. The error counters are modified according to the CAN 2.0B Specification. A node is in ‘error active’ state if the Transmit Error Counter or the Receive Error Counter are less than or equal to 127 decimal. A node is in ‘error passive’ state if the Transmit or Receive Error Counter value exceeds or equals 128 decimal. A node is in ‘Bus Off’ state if the Transmit Error Counter exceeds or equals the value of 256 decimal. An ‘error passive’ node becomes ‘error active’ again when both the Transmit Error Count and the Receive Error Count are less than or equal to 127. form_err. A CAN message format error detected. ack_err. A CAN message acknowledge error detected. stuff_err. A bit stuffing error detected. bit_err. A bit error detected. ovr_load. An overload frame received. arb_loss. The arbitration lost while sending a message. A node which is in ‘Bus Off’ state becomes ‘error active’ with its error counters both set to ‘0’ after 128 occurrences of 11 consecutive ‘recessive’ bits are monitored on the bus. There are two bits in the error status register: ■ ‘txgte96’ (CAN_CSR_ERR_SR[18]) ■ ‘rxgte96’ (CAN_CSR_ERR_SR[19]) PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 239 Controller Area Network (CAN) 23.9 Operating Modes in CAN The CAN module operates mainly in three different modes. The command register CAN_CSR_CMD is used to select the operating modes by setting the corresponding bit for each mode. The three operating modes are as follows: ■ SRAM Test Mode: CAN_CSR_CMD[2] ■ Listen Only Mode: CAN_CSR_CMD[1] ■ Run/Stop Mode: CAN_CSR_CMD[0] 23.9.1 Listen Only Mode In Listen Only mode, the CAN controller only listens to the CAN receive line without acknowledging the received messages on the bus. It does not send any messages in this mode. However, the error flags are updated so that the bit timing is adjusted until no error occurs. The various steps involved in automatic baud rate detection are as follows. 1. The CAN controller is initialized for acceptance of all messages (i.e., the global/local mask is set to ‘0’). 2. The bit timing values of the first possible CANOpen bit rate (10 Kbps) is loaded and the controller is switched into “Listen Only” mode. 3. Assuming that there is traffic on the network and the bit rate is correct, the message is accepted. 4. The error registers will not change and the flag for message reception is set inside the CAN controller. This means the correct bit rate is detected. 5. Assuming the bit rate is not correct, the error flags are updated (stuff-, CRC, or form-error). 6. In this scenario, the CAN controller is switched off and the next possible bit timing values are loaded from the bit rate table. 23.9.2 Run/Stop Mode The CAN controller is in Run mode when it is operating normally. The CAN controller is stopped while it is in the SRAM Test mode. 240 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 24. USB The PSoC® USB block acts as a USB device that communicates with a USB host. The USB block is available as a fixed function digital block in the PSoC device. It supports full speed communication (12 Mbps) and is designed to be compliant with the USB Specification Revision.2.0. USB devices can be designed for plug and play applications with the host and also support hot swapping. This chapter details the PSoC USB block and transfer modes. For details about the USB specification, see the USB Implementers Forum web site. 24.1 Features The PSoC USB has these features: ■ Complies with USB Specification 2.0 ■ Supports full speed peripherals device operation with a signaling bit rate of 12 Mbps ■ Supports 8 data endpoints and 1 control endpoint ■ Supports four types of transfers – bulk, interrupt, isochronous, and control ■ Supports Plug and Play ■ Supports two types of logical transfer modes: ❐ Store and Forward mode ❐ Cut Through mode ■ Differential signal (D+ and D-) output ■ Supports maximum packet size of 64 bytes using the Store and Forward mode and maximum packet size of 1023 for Isochronous transfer using the Cut through mode ■ Capable of supplying PS/2 and CMOS signals ■ Supports two operating voltage ranges, with a nominal voltage of 3.3 V PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 241 USB 24.2 Block Diagram Figure 24-1 illustrates the architecture of the USB block. It consists of the Serial Interface Engine (SIE) and Arbiter. Figure 24-1. USB Block Diagram clk_bus clk_usb Frequency Tuning data USB Block Arbiter 512 Bytes SRAM CPU Interface Memory Interface CPU Arbiter Logic SIE Interface DMA Interface SIE Transceiver D+ 242 D- PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB 24.2.1 Serial Interface Engine (SIE) The Serial Interface Engine (SIE) is responsible for handling the decoding and creating of data and control packets during transmit and receive. It decodes the USB bit streams into USB packets during receive and creates USB bit streams during transmit. The following are the features of the SIE block: ■ Conforms to the USB 2.0 Specification ■ Supports 1 device address ■ Supports 8 data endpoints and 1 control endpoint ■ Supports interrupt for each endpoint ■ Operates at Full Speed with a 48 MHz Clock (maximum permitted tolerance is ±0.25%) ■ Integrates an 8-byte buffer in the Control endpoint interrupt for an endpoint USB_SIE_INT_SR register. is obtained from the The SIE registers CNT0 and CNT1 hold the count value for each endpoint which reports the number of data bytes in a USB transfer. In the case of an OUT endpoint, the firmware programs the maximum number of bytes that can be received for the endpoint. The SIE updates the register with the number of bytes received. In the case of an IN endpoint, it holds the number of bytes that will be transmitted. The SIE Control register for each endpoint, USB_SIE_EPx_CR0, holds the mode value. The mode value determines the response of the USB block to the host. See Table 24-1 for the different mode values. The table describes the mode values corresponding to each type token: the SETUP, IN and OUT tokens. The registers for this block are mainly used to configure the data endpoint operations and the Control Endpoint Data buffers. The register also controls the interrupt available for each endpoint. The SIE generates an interrupt at the end of a transfer. The interrupt enabling and disabling for an endpoint can be done using the USB_SIE_INT_EN register. The status of the Transition error is also reported by the SIE. The bit “err_in_txn” in the USB_SIE_EPx_CR0 register indicates the occurrence of an error. When this bit is set and the USB block is in Store and Forward mode, the hardware automatically retransmits the same data when it receives another IN token from the host. In Cut Through mode, this bit can be read by the firmware to determine if the data should be retransmitted. Table 24-1. Mode Values in the MODE bits of the SIE_EPx_CR0 Register Encoding SETUP IN OUT Disable Mode 0000 Ignore Ignore Ignore Comments NAK IN/OUT 0001 Accept NAK NAK Status OUT Only 0010 Accept STALL Check When this mode is set, it accepts a SETUP token, STALLs in case of IN token and ACKs with a zero length packet in case of OUT token. Used for control endpoint STALL IN/OUT 0011 Accept STALL STALL When this mode is set, it accepts a SETUP token, STALLs in case of IN and OUT token. Used for control endpoint Ignore all USB traffic to this endpoint NAK IN and OUT token Reserved 0100 Ignore Ignore ISO OUT 0101 Ignore Ignore Always Isochronous OUT Status IN only 0110 Accept TX 0 byte STATLL When this mode is set, it accepts a SETUP token, STALLs in case of OUT token and ACKs with a zero length packet in case of IN token. Used for control endpoint ISO IN 0111 Ignore TX Count Ignore NAK OUT 1000 Ignore Ignore NAK Send NAK handshake to OUT token ACK OUT (STALL = 0) 1001 Ignore Ignore ACK This mode is changed by the SIE to mode 1000 on issuance of ACK handshake to an OUT STALL ACK OUT (STALL = 1) 1001 Ignore Ignore Reserved 1010 Ignore Ignore Ignore STALL the OUT transfer Ignore ACK OUT – STATUS IN 1011 Accept TX0 byte ACK NAK IN 1100 Ignore NAK Ignore PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Isochronous IN ACK the OUT token or send zero length data packet for IN token. Send NAK handshake for IN token 243 USB Table 24-1. Mode Values in the MODE bits of the SIE_EPx_CR0 Register (continued) Mode Encoding IN TX Count Ignore 1101 Ignore STALL Ignore STALL the IN transfer 1110 Ignore Ignore 1111 Accept TX Count Check Respond to IN data or Status OUT ACK IN (STALL = 1) Reserved ACK IN – Status OUT Arbiter The Arbiter is the block which handles access of the SRAM memory by the endpoints. The SRAM memory can be accessed by the CPU or the SIE. The Arbiter handles the arbitration between the CPU and the SIE. The Arbiter consists of the following blocks: SIE Interface Module ■ CPU Interface Module ■ Memory Interface ■ DMA Engine ■ Arbiter Logic ■ Synchronization Module The Arbiter registers are used to handle the endpoint configurations, the Read address, and the Write address for the endpoints. It also configures the logical transfer type required for each endpoint. The types of logical transfers are discussed below. Also, each endpoint supports interrupt. The Arbiter has only one interrupt line for the Interrupt Controller. The Arbiter registers handle the enabling/disabling of the interrupts for the endpoints and hold the status of the interrupts. The Arbiter is also responsible for the memory management (i.e., sharing the available 512 bytes of SRAM among the data endpoints). 24.2.2.1 SIE Interface Module This module handles all the transactions with the SIE block. The SIE reads data from the SRAM memory and transmits to the host. Similarly, it writes the data received from the host to the SRAM memory. These requests are registered in the SIE Interface module and are handled by this block. 24.2.2.2 CPU Interface Block This module handles all the transactions with the CPU. The CPU makes requests for the reads and writes to the SRAM memory for each endpoint. These requests are registered in the CPU Interface block and are handled by the block. 24.2.2.3 Memory Interface The memory interface is used to control the interface between the USB block and the SRAM memory unit. The 244 Comments Ignore 1101 ■ OUT This mode is changed by the SIE to mode 1100 after receiving ACK handshake to an IN data ACK IN (STALL = 0) 24.2.2 SETUP Ignore maximum memory size supported is 512 bytes organized as 256 x a 16-bit memory unit. This is a dedicated memory for the USB. All the control and data lines, including the Data In lines, Data Out lines, Enable line, Address lines, and Direction Control line between the USB and the memory unit, are handled by the memory interface. The memory access can be requested by the SIE or by the CPU. The SIE Interface block and the CPU Interface block handle these requests. 24.2.2.4 DMA Interface When Direct Memory Access (DMA) is configured, the DMA interface is responsible for all transactions back and forth between the DMA and USB. The block supports the DMA request line for each data endpoint. The behavior of the DMA depends on the type of logical transfer mode configured in the Configuration register. Note that DMA transfers from UDBs to the USB block must first go through SRAM to ensure that proper timing is kept. An additional transaction descriptor should be used to transfer from UDBs to SRAM, and then from SRAM to USB. Other applicable DMA transfers from sources besides UDBs are not constrained to this path. 24.2.2.5 Arbiter Logic This is the main block of the Arbiter. It is responsible for arbitrations for all the transactions that happen in the Arbiter. It arbitrates the CPU, DMA, and SIE access to the memory unit and the registers. This block also handles the memory management. The memory management is either “Manual” or “Automatic.” In the case of Manual Memory Management, the read and write address manipulations are done by the firmware. In the case of Automatic management, all the memory handling is done by this block itself. This block takes care of the buffer size allocation, depending on the programmed buffer size (using the USB_BUF_SIZE). It also does the handling of common memory area. This block also handles the interrupt requests for each endpoint. Each endpoint can have interrupts due to: ■ DMA Grants ■ IN Buffer Full ■ Buffer Overflow ■ Buffer Underflow PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB These arbiter interrupt requests are routed to only one interrupt line which acts as a signal to the interrupt controller. 24.2.2.6 Synchronization Block The USB block uses 2 clocks: the System Clock and the USB Clock. The System Clock is used by the Arbiter. The USB Clock is used by the SIE and the OsClock module. Because these two are different clocks, synchronization is required between the blocks. The handling of the synchronization is done by this block. 24.3 How it Works The USB Block operates at a certain frequency and voltage range. For proper operation of the USB block, the user must ensure that the operating ranges are within tolerances. The following sections discuss the operating ranges required for the PSoC USB. 24.3.1 Operating Frequency The USB block needs two different clocks to work: the System Clock which controls the Arbiter, memory and the register block, and the USB clock which controls the SIE and the OsClock. ■ Minimum system clock – 24 MHz ■ USB Clock for Full Speed operation – 48 MHz (+0.25% tolerance) The USB needs a 48 MHz clock to function. The clock to the USB is called the clk_usb. The clk_usb can be derived from either IMOCLK, doubler clock (IMOCLK * 2), PLL, or the DSI clock. For further details on the clock for this block, see the Clocking System chapter on page 109. The OsClock block of the USB trims the USB clock to lock to the frequency of the USB packets. The USB clock is clocked to the USB token as per the USB 2.0 Specification. When the frequency is locked with other USB bit streams, the block will locate a particular edge in the USB packet. The number of clock periods between these edges is measured to lock the internal oscillator frequency with the frequency of the USB packet. The frequency tuning value is sent to the Clocking system by the USB Block to lock the frequency. The locking of the frequency is done by the hardware and needs no user intervention. The Synchronization Block of the Arbiter handles the synchronization of the USB Clock and System Clock. 24.3.2 Operating Voltage The USB block can operate in two voltage ranges: ■ Standard voltage range – 4.35 V to 5.25 V ■ Low voltage range – 3.15 V to 3.45 V PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB needs a nominal voltage of 3.3 V for its operation. It supports an internal regulator which is used for voltage regulation. While in the Standard Voltage Range, the voltage is regulated to 3.3 V by the internal regulator. While in the Low Voltage Range, the internal regulator should be bypassed. The “reg_enable” bit in the USB_USB_CR1 register is used to control the regulator. 24.3.3 Transceiver The USB block includes the transmitter and the receiver. The signal between the USB device and the host is a differential signal. The receiver receives the differential signal and converts it to a single ended signal. The single ended input is given to the USB block at a nominal voltage range of 1.55 V to 1.95 V. The transmitter converts the single ended signal to the differential signal and transmits it to the host. The differential signal is given to the upstream devices at a nominal voltage range of 0 V to 3.3 V. The transceiver also supports the PS/2 signals. It can receive and transmit PS/2 signals at a nominal voltage of 0 V to 5 V. The transceiver has pull-up resistors to support the PS/2 signals. In addition to the PS/2 signals, the transceiver also supports CMOS signal levels. The PS/2 and the CMOS modes can be selected using the registers USB_USBIO_CR1 and USB_USBIO_CR2. The Transmitter can be manually forced to transmit signals. The register USB_USBIO_CR0 is used to manually transmit the signals. Examples are as follows: ■ When the manual transmission is enabled, the register can be configured to transmit Single Ended Zero signal (that is, D+ and D- are low). ■ Configurable to transmit the USB signals. The USB signals can be two types: ■ ❐ D+ low and D- high = J ❐ D+ high and D- low = K The register also has a bit which is used to read the received signal levels. The bit can show if D+ < D– or D+ > D–. 24.3.4 Endpoints The SIE and Arbiter support 8 data endpoints (EP1 to EP8) and one control endpoint (EP0). The data endpoints share the SRAM memory area of 512 bytes. The endpoint memory management can be either “Manual” or “Automatic.” The endpoints are configured for direction and other configuration using the SIE and arbiter registers. The endpoint “read address” and “write address” registers are accessed through the Arbiter. Each endpoint supports a set of inter- 245 USB rupts. The interrupts can be enabled or disabled for an individual endpoint. The interrupts for each endpoint can also be collectively enabled or disabled. The endpoints can be individually made active. In the Auto Management mode, the register USB_EP_ACTIVE is written to control the active state of the endpoint. The endpoint activation cannot be dynamically changed during runtime. In Manual Memory Management mode the firmware decides the memory allocation, so it is not required to specify the active endpoints. The EP_ACTIVE register is ignored during the manual memory management mode. The USB_EP_TYPE register is used to control the transfer direction (IN, OUT) for the endpoints. The control endpoint has a separate 8 bytes for its data. 24.3.5 Transfer Types ■ Separate interrupt line for each data endpoint and control endpoint. ■ The register SIE_EP_INT_EN and SIE_EP_INT_SR control/ show the status of both the SIE and the Data Endpoint interrupts. Arbiter Interrupt Line The arbiter generates interrupts for the endpoints during these events: ■ Buffer overflow ■ Buffer underflow ■ DMA grant ■ IN endpoint local buffer full This information applies to the arbiter interrupts. ■ These interrupts can be generated by every endpoint. The register USB_ARB_EPx_INT_EN (where x = 1 to 8 for each endpoint) is used to enable or disable each interrupt for the endpoint. ■ The Status of each interrupt for every endpoint can be read using the USB_ARB_EPx_INT_SR (where x = 1 to 8 for each endpoint) register. ■ The interrupt for an endpoint can be collectively enabled or disabled using the USB_ARB_INT_EN register_ Each bit in this register corresponds to each endpoint. ■ The status of the Arbiter interrupt for an endpoint can be read using the USB_ARB_INT_SR register. The PSoC USB supports Full Speed transfers and is compliant with the USB 2.0 Specification. It supports four types of transfers: ■ Interrupt Transfer ■ Bulk Transfer ■ Isonchronous Transfer ■ Control Transfer For further details about these transfers, refer to the USB Specification 2.0. 24.3.6 Interrupts There is only one arbiter line common for all the endpoints. The interrupts are generated by the SIE and the Arbiter. The following interrupt lines are available for the interrupt controller: ■ Nine SIE interrupt lines (one for each endpoint and control endpoint) ■ Arbiter interrupt line ■ SIE interrupt line for SOF ■ ■ ■ Generated whenever the SOF is received. SIE Data Interrupt ■ Interrupt generated for the data valid or error in transaction. SIE data endpoints interrupt line ■ One interrupt line common for all endpoints. Reset interrupt line ■ The sticky bit “data_valid,” in the USB_SIE_EPx_CNT0 register, indicates the data valid state. ■ The sticky bit “err_in_txn” in the USB_SIE_EPx_CR0 register indicates the error in transaction state. Nine SIE Interrupts ■ SIE Interrupt for SOF Generated after the completion of packet transmission. ❐ Automatic for acknowledged transfer ❐ Can be enabled for non-acknowledged transfer ■ The register USB_SIE_EP_INT_EN is used to enable the SIE interrupt for each endpoint. Each bit in the register corresponds to each endpoint. ■ The status of the SIE interrupt can be read using the USB_SIE_EP_INT_SR register. These bits are sticky bits and need firmware to clear the status. 246 24.4 Logical Transfer Modes The USB block in PSoC devices supports two types of logical transfers. The logical transfers can be configured using the register setting for each endpoint. Any of the logical transfer methods can be adapted to support the three types of data transfers (Interrupt, Bulk, and Isochronous) men- PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB tioned in the USB 2.0 Specification. The Control transfer is mandatory in any USB device. The logical transfer mode is a combination of memory management and DMA configurations. The Logical Transfer modes are related to the data transfer within the USB block (i.e., to/ from the SRAM memory unit for each endpoint). It does not represent the transfer methods between the device and the host (i.e., the transfer types specified in the USB 2.0 Specification). The USB block supports two basic types of transfer modes and are detailed in Table 24-2 on page 247. ■ Store and Forward mode ■ Cut Through mode Table 24-2. USB Transfer Modes Feature Store and Forward Mode Cut Through Mode SRAM Memory Usage Requires more memory SRAM Memory Management Manual Auto SRAM Memory Sharing 512 bytes of SRAM shared between endpoints. Sharing is done by firmware. Each endpoint is allocated less share of memory automatically by the block. Rest of memory is available as “Common Area.” This Common Area is used during the transfer. IN Command Entire packet present in SRAM memory before the IN command is received. Memory filled with data only when SRAM IN command is received. Data is given to host when enough data is available (based on DMA configuration). Does not wait for the entire data to be filled. OUT Command Entire packet is written to SRAM memory on OUT command. After entire data is available, it is copied from SRAM memory to the USB device. Waits only for enough bytes (depends on DMA configuration) to be written in SRAM memory. When enough bytes are present, it is immediately copied from SRAM memory to the USB device. Transfer of Data Data is transferred when all bytes are written to the memory. Data is transferred when enough bytes are available. It does not wait for the entire data to be filled. Types Based on DMA Supported Transfer Types Requires less memory No DMA mode Only Auto DMA mode Manual DMA mode Best suited for Interrupt and Bulk transfers Best suited for Isochronous transfer Every endpoint has a set of registers that need to be handled during the modes of operation, as detailed in Table 24-3. Table 24-3. Endpoint Registers Register Comment Content Usage ARB_RWx_WA Endpoint Write Address register Address of the SRAM This register indicates the SRAM location to which the data in the Data register is to be written. ARB_RWx_RA Endpoint Read Address register Address of the SRAM This register indicates the SRAM location from which the data must be read and stored to the Data register. Data register is read/ written to perform any transaction. ARB_RWx_DR Endpoint Data Register IN command: Data written to the Data register is copied to the SRAM location specified by the WA register. After write, the WA value is automatically incremented to point to the next memory location. 8-Bit Data OUT command: Data available in the SRAM location pointed by the RA register is read and stored to the DR. When the DR is read, the value of RA is automatically incremented to point to the next SRAM memory location that must be read. Holds the number of bytes that can be transferred. IN command: Holds the number of bytes to be transferred to host. SIE_EPx_CNT0 and SIE_EPx_CNT1 Endpoint Byte Count Register Number of Bytes “Mode” bits in SIE_EPx_CR0 Mode Values Response to the Host In the Manual Memory Management case, the endpoint read and endpoint write address registers are updated by the firmware. So the memory allocation can be done as PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F OUT command: Holds the maximum number of bytes that can be received. The firmware programs the maximum number of bytes that can be received for that endpoint. The SIE updates the register with the number of bytes received for the endpoint. Controls how the USB device responds to the USB traffic and the USB host. Some examples of mode include ACK, NAK, STALL, etc. See Table 24-1 on page 243 for additional details. required by the user and the memory allocation decides which endpoints are active. (i.e., the user can decide to 247 USB share the 512 bytes for all the 8 endpoints or a lesser number of endpoints). Figure 24-2. No DMA Access IN Transaction Write WA register (based on required memory allocation) In the Automatic memory management case, the endpoint read and endpoint write address registers are updated by the USB block. The block assigns memory to the endpoints that are activated using the EP_ACTIVE register. The size of memory allocated depends on the value in the BUF_SIZE register. The rest of the memory, after allocation, is called the “Common Area” memory and used for the transfer of data. Write packet size to Byte Count register Write data to Data register No In the following text, the algorithm for the IN and OUT transaction for each mode is discussed. An IN transaction is when the data is read by the USB host (for example, PC). An OUT transaction is when the data is written by the USB host to the USB device (in this case, PSoC 5LP). The choice of using the DMA and memory management can be configured using the USB_ARB_CFG register and the mode is common to all endpoints. 24.4.1 24.4.1.1 Value automatically written to the SRAM specified by WA location. WA++ Is all data written to SRAM? Yes Write the RA register (same as initial WA register) Store and Forward Mode Set mode value in CR0 register No DMA Access This is the Manual Memory Management mode with no DMA access. IN Transaction (CPU Write, SIE Read). The steps for an IN transaction on an IN endpoint are shown in Figure 24-2. Wait Is IN command received? No Yes Responds automatically with ACK (configured as Mode value) USB Block reads value from RA and transmits to host. RA++ No Is all data transmitted? Yes Interrupt Generated Set the mode as NAK for the last byte in transfer. Status bits set by the block OUT Transaction (CPU Read, SIE Write). The steps for an OUT transaction on an OUT endpoint are shown in Figure 24-3. 248 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB 24.4.1.2 Figure 24-3. No DMA Access OUT Transaction This is the Manual Memory Management mode with Manual DMA Access. This mode requires the configuration of the DMA controller. See DMA Interface on page 244 for details and constraints regarding DMA transfers to the USB block. Write WA register (based on required memory allocation) To inhibit CRC set the “crc_bypass” bit in the ARB_EPx_CFG register Write maximum packet size to Byte Count register Set mode value in CR0 register Wait Is OUT command received? No Yes Manual DMA Access Responds automatically with ACK (configured as mode value) Data received from host written to SRAM location WA This mode is similar to the No DMA Access except that the write/read of packets is performed by DMA. A DMA request for an endpoint is generated by setting the DMA_CFG bit in the ARB_EPx_CFG register. When the DMA service is granted and is done (DMA_GNT), an arbiter interrupt can be programmed to occur. The transfer is done using a single DMA cycle or multiple DMA cycles. After completion of every DMA cycle the arbiter interrupt (DMA_GNT) is generated. Similarly, when all the bytes of data (programmed in the byte count) are written to the memory, the arbiter interrupt occurs and the IN_BUF_FULL bit is set. IN Transaction (CPU Write, SIE Read). The steps for an IN transaction on an IN endpoint are shown in Figure 24-4. WA++ No Is all data written to SRAM? Yes SIE sets mode to NAK. Updates Byte Count with actual number of data received and sets the data valid bit Write the RA value (same as initial WA) Data in Data register is read by CPU and given to device. RA++ is done automatically. SIE Data Interrupt Generated USB Block reads the data at location RA and writes to Data register No Is all data read from SRAM? Yes End PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 249 USB Figure 24-5. Manual DMA OUT Transaction Figure 24-4. Manual DMA IN Transaction Write WA register (based on required memory allocation) Write WA register (based on required memory allocation) Write Packet size to Byte Count register Set the DMA request Value automatically written to the SRAM specified by WA location. WA++ DMA writes data to Data register No DMA_GNT interrupt generated for every DMA cycle. IN_BUF_FULL interrupt generated after full data write. Is all data written to SRAM? To inhibit CRC set the “crc_bypass” bit in the ARB_EPx_CFG register. Write maximum packet size to Byte Count register Set mode value in CR0 register Yes Wait Write the RA register (same as initial WA register) Is OUT command received? No Set Mode value in CR0 register Wait No Is IN command received? Yes Yes Responds automatically with ACK (configured as Mode value) Data received from host written to SRAM location WA WA++ USB Block reads value from RA and transmits to host. RA++ No Responds automatically with ACK (configured as Mode value) No Is all data written to SRAM? Is all data transmitted? Interrupt Generated Yes Yes Set the mode as NACK for the last byte in transfer. Status bits set by the block. SIE sets mode to NACK. Updates Byte Count with actual number of data received and sets the data valid bit Write the RA value (same as initial WA) OUT Transaction (CPU Read, SIE Write). The steps for an OUT transaction on an OUT endpoint are shown in Figure 24-5. Data in Data register is read by DMA and given to device. RA++ SIE Data Interrupt Generated Configure the DMA request USB block reads the data at location RA and writes to Data register No Is all data read from SRAM? Yes End 250 At the end of every DMA cycle, DMA_GNT interrupt is generated. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB 24.4.2 Cut Through Mode This is the Auto Memory Management mode with Auto DMA Access. The CPU programs the initial buffer size requirement for IN/OUT packets and informs the Arbiter block of the endpoint configuration details for the particular application being considered. The block then controls memory partitioning and handling of all memory pointers. During the memory allocation, each active IN endpoint (set by the EP_ACTIVE and EP_TYPE registers) is allocated a small amount of memory configured using the BUF_SIZE register. The remaining memory is left as “Common Area” and is common for all endpoints. In this mode, the memory requirement is less and it is suitable for the Full Speed “Isochronous Transfer” up to 1023 bytes. When an IN command is sent by the host, the device responds with the data present in the dedicated memory area for that endpoint. It simultaneously issues a DMA request for more data for that EP. This data fills up in the Common Area. The device does not wait for the entire packets of data to be available. It only waits for the (USB_DMA_THRES_MSB, USB_DMA_THRES) number of data available in the SRAM memory and begins the transfer from the common area. See DMA Interface on page 244 for details and constraints regarding DMA transfers to the USB block. Similarly, when an OUT command is received, the data for the OUT endpoint is written to the common area. When some data (greater than (USB_DMA_THRES_MSB, USB_DMA_THRES)) is available in the common area, the Arbiter block initiates a DMA request to the PHUB and the data is immediately written to the device. The device does not wait for the common area to be filled. This mode requires the configuration of the DMA_THRES and DMA_THRES_MSB registers to hold the number of bytes that can be transferred in one DMA transfer. Similarly, the PHUB register must be configured for the BURSTCNT values. The BURSTCNT value must always be equal to the value set in the DMA_THRES registers. The block sends the Termin signal to the PHUB along with the last data byte of the packet. Apart from the DMA registers, this mode also needs the configuration of the BUF_SIZE for the IN and the OUT buffers and the EP_ACTIVE and the EP_TYPE registers. IN Transaction (CPU Write, SIE Read). The steps for an IN transaction on an IN endpoint are shown in Figure 24-6 on page 252. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 251 USB Figure 24-6. Cut Through Mode IN Transaction Write packet size of the endpoint to Byte Count register Set IN_DATA_RDY for the endpoint in ARB_EP1_CFG register This memory location is very limited. The memory location is filled initially to make sure the host does not stall when an IN command is sent. When an IN command is received the PHUB initiates the copy of data from device to common area. This initialization would take some time. The data in the end point buffer is transmitted until the data is copied to the common area. Block automatically raises interrupt for DMA Data automatically read and written to SRAM pointed by WA. WA++ DMA writes to Data register. No Is the endpoint buffer filled? IN_BUF_FULL Interrupt generated Yes Update Mode value in the Mode register Wait Is IN command received No Block automatically sends the ACK. (Configured as Mode value) Yes Is the complete data available in the memory Yes No SIE reads data from SRAM (specified by location RA) and transmits to host RA++ Raise a DMA request No SIE reads data from SRAM (specified by location RA) and transmits to host Is all data in buffer transmitted? RA++ No Set the data valid bits Is all data in buffer transmitted? End Yes In the mean time, the PHUB initiates the transaction. The data from the device is copied to the common area. The data from the USB is written to the SRAM by the DMA. Wait Is data in Common Area > {DMA_THRES, DMA_THRES_MSB) No Yes The process is continued till all the data is transferred Initiate PHUB transfer. Block transfer data available in Common Area Set the data valid bits End 252 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB OUT Transaction (CPU Read, SIE Write). The steps for an OUT transaction on an OUT endpoint are shown in Figure 24-7. Figure 24-8. IN Transaction Set the mode bits to ACK the IN token Figure 24-7. Cut Through Mode OUT Transaction No Is SETUP token received Write maximum bytes to Byte Count register Yes Program the Mode register for the endpoint The block ACKs it Wait Is the OUT command received from host? No Generates Interrupt and sets the bit to indicate that IN token is received. Yes Read the status bit and the Data valid bit The DMA writes the received data to the SRAM in location specified by WA WA++ Is Data Valid? No Yes Is data in SRAM > (DMA_THRES, DMA_THRES_MSB)? Read the EP0_DRx register to find the type of request Yes Copy the required data to the EP0_DRx registers DMA request is raised The process is continued till all the data is transferred No USB Block writes the data from SRAM to the Data register Data in the Data register is read and given to the USB device by the DMA. RA is incremented automatically. Set the data valid bit and the mode bits. Also update the byte count value No Is all the data from SRAM copied to device? Is IN token received? No Yes Set the data valid bits End 24.4.3 Control Endpoint Logical Transfer The control endpoint has a special logical transfer mode. It does not share the 512 bytes of memory. Instead it has dedicated 8 byte register buffer. The IN and OUT transaction for the control endpoint is detailed below: No Yes The block transmits the data from the EP0_DRx registers The block sets the mode value to NAK all further IN tokens. Blocks generates interrupt on receiving ACK from host and sets the IN byte received bit. Are all bytes transferred? Yes End PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 253 USB 24.5 Figure 24-9. OUT Transaction Program the mode bits for ACK_OUT PS/2 and CMOS I/O Modes The USB transceiver is designed in such a way that, apart from the USB signals, it can also transmit other signal levels. The pull up resistors are available at the transmitter end, which enables additional signal levels. The registers USB_USBIO_CR1 and USB_USBIO_CR2 must be configured to get different signal levels. No Is the SETUP token received? Yes The “test_res” bit in the USBIO_CR2 register puts the transmitter in pull up mode where the pull up resistors are connected. The block ACKs the SETUP token Generates Interrupt and sets a bit in EP0_CR register to indicate that SETUP token was received. The I/O mode bit in the USBIO_CR1 register puts the USB in either USB mode or Drive mode. When put in Drive mode, the USB signals are disabled and the bits DMI and DPI are used to drive D- and D+, respectively. There are two different drive modes. In CMOS Drive mode, D+ follows the DPI and D- follows the DMI. In the case of Open Drain mode, the pull up resistors play a role. In this state, when the DPI and DMI bits are set to high, D+ and D- are high impedance. Read the Data valid bit in EP0_CNT Is data valid? Yes In all voltage ranges from 1.7 V to 5.5 V, the "suspend," "pull up," and "high impedance” drive modes are available. The drive modes can be selected using the USB_USBIO_CR1 and USB_USBIO_CR2 registers. Read EP0_DRx to read the type of request Update the mode bits to ACK an OUT token The pull up resistors can be connected between Vdd and D+ and D-, independent of the Drive modes. The bit “p2puen” is used for this. No Is OUT token received? An internal pull up of 1.5 k is also supported and can be enabled using the register USBIO_CR1. The USBIO_CR1 register is also used to poll the state of the D+ and D- pins. Yes The block stores the received byte to the EP0_DRx register and ACK the received byte. Interrupt generated. No No Read the status and data valid bits Is data valid? No Yes Set the mode bit to NAK all OUT tokens till all bytes have been received. Are all bytes received? Yes End 254 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F USB 24.6 Register List Table 24-4. USB Register List Register Name Comments Features General Registers USB_CR0 USB Control register 0 To enable the USB and store the USB Device address USB_CR1 USB Control register 1 To monitor the bus activity and control the regulator operation USBIO_CR0 USB I/O Control register 0 To control the operation on D+ and D- signals USBIO_CR1 USB I/O Control register 1 To configure the pull up registers USBIO_CR2 USB I/O Control register 2 To control in test modes USB_BUF_SIZE Dedicated endpoint buffer size register Stores the dedicated buffer size for each endpoint USB_EP_ACTIVE Endpoint active register Stores the status of active endpoints USB_EP_TYPE Endpoint Type register Stores the type of endpoint either IN/OUT USB_EP0_DRx x= 0 -7 Control endpoint Data register The endpoint 0 is the control endpoint USB_EP0_CR Endpoint 0 Control register USB_EP0_CNT Endpoint 0 Count register SIE Registers USB_SIE_EP_INT_EN Interrupt enable register To enable the interrupts for each endpoint USB_SIE_EP_INT_SR Interrupt status register To find the status of interrupt for each endpoint USB_SIE_EPx_CNT0 x= 1- 8 Non control endpoint Count register Handles the Data toggle state and MSB of the 11 bit counter USB_SIE_EPx_CNT1 x= 1- 8 Non control endpoint Count register LSB of the 11 bit counter USB_SIE_EPx_CR0 x=1-8 Non control endpoint Control register Controls the mode for the endpoint and stores the state of error, ACK and NACK for the endpoint. OSCLK_DR0 OsClock Lock register 0 The LSB of the Oscillator locking circuit output OSCLK_DR1 OsClock Lock register 1 The MSB of the Oscillator locking circuit output USB_ARB_EPx_CFG x=1–8 Endpoint configuration register Stores the configuration for the transfer modes, reset of pointers and CRC USB_ARB_Epx_INT_EN x=1–8 Endpoint Interrupt enable register To enable the required interrupts USB_ARB_Epx_SR x = 1- 8 Endpoint status register To indicate status like overflow, underflow, DMA grant and Local buffer full USB_ARB_RWx_WA x=1–8 Endpoint Write address register Stores the LSB 8 bits of the Write address pointer USB_ARB_RWx_WA_MSB x=1–8 Endpoint Write address register Stores the MSB 1 bit of the Write address pointer USB_ARB_RWx_RA x=1–8 Endpoint Read address register Stores the LSB 8 bits of the Read address pointer USB_ARB_RWx_RA_MSB x=1–8 Endpoint Read address register Stores the MSB 1 bit of the Read address pointer USB_ARB_CFG Arbiter Configuration register USB_ARB_INT_EN Arbiter Interrupt Enable register To enable the interrupt for each endpoint USB_ARB_INT_SR Arbiter Interrupt Status register To store the interrupt status for each endpoint OsClock Registers Arbiter Registers USB_CWA Common Area Write Address register The LSB 8 bits of the Write address pointer USB_CWA_MSB Common Area Write Address register The MSB 1 bit of the Write address pointer USB_DMA_THRES DMA Threshold Count register The LSB 8 bits of the DMA threshold count register PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 255 USB Table 24-4. USB Register List (continued) Register Name Comments Features USB_DMA_THRES_MSB DMA Threshold Count register The MSB 1 bit of the DMA threshold count register USB_SOF0 Start of Frame register 0 LSB 8 bits of the Start of Frame counter USB_SOF1 Start of Frame register 1 MSB 3 bits of the Start of Frame counter USB_BUS_RST_CNT Bus reset count register The reset counter for the USB 256 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 25. Timer, Counter, and PWM Timer blocks in PSoC® devices are 8/16 bits and configurable to act as Timer, Counter, or Pulse Width Modulator (PWM) blocks that play important roles in embedded systems. PSoC devices give a maximum of four instances of the block. If additional blocks are required, they can be configured in the UDBs using PSoC Creator™. Timer blocks have various clock sources and are connected to the General Purpose Input/Output (GPIO) though the Digital System Interconnect (DSI). 25.1 Features ■ 8/16-bit timer/counter/PWM that acts as a down counter ■ Supports the following modes: ■ ❐ Timer ❐ Gated Timer ❐ Pulse-width Modulator (PWM) ❐ One Shot Supports interrupts upon: ❐ Terminal count – the final value in the Count register is reached ❐ Compare true – the timer value matches with the Compare register ❐ Capture – capture of timer value on edge detection in the Capture signal ■ Counts when Enable signal is asserted ■ Supports the free running timer ■ Period reload on start, reset, and terminal count ■ Selectable clock source ■ Supports kill and dead band features 25.2 Block Diagram Figure 25-1 on page 258 shows one timer block. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 257 Timer, Counter, and PWM Figure 25-1. Timer Block Diagram Clock Timer Block DSI Configuration registers CFG0 and CFG1 DSI Capture signal Timer Reset pin Enable pin Timer Enable signal Kill pin Period registers PER1 (MSB) PER0 (LSB) Capture registers CAP1 (MSB) CAP0 (LSB) Count registers CNT_CMP1 CNT_CMP0 (MSB) (LSB) Compare registers Capture pin Kill signal 25.3 Compare Output signal Compare Output pin Terminal Count Interrupt signal Capture/Compare Interrupt signal Interrupt signal to Interrupt Controller Timer Enable Interrupt signal Status Register SR0 How It Works The block receives a clock signal that is selectable from different sources. The block in PSoC devices is a down counter and counts for every rising edge of the input clock. It counts down from the period value to zero. When it reaches zero (terminal count) the period value is reloaded into the count register, and the timer continues to count. If the timer is configured for One Shot mode, the timer stops when it reaches the terminal count. The timer block can act in various modes, depending on appropriate configuration of the registers: ■ Terminal Count Output pin Timer Stop Interrupt signal External Route Registers RT0 and RT1 Timer Reset signal Terminal Count Output signal Timer ❐ Free Run ❐ Gated Timer – Pulse Width – Period 25.3.1 Clock Selection The block supports the flexibility to select the required clock source. As shown in Figure 25-2 on page 259, the block uses the CLK_BUS frequency, or it is routed through one of the eight selectable clock lines CLK_BUS_EN 0…7, which are synchronous to the clock bus. Clock selection is done through the Configuration register CFG1. If the BUS_CLK_SEL bit in register CFG1 is set, the block uses the CLK_BUS frequency, instead of the eight selectable digital clock lines. If the BUS_CLK_SEL bit is set to 0, one of the eight selectable lines is used for the clock. The bits CLK_BUS_EN_SEL in Configuration register CFG1 are set to choose one of eight selectable digital clock lines. The clock for the digital clock lines can be derived from the CLK_BUS or it can be another UDB signal or external clock signal. – Stop on Interrupt ■ PWM ■ One Shot The block can be used as a timer to capture time of external event, to measure period and pulse width of the input signal, and to find the time of occurrence of interrupt and as a PWM generation unit. 258 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Timer, Counter, and PWM Figure 25-2. Clock Selection CLK_BUS_SEL CFG1 Latch CLK_BUS_EN[7:0] CLK FF 3 Rising Edge Detect CLK_BUS_EN_SEL[2:0] CFG1 CLK_BUS 25.3.2 Enabling and Disabling Block The block is enabled or disabled by setting the Enable bit EN in Configuration register TMRx_CFG0. All the required configurations for the block must be done before it is enabled. When the block is enabled, it functions in the configured mode (Timer or PWM). Enabling a block updates the registers with the new configured value. Disabling a block retains the values in the registers until it is enabled again. ■ When the EN bit is set, the previous state is cleared and the count register is loaded with the reload value from the period register. The block starts to count. ■ When Configuration and Period registers are modified with the EN bit set to ‘1’, the changes go into effect only after the completion of the current running period (at the terminal count). ■ When Configuration and Period registers are modified with the EN bit set to ‘0’, the changes go into effect immediately after the EN bit is set to ‘1’. ■ When the block is enabled, the count value is loaded with the new reload value, regardless of the state of the register before setting EN = ‘0’. Input signals are connected to the GPIO through the Digital System Interconnects (DSI). The user maps the input pins to the DSI routing through External Routing register RT0. DSI 1 through DSI 4 within any block can be routed to as any of the above input signals, depending on user mapping. Mapping between DSI routing and the input pins is not fixed. See Figure 25-1 on page 258. The block has two outputs, terminal count and compare output. They are synchronized to the clock signal. This is done by setting the bits in the external routing register RT1. When the pins are set as asynchronous, the changes go into effect immediately. If synchronous, the changes go into effect during the next clock cycle. When the register values are changed after setting EN = ‘0’, the changes go into effect immediately. This is useful during the PWM mode, where the user can change the PWM period or duty cycle immediately. 25.3.3 Input Signal Characteristics The block has four input signals separate from the clock signal: ■ Enable ■ Capture ■ Timer Reset ■ Kill PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 259 Timer, Counter, and PWM 25.3.3.1 Enable Signal The effect of the enable signal is explained in the timing diagram for each mode. The following characteristics apply: ■ Gated timer pulse width mode and period mode take the Enable signal as input. ■ Gated timer stop at interrupt mode and PWM mode need an asserted Enable signal to function properly. ■ Free run mode is independent of the enable signal. ■ Enable signal polarity is reversed by setting the bit INV in configuration register CFG0. ■ Use of the capture signal to capture a time instance is valid only when the enable signal is asserted. 25.3.3.2 Capture Signal The capture signal is useful to find the time when an event occurs. The capture signal is usually combined with the free run timer mode. For the timer block to respond to the capture signal, the enable signal must be asserted before asserting the capture signal. The following describes the process: ■ The time value is captured in the capture register by assertion of the Capture signal for the block. ■ Whenever the rising edge of the Capture signal is detected, the count value is captured in the Capture register. ■ The capture register is read to find the time when the assertion of Capture signal occurred. ■ With every assertion of the Capture signal, a new value is captured to the Capture register. ■ An interrupt can be configured to occur at the assertion of the Capture signal. The interrupt bit in the Status register should be unmasked for the capture interrupt to occur. The Capture register value can be read in the capture ISR. ■ When using a fixed function timer with interrupt on capture enabled, read the capture register twice. The first reading yields an incorrect value (0xff)” Figure 25-3 shows the effect of the capture signal (period register = 0xFFFF). 260 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Timer, Counter, and PWM Figure 25-3. Capture Mode Timing Diagram Clock EN Bit Enable 0xFFFF 0xFFFE 0xFFFF 0xFFFE 0xFFFD 0xFFFD 0x1000 0x1 0x1 Count Value 0x0 0x0 Capture Input Pin Capture Register 0xFFFD 0x1000 Capture Interrupt PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 261 Timer, Counter, and PWM 25.3.3.3 Timer Reset Signal When the timer reset pin is asserted, the count value in the Count register (TMRx_CNT) is set to 0x00. When the timer reset pin is deasserted, the TMRx_CNT register is reloaded with the period value, and it functions in the configured mode. This signal stops the block operation for the time during which the timer reset signal is high and then restarts the operation from the beginning. Figure 25-4 is a timing diagram for the timer reset signal (Period register = 0xFFFF). Figure 25-4. Timer Reset Signal Timing Diagram Clock EN Bit Enable Timer Reset 0xFFFF 0xFFFE 0xFFFF 0xFFFE 0xFFFD 0xFFFD 0x0100 0x0100 0x00FF 0x00FF Count Value 0x1 0x0 25.3.3.4 Kill Signal The Kill signal is valid only during PWM mode. The effect of the kill signal is explained in PWM mode in the sections ahead. 262 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Timer, Counter, and PWM 25.3.4 Operating Modes 25.3.4.1 Timer Mode – Free Run Mode ❐ A terminal count output signal that generates a pulse at the terminal count – The terminal count output signal can be routed to any GPIO through the DSI. ❐ An interrupt at the terminal count – To initiate an interrupt, the terminal count interrupt in the Status register must be unmasked. The register configuration for Timer mode is: ■ Registers to set – TMRx_CFG0, TMRx_CFG1, TMRx_CFG2 ■ Bit MODE in TMRx_CFG0 = 0 – Timer mode ■ TMRx_CFG2[1:0] = 0 – Timer runs in continuous mode ■ The current timer value is read from the 8-bit Count registers CNT0 and CNT1. In the case of the 32-bit controller, a 16-bit read of the Capture register can be done. The Free Run mode is mainly used to obtain the current system time. Timer operation, automatically forced into the Free Run mode, occurs independent of the state of the Enable pin. This mode is called Free Run because the timer runs even if the state of the Enable pin is low. ■ In the case of the 8-bit controller, the 8-bit read is done. When an 8-bit read is done for the CNT0 register (LSB) the values of LSB and MSB are automatically captured in the Capture registers. The user can read the Capture register to obtain the 16-bit time value. The following describes the process: Figure 25-5 shows the terminal count output signal and the terminal count interrupt behavior in the Free Run mode (Period register value = 0xFFFF) and illustrates the following behavior. Note These bits cannot be modified by the user. ■ The timer is a down counter, and the current time value is stored in the TMRx_CNT registers. ■ The reload value for the timer is stored in the Period registers TMRx_PER0 and TMRx_PER1. ■ After the count reaches zero (terminal count), the period value is reloaded automatically to the Count registers for the timer to count. The reload value determines the period for the timer. Two types of output result when the terminal count is reached: ■ Independence of the Timer from the Enable signal for the block ■ The effect of changing the Period register with both EN = ‘1’ and EN = ‘0’ ❐ When the Period register is changed with EN = 1, the effect takes place only after the terminal count. ❐ When the Period register is changed with EN = 0, the effect takes place immediately after setting EN = 1. Figure 25-5. Free Run Mode Timing Diagram Clock EN Bit Period Value Period changed without changing EN 0xFFFF 0x00FF Period changed after EN = 0 0xE000 Automatic reload of period 0xFFFF Immediate effect after EN = 1 0x00FF 0xFFFF 0xFFFE 0xFFFD 0xE000 0xE000 0x00FE 0xFFFE 0xDFFF 0xDFFF 0xDFFE 0xFFFD 0xDFFE 0x00FD 0x1 Count Value 0x0 0x1 0x0 0x1 0x0 0x1 0x0 Terminal Count Output Pin Terminal Count Interrupt PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 263 Timer, Counter, and PWM 25.3.4.2 Gated Timer Mode In the Gated Timer mode, the timer does not run continuously; it starts and stops, based on certain criteria. The Gated Timer mode measures some parameters of the input signal, including the period of the input signal, the pulse width of the input signal, and the time after which an interrupt occurs. Depending on the configuration of the register, the following modes are supported: ■ Pulse Width ■ Period ■ Stop on Interrupt The following describes the process: ■ When the EN bit is set to ‘1’, the Count register is loaded with the period value from the Period register. ■ The timer begins counting whenever a rising edge occurs in the enable input. The Count register counts for every clock cycle. ■ When the next edge is reached (falling edge in the case of a Pulse Width count and the next rising edge in the case of a Period count), the timer stops to count. ■ On reaching the terminal count, the TMRx_CNT register is automatically reloaded with the period value. The timer stop interrupt can be configured to occur when the timer stops to count. The timer stop interrupt enable bit should be unmasked for the interrupt to occur. ■ The state of the timer is obtained from the TSTOP bit in the Status register. This sticky bit shows whether the timer has stopped counting; the user must clear the bit. The register configuration for the Counter mode is: ■ Registers to set – TMRx_CFG0, TMRx_CFG1, TMRx_CFG2 ■ Bit MODE in TMRx_CFG0 = 0 – block acts in gated timer mode ■ Two bits of TMRx_CFG2[1:0] – gated timer runs in various modes These modes are shown in Table 25-1. Table 25-1. TMRx_CFG2[1:0] Bit Settings in Gated Timer Mode TMRx_CFG2[1:0] Comments 00 Timer runs while EN bit of CFG0 register is set to ‘1’ 01 Pulse width count – counts from positive edge to negative edge of TIMEREN 10 Period count – counts from one positive edge to the next positive edge of TIMIEREN 11 Counts from enabled to IRQ The signal for which the pulse or period is measured is given to the Enable pin. 264 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Timer, Counter, and PWM Pulse Width Mode The count value is read using 16-bit read in case of a 32-bit controller and 8-bit read in case of a 8-bit controller. During 16-bit read, the count values are read as one 16-bit value and the value is captured in the Capture register. During the 8-bit read, a read of the CNT0 (LSB value) captures the LSB and MSB in the Capture register. The user can read the Capture register to obtain the time value. The input signal is given to the Enable pin. The timer begins counting at the rising edge of the Enable signal and stops counting at the falling edge of the Enable signal. There is a latency of one clock cycle for the block to detect the edges. The difference in the count value before and after the count is equal to the pulse width of the input signal in terms of counts. Figure 25-6 shows the Gated Timer in Pulse Width mode. In this figure, the One Shot mode is disabled, so the timer will start to count when the next rising edge is encountered. When the One Shot mode is enabled, the timer stops after the falling edge and should be enabled again. Figure 25-6. Gated Timer in Pulse Width Mode Clock EN Bit Enable Pin 0xFFFF 0xFFFE 0xFFFD 0x1001 0x1000 0x1000 0x0FFF 0x0FFE Count Value 0x0100 0x00FF Final CNT.reg Value 0x1000 0x00FF Timer Stop Interrupt PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 265 Timer, Counter, and PWM Period Mode The count value is read, using a 16-bit read in the case of a 32-bit controller and an 8-bit read in case of an 8-bit controller. During a 16-bit read, the count values are read as one 16-bit value, and the value is captured in the Capture register. During the 8-bit read, a read of the CNT0 register (LSB value) captures the LSB and MSB in the Capture register. The user can read the Capture register to obtain the time value. The input signal is given to the Enable pin. In this mode, the timer begins counting at the rising edge of the Enable signal and stops counting at the next rising edge. There is a latency of one clock cycle for the block to detect the edges. The difference in the count value between the start and the end of the count is equal to the period (in counts) of the input signal. Figure 25-7 shows the Gated Timer in Period mode. In this figure, the One Shot mode is disabled; the timer starts to count when encountering the next rising edge after the period calculation. When the One Shot mode is enabled, the timer stops after the second rising edge and should be enabled again. Figure 25-7. Gated Timer in Period Mode Clock EN Bit Enable Pin 0xFFFF 0xFFFE 0xFFFD 0x1001 0x1000 0x1000 0x0FFF 0x0FFE Count Value 0x0100 0x00FF Final CNT.reg Value 0x1000 0x00FF Timer Stop Interrupt 266 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Timer, Counter, and PWM Stop on Interrupt Mode The Stop on Interrupt mode is useful to stop the timer on occurrence of a specific event for the block. In this mode, the timer starts counting when the EN bit is set to ‘1’ and stops counting when an Interrupt Request (IRQ) is received. The IRQ is any configured interrupt (Terminal Count/Capture, Compare/Timer Stop) of the block. When the IRQ is received, the timer is automatically disabled. The timer should be enabled (EN = ‘1’) to start the timer again. The timer begins to run only after it is disabled and enabled again. The count value is read using a 16-bit read in case of 32-bit controller and an 8-bit read in case of 8-bit controller. During a 16-bit read, the count values are read as one value and the value is also captured in the Capture register. During the 8-bit read, a read of the CNT0 register (LSB value) captures the LSB and MSB in the Capture register. The user can read the Capture register to obtain the time value. Figure 25-8 shows the Gated Timer in IRQ mode. Figure 25-8. Gated Timer in IRQ Mode Clock EN Bit IRQ Pin 0xFFFF 0xFFFE 0xFFFD 0x1001 0x1000 Count Value Final CNT.reg Value 0x1000 Timer Stop Interrupt PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 267 Timer, Counter, and PWM 25.3.4.3 Pulse-width Modulator Mode The Pulse-width Modulator (PWM) mode is also called the Comparator mode, because the comparison output is a PWM output with a varying duty cycle and a varying period. The duty cycle depends on the compare type and compare value. The period depends on the Period register. For example, consider a 16-bit PWM block with a clock of 48 MHz. The period value is set to 0x8000 (32768 in decimal). This block gives a PWM period as follows: During the Comparator mode alone, the terminal count output pin acts as the complement to the compare output pin. To use this feature, enable the dead band mode (see Dead Band Feature on page 270). Enable the dead band feature by setting ‘1’ in the DB bit of CFG0. In the Comparator mode, the CNT register cannot be read. Compare Types The following is a description of various compare types. PWM Period = (Period Value * 1/Clock frequency) CMP_CFG = 000 PWM period for this example = (32768 * 1/48MHz) = 682.7 microsecond The compare output pin generates a pulse when the timer value = the comparator value. In this case, the width of the pulse = one clock cycle. The compare output interrupt signal occurs when the compare value = Timer Value. The register configuration for the Comparator mode is: ■ Registers to set – TMRx_CFG0, TMRx_CFG1, TMRx_CFG2 CMP_CFG = 001 ■ Bit MODE in TMRx_CFG0 = 1 – block acts as Comparator ■ Three Bits CMP_CFG in TMRx_CFG2 – Comparator runs in various compare modes The compare output pin generates a pulse when the timer value is less than the comparator value. The following describes the event: The following table lists appropriate register settings. ■ The width of the pulse = one clock cycle x Comparator value. ■ The rising edge occurs when the timer value becomes less than the comparator value, such as when the less than condition is met. ■ The falling edge of the pulse occurs when the terminal count is reached, such as when the condition changes to false. ■ When the comparator is disabled (EN = ‘0’) before the terminal count, the output remains high. ■ The Compare output interrupt signal occurs when the timer value is less than the Compare value. Table 25-2. Register Settings for Compare Type CMP_CGF Comments 000 Timer Value == Comparator Value 001 Timer Value < Comparator Value 010 Timer Value <=Comparator Value 011 Timer Value > Comparator Value 100 Timer Value >= Comparator Value The Comparator mode compares the timer value and the Compare register value, using either “==”, “<”, “<=”, “>” or “>=” depending on the mode configuration in the CFG2 register. The following describes the compare process: 1. The timer value begins to count when EN = ‘1’. 2. When the compare is true, the compare output signal is asserted or the compare interrupt signal is asserted. The block continues to count. 3. The CNT register is reloaded with the period value when the terminal count is reached and begins to count compare again. 4. The output of the compare is either the compare output signal or interrupt at the compare. 5. The interrupt occurs when the compare interrupt enable bit is unmasked in the Status register. 6. The compare output signal is routed to the GPIO pin using the DSI. 268 CMP_CFG = 010 The compare output pin generates a pulse when the timer value is less than or equal to the comparator value. The following describes the event: ■ The width of the pulse = one clock cycle x (Comparator value + 1). ■ The rising edge occurs when the timer value becomes equal to the comparator value, such as when the less than or equal to condition is met. ■ The falling edge of the pulse occurs when the terminal count is reached, such as when the condition changes to false. ■ When the comparator is disabled (EN = ‘0’) before the terminal count, the output remains high. ■ The Compare output interrupt signal occurs when the timer value = Compare value. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Timer, Counter, and PWM CMP_CFG = 011 CMP_CFG = 100 The compare output pin generates a pulse when the timer value is greater than the comparator value. The following describes the event: The compare output pin generates a pulse when the timer value is greater than or equal to the comparator value. The following describes the event: ■ The width of the pulse = one clock cycle x (Period – Comparator value). ■ The width of the pulse = one clock cycle x (Period – Comparator value + 1). ■ The rising edge occurs when the Count register is reloaded with the period value, such as when the greater than condition is met. ■ The rising edge occurs when the Count register is reloaded with the period value, such as when the greater than or equal to condition is met. ■ The falling edge of the pulse occurs at the end of count value = (Comparator value + 1), such as when the condition changes to false. ■ The falling edge of the pulse occurs at the end of count value = Comparator value, such as when the condition changes to false. ■ When the comparator is disabled (EN = ‘0’) before the condition changes to false, the output remains high. ■ When the comparator is disabled (EN = ‘0’) before the condition changes to false, the output remains high. ■ The Compare output interrupt signal occurs after the reload of the period value. ■ The Compare output interrupt signal occurs after the reload of the period value. Figure 25-9 shows the compare output for various Compare types. The Period register is loaded with 0xFFFF, and the Compare register is loaded with 0x1000. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 269 Timer, Counter, and PWM Figure 25-9. Compare Output for Various Compare Types Clock EN Bit Enable Compare Register 0x0100 0xFFFF 0xFFFF 0xFFFE 0xFFFE 0xFFFD 0xFFFD 0x0100 0x0100 0x00FF 0x00FF Count Value 0x1 Compare Output Pin 0x1 0x0 Timer = Compare 0x0 Timer < Compare Timer <= Compare Timer > Compare Timer >= Compare On the Fly Duty Cycle Update Dead Band Feature Support for multiple comparisons depends on the bit CMP_BUFF in Configuration register CFG0. The following describes the process: The dead band feature is used only in Comparator mode. To enable the dead band feature, set the DB bit in Configuration register TMRx_CFG0 to ‘1’. In the dead band mode, the terminal count output pin complements the comparator output pin. ■ ■ When the CMP_BUFF is set to ‘1’; the updated comparator value takes effect only after completion of the currently running period. After the terminal count, the new compare value is taken for further comparison. When this mode is used, the PWM block detects only one compare during a period. When the CMP_BUFF is set to ‘0’; the updated comparator value takes effect immediately even before the completion of the current running period. This may result in another toggling of the pin even before the completion of current period, thus supporting multiple comparisons. 270 During the dead band period, both compare output and complement compare output are low for a period, determined by the DEADBAND_PERIOD bits in the TMRx_CFG0 register. The dead band feature allows generation of two PWM pulses with non-overlapping outputs. The dead band feature uses a counter. The following describes the process: ■ When the comparator asserts the comparator output, it negates the asserted output for the dead band period. ■ The dead band period is loaded and counted for the period configured in the DEADBAND_PERIOD bits. ■ When the dead band period has completed, the signal is asserted, and the complement is negated. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Timer, Counter, and PWM ■ A dead band period of zero has no effect. ■ When the rate of change in the compare output is less than the dead band period, the immediate change is ignored. Transitions in the compare and complement compare output occur only for the next change in the compare output. ■ When the rate of change in the compare output is more than the dead band period, the transitions occur at both compare output changes. The following table shows end criteria (where the block stops) for each mode. When an end criterion is met, the block stops running and the EN bit is cleared. If the user wants to run the block again, then the block must be enabled (EN = ‘1’): Table 25-3. Block Stops Modes Counter The following describes the process: ■ ■ ■ When the Kill signal is asserted, the compare output and the complement of the compare output (if it exists) go to its unasserted state. The terminal count output acts as the complement of the compare output when the dead band feature is enabled. Free Run Mode Terminal Count Capture Mode Terminal Count Pulse Width Mode Negative Edge Period Mode Second Positive Edge IRQ Mode IRQ CMP_CFG = 000 Terminal Count CMP_CFG = 001 PWM Terminal Count CMP_CFG = 010 Terminal Count 25.3.5 CMP_CFG = 011 Terminal Count CMP_CFG = 100 Terminal Count Interrupt Enabling When the Kill signal is reasserted, the output signal is restored to its default state. Kill signal duration should be at least one full clock cycle for proper stopping and restoration of the output signal. There is a latency of two clock cycles before the output signal is restored. The block supports four types of interrupt: When the Kill signal is asserted, any change in the compare output is ignored, and the deassertion of the Kill signal results only in the previous default state. 25.3.4.4 One Shot Mode The One Shot mode works in combination with all of the modes specified above. The only difference is that the automatic reload of the Count register with the period does not occur. The block stops working when the required criteria are reached; there is no further reload and running of the block. ■ Terminal Count ■ Capture/Compare ■ Timer Enable ■ Timer Stop These interrupts are enabled by setting the corresponding bits in the Status register; occurrences are stored in the Status registers. Because these Status register bits are sticky, the interrupt request bits must be cleared explicitly by the software on occurrence of the interrupt. See Figure 25-1 on page 258. The process is described as follows: ■ Interrupt signals are sent to the Interrupt controller block, where execution is decided and processed. ■ The blocks are configured to support any combination of the four interrupts; only one interrupt is supported at a time. ■ When another interrupt signal comes during the execution of one interrupt, the new interrupt request is held pending until the previous interrupt execution is completed. ■ After the completion of the previous interrupt the new interrupt begins the execution. The register configuration for the One Shot mode is: ■ Criteria Timer Kill Feature The Kill signal is mainly used to deactivate the PWM signal in case of fault. Used only in Comparator mode, this signal places the output signals of the block in an unasserted state. Sub-Types Bit ONESHOT in Configuration register TMRx_CFG0 = 1 – enabled One Shot mode PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 271 Timer, Counter, and PWM Interrupt signals can be of two types: ■ ■ Raw Interrupt – Sent whenever the interrupt occurs. These interrupt signals do not wait for the execution of the previous interrupt request; they are continuously sent whenever the interrupt occurs. This type of interrupt signal is called "pulse input" because for every interrupt occurrence, a pulse is sent on the interrupt signal and does not wait for acknowledgement from the CPU. Status Interrupt – Sent depending on the status bits in the Status register. When the status bit is set to ‘1’, the interrupt signal is sent. The next interrupt signal is sent only after the status bit is cleared. The clearing of the status bit is handled by the software inside the interrupt service routine. The interrupt signal is not sent for every interrupt occurrence, but for every new setting of the status bit in the Status register. These types of interrupt allow the user control over the execution of the interrupt. This type of interrupt signal is called "level input" because the signal is asserted and remains asserted until the bit is cleared by the software. The selection of the interrupt signal type is decided using the bit IRQ_SEL, available in the configuration register TMRx_CFG1. 25.3.6 Sleep Mode Behavior The block supports the following two power saving features: ■ When the timer blocks are not accessed by the PHUB, the clock to the AHB interface is gated off, preventing all registers in the block from accessing the clock. ■ When the EN bit for a block is not asserted, the clock for that particular block is gated off. The block retains the values of the Period, Configuration, and Compare registers during the sleep and hibernate states. The Count register value is not retained during the sleep and hibernate states. 25.4 Register Listing The following table lists the registers. Table 25-4. Registers Register Names Comments Features TMRx_CFG0 Configuration Register Configures Enable of block, One Shot mode, mode of block, Enable pin inversion and dead band features TMRx_CFG1 Configuration Register Configures clock, deadband mode, disable on clear, first terminal count, IRQ selection TMRx_CFG2 Configuration Register Configures each of the modes, reset on disable, clear on disable, timer enable TMRx_PER0, TMRx_PER1 Period Register Retains the reload value TMRx_CNT_CMP0, Count/Comparator TMRx_CNT_CMP1 Registers 272 In the Comparator mode, the Count register cannot be read. So the Compare and Count register share the same address space. TMRx_CAP0, TMRx_CAP1 Capture Register TMRx_SR0 Status Register Hold the status of interrupts and controls the interrupt masking TMRx_RT0, TMRx_RT1 External Routing Registers Controls synchronization of the signals and routing of the signals to the DSI PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 26. I2C PSoC® 5LP devices include a fixed block I2C peripheral designed to interface the PSoC device with an I2C communications bus. Additional I2C interfaces can be created using universal digital blocks (UDBs) and PSoC Creator™. This chapter describes the fixed block I2C interface. For details on the UDB-based interface, see the component datasheet in PSoC Creator. Users not familiar with the I2C interface and the basics of an I2C transaction should see 26.2 Background Information. 26.1 Features The I2C is a serial communication block, which is used to interface the PSoC device to a two-wire I2C bus. This block eliminates an excessive CPU intervention and overhead with a hardware-based status detection and framing bit generation. This block can operate as a slave, master, or multimaster. Basic I2C features include: ■ Slave/master/multimaster, transmitter and receiver operation ■ Process bytes for low CPU overhead ■ Provides hardware support for bus status detection and frame bit generation ■ Generates interrupts for a variety of bus events ■ Interrupt or polling CPU interface ■ Supports bus stalling ■ Support for clock rates of up to 1MHz(Fast-mode plus) ■ 7 or 10-bit addressing (10-bit addressing requires firmware support) ■ SMBus operation (requires additional firmware layers on top of this block) ■ Routes SDA and SCL connection directly to one of two pairs of assigned pins on the SIO port, or through the DSI to any pair of GPIO or SIO pins ■ Provides HW address compare, and wake from sleep on address match ■ Provides 50 ns glitch filtering PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 273 I2C 26.2 For more information, see the I2C-Bus Specification, and User Manual, Version 03 at http://www.nxp.com/ acrobat_download/usermanuals/UM10204_3.pdf. Background Information The following paragraph introduces the basics of I2C bus communication. 26.2.1 Typical I2C Data Transfer 26.2.2 2 I C Bus Description In a typical I2C transaction, the following sequence takes place: I2C, The Inter IC, or bus was developed by Philips Semiconductors (now NXP) to provide a simple means to allow multiple ICs to communicate directly with each other over a common bus. Features of the I2C bus include: 1. A master device controls the SCL line and generates a Start condition followed by a data byte. The data byte contains a 7-bit slave address and a Read/Write (RW) bit. The Read/Write bit defines the direction of the data transfer with respect to the master. It is high for read and low for write. ■ Only two bus lines are required: (1) serial data (SDA) and (2) serial clock (SCL). ■ Serial, 8-bit, bi-directional data transfers can be made at up to 100 kbps in the standard mode, up to 400 kbps in the fast mode and up to 1 Mbps in the fast mode plus. See Figure 26-2 for an example I2C bus transfer. ■ Devices are connected to the bus using open collector or open-drain output stages; with pull up resistors for wired AND functions. ■ Each slave device connected to the bus has a unique software address. 3. The master transmits or receives an indeterminate number of bytes, depending on the RW direction. ■ Simple master/slave relationships exist; masters and slaves can operate as either transmitters or receivers. 4. When the transfer is complete, the master generates a Stop condition. ■ In the Multi-Master mode, to settle the contention for the bus between two masters, collision detection and arbitration techniques are used. 2. The slave device recognizes its address and acknowledges (ACK) the byte by pulling the data line low during the ninth bit time. If the slave does not respond to the first data byte with an ACK, a Stop condition is generated by the master to terminate the transfer. A Repeated Start condition may also be generated for a retry attempt. Figure 26-1. I2C Transfer of a Single Data Byte, With Clock Stretching by a Non-PSoC Slave P SDA Acknowledgement signal from Slave MSB Byte complete, interrupt within Slave S or SR SCL 1 2 7 8 9 How It Works The PSoC 5LP I2C interface provides support for bus status detection and generation of framing bits. It can operate at up to fast mode plus speeds, in these modes: ■ Slave – The interface listens for Start and Stop conditions to begin and end data transfers. 274 Clock line held low while interrupts are serviced. 1 2 3-8 9 ACK ACK SR or P STOP or Repeated START condition START or Repeated START condition 26.3 Acknowledgement SR signal from Receiver ■ Master – The interface generates the Start and Stop conditions and initiates data transfers by transmitting a slave address. ■ Multi-Master – The interface provides clock synchronization and arbitration to allow multiple masters on the same bus. Slave mode can be enabled at the same time as master mode. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I2C For details about the operation of these three modes, see 26.3.6 Operating the I2C Interface on page 275 and sections 26.6 Slave Mode Transfer Examples on page 280, 26.7 Master Mode Transfer Examples on page 283, and 26.8 Multi-Master Mode Transfer Examples on page 285. The I2C interface supports either 7-bit or 10-bit addressing. The hardware supports 7-bit address compare. In slave mode, 7-bit address detection is done by using either a hardware address compare or by the CPU in firmware. A 10-bit address detection must be done by the CPU in firmware. In master mode, 10-bit address generation must be done by the CPU in firmware. 26.3.1 Bus Stalling (Clock Stretching) After a byte is transferred on the I2C bus, a slave device may need time to store the received byte or to prepare another byte to be transmitted. In that case, the slave can hold the SCL line low before or after acknowledgment of a byte, which forces the master into a wait state until the slave is ready. This operation is known as stalling the I2C bus. Some devices in master mode may not support bus stalling; the system design should be checked before using bus stalling in slave mode. I2C The slave interface can stall the bus on every received address and on every completed byte transfer. After a byte is transferred, the master has half a clock cycle to start another read/write before the slave stalls the SCL bus again. 26.3.2 26.3.4 Pin Connections The I2C block controls the data (SDA) and the clock (SCL) to the external I2C interface, through direction connections to the GPIO/SIO pins. When I2C is enabled, these GPIO/ SIO pins are not available for general purpose use. The SDA and SCL connections of the I2C interface can be directly routed to one of two pairs of assigned pins on the SIO port. The connections can also be routed through the DSI to any other pair of GPIO or SIO pins. In all cases, the GPIO or SIO pins must be configured for “Open Drain, Drives Low” mode (see 19.3.2.5 Open Drain, Drives High and Drives Low on page 156). PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I2C Interrupts The I2C interface generates interrupts for these conditions: ■ Byte transfer (receive or transmit) complete ■ I2C bus Stop condition detected ■ I2C bus error detected The I2C interface cannot generate DMA requests. 26.3.5 Control by Registers The I2C interface is controlled by reading and writing a set of configuration, control, and status registers listed in the following table. These 8-bit wide registers are used to turn the I2C interface on or off, connect to I/O pins, set the baud rate, provide status and control for the data transfer processes, and monitor for exceptions. Table 26-1. I2C Registers Register I2C_CFG Usage Configuration – basic operating modes, oversample rate, and selection of interrupts. I2C_XCFG Configuration – configures enhanced features. I2C_CLK_DIV1 I2C_CLK_DIV2 Clock Divide – sets baud rate (along with oversample rate in I2C_CFG). I2C_CSR Control / Status – used to control the flow of data bytes and to keep track of the bus state during a transfer. I2C_MCSR Master Mode Control / Status – implements I2C framing controls and provides bus status. I2C_ADR Slave Address – for slave address recognition in hardware, holds the 7-bit slave address. I2C_D Data – provides read / write access to the data shift register. System Management Bus The System Management Bus (SMBus) is a bus definition based on the I2C bus. It is similar to, and generally a subset of, the I2C bus. For more information, see the SMBus Specification, Version 1.1. The I2C interface generally supports SMBus, although additional firmware support may be required. 26.3.3 To use the I2C in sleep mode, it must be routed through one of the dedicated pairs of SIO lines. 26.3.6 Operating the I2C Interface Operate the I2C interface in this manner: 1. Turn on the I2C interface by setting the I2C_XCFG bit 7, csr_clk_en. 2. To route the SDA and SCL to the desired pin pair, set up I2C_CFG as described in Table 26-2. 3. Select the baud rate (SCL clock frequency) by setting the I2C_CFG register, bit 2 and the I2C_CLK_DIV1 and I2C.CLK_DIV2 registers as shown in Table 26-3. The formula to determine the baud rate is: Baud Rate = Bus clock frequency / (Clock Division Factor * Oversample Rate) 4. Enable the desired mode of operation, following the instructions in 26.3.6.1 Slave Mode on page 277, 26.3.6.2 Master Mode on page 278, or 26.3.6.3 MultiMaster Mode on page 279. 275 I2C Table 26-2. Configuration of the I2C_CFG Register, Bit 7 Port Pinsa Pin Pair Register Settings I2C0 P12[4,5] I2C_CFG[6] = 1, I2C_CFG[7] = 0 I2C1 P12[0,1] I2C_CFG[6] = 1, I2C_CFG[7] = 1 Any other GPIO / SIO pin pair Selectable I2C_CFG[6] = 0, other DSI and GPIO registers according to pin pair selected a. The port pins used must be configured to “Open drain, Drives Low” mode (mode 4). The SIO pins are more suited for this purpose than the GPIO pins as the SIO pins have higher current sink capability and over voltage tolerance. Table 26-3. Configuration For I2C Baud Ratea Bus Clock (MHz) 3 I2C Mode Oversample Rate Divide Factor I2C_CLK_DIV2[1:0] SCL (kHz) I2C.CLK_DIV1[7:0] Standard 16 (0)2'b00 (2)8'b00000010 93.75 6 Standard 32 (0)2'b01 (2)8'b00000010 93.75 6 Fast 16 (0)2'b02 (1)8b'00000001 375 12 Standard 32 (0)2'b03 (4)8'b00000100 93.75 12 Fast 16 (0)2'b04 (2)8'b00000010 375 24 Standard 32 (0)2'b05 (8)8b'00001000 93.75 24 Fast 16 (0)2'b06 (4)8'b00000100 375 48 Standard 32 (0)2'b07 (16)8b'00010000 93.75 48 Fast 16 (0)2'b08 (8)8b'00001000 375 48 Fast plus 16 (0)2'b09 (3)8b'00000011 1000 67 Standard 32 (0)2'b10 (21)8b'00010101 99.7 67 Fast 16 (0)2'b11 (11)8b'00001011 381 67 Fast plus 16 (0)2'b12 (4)8'b00000100 1046 80 Standard 32 (0)2'b13 (25)8b'00011001 100 80 Fast 16 (0)2'b14 (13)8b'00001101 385 80 Fast plus 16 (0)2'b15 (5)8b'00000101 1000 a. Other values of bus clock, oversample rate and clock divider cause the baud rate to be scaled accordingly. 276 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I2C 26.3.6.1 Slave Mode To enable slave mode operation, set I2C_CFG bit 0, Enable Slave. See Figure 26-2. Figure 26-2. Slave Mode Operation Master transmits another byte Successful Slave Transmitter/Reciever CPU writes (ACK) to I2C_CSR register. An interrupt is generated on byte complete. START 7-Bit Address 7 8 ACK = Slave OK to receive more. Master may send more or issue stop. STOP Write (RX) ACK/ NACK 1 7 8 CPU reads the received byte from the I2C_D register and checks for “Own Address” and R/W. 9 NACK = Slave says no more CPU reads the received byte from the I2C_D register. Read (TX) 1 R/W SCL line is held low. CPU issues ACK/ NACK command with a write to the I2C_CSR register. 8-Bit Data ACK A byte interrupt is generated. SCL line is held low. CPU writes the byte to transmit to the I2C_D register. CPU writes (ACK | TRANSMIT) to I2C_CSR register. An interrupt is generated on a complete byte + ACK/NACK. SCL line is held low. 8-Bit Data ACK ACK/ NACK 9 1 7 8 NACK = Master says end-of-data STOP 9 ACK = Master wants to read another byte. CPU writes a new byte to the I2C_D register and then writes a TRANSMIT command to I2C_CSR to release the bus. In slave mode, the I2C interface continually monitors the bus for a Start condition. When a Start condition is detected, the following ensues. 1. The first byte, which is the Address / RW byte, starts to be shifted in. When all eight bits are received, a Byte Complete status is generated. 2. On the following low of the clock, the bus is stalled by holding SCL low, until the address byte is read and compared. An ACK or NACK is then issued, based on that comparison. 3. If there is an address match, the RW bit determines the direction of the data transfer, as shown in the two branches of Figure 26-2. After each byte is received, or when a new byte can be transmitted, a Byte Complete status is generated, and SCL is held low to stall the bus until the CPU handles the interrupt and prepares itself for the next transfer. 4. When transmitting bytes, the slave receives an ACK / NACK from the master for each byte sent. ACK is a signal that the master wants another byte. NACK or a Stop condition is a signal that the master does not want any more bytes – the CPU should let the I2C interface go to an idle state. 5. When receiving bytes, the slave ACKs / NACKs each byte received from the master. ACK is a signal that the slave can accept another byte. NACK is a signal that no more bytes can be accepted – after generating a NACK the CPU should then let the I2C interface go to an idle state. 6. Data transfer is complete when the master generates a Stop condition. 7. At anytime when a Stop condition or Bus Error is detected, the I2C interface is automatically reset to an idle state. Slave Address Recognition The slave address recognition feature can be enabled in hardware to reduce CPU usage. To enable hardware address recognition: 1. Set the 7-bit slave address in I2C_ADR, bits 0 to 6. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 277 I2C immediately, instead of waiting for the CPU to write the correct data to the I2C_D.data register. As a result, garbage is read by the external master. 2. Set I2C_XCFG, bit 0, HW Addr En. When hardware address recognition is enabled, the address portion of the first byte received after a bus Start condition is compared to the value in I2C_ADR. ■ If no match is detected, the byte is automatically NAKed. ■ The use case mentioned above is very rare and the workaround in that case is to use the software address matching instead (XCFG.hw_addr_en=0) of hardware address matching. However, the workaround prevents the wake-onaddress capability. Fortunately, a different piece of logic handles the slave operation during power-down and that is not vulnerable to the issue mentioned above. So, when entering power-down mode, set the XCFG.hw_addr_en to ‘1’ (enable the hardware address decoding) and disable it when leaving the power-down mode. If a match is detected, the byte is automatically ACKed, a byte complete interrupt is generated, and the remainder of the transfer is performed as described above. Note If an external I2C master addresses another I2C slave and then issues a repeated start (instead of a stop) to this slave, this I2C slave will return garbage data on read. The problem is that the block does not clear a critical internal signal (the I2C_CSR.address bit) on a repeated start. This causes the ensuing data access to this slave to proceed 26.3.6.2 Master Mode To enable master mode operation, set the I2C_CFG bit 1, Enable Master. See Figure 26-3. Figure 26-3. Master Mode Operations Successful Master Transmitter/ Receiver START A Start/Address compete interrupt is generated. 7-Bit Address R/W ACK CPU issues ACK/ NACK command to the I2C_CSR register. ACK =Master wants more data bytes 8-Bit Data CPU issues a command to the I2C_CSR register to release SCL The SCL line is held low. The SCL line is held low. STOP ACK/ NACK 1 7 8 9 Read (RX) CPU issues generate START command to I2C_MCS. An interrupt is generated on byte complete. CPU reads the received byte from I2C_D register. NACK = Master indicates end-ofdata 1 7 8 CPU writes address byte to the I2C_D register. 9 Write (TX) CPU checks Read/ Write Bit CPU issues TRANSMIT command to the I2C_CSR register. CPU writes a byte to transmit I2C_D register. An interrupt is generated on completion of the byte + ACK/NACK. 8-Bit Data ACK/ NACK 1 7 Master wants to send more bytes. If Enable Slave is not set, the I2C interface is in Master Only mode and ignores all externally generated Start conditions. Operation in master mode is as follows: 278 8 The SCL line is held low. CPU issues STOP command NACK = Slave says no more. STOP 9 ACK = Slave says OK to receive more. Master can send more or Stop 1. To start a transfer, the master writes the slave address/ direction byte to I2C_D and sets I2C_MCSR bit 0, Start Gen (or bit 1, Restart Gen). In a single-master environment the Start condition is generated; the byte is transmitted; and a Byte Complete interrupt is generated. If the byte is ACKed by the slave, PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I2C a slave. The firmware may then ACK the address to continue the transfer as a slave, or NACK the address. data bytes can be sent or received as shown in the two branches of Figure 26-3. 2. When transmitting bytes, the master receives an ACK/ NACK from the slave for each byte sent. ❐ ACK is a signal that the slave can accept another byte. NACK is a signal that no more bytes can be accepted. 3. When receiving bytes, the master ACKs/NACKs each byte received from the slave. ACK is a signal that the master wants another byte. NACK is a signal that the master is done accepting bytes. 4. When data transfer is complete, the master issues a STOP command and enters an idle state. Instead of a Stop condition, the CPU can issue a Restart command, and another transfer is immediately started. 26.3.6.3 Multi-Master Mode Multi-master mode becomes enabled when Master mode is enabled by setting the I2C_CFG bit 1, Enable Master. In Multi-master mode, the CPU starts the transfer in the same manner as in a single-master environment. However, before generating a Start condition, the master must monitor the Bus Free bit in I2C_MCSR, and wait until the I2C bus is free. When a Start condition is initiated, one among the following may occur: ■ ■ Another master in a multimaster environment has generated a valid Start, and the bus is now busy. The Start condition is not generated. The resulting behavior depends upon whether Slave mode is enabled. ❐ Slave mode is enabled – A Byte Complete interrupt is generated. When reading I2C_MCSR, the master sees that the Start Gen bit is still set and that I2C_CSR has the Address bit set, indicating that the block has been addressed as a slave. The firmware may then ACK the address to continue the transfer as a slave, or NACK the address. ❐ Slave mode is not enabled – The Start Gen bit remains set, and the transfer is delayed until the bus becomes free. A Byte Complete is generated when the Start condition is generated and the address byte is transmitted. The Start condition is generated, but the master loses arbitration to another master. The resulting behavior depends upon whether Slave mode is enabled. ❐ Slave mode is enabled – A byte complete interrupt is generated. When reading I2C_MCSR, the master sees that the Start Gen bit is clear, indicating that the Start condition was generated. However, the Lost Arb bit is set in I2C_CSR. The Address status is also set, indicating that the block has been addressed as PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 26.4 Slave mode is not enabled – A Byte Complete interrupt is generated. The Start Gen bit is clear and the Lost Arb bit is set. The hardware waits for a command from the CPU, stalling the bus if necessary. The master clears I2C_CSR to release the bus and allow the transfer to continue, and the I2C interface goes back to idle mode. The firmware can then retry the transfer when the bus becomes free again. Hardware Address Compare The hardware has the ability to compare the seven address bits received on the SDA line with that configured in the I2C_ADR register. On a true compare, the address is automatically ACKed, the SCL line held low, and a byte complete interrupt is issued. On reception of the byte complete interrupt from the hardware, the firmware needs to read bit [0] of the data register to determine Read/Write direction for the transfer. The firmware must then set the transmit bit in the I2C_CSR register to release the SCL. On a mismatch, the address is automatically NAKed and the hardware reverts to an idle state, waiting for the new START detection. 26.5 Wake from Sleep When the HW address compare is enabled and the device is put to sleep, the slave can be used to wake the device on an I2C HW address match (only when either of the SIO pairs are used as I2C pins). While in sleep, the master clock is disabled. The incoming SCL clock is used to latch the address into the block. When the address matches, the wakeup interrupt is asserted to wake the system up, and the SCL is pulled low until the master clock is operational. After the system wakes, the I2C block is switched back to normal operation, and all other transaction proceed. The I2C block responds to a transaction during sleep, only when the following conditions are met: ■ I2C is configured as a slave ■ Hardware address compare is enabled ■ There is an address value written in the I2C_ADR ■ The I2C_ON bit in I2C_XCFG is set to 1'b1 To ensure proper sleep mode operation, follow these instructions: 1. The CPU must set the Force NACK bit of the I2C_XCFG register to put the part to sleep. 2. The firmware must poll the Ready_To_Sleep bit in the I2C_XCFG register; when the bit is high, the device can be put to sleep. 279 I2C the power down signal to the I2C block, which is required for an I2C address match wake up. If I2C address match wake up is required in standby mode, then clear the EN_I2C bit in the PM_STBY_CFG5 register before entering the standby mode. Clearing this bit asserts 26.6 Slave Mode Transfer Examples Slave mode receives or transmits data, as described in this section. 280 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I2C 26.6.1 Slave Receive A slave receive operation is accomplished as shown in Figure 26-4. Figure 26-4. Slave Receive Operation Sequence Start NO Write ‘1’ to I2C_XCFG[7], csr_clk_en, to start up the I2C interface hardware. Byte Complete, I2C_CSR[0] == 1, or Error, (I2C_CSR & 0xA0) != 0? ERROR E YES Write ‘0’ to I2C.CSR to clear all bits. I2C_CSR[3] == 1, Address? Set I2C_CFG[3:2], Clock Rate, and I2C_CLK_DIV to set the SCL frequency. NO E YES Write ‘0’ to I2C_CSR[3] to reset address. Set I2C_CFG[7:6], SIO Select and PSelect, to connect SDA and SCL to the appropriate pins. Write ‘0’ to I2C_CSR[4 to NACK NO Set I2C_CFG[0], Enable Slave, to start Slave mode. I2C_D[7:1] == MyAddr? YES Write ‘1’ to I2C_CSR[4] to ACK. YES I2C_D[0] = 0, I2C Write? NO Go do slave transmit functions. Write ‘0’ to I2C_CSR[2] to set Receive mode Write ‘0’ to I2C_CSR[4] to NACK ERROR Byte Complete, I2C_CSR[0] == 1, or Error, I2C_CSR[7] != 0, or Stop, I2C_CSR[5] != 0? E NO STOP NO YES Stop Status, I2C_CSR[5] == 1, or Error, I2C_CSR[7] != 0? ERROR E Copy I2C_D to receive data buffer. YES Done? NO Write ‘1’ to I2C_CSR[4] to ACK E Report a successful transfer. YES Report and handle error. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F End 281 I2C 26.6.2 Slave Transmit A slave transmit operation is accomplished as shown in Figure 26-5. Figure 26-5. Slave Transmit Operation Sequence Flow Chart for Slave Transmit Start NO Write ‘1’ to I2C_XCFG[7], Clk Gate En, to start up the I2C interface hardware. Byte Complete, I2C_CSR[0] == 1, or Error, (I2C_CSR & 0xA0) != 0? ERROR E YES Write ‘0’ to I2C_CSR to clear all bits I2C_CSR[3] == 1, Address? Set I2C_CFG[3:2], Clock Rate, and I2C_CLK_DIV to set the SCL frequency. NO E YES Write ‘0’ to I2C_CSR[3] to reset address. Set I2C_CFG[7:6], SIO Select and PSelect, to connect SDA and SCL to the appropriate pins. Write ‘0’ to I2C_CSR[4] to NACK. NO Set I2C_CFG[0], Enable Slave, to start slave mode. I2C_D[7:1] == MyAddr? YES Write ‘1’ to I2C_CSR[4] to ACK. NO I2C_D[0] = 0, I2C Write? YES Go do slave receive functions. Copy first/next byte from transmit data buffer to I2C_D. Write ‘1’ to I2C_CSR[2] to start transmitting byte. NO ERROR NO Byte Complete, I2C_CSR[0] == 1, or Error, (I2C_CSR & 0xA8) != 0? E Stop Status, I2C_CSR[5] == 1, or Error, I2C_CSR[7] != 0? ERROR E YES YES Report a successful transfer. ACK Byte ACK’ed or NACK’ed? I2C_CSR[1] End E NACK Report and handle error. Note that, instead of waiting for Byte Complete or Error, an interrupt can be generated for each of these conditions, as well as for the I2C Stop condition. The interrupt handler can then do some or all of the functions shown. 282 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I2C 26.7 Master Mode Transfer Examples Master mode receives or transmits data, as described in this section. 26.7.1 Single Master Receive A master receive operation in a single-master system is accomplished as shown in Figure 26-6. Figure 26-6. Single Master Mode Receive Operation Flow Chart for Single Master Receive Write ‘1’ to I2C_MCSR[0], Start Gen, to start the transfer. Start Write ‘1’ to I2C_XCFG[7], Clk Gate En, to start up the I2C interface hardware. NO Byte Complete, I2C_CSR[0] == 1, or Error, (I2C_CSR & 0xA0) != 0 or I2C_MCSR[2] != 1? Write ‘0’ to I2C_CSR to clear all bits. ERROR E YES Set I2C_CFG[3:2], Clock Rate, and I2C_CLK_DIV to set the SCL frequency. NACK * If address byte is NACK’ed, instead of retry, an error can be reported. Set I2C_CFG[7:6], SIO Select and PSelect, to connect SDA and SCL to the appropriate pins. Byte ACK’ed or NACK’ed? I2C_CSR[1] ACK Write ‘0’ to I2C_CSR[2] to set Receive mode. Set I2C_CFG[1], Enable Master, to start master mode. Set I2C_D = Slave Addr/Read Write ‘0’ to I2C_CSR[4] to NACK. ERROR Byte Complete, I2C_CSR[0] == 1, or Error, (I2C_CSR & 0xA0) != 0 or I2C_MCSR[2] != 1? E NO NO YES Stop Status, I2C_CSR[5] == 1, or Error, I2C_CSR[7] != 0? ERROR E Copy I2C_D to receive data buffer. YES E Done? NO Write ‘1’ to I2C_CSR[4] to ACK. Report a successful transfer. YES Report and handle error. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F End 283 I2C 26.7.2 Single Master Transmit Figure 26-7 illustrates the process by which you generate a master transmit operation in a single master system. Figure 26-7. Single Master Mode Transmit Operation Flow Chart for Single Master Transmit Write ‘1’ to I2C_MCSR[0], Start Gen, to start the transfer. Start Write ‘1’ to I2C_XCFG[7], Clk Gate En, to start up the I2C interface hardware. NO ERROR Byte Complete, I2C_CSR[0] == 1, or Error, (I2C_CSR & 0xA0) != 0 or I2C_MCSR[2] != 1? Write ‘0’ to I2C_CSR to clear all bits. E YES Set I2C_CFG[3:2], Clock Rate, and I2C_CLK_DIV to set the SCL frequency. NACK * If address byte is NACK’ed, instead of retry, an error can be reported. Set I2C_CFG[7:6], SIO Select and PSelect, to connect SDA and SCL to the appropriate pins. Byte ACK’ed or NACK’ed? I2C_CSR[1] ACK Write ‘1’ to I2C_CSR[2] to set transmit mode. Set I2C_CFG[1], Enable Master, to start master mode. Set I2C_D = Slave Addr/Write. NO Done? YES Copy first/next byte from transmit data buffer to I2C_D. Write ‘1’ to I2C_CSR[2] to start transmitting byte. ERROR E NO Byte Complete, I2C_CSR[0] == 1, or Error, (I2C_CSR & 0xA0) != 0 or I2C_MCSR[2] != 1? Write ‘0’ to I2C_CSR[2] to generate a Stop condition. NO Stop Status, I2C_CSR[5] == 1, or Error, I2C_CSR[7] != 0? ERROR E YES YES ACK Report a successful transfer. Byte ACK’ed or NACK’ed? I2C_CSR[1] NACK End E Write to I2C_CSR[4]. A Stop condition is automatically generated by the hardware, regardless of the value written. 284 Report and handle error. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F I2C Defining single master operations allows the following assumptions to be made: ■ ■ There is no need to check for bus busy (I2C_MCSR[3]) or Lost Arb (I2C_CSR[6]). 26.8 There is no need to Enable Slave (I2C_CFG[0]) when enabling the master mode, as the interface will never be forced into slave mode due to bus busy or lost arbitration. Multi-Master Mode Transfer Examples In multi-master mode, data transfer can be achieved with the slave mode not enabled or with the slave mode enabled. 26.8.1 Multi-Master, Slave Not Enabled A master data transfer operation in a multi-master system, where the slave mode is not enabled is shown in Figure 26-8. Figure 26-8. Multi-Master Mode, Slave Not Enabled Sequence Flow Chart for Multi-Master, Slave Not Enabled Start Write ‘1’ to I2C_XCFG[7], Clk Gate En, to start up the I2C interface hardware. Bus Busy? I2C_MCSR[3] YES NO Write ‘1’ to I2C_MCSR[0], Start Gen, to start the transfer. Write ‘0’ to I2C_CSR to clear all bits. Set I2C_CFG[3:2], Clock Rate, and I2C_CLK_DIV to set the SCL frequency. Start condition? I2C_MCSR[0] == 0 Set I2C_CFG[7:6], SIO Select and PSelect, to connect SDA and SCL to the appropriate pins. Set I2C_CFG[1], Enable Master, to start Master mode with slave not enabled. NO Bus became busy. YES ERROR NO Byte Complete, I2C_CSR[0] == 1, or Error, I2C_CSR[7] != 0 or I2C_MCSR[2] != 1? YES Set I2C_D = Slave Addr/Read or Write Lost arbitration? I2C_CSR[6] == 1 YES Lost arbitration, restart transfer. NO Report and handle error. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Continue with data transfer as in single master. 285 I2C 26.8.2 Multi-Master, Slave Enabled A master data transfer operation in a multi-master system, where the slave mode is enabled is shown in Figure 26-9. Figure 26-9. Multi-Master Mode, Slave Enabled Sequence Flow Chart for Multi-Master, Slave Enabled Start Write ‘1' to I2C_XCFG[7], Clk Gate En, to start up the I2C interface hardware. Bus Busy? I2C_MCSR[3] YES NO Write ‘0’ to I2C_CSR to clear all bits. Write ‘1’ to I2C_MCSR[0], Start Gen, to start the transfer. Set I2C_CFG[3:2], Clock Rate, and I2C_CLK_DIV to set the SCL frequency. Set I2C_CFG[7:6], SIO Select and PSelect, to connect SDA and SCL to the appropriate pins. ERROR Byte Complete, I2C_CSR[0] == 1, or Error, I2C_CSR[7] != 0? NO YES Set I2C_CFG[1], Enable Master, and I2C_CFG[0], Enable Slave, to start both modes. YES Bus became busy, or lost arbitration? I2C_MCSR[0] == 1 and I2C_CSR[3] == 1 Set I2C_D = Slave Addr/Read or Write. NO Report and handle error. 286 Continue with address recognition as a slave. Continue with data transfer as in single master. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 27. Digital Filter Block (DFB) Some PSoC® devices have a dedicated hardware digital filter block (DFB) used to filter applications. The heart of DFB is a multiply and accumulate unit (MAC), which can do 24 bit * 24 bit multiply and 48 bit accumulate in one system clock cycle. In addition, there are data RAMs to store data and coefficients of digital filters. 27.1 Features ■ Two 24-bit wide streaming data channels ■ Two sets of data RAMs each that can store 128 words of 24-bit width each ■ One interrupt and two DMA request channels ■ Three Semaphore bits to interact with system software ■ Data alignment and coherency protection support options for input and output samples 27.2 Block Diagram The digital filter block (DFB) is a 24-bit fixed point, programmable limited scope DSP engine. The DFB is made up of four primary subfunctions as shown in the DFB Basic Block diagram in Figure 27-1. ■ Controller ■ Datapath ■ Address Calculation Units (ACUs) ■ Bus Interface The Controller consists of a small amount of digital logic and memories. The memories in the controller are filled with assembled code that make up the data transform function the DFB is intended to perform. The Datapath subblock is a 24-bit fixed point, numerical processor containing a Multiply and Accumulator (MAC), a multifunction Arithmetic Logic Unit (ALU), sample and coefficient and data RAM (data RAM is shown in Figure 27-1) as well as data routing, shifting, holding, and rounding functions. The datapath block is the calculation unit inside the DFB. The addressing of the two data RAMs in the datapath block are controlled by the Address Calculation Units (ACUs). There are two (identical) ACUs, one for each RAM. These three subfunctions make up the core of the DFB block and are wrapped with a 32-bit DMA-capable AHB-Lite Bus Interface with Control/Status registers. Each of these four subfunctions are discussed in the following sections. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 287 Digital Filter Block (DFB) Figure 27-1. Digital Filter Block Diagram D a ta p a th S ta g e R e g is te r A MAC ALU S h ift H o ld R ound D a ta RAM A S ta g e R e g is te r B In p u t fr o m CPU/ DM A Bus In te r fa c e H o ld R e g is te r A D a ta RAM B H o ld R e g is te r B O u tp u t to CPU/ DM A C o n tr o l d fb _ in tr C o n tr o lle r ACU A d fb _ d m a r e q 1 d fb _ d m a r e q 2 ACU B A d d r e s s C a lc u la tio n U n it d fb _ g lo b a li1 & 2 d fb _ g lo b a lo 1 & 2 D S I s ig n a ls F ig u re 3 0 -1 . D ig ita l F ilte r B lo c k D ia g r a m 27.3 27.3.1 How It Works Controller The controller consists of a RAM-based state machine, a RAM-based control store, program counters, and next state control logic (see Figure 27-2 on page 289). Its function is to control the address calculation units and the datapath, and to communicate with the bus interface to move data in and out of the datapath. 288 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital Filter Block (DFB) Figure 27-2. Controller Block Diagram fstate[4:0] loop nstate[4:0] fsm_addr[4:0] FSM RAM Fjump addr Fjump limit Jump addr Loop PC A csa_addr[5:0] PC B csb_addr[5:0] Control Store RAM A Conditions Next State Logic RAM Selection eob The contents of FSM RAM, the two control store RAMs, the ACU RAM, and potentially the two datapath RAM (if initial conditions are required) must be loaded by the system before use. The contents of the DFB RAMs are stored in flash memory from where they are written into the RAM before the DFB operation is enabled. The next state decode logic and the FSM RAM comprise the main DFB branch control. The next state decoder generates the FSM RAM’s address and the RAM produces next state information as well as branch flag masks. These masks enable the use of flags as jump conditions for conditional branching. This state machine controls the program counter to produce the address for the Control Store RAMs. There are two identical Control Store (CS) RAMs and an associated Program Counter to allow an interleaving methodology for CS opcode fetches. The CS RAMs are 64x32 each. Both CS RAMs are sometimes filled with identical data. It is possible to effectively double the control store instruction space by using different contents in each RAM. It is during branch conditions that next state address calculations happen. Hence, the two possible branch addresses are supplied – one to each RAM. When the branch condition is determined, late in the cycle, the controller simply picks the PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Control Store RAM B To Datapath, ACU, and Bus Interface correct CS RAM output. Opcode execution then switches to and stays with the CS RAM until the next jump condition. 27.3.1.1 FSM RAM FSM RAM is 64x32 RAM. It is used as ROM. The FSM RAM is filled with control flow information implementing the desired function of the DFB prior to use. This RAM is loaded typically at system boot time, but is not restricted to any particular time as long as the DFB is not running (run is deasserted in DFB_CR[0]). The code in this and the Control Store RAMs can be altered at anytime to change the function performed by the DFB. In fact, some applications have the algorithm loaded routinely and swapped out when several channels of data need processing or when one channel needs multiple transforms – when the code is too large to fit in the available space. The FSM RAM is addressed as two banks of 32x32. The Bank selection is achieved using the CSR bit (DFB_CR[1]). The primary use of the two banks is to allow two separate code stores to load and jump between without incurring the reload penalty of the FSM RAM. 289 Digital Filter Block (DFB) Table 27-1 shows the bit fields used for the controller by the 32-bit FSM RAM. Table 27-1. FSM RAM Bit Field Mapping False Jump Limit False Jump Address Next State jaddr fjlim fjaddr nstate 22:17 16:11 10:5 4:0 Name Enables Loop Jump Address Signal enables loop Bits 31:24 23 Description Enables for the top 8 input branching conditions Signifies a code loop Jump address for CS RAMs on TRUE Address loop limit Jump address for CS RAMs on FALSE* Next state address for FSM * This false jump address is for use only in a loop state, where the controller moves back to the start of the loop on a false condition. If the state is not a loop state, then this address is used for the next state on false value. 27.3.1.2 Program Counter The primary purpose of the program counter (PC) is to supply correct addresses to the Control Store (CS) RAMs. This is not as simple as providing a direct address from the FSM RAM because jump addresses must be determined and held in the PC before branches are taken. The PC also controls the incrementing and wrapping of addresses for loops, allowing the FSM to sit in one state during looping processes. For this reason the FSM RAM sends out the jump address and loop conditions to the PC 27.3.1.3 Control Store The term Control Store (CS) refers to a bank of two interleaved RAMs used to hold control opcodes for the ACUs and the Datapath unit. These RAMs are addressed by the FSM RAM indirectly through the Program Counters and set the per-cycle operation state of the DP and ACUs. The outputs of these two 32-bit wide RAMs are muxed to one control bus (based upon which is presently the active RAM denoted by DFB_SR[0]) and provide the following bit-fields to the ACUs and Datapath unit listed in Table 27-2. Table 27-2. Control Store RAM Bit Field Mapping Name DP CTRL Bus WR ACU-A Opcode ACU-B Opcode ACU Addr End of Block Signal dp_ctrl buswr acua_op acub_op acu_addr eob Bits 31:14 13 12:9 8:5 4:1 0 Description 27.3.1.4 Control bus to the Datapath Unit Signifies a data output condition to the bus ACU A’s opcode Next State Decoder The Next State Decoder is combination logic that controls the state transitions in the FSM RAM. The next state decoder is the logic that gives the address (state address) to the FSM. The result of the next state decoder is governed by the branching signal conditions. You get a state transition when one of these two conditions exist: ■ EOB is high and the signal condition goes high. This is the jump on true branch. ■ Loop (cfsmram[23]) is low meaning no loop, EOB is high, and condition is low. This is the flow through condition for a false condition. The branching conditions are: ACU B’s opcode ACU RAM’s address End of Block marker 2. Datapath status inputs such as sign, threshold, and equal. ❐ Dpsign – A jump based on the MSB of the ALU output. If ALU output goes negative, assert. ❐ Dpthresh – Datapath Threshold – Asserted when the ALU detects a sign change, such as a zero crossing detection. ❐ Dpeq – Datapath Equity – Asserted when the ALU hardware detects an output value of zero. 3. Acueq – ACU A or B REG is equal to MREG or LREG, if modflag is set. ACU A or B REG is equal to 127 or 0, if modflag is cleared. This means that the pointer to the DP Data registers has reached its upper/lower limit. See 27.3.2 Datapath on page 291 for clarity. 1. End of block is encountered for a control store block – a condition for a jump because a jump instruction signifies the end of the block. 290 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital Filter Block (DFB) 27.3.2 4. IN1 or IN2 – When new data is available in one of the staging registers A or B. Signals a new input cycle and is available for consumption. Remains asserted until cleared by a bus read command. Datapath (DP) is the name used to refer to the numerical calculation unit of the DFB. The datapath subblock is a 24bit fixed-point numerical processor containing a 48-bit MAC, a multi-function ALU, sample and coefficient data RAMs as well as data routing, shifting, holding and rounding functions. 5. globali1 – Branch control input from DSI port. 6. globali2 – Branch control input from DSI port. 7. The sat_det flag (Saturation) from ALU – This flag is set when saturation occurs in MAC, ALU, or Shifter. The DP architecture makes use of two 128x24 single-port RAMs (RAM A and RAM B). The RAMs can be loaded from the bus or from the datapath output (feedback). These RAMs hold data and coefficients with size and location under full DFB controller control. 8. Any of the semaphores (see the PHUB and DMAC chapter on page 55). For branching, the branching conditions must be enabled. The ENGLOBALS, ENSATRND, ENSEM, SETSEM, and CLEARSEM commands are used. The heart of the DP unit is a 48-bit Multiply and Accumulator (MAC). Two 24-bit values can be multiplied and the result added to the 48-bit accumulator in each clock cycle. This accumulator or any memory value can be routed to the ALU. Results from the ALU can then be stored in either Data RAM. The MAC is the only portion of the DP that is wider than 24 bits. All results from the MAC are passed on to the ALU as 24-bit values representing the high-order 24 bits in the accumulator shifted by one (bits 46:23). The MAC assumes an implied binary point after the MSB which shifts the result down a bit in the output of the MAC. For this reason, bits 46:23 are used instead of 47:24. If the ALU command is ENSEM, then the data on acu_addr[2:0] is written to the register sem_en for enabling semaphores to be branching conditions. The acu_addr[2:0] is converted bitwise to enable each of the three semaphores. The SETSEM and CLEARSEM are used to set or clear the semaphores based on the semaphore selected in acu_addr[2:0]. Acu_addr[2] -> semaphore2 Acu_addr[1] -> semaphore1 Acu_addr[0] -> semaphore0 The DP unit also contains an optimized ALU that supports add, subtract, comparison, threshold, absolute value, squelch, saturation, and other functions. The ENGLOBALS command is used to enable the use of external dsi inputs and datapath saturation flags as branching conditions. ENGLOBALS shares an ALU opcode with ENSATRND. They are differentiated by the acu_addr[3] bit as shown in Table 27-3. With the exception of the DP RAM addresses, the DP unit is completely controlled by seven control fields totaling 18 bits coming from the DFB Controller as the DP_CTRL control bus (Table 27-2 on page 290). These 18 bits of control are listed in Table 27-4. Table 27-3. ENGLOBALS and ENSATRND Commands Acu_addr[0]: enables globali1 Englobals Acu_addr[3]=0 Datapath Acu_addr[1]: enables globali2 Acu_addr[2]: enables sat_det Acu_addr[0]: writes to rnd_flag Ensatrnd Acu_addr[1]: writes to sat_flag Acu_addr[3]=1 Acu_addr[2]: creates strobe to clear saturation flag Table 27-4. Datapath Opcode Bit Field Mapping Name B Mux Ctrl A Mux Ctrl MAC Opcode ALU Opcode Shift Opcode RAM A WR RAM B WR Signal muxb*_ctrl muxa*_ctrl mac_op alu_op shift dpa_r_wb dpa_r_wb Width 3 3 2 5 3 1 Description mux1b mux2b mux3b mux1a mux2a mux3a MAC opcode ALU opcode DP output shifter opcode Write signal to RAM A 1 Write signal to RAM B Note how the different signals from Table 27-4 affect the functioning of the different elements in the datapath. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 291 Digital Filter Block (DFB) Figure 27-3. Datapath mux2a mux1a mux1a mux2a mux3a alu_op[4:0] mux3a RAM A 128 x 24 mac_op[1:0] shift_op[2:0] Pass Mux A AHB Bus MAC mux3b mux1b B mux2b RAM B 128 x 24 mux1b dp_out A Hold Shift Round or Pass AHB Bus B Round Flag mux3b mux2b Round Mode – If DP is in Round mode, any result passing out of the DP unit is being rounded to a 16-bit value. This feature status is shown in the register setting, DFB_SR[2]. Saturation Mode – If DP is in Saturation mode, any mathematical operation that produces a number outside the range of a 24-bit 2’s complement number is clamped to the maximum positive or negative number. Enabling and disabling saturation and rounding is under the control of DFB controller. See the ALU instruction set. The status is visible at DFB_SR[1]. 27.3.2.1 MAC The multiply add function takes two 24-bit signed numbers and calculates a 48-bit signed result, then adds a signed 48bit value ((a*b)+c). The accumulator consists of a 48-bit register and the multiply adder. Together these two functions, along with some control logic, make up the MAC. Based on the opcode (mac_op) coming from the DFB controller it can do one of the following operation: ■ Multiply and accumulate with previous Values ■ Clear Accumulator and load with current product. ■ Hold accumulator, no multiply (no power in mult) ■ Add ALU value to product and start new accumulation The output of MAC is higher order 24 bits of multiply accumulate operation. The MAC assumes an implied binary point after the MSB, which shifts the result down a bit in the output of the MAC. For this reason, bits 46:23 are used 292 instead of 47:24. The instruction set for the MAC, ALU and Shifter is listed in Table 27-7 on page 299, Table 27-6 on page 299, and Table 27-8 on page 300. 27.3.2.2 ALU The ALU provides data control on the output end of the data path. ALU supports add, subtract comparison, threshold, absolute value, squelch, saturation, and other functions. See Table 27-6 for various instructions supported by ALU. The ALU commands as well as inputs are pipelined. This pipelining can be made use for data movement in some filtering applications. This pipelining causes a delay of two clock cycles for the ALU input to reach the output. 27.3.2.3 Shifter and Rounder The shifter at the ALU output can be used to shift the ALU results as required. See Table 27-8 for various shifter commands. Rounder rounds the results to a 16 bit value when the data path is operating in round mode. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital Filter Block (DFB) 27.3.3 Address Calculation Unit The Address Calculation Units (ACUs) generate addresses for each DP RAM. There are two address calculation unit for supporting sophisticated branching operations. The ACU is capable of saving and restoring address, incrementing or decrementing address by 1 or n (n is a constant value stored in FREG), flagging a programmable terminal count, and a number of other functions. REG – Stores the current value that the ACU is operating on and outputs it on every cycle, unless a command specifies otherwise. FREG – Loads with a value to increment or decrement by, when using the ADDF and SUBF commands. For example, load two into FREG and then it is possible to increment through the data RAMs by two. MREG – Stores the maximum value before a wraparound if modulo arithmetic is turned on. When the address calculated by the ACU exceeds MREG value, it will wraparound to LREG value, if modulo arithmetic is turned on. LREG – Stores the minimum value before a wraparound to the MREG value when modulus arithmetic is turned on. Modulus arithmetic is enabled using the SETMODE ACU command and disabled using the UNSETMOD command. Modulus arithmetic prevents the ACU from incrementing past the value of MREG and from decrementing below the value of LREG. Make sure the REG value is within the LREG:MREG range at the time modulus arithmetic is turned on to avoid unexpected results. The ACU (including the ACU RAM) is initialized whenever a hard reset event occurs or when the RUN bit in the DEC_CR register is ‘0’. Initialization is as follows: ACU RAM Contents=0, MREG=127, LREG=0, FREG=2. The current address and state of the register of both ACUs can be stored or retrieved from memory with assembly instructions. This is used in context switching. A 16x14 ACU RAM is used for this purpose. The 16x14 RAM is used by both ACUs. The upper seven bits are for ACU B and the lower seven bits are for ACU A. Thus, each ACU can store 16 addresses or state elements. The ACU instructions perform incrementing/decrementing of the data RAM addresses by one or the value in FREG. Apart from this, the modulus arithmetic is used to enable a wrap around at user defined limits. dence requires the same value on the ACU_addr for all commands involved. 27.3.4 Bus Interface and Register Descriptions The DFB block is wrapped with a 32-bit AHB-Lite Slave bus interface. A 32-bit bus was chosen to accommodate the fact that the RAMs in the DFB are all 24 bits and most of the bus transfers to the DFB are 24 bits. The DFB has a set of expanded Control and Status Registers (CSR) that are accessible through the system bus at all times. The registers containing CSR bit information are address mapped as 32-bit registers with active bits only in the low byte. This arrangement works well for both 8-bit and 32-bit MCUs. The CSRs that hold sample data are 24-bits wide (Staging and Hold register) and coherency interlocking HW is included to allow 8-bit and 16-bit accesses. In normal mode of operation, the DFB RAMs (except the input staging and output holding registers) is controlled by the DFB controller and is not accessible to CPU/DMA. If CPU/DMA needs control of this DFB RAM memory it should make use of DFB_RAM_DIR control bits (one per RAM) to give the RAM control to system bus. 27.3.4.1 Streaming Mode In streaming mode the filter coefficients and historic data are loaded into DFB before starting the DFB operations. Runtime data movement is through the staging and holding registers. The DFB has: ■ Two 24-bit input staging registers ■ Two 24-bit output holding registers These registers can be accessed by both DFB as well as AHB Bus (CPU/DMA). In reality, these registers are double buffered, but to the DFB controller and the system bus, they appear as single registers. In streaming mode data to be filtered is streamed in to staging registers. Filter output is streamed out through DFB holding registers. The two sets of input and output registers aid stereo data processing applications. Applications requiring more than two concurrent channels must use block mode. Note Apart from addressing the ACU RAM, the ACU_addr is also used as an argument for other ALU and branching commands. The single ACU_addr value can be used simultaneously for different commands (ACU, ALU...) if coinci- PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 293 Digital Filter Block (DFB) Figure 27-4. Streaming Mode Transfer Staging Reg A 128 X 24 RAM A Staging Reg B 128 X 24 RAM B Holding Reg A2 24 Bit Holding Reg A1 24 Bit Holding Reg B2 24 Bit Holding Reg B1 24 Bit A MAC, ALU, Shift, etc. B Input Select In input Streaming mode, the sample rate is determined by the ADC or other sampling resources providing the input samples. By definition the DFB must be running (processing) samples faster than or at the exact same rate as the sample source to function properly. Therefore, the DFB knows how to stall and wait for subsequent input data or postpone operation on that channel and switch to another channel (if in use). When the calculation engine is finished processing a sample, a bus read instruction can be issued. At this point, the next staged sample is read or, if not present yet, the DFB controller stalls while waiting for the next input sample. If two streaming channels are being processed, the DFB controller, upon completion of a calculation, can jump to the other channel. The full or empty status of the two Staging registers is visible to the DFB controller and it can branch based on the status information, allowing it control of which channel it is working on. When the bus read instruction is issued by the DFB controller, it does not request the bus, generate an interrupt, or DMA request. It simply tells the DFB bus interface that it wants the next sample and will wait until it arrives. In this state, the DFB controller waits until the bus interface signals that the sample has arrived. A one 24-bit word Staging register is used for a sample rate at or below 1 Msps and guaranteed bus latency lower than the sample period. There are two Staging registers: one for each supported channel. In streaming mode new samples arrive in the staging registers. The DFB controller checks for new data write to staging registers and branch to process data depending on the CFSM code. ters with the low order ACU RAM address bit (acu_addr[0]). If the address bit is low, Staging register A is read; if the address bit is high, B is read. When read, the associated Stage Valid signal is automatically cleared by the hardware. Apart from this, the Staging register also has a key coherence byte setting. This setting is available to reduce errors due to bus access being less wide as compared to the register width. The staging registers are protected on writes, so the underlying hardware does not incorrectly use the field when it is partially updated by the system software. If the system software is in the middle of reading from the holding registers, the DFB will not update the holding registers until the coherency key byte is read. The Key Coherency byte is basically the user (software) telling the hardware which byte of the field is written or read last when an update to the field is desired. In the Staging register the new value availability is flagged only when the key coherency byte is written to. 27.3.4.2 Block Transfer Modes Block mode is defined by the system software moving sets of samples or coefficient data in and out of the DFB data RAMs in blocks. This method of using the DFB supports such features as multi-channel processing and deeper filters than the embedded data RAMs will support. It can also be used to initialize the DFB RAMs for streaming mode operation. The DFB datapath block has two 128x24 embedded data RAMs. These hold the data (signal or coefficients) used in the calculation of numerical processes. These two RAMs are completely separate memories from the bus’ point of view. The DFB views these two RAMs as a working set, as shown in Figure 27-5 on page 295. The input staging registers are read by the DFB controller by asserting the bus read signal and addressing the two regis- 294 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital Filter Block (DFB) Figure 27-5. Block Mode Transfer 128 X 24 RAM A A 128 X 24 RAM B B Bus Control Logic MAC ALU Shift etc RAM Control from CSR The primary concept of Block mode is to allow the system software full control of what is in the data RAM for each calculation cycle of the DFB. In general, this extends the functionality of the DFB by trading performance for fundamental features such as the ability to implement filters with more taps than 128 or to time division multiplex the processing of more than two low sample rate channels. The system software burden of Block mode is in the management of the RAM’s contents. Both system and DFB performance is lost due to software servicing of the DFB and because the DFB must stall while the system software reads/writes the data RAMs. Block mode also creates more bus traffic on the system bus for a given sample rate. Typically, results of DFB applications are streaming in nature. However, in cases where results are created as data sets, Block mode can be used to move the resultant data sets out of the DFB data RAMs. 27.3.4.3 Result Handling Frequently DFB block output results are generated at periodic intervals after a series of mathematical calculations. This also happens after a wait for the input sample stream. The generation rate of these result elements will vary radically based on the function being programmed and run on the DFB. The system software takes control of memory by putting it on the system bus with the use of (DFB_RAM_DIR) control bits (one per RAM). It then reads/writes the data and “passes” the memory back to the DFB by toggling the control bit back. While this is happening, the DFB must stall, unless it is performing some function that only requires one of the two data RAMs. The two data RAMs are individually controlled by the system software as to which resource has control of them – the bus or the DFB. To assist system software with the handling of resultant data, the DFB implements two Holding registers, 24 bits wide, for output results. In reality, these two Holding registers are double buffered, but to the DFB controller and the system bus, they appear as a single register. They are referred to as a single register hereafter, but keep in mind there are really two registers to deal with bus latency issues. The fact they are double buffered is transparent to both the bus and the DFB controller. Hardware automatically manages the fact that they are double buffered. Any number of data channels can be supported with Block mode (within reason). With each added data channel, the system software has the additional burden of tracking and managing and sample rates supported reduces considerably because the DFB must be stalled for data movement operations. The intent of having two fully addressed Holding registers is primarily to allow the controller and system software to map filter channels so that DMA requests are much easier to support. The two Holding registers are addressed with the low-order ACU address bit out of the Control Store. The DFB controller provides a semaphore methodology to communicate with the system software as to the status of the data RAMs when being passed back and forth for block transfers. Optional interrupt support can be associated with the setting and clearing of semaphores. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F When bus write is asserted in the CS word and the loworder ACU address bit (acu_addr[0]) is low, Holding register A is written and Holding register B is written when the loworder address bit is high. There are a couple of methods provided to read the Holding registers on the system bus. These registers are generic 295 Digital Filter Block (DFB) read only CSRs. They can be read manually by software running on the MCU under poled or interrupt control (DFB_INTR_CTRL), or each can be associated with a DMA request signal and read by the system DMA controller (DFB_DMA_CTRL). Pending interrupts from the Holding register update is monitored from the DFB_SR register. Operations on the Holding registers are protected. The nature of the protection is set by the coherence bits (DFB_COHER). The Holding registers are protected on reads so that the underlying HW does not update it when partially read by the System SW or DMA. The key coherency byte is selected in the Coherency register. The Key Coherency byte is basically the user (software) telling the hardware which byte of the field is read last. The Holding registers are considered read when the key coherency byte is read. Note 1 In Block mode, when more than two channels are being processed, management of the output results is more burdensome to the system software as it can no longer be constantly mapped one-to-one with a Holding register or DMA request. Note 2 In 8-bit devices, reading the Holding registers manually results in a multi-cycle operation. Figure 27-6 explains DFB control signals can be used for data streaming and result handling. Figure 27-6. Control Signals for Data Streaming and result handling Dfb_intr INTR_CTRL[0] Stage A Valid, in1 Dfb_dmareq1 Stage A Coherency Key DFB_COHER[1:0] DMA_CTRL[1:0] Data Write Strobe Stage A LOW [8 Bits] Stage A MED [8 Bits] Stage A HIGH [8 Bits] Stage B LOW [8 Bits] Stage B MED [8 Bits] Stage B HIGH [8 Bits] 128 X 24 RAM A Hold A A MAC, ALU, Shift, etc. DMA CPU 128 X 24 RAM B DMA CPU B Hold B Data Write Strobe Stage B Coherency Key DFB_COHER[3:2] Input Select Dfb_intr INTR_CTRL[1] Stage B Valid, in2 Dfb_dmareq2 DMA_CTRL[3:2] 296 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital Filter Block (DFB) 27.3.4.4 Data Alignment The hardware provides a data alignment feature in the input Staging registers and in the output Holding registers for system software convenience. Both staging and holding registers support byte accesses that addresses alignment issues for input and output samples of 8 bits or less. Also, all four of these registers are mapped as 32-bit registers (only three of the four bytes are used) so there are no alignment issues for samples between 17 and 24 bits. However, for sample sizes between 9 and 16, it is convenient to read and write these samples on bus bits 15:0, while they source and sink on bits 23:8 of the Holding and Staging registers. The CSR DALIGN provides bits that enable an alignment feature which allows bus bits 15:0 to either be sourced from Holding register bits 23:8 or sink to Staging register bits 23:8. Each Staging and Holding register can be configured individually with a bit in the DALIGN register. If the bit is set high, the effective byte shift occurs. For example, if an output sample from the Decimator is 12 bits wide, aligned to bit 23 of the Decimator Output Sample register, and is desired to stream this value to the DFB, the similar data alignment feature of the Decimator can be enabled, allowing the 16 bits of the Decimator Output Sample register to be read on bus bits 15:0. Setting the alignment feature in the DFB for the Staging A input register, these 16 bits can be written on bus bits 15:0 and will be written into bits 23:8 of the Staging A register when required. 27.3.4.5 DMA and Semaphores To set and clear semaphores bits, two DP ALU commands are available: SEM_SET and SEM_CLR. For each active high bit of the ACU address, the corresponding semaphore bit is either set or cleared. For system software to write into a semaphore bit the register DFB_SEMA is used. The mask bit is set when the corresponding semaphore bit in the register is updated. Any of the semaphore bits can be optionally (programmable) associated with the system interrupt signal (DFB_INTR_CTRL) or either of the DMAREQ (DFB_DMA_CTRL) outputs leaving the DFB, and/or either of the outgoing Global signal. Pending semaphore interrupts are monitored from the DFB_SR register. 27.3.4.6 DSI Routed Inputs and Outputs The DFB has the option to take two DSI global inputs (globali1 and globali2) and two DSI global outputs (globalo1 and globalo2). Use of the global outputs is optional. If needed, they can be programmed to carry one of four different DFB internal status/control signals. These can be routed to the DSI and used as inputs to other circuits. The global outputs can be configured to carry semaphore, an interrupt, or DP status signals as listed in Table 27-5 on page 298. This is done using the DFB_DSI_CTRL register. The DSI inputs into the DFB to control operations of the FSM are optionally used as branching inputs to the Controller's next state decoder. See the section on 27.3.1.4 Next State Decoder on page 290 for more details. The DFB bus interface supports two DMA request signals. These can be associated with the two Holding registers (optional) or associated with the semaphore bits (see register DFB_DMA_CTRL). The DFB provides three generic semaphore register bits that the system software and the DFB controller can use to communicate. The intent of these three semaphores is to allow the system software and the DFB controller to communicate the status of data movement in and out of the DFB and, in particular, the handling of block data transfers. The definition of these three bits is left to the system and controller software architects. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 297 Digital Filter Block (DFB) 27.4 DFB Instruction Set Each control word for the DFB is 32 bits long. The fields in the control word are as follows: ■ Datapath Mux Control – 6 bits ■ Data RAM R/W – 2 bits ■ Bus R/W – 1 bit ■ ALU Control – 5 bits ■ MAC Control – 2 bits ■ Shifter Control – 3 bits ■ ACU Control – 8 bits ■ ACURAM Address – 4 bits ■ End of Code Block – 1 bit The mux control bits are split equally between the A and B paths each having 3 bits. Three bits are allocated and encode the control of the mux1, mux2, and mux3 functions as shown in Table 27-5. Table 27-5. Mux Functions Assembly Name Code 298 Function Function Function MUX1 MUX2 MUX3 0 BA mux1 = AHB Bus mux2 = mux1 mux3 = mux2 1 SA mux1 = dp_out mux2 = mux1 mux3 = mux2 Function AHB ->ALU dp_out->ALU AHB->RAM 2 BRA mux1 = AHB Bus mux2 = RAM out mux3 = mux2 3 SRA mux1 = dp_out mux2 = RAM out mux3 = mux2 4 BM mux1 = AHB Bus mux2 = mux1 mux3 = MAC AHB->MAC->ALU 5 SM mux1 = dp_out mux2 = mux1 mux3 = MAC dp_out->MAC->ALU 6 BRM mux1 = AHB Bus mux2 = RAM out mux3 = MAC 7 SRM mux1 = dp_out mux2 = RAM out mux3 = MAC AHB->ALU dp_out->RAM dp_out ->ALU AHB->MAC->ALU AHB->RAM dp_out->MAC->ALU dp_out->RAM PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital Filter Block (DFB) ALU functions are programmed as shown in Table 27-6 and are encoded in 5 bits. Table 27-6. ALU Functions Code Assembly Name Function 0 SET0 1 SET1 Set ALU output to 1 2 SETA PASS A to ALU output 3 SETB PASS B to ALU output 4 NEGA Set ALU output to –A 5 NEGB Set ALU output to –B 6 PASSRAMA Pass RAM A output directly to ALU output 7 PASSRAMB Pass RAM B output directly to ALU output 8 ADD Add A and B and put result on the ALU output 9 TDECA Put A-1 on the ALU output, set threshold detection 10 SUBA Put B-A on the ALU output 11 SUBB Put A-B on the ALU output 12 ABSA Put |A| on the ALU output 13 ABSB Put |B| on the ALU output 14 ADDABSA Put |A| + B on the ALU output 15 ADDABSB Put A + |B| on the ALU output 16 HOLD Hold ALU output from previous cycle 17 ENGLOBALS, - Enables global and saturation jump conditions using a 3-bit field to specify which events are active jump conditions 17 ENSATRND, - Writes to the saturation and rounding enable register using a 3-bit field to enable and disable them 18 ENSEM, --- Enables semaphores as jump conditions using a 3-bit field to specify which are active 19 SETSEM, --- Set the semaphores high using the 3-bit mask 20 CLEARSEM, --- Set the semaphores low using mask, addr[2:0] 21 TSUBA Put B-A on the ALU output, set threshold detection 22 TSUBB Put A-B on the ALU output, set threshold detection 23 TADDABSA Put |A| + B on the ALU output, set threshold detection 24 TADDABSB Put A + |B| on the ALU output, set threshold detection 25 SQLCMP Load squelch comparison register with a value from side A, Pass Side B 26 SQLCNT Load squelch count register with value from side A, Pass Side B 27 SQA Squelch side A: If value is above threshold pass it. If value is below threshold and the squelch count register is zero, pass. zero. Otherwise pass A 28 SQB Squelch side B: If value is above threshold pass it. If value is below threshold and the squelch count register is zero, pass. zero. Otherwise pass B UNDEFINED Undefined Opcodes 29-31 Set ALU output to 0 MAC functions are programmed as shown in Table 27-7 and are encoded in 2 bits. Table 27-7. MAC Functions Code 0 Assembly Name LOADALU Function Add ALU value to product and start new accumulation 1 CLRA Load accumulator with product but a 0 sum 2 HOLD Hold accumulator, no multiply (no power in mult) 3 MACC Default – just accumulate PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 299 Digital Filter Block (DFB) Shifter functions are programmed as shown in Table 27-8 and are encoded in 3 bits. If deeper shifts are required, data can be passed through the ALU on multiple cycles. Table 27-8. Shifter Functions Code Assembly Name Function 0No shift 1 shift(right,1) Shift right 1 (divide by 2) 2 shift(right,2) Shift right 2 (divide by 4) 3 shift(right,3) Shift right 3 4 shift(right,4) Shift right 4 5 shift(right,8) Shift right 8 6 shift(left,1) Shift left 1 (multiply by 2) 7 shift(left,2) Shift left 2 (multiply by 4) Two ACUs are supplied. There are 16 functions per ACU as shown in Table 27-9 and are encoded in 4 bits. This RAM is useful when parallel filters or algorithms are implemented and control flow needs to shift from one to the other, while still maintaining the relative addresses for each filter. Table 27-9. ACU Functions Code Assembly Name 0 HOLD 1 INCR Function Put REG on output, hold REG in REG If (modflag && REG = MREG Put LREG on output, write to REG else If (!modflag && REG = 127) Put 0 on output, write to REG else Put REG+1 on the output, write to REG If (modflag && REG = LREG) Put MREG on output, write to REG 2 DECR else If (!modflag && REG = 0) Put 127 on output, write to REG else Put REG-1 on the output, write to REG Read ACU RAM and put value on output 3 READ 4 WRITE 5 LOADF Load FREG from ACU RAM, put REG on output 6 LOADL Load LREG from ACU RAM, put REG on output 7 LOADM 8 WRITEL 9 SETMOD 10 UNSETMOD Write to REG Put REG on output, write output to RAM Hold REG in REG Load MREG from ACU RAM, put REG on output Put LREG on output, assert RAM write enable Hold REG in REG Set modflag true, put REG on output Hold REG in REG Set modflag false, put REG on output Hold REG in REG If (modflag) 11 CLEAR Put LREG on output, write to REG else Put 0 on output, write to REG 300 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital Filter Block (DFB) Table 27-9. ACU Functions (continued) Code 12 Assembly Name ADDF Function If (modflag && REG+FREG>MREG) Put ((((REG+FREG)-MREG)-1)+LREG) on output else If (!modflag && REG+FREG>127) Put (((REG+FREG)-127)-1) on output else Put REG+FREG on output, write to REG If (modflag && REG-FREG 1. 25 fF – 5 pF Vout V out V ref TIA : 850 fF Mod Gain 0 : 425 fF Mod Gain 1 : 850 fF Vref Vgnd V ref Vin CT, TIA : 850 fF SC: 850 fF Trk Hld: 12.0 pF 306 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Switched Capacitor/Continuous Time 28.3 V I load = C load ------t How it Works Each instance of the SC/CT block is able to implement any of the available configurations. Selection of the mode bits configures most of the resources required to implement these configurations. 28.3.1 Operational Mode of Block is Set The operational mode of the SC/CT block is selected by setting the MODE[2:0] bits in the SC[0..3]_CR0 register, bits [3:1]. Table 28-1. SC/CT Block Operational Mode Settings SC_MODE[2:0] Operational Mode [000] Naked Opamp Mode [001] Trans Impedance Amplifier [010] Continuous Time Mixer [011] Discrete Time Mixer -- NRZ S/H [100] Unity Gain Buffer [101] First-Order Modulator [110] Programmable Gain Amplifier [111] Track and Hold Amplifier Equation 1 where Cload includes the total internal capacitance at the output node of the amplifier plus any external capacitive loads. A value of 10 pF should be used for the internal load from analog bus routing. Set the drive controls, SC_DRIVE[1:0], according to the slew requirements at the output in SC[0..3]_CR1[1:0] register bits. Table 28-2. Output Load Current by Drive Setting SC_DRIVE[1:0] I_load (µA) 2'b00 175 2'b01 250 2'b10 330 2'b11 400 Figure 28-3. Naked Opamp Drive Control I_LOAD 90 80 70 60 50 40 dB 30 28.4 20 Naked Opamp 175 uA 10 The naked opamp mode provides direct access to the input and output terminals of the opamp. All of the other circuitry (resistors and capacitors) is disconnected in this mode. This mode is used for applications that require a general purpose opamp with external components. Figure 28-2. Naked Opamp Configuration V IN + V OUT V IN - The naked opamp is selected by setting the MODE[2:0] bits in the SC[0...3]_CR0 to 000. The opamp is a two stage design with a rail-to-rail input folded cascade first stage and a class A second stage. The opamp is internally compensated. To accommodate varying load conditions, the compensation capacitor and output stage drive strength is programmable. The setting to apply is determined from the minimum required slew rate determined from the signal swing and time, and load capacitance. This is primarily a consideration for the stability reasons. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 0 -10 250 uA 330 uA 400 uA -20 0.001 28.4.1 0.01 0.1 1 10 100 1000 10000 Bandwidth/Stability Control This block has three control options for modifying closed loop bandwidth and stability that apply to all configurations: current through the first stage of the amplifier (BIAS_CONTROL), Miller capacitance between the amplifier input and the output stage (SC_COMP[1:0]), and feedback capacitance between the output stage and the negative input terminal (SC_REDC[1:0]). 28.4.1.1 BIAS_CONTROL The bias control option doubles the current through the amplifier stage. AC open loop stability analysis for all continuous time modes shows that leaving this option set to ‘1’ and then controlling the bandwidth/stability using the capacitor options results in a greater overall bandwidth when the circuit is stabilized than using the option of less current in the first stage. The bias current is doubled by setting the SC[0..3]_CR2[0] register bit. 307 Switched Capacitor/Continuous Time 28.4.1.2 SC_COMP[1:0] SC_COMP bits set the amount of compensation capacitance used in the amplifier. This directly affects the gain bandwidth of the amplifier and is an important tool in tuning the circuit stability. Follow the recommendations in the upcoming tables for this setting. The Miller capacitance is set to one of the four values in the SC[0..3]_CR1[3:2] register bits. Table 28-3. Miller Capacitance between Amplifier Output and Output Driver SC_COMP[1:0] 1.30 01 2.60 10 3.90 11 5.20 28.4.1.3 CFB (pF) SC_REDC[1:0] 00 0.00 01 1.30 10 0.85 11 2.15 Recommended Settings by Mode CMiller (pF) 00 Table 28-4. CFB in CT Mix, PGA, Opamp, Unity Gain Buffer, and T/H Modes Stability settings for each mode are listed in Table 28-5 on page 308. These are the settings used to simulate each mode. For the transimpedance amplifier (TIA) mode, the analog global load was modeled at the input as 10 pF between two 150- switch impedances with an additional 40 pF added to the input to model the input diode capacitance. SC_REDC[1:0] The capacitance option between the output driver and the negative input terminal is another stability control option. Depending on the continuous time configuration, this capacitor option generally contributes to a higher frequency zero and a lower frequency pole, thus reducing the overall bandwidth and gaining some phase margin at the unity gain frequency. This capacitance is set to one of the four values in SC[0..3]_CR2[3:2] register bits. For all continuous time modes, the output is modeled with two 150 switches with an 8-pF load in between, then followed b ya 300 impedance and a 50-pF external load. The modulator mode is simulated with a 0.5-pF load at the output. Table 28-5. Recommended Stability Settings by Mode SC_MODE[2:0] Operational Mode BIAS_CONTROL SC_COMP[1:0] SC_REDC[1:0] [001] Trans Impedance Amplifier 1 3 3 [010] Continuous Time Mixer 1 2 1 [011] Discrete Time Mixer -- NRZ S/H 1 2 0 [100] Unity Gain Buffer 1 2 0 [101] First-Order Modulator 1 1 0 [110] Programmable Gain Amplifier [111] Track and Hold Amplifier 308 See Table 28-7 on page 310 1 2 0 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Switched Capacitor/Continuous Time 28.5 Continuous Time Unity Gain Buffer The continuous time unity gain buffer is a naked opamp with the inverting input locally connected to the output. Use of routing features external to the block is not required to implement this function. Figure 28-4. Unity Gain Buffer Configuration V IN V OUT The unity gain buffer is used when an internally generated signal with high output impedance, such as a voltage DAC output, is required to drive a load; or when an external source with a high impedance is required to drive a significant on-chip load, such as the Continuous Time Mixer. 28.6 Continuous Time Programmable Gain Amplifier The programmable gain amplifier (PGA) is a continuous time opamp with selectable taps for input and feedback resistances. The PGA is selected by setting the MODE[2:0] bits in the SC[0...3]_CR0 register to ‘110’. Figure 28-5. PGA Configuration Rfb = 20 k to 1 Mohm rval[000]:rval[101] R in =20k or 40k rval< 110 > : rval< 111> V in 0 R in = 9.6k or 19.6k 1 V out 1 V ref 0 0.3k pga _ rlad sc_pga_ gndVref sc _ gain The PGA can be implemented as either a positive gain or negative gain topology, or as half of a differential amplifier. The specific gain configuration is selected by the SC_GAIN bit [5] in register SCL[0..3]_CR1. Any added input resistance from analog routing affects the PGA gain. Direct voltage reference (vref) connections to the PGA inverting input are not recommended. The low input impedance of the PGAs gain resistor ladder creates an error to the higher output impedance of vref if the two are directly connected. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Table 28-6. PGA Gain Configuration SC_GAIN Gain 0 Inverting (-RFB/RIN) 1 Non-inverting (1+ RFB/RIN) The positive gain (non-inverting) topology is shown in Figure 28-6. 309 Switched Capacitor/Continuous Time Figure 28-6. PGA Positive Gain (Noninverting) Topology RLAD is at very high impedance to minimize gain errors. The output of the differential amplifier is VIN VOUT+ - VOUT- = Gain*(VIN+ - VIN-). VOUT The common mode voltage of the output remains at the common mode voltage of the input. VCM = (VIN+ + VIN-)/2. VREF RIN Equation 2 Equation 3 Because of capacitive loading, each gain step has a different requirement for compensation capacitors. RFB Table 28-7. PGA Stability Settings by Gain Figure 28-7. PGA Negative Gain (Inverting) Topology SC_RVAL [2:0] R20_40B Bin Bin RFB VIN RIN VREF Figure 28-8. PGA Differential Amplifier Topology VIN+ VOUT+ RIN RLAD RLAD RIN RFB NonInverting BIAS_ SC_COMP SC_REDC Gain (AC) CONTROL [1:0] [1:0] Lin 0 0 1 1 2 0 0 1 1 1 2 0 1 0 2 1 2 1 1 1 2 1 2 1 10 0 4 1 0 1 10 1 4 1 0 1 11 0 8 1 0 1 11 1 8 1 0 1 100 0 16 1 1 1 100 1 16 1 1 1 101 0 24 1 1 3 101 1 32 1 1 3 110 0 24 1 0 2 110 1 48 1 1 0 111 0 25 1 0 2 111 1 50 1 1 0 The negative gain (inverting) topology is shown in Figure 28-7. 28.7 RFB VOUTVIN- Continuous Time Transimpedance Amplifier The transimpedance amplifier (TIA) is a continuous time opamp with dedicated and selectable feedback resistor. The TIA is selected by setting the MODE[2:0] bits in the SC[0..3]_CR0 register to ‘001’. The differential amplifier two PGAs in parallel. The connection (RLAD) is external to the SC blocks and has very low impedance to reduce gain error. When not in differential mode, RIN is connected to the analog or global routing and 310 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Switched Capacitor/Continuous Time C FB The CFB options for TIA mode are larger than for the other continuous time modes, as shown in Table 28-9. The feedback capacitance is set in bits [3:2] of the SCL[0..3]_CR2 register. R FB Table 28-9. Feedback Capacitance Settings Figure 28-9. Transimpedance Amplifier Configuration CFB (pF) SC_REDC[1:0] V IN V REF The output of the transimpedance amplifier is a voltage that is proportional to input current; the conversion gain is a resistor value, where: V OUT = V REF – I IN R FB 00 0.00 01 1.30 10 3.30 11 4.60 A large source capacitance causes instability in the TIA with the small feedback resistor settings. Therefore, in applications where the internal capacitance is not sufficient to stabilize the TIA, an external capacitance is necessary. This is connected using the analog global routing. Equation 4 The output voltage is referenced to VREF, which is routable to the analog globals or through local analog routing to any selected reference. The feedback resistor can be programmed from 20 k to 1.0 M in eight steps, selected in bits [6:4] of the SCL[0..3]_CR2 register. Table 28-8. Feedback Resistor Settings Nominal RFB (k) SC_RVAL[2:0] 000 20 001 30 010 40 011 80 100 120 101 250 110 500 111 1000 The feedback resistor is untrimmed polysilicon, so the absolute resistance value varies largely with process and temperature. Calibration of the TIA gain is expected to be done by the user using the precise outputs of the current output DAC combined with measurements in the ADC. Stability of this opamp topology in general is affected by shunt capacitance on the inverting input. This capacitance is determined largely by parasitic capacitances in the analog global routing and at the input pin. An internal shunt feedback capacitor is used to maintain stability. Because the input capacitance is larger in the TIA than in other modes, the stability capacitance is somewhat larger. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 311 Switched Capacitor/Continuous Time 28.8 Continuous Time Mixer Figure 28-10. Continuous Time Mixer Waveforms The continuous time mixer uses input switches to toggle a PGA between an inverting PGA gain of –1 and a noninverting PGA gain of +1. The maximum toggle frequency is 1 MHz. The continuous time mixer is selected by setting the MODE[2:0] bits in the SC[0..3]_CR0 register to ‘010’. The continuous time mode is chosen to achieve up conversion because it provides higher conversion gain relative to the sampled mixer. In the CT mixer, the magnitude of the FCLK + FIN and FCLK – FIN are equal, while in the sampled case, there is attenuation between the two configurations. Example waveforms where the input is at 200 kHz and the carrier is at 255 kHz, are shown in Figure 28-10. Sign al Ca rr ier Mult 0 10 20 30 40 The output spectrum of the mixer includes terms at 455 kHz, 55 kHz, at 3 × fCARRIER ± fSIGNAL, 5 × fCARRIER ± fSIGNAL, 7 × fCARRIER ± fSIGNAL, and so on. The up conversion is ultimately achieved by filtering out the desired harmonic of the mixed product of the input frequency and modulating frequency using gain toggling. Usage options for the continuous time mixer mode include controlling the sampling function and setting the value of the resistor in the inverting gain configuration. Figure 28-11 shows the continuous time mixer configuration. Figure 28-11. Continuous Time Mixer Configuration R mix = 20 k or 40k sc_clock V in !sc_clock Rmix = 20 k or40k V out 1 0 V ref sc_clock Table 28-10. Sampling Configurations for CT Mixer SC_DYN_CNTRL Configuration 0 Inverting Amplifier with Gain of 1 1 Unity Gain Buffer Table 28-11. Input Resistor Settings for CT Mixer Inverting Mode RMIX R20_40B 0 40 k 1 20 k 312 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Switched Capacitor/Continuous Time 28.9 Sampled Mixer The sampled mixer is a nonreturn-to-zero (NRZ) sample and hold circuit with very fast response. The mixer is selected by setting the MODE[2:0] bits in the SC[0..3]_CR0 register to ‘011’. The discrete time mode has a maximum FCLK of 4 MHz. The maximum input frequency in discrete time mode is 14 MHz. The mixer output is designed to either drive an off-chip ceramic filter (455 kHz Murata Cerafil) or the internal ADC through the on-chip analog routing. For the ADC to correctly sample the mixer output, the sample clock for the ADC and mixer must be the same. This example has a 500 kHz down-converted signal, but we are sampling it at 2 MHz. Because the ADC and the switched capacitor block can both run at the same 2 MHz sample rate, there is no need to low-pass filter the output of the switched capacitor block. Its output can be fed directly into the ADC input. A few examples illustrate the frequency shifting capabilities of the mixer. For a signal frequency at 1.36 MHz, and a carrier at 1.28 MHz, the output frequency is the difference between the two frequencies, as shown in Figure 28-12. Figure 28-12. Sampled Mixer N = 1 The sample and hold mixer is primarily used for down-conversion mixing. The down conversion is achieved by filtering the desired harmonics of the mixed product of the input frequency and sample clock frequency. Correct frequency planning is required to achieve the desired results. For a given input carrier frequency, FIN, a sample clock frequency, FCLK, can be chosen to provide the desired IF frequency, FIF, for the system. Signal Provided that FCLK is less than 4 MHz, and FIN is less than 14 MHz: Carrier Dif f 0 If 2N – 1--------------F CLK F IN N F CLK 2 Equation 5 F IF = N F CLK – F IN Equation 6 2N + 1 N F CLK F IN ---------------- F CLK 2 Equation 7 F IF = F IN – N F CLK Equation 8 5 10 15 20 25 30 35 For a higher frequency signal at 13.6 MHz, and the carrier at 3.2 MHz, the output is at the same frequency, but longer separation between the samples, as shown in Figure 28-13. then Figure 28-13. Sampled Mixer N = 3 If then Equation 1 and Equation 2 can be summarized as: F IF = abs N F CLK – F IN Carrier Equation 9 Consider an example using an input carrier frequency of 13.5 MHz and a desired IF frequency of 500 kHz. Set the sample clock frequency and ADC sample frequency to be 2 MHz. From the down conversion equations above, calculate the IF frequency with N = 7. F IF = 7 F CLK – F IN = 500kHz Signal Equation 10 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Dif f 0 0.5 1 1.5 2 2.5 3 3.5 There is no increase in harmonic distortion, only an increase in the level of the sampling aliases. When the mixer output is sampled at the same rate as the carrier frequency, the aliases are suppressed. The discrete time mixer configuration (NRZ S+H) is shown in Figure 28-14 on page 314. The options specific to this 313 Switched Capacitor/Continuous Time configuration are the reference option and the clock division option. Figure 28-14. Switched Capacitor Discrete Time Mixer Configuration C1 1 Vin The architecture of the discrete mixer is such that the output changes with a new hold value on both the rising and falling edge of the input clock. The SC_DIV control signal can be used to designate that output only change on the rising edge of the input clock. This is achieved by resetting the SC[0..3]_CR1[4] bit. 2 1&!sc_gndVref The use of the internal ground can cause different step sizes up versus down because the amplifier does not respond identically when the negative terminal jumps below ground. To avoid this distortion, use the external reference option and set it to 500 mV or greater. Vref 1&sc_gndVref 2 1 V out sc_gndVref Table 28-13. Clock Division Option for Sample and Hold Mixer SC_DIV !sc_gndVref Vref 2&!sc_gndVref 0 1 SC_CLOCK should be set to the desired sample frequency Vref 2&sc_gndVref 1 C4 SC_CLOCK Requirements SC_CLOCK should be set to half the desired sample frequency Vin 2 The option exists to either use an external reference voltage or to have the reference grounded internally. This option is controlled by the SC_GNDVREF SC[0..3]_CR2 signal as described in Table 28-12. Table 28-12. External Reference Option for Sample and Hold Mixer SC_GNDVREF Amplifier/Capacitor Reference 0 External Voltage 1 Internal Ground 314 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Switched Capacitor/Continuous Time 28.10 Delta Sigma Modulator The SC/CT block can be programmed to function as a switched capacitor integrator to use in a first-order modulator loop at high oversampling ratios. The Delta Sigma Modulator is selected by setting the MODE[2:0] bits in the SC[0..3]_CR0 register to ‘101’. The integrator output is compared to a reference level and fed back to the input in a feedback loop. The modulator output is clocked at the high sampling rate, and needs to be decimated down to the signal band of interest using a decimation filter. Figure 28-15. Discrete Time Delta Sigma Modulator Block Diagram + Out Sample / Hold Integrator Comparator Input - The modulator can also be used as an incremental modulator by using a reset switch that is placed across the integrating capacitor. The accuracy of the sampled data from the first-order modulator is determined from several factors: the maximum input signal bandwidth, oversampling ratio, and the sampling clock jitter. The oversampling clock is limited to a maximum of 4 MHz. Oversampling below x64 does not produce a stable output. Table 28-14 below shows the expected performance from a system simulation. Table 28-14. Incremental Modulator Expected Performance from System Simulation Oversampling Rate OSR (fsamp/fsig/2) Maximum Input Signal Frequency Sampling Clock Frequency (MHz) Signal-to-Noise Ratio After Decimation by OSR (at Maximum Input Signal) 16 kHz 64 2.048 54 dB 8 kHz 128 2.048 64 dB 32 kHz 64 4.096 54 dB 16 kHz 128 4.096 64 dB The signal-to-noise ratio (SNR) values include the effects of limit cycle oscillations. The configuration diagram of the discrete time first-order modulator is shown in Figure 28-16 on page 316. There are two mode-specific usage options: a reset switch placed across the integrating capacitor and a gain setting to adjust the allowable input amplitude range. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 315 Switched Capacitor/Continuous Time Figure 28-16. Switched Capacitor First-Order Modulator Configuration Sampling Phase C 5 = 1.7pF C 4 = 850fF Gain 0: C 1 = 850F Gain 1: C 1 = 425F C 2 = 850fF Vin sc_dyn_cntrl Vref V out Vout_mod = 0 V ref Vout_mod = 1 Gain 0: C3 = 425fF Gain 1: C3 = 850fF Comparator V ref Vout _mod Integrating Phase C 5 = 1.7pF C 4 = 850fF Gain 0: C1 = 850F Gain 1: C1 = 425F Gain 0: C3 = 425fF Gain 1: C3 = 850fF V ref C 2 = 850fF sc_dyn_cntrl V out Vout_mod = 0 Vout_mod = 1 V ref V re f Comparator 28.10.1 V ref Vout _mod First-Order Modulator, Incremental Mode The dynamic control input SC[0..3]_CR1[5] can be used to reset the integrating capacitor if to perform an incremental conversion: Table 28-15. First-Order Modulator, Integrating/Incremental Mode SC_DYN_CNTRL State 0 Integrating 1 Reset. VOUT is connected to amplifier negative terminal. The range of the allowed input amplitude can be set using the SC_GAIN SC[0..3]_CR1[5] control signal as shown in Table 28-16. Table 28-16. First-Order Modulator, Input Amplitude SC_GAIN Maximum Input Amplitude 0 ± half VREF 1 ± 2 VREF 316 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Switched Capacitor/Continuous Time 28.11 Track and Hold Amplifier Track and hold amplifier mode is derived using the unity gain buffer amplifier. Implementation is shown in Figure 28-17. Figure 28-17. Track and Hold Block Diagram VOUT !sc_dyn_cntrl VIN Ctrk_hld = 12.0 pF (pfet gate cap) Track and hold mode tracks to 1% of a 5.5 V input step in less than 1 µs. The charge injection error from the sample switch is < 1.1 mV. The hold loss is < 0.2 mV. The control of the amplifier between track and hold is done using the SC_DYN_CNTRL input as shown in Table 28-17. This feature is enabled by setting the register bit value SC[0..3]_CLK[5]. Table 28-17. Track and Hold Amplifier Control SC_DYN_CNTRL Output 0 Track VIN 1 Hold sampled value PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 317 Switched Capacitor/Continuous Time 318 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 29. Analog Routing PSoC® 5LP has a flexible analog routing architecture to route signals between GPIOs and analog resource blocks such as the ADC, switched capacitor, and DAC. One of the strong points of this flexible routing architecture is that it allows dynamic configuration of input/output connections to the different analog blocks. For example, the comparator input can be switched between two GPIOs, on the fly, by DSI control signals and register settings. Knowing and understanding the architecture enables efficient and optimal utilization of the device analog routing resources. 29.1 Features PSoC analog routing has the following features: ■ Flexible, configurable analog routing architecture ■ Dedicated routing options for LCD drive capability ■ Eight analog globals (AGs) and one analog multiplexer bus (AMUXBUS) for GPIOs on each side ■ Flexible routing options within the analog core to interconnect analog resource blocks using analog local bus (abus) 29.2 Block Diagram The PSoC 5LP analog system block diagram is shown in Figure 29-1 on page 320. In Figure 29-1, the CapSense® system is limited to the GPIO controls, there are no separate blocks. Figure 29-2 on page 321 shows detailed analog routing architecture. All the figures used to explain analog routing in this chapter are derived from Figure 29-2. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 319 Analog Routing Figure 29-1. Analog System Block Diagram ANALOG SYSTEM 5 LCD 8 Left Side ARBs Right Side ARBs Delta Sigma Channel (1x) LCD LPF LPF Switched Capacitor (2x) Switched Capacitor (2x) DAC (2x) DAC (2x) Opamp (2x) Opamp (2x) Comparator (2x) Comparator (2x) CapSense Refbuf CapSense Refbuf SAR0 SAR1 CY8C55 only 4 Analog Analog Analog Mux Global Local Bus Bus Bus Precision Reference Right Side Analog Globals and Muxes Right Side GPIO Left Side GPIO Left Side Analog Globals and Muxes 4 8 Analog Analog Analog Local Global Mux Bus Bus Bus 5 LCD Analog Interface 320 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Analog Routing Figure 29-2. Analog Interconnect Vssd Vcca * Vssa Vdda * * * ExVrefL2 opamp2 swinp 01 2 34 56 7 0123 * opamp1 swfol swfol GPIO P3[5] GPIO swinp P3[4] GPIO swinn P3[3] GPIO P3[2] GPIO P3[1] GPIO P3[0] GPXT *P15[1] GPXT *P15[0] swinn swfol swfol opamp3 3210 76543210 swinn * + - comp2 refsel[1:0] sc0 Vin Vref out vssa sc0_bgref (1.024V) sc2_bgref (1.024V) refsel[1:0] Vssa sc1_bgref (1.024V) sc3_bgref (1.024V) Vin Vref out sc3 ABUSL0 ABUSL1 ABUSL2 ABUSL3 v0 DAC0 i0 DAC1 v1 i1 v2 DAC2 i2 DAC3 v3 i3 USB IO USB IO * P15[6] GPIO P5[7] GPIO P5[6] GPIO P5[5] GPIO P5[4] SIO P12[7] SIO P12[6] GPIO *P1[7] GPIO *P1[6] dac_vref (0.256V) vssd dsm0_vcm_vref1 (0.8V) dsm0_vcm_vref2 (0.7V) + DSM0 - vssa DSM vcm refs qtz_ref vref_vss_ext dsm0_qtz_vref2 (1.2V) dsm0_qtz_vref1 (1.024V) Vdda/3 Vdda/4 ExVrefL ExVrefR refmux[2:0] Vp (+) Vn (-) SAR0 Vrefhi_out refs SAR_vref1 (1.024V) SAR_vref2 (1.2V) (+) Vp SAR1 (-) Vn Vrefhi_out refs SAR_vref1 (1.024V) SAR_vref2 (1.2V) SAR ADC Vdda Vdda/2 ExVrefL1 en_resvda ExVrefL2 refmux[2:0] 01 23456 7 0123 3210 76543210 AGR[3] AGR[2] AGR[1] AGR[0] AMUXBUSR Vssd XRES Vbat Ind Vboost * * Vssb GPIO P2[5] GPIO P2[6] GPIO P2[7] SIO P12[4] SIO P12[5] GPIO P6[4] GPIO P6[5] GPIO P6[6] GPIO P6[7] * * Large ( ~200 Ohms) * * Switch Resistance Small ( ~870 Ohms ) * * Connection * Vddio1 * AMUXBUSL Mux Group Switch Group * * AGL[1] AGL[0] PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F * AGL[3] AGL[2] AGR[0] AMUXBUSR AGR[3] AGR[2] AGR[1] LPF GPIO P5[0] GPIO P5[1] GPIO P5[2] GPIO P5[3] GPIO P1[0] GPIO P1[1] GPIO P1[2] GPIO P1[3] GPIO P1[4] GPIO P1[5] AGL[1] AGL[2] AGL[3] VBE Vss ref * TS ADC AMUXBUSR ANALOG ANALOG BUS GLOBALS * AMUXBUSL AGL[0] ANALOG ANALOG GLOBALS BUS : Vdda Vdda/2 en_resvda refmux[2:0] AMUXBUSL Vssd Vddd * P15[7] VIDAC vcmsel[1:0] Vccd ABUSR0 ABUSR1 ABUSR2 ABUSR3 * * Vddio2 refbuf_vref1 (1.024V) refbuf_vref2 (1.2V) sc1 Vin Vref out SC/CT Vin Vref out sc2 out ref in * * Vddd refbufr AGR[4] AMUXBUSR CAPSENSE out ref in refbufl refbuf_vref2 (1.2V) GPIO P6[0] GPIO P6[1] GPIO P6[2] GPIO P6[3] GPIO P15[4] GPIO P15[5] GPIO P2[0] GPIO P2[1] GPIO P2[2] GPIO P2[3] * GPIO P2[4] * i1 bg_vda_swabusl0 refbuf_vref1 (1.024V) Vssd + - cmp0_vref (1.024V) cmp1_vref Vdda Vdda/2 Vccd comp3 ExVrefR i3 refbufr_ cmp refbufl_ cmp vref_cmp1 (0.256V) bg_vda_res_en comp1 + - COMPARATOR cmp_muxvn[1:0] abuf_vref_int (1.024V) swin AGR[7] AGR[6] AGR[5] GPIO P4[2] GPIO P4[3] GPIO P4[4] GPIO P4[5] GPIO P4[6] GPIO P4[7] swout out1 comp0 + - cmp1_vref cmp0_vref (1.024V) in1 out0 swin i2 * LPF in0 swout abuf_vref_int (1.024V) cmp1_vref i0 * * * opamp0 * * AGL[6] AGL[7] AGR[6] AGR[7] * * * * AMUXBUSL AGL[4] AGL[5] * * * AGR[4] AGR[5] AGL[6] AGL[7] swinp GPIO P0[4] GPIO P0[5] GPIO P0[6] GPIO P0[7] * * AMUXBUSR AMUXBUSL AGL[4] AGL[5] ExVrefL ExVrefL1 Vddio3 GPIO P3[6] GPIO P3[7] SIO P12[0] SIO P12[1] GPIO P15[2] GPIO P15[3] SIO P12[2] SIO P12[3] GPIO P4[0] GPIO P4[1] GPIO P0[0] GPIO P0[1] GPIO P0[2] GPIO P0[3] Vddio0 swinp swinn Notes: * Denotes pins on all packages LCD signals are not shown. Rev #60 13-Feb-2012 321 Analog Routing 29.3 How it Works Analog routing resources in PSoC 5LP devices include analog globals (AGs), analog mux bus (AMUXBUS), liquid crystal display bias bus (LCDBUS), and local analog buses (abus). The analog globals and AMUXBUS go to the GPIOs and provide a way to route signals between the GPIOs and the analog resource blocks (ARBs). The LCDBUS is used for LCD bias signal routing. Analog resource blocks include the following: DACs, comparators, CapSense, switched capacitors, Delta Sigma ADC, and opamps. The analog local buses (abus) are local buses used for connections between ARBs. In addition, there is a VREF bus, as shown in Figure 29-2 on page 321. This VREF bus carries the reference voltages for different analog blocks that are generated by the precision reference block. See the Precision Reference chapter on page 373 for details on these reference voltages. ■ In the lower right half, Px[3:0] maps to AGR[3:0] and Px[7:4] maps to AGR[3:0] ■ In the upper right half, Px[3:0] maps to AGR[7:4] and Px[7:4] maps to AGR[7:4] This means that two pins on each port are connectable to the same global, as shown in the diagram. The analog global bus connects to inputs and/or outputs of the following ARBs: DAC, comparator, output buffer, switched capacitor, Delta Sigma ADC, and CapSense (which is a virtual block). These connections are made through switches and muxes. PRT[x]_AG registers are used to configure the analog globals (AGs) for each GPIO port pin. See 29.6 Analog Routing Register Summary on page 333 for register details. All these analog routing resources are explained in detail in the following sections. Port 12 contains the Special Input/Output (SIO) pins. These pins are grouped in pairs for each quadrant of the device (lower right: P12[6] and P12[7], lower left: P12[4] and P12[5], upper left: P12[2] and P12[3], upper right: P12[0] and P12[1]), with each pair sharing a reference generation (REFGEN) block. The SIO REFGEN block can select from one of two analog globals routed to the pair shown in Figure 29-2 on page 321. The mux selection is controlled by the {PRT12_AG} register. See the I/O System chapter on page 151 for details about SIO operation. Figure 29-4 on page 324 illustrates the difference between switches and muxes. 29.3.2 Analog switches and muxes establish connections between the above mentioned analog routing buses and the ARBs. 29.3.1 Analog Globals (AGs) The PSoC 5LP die is divided into four quadrants, as shown in Figure 29-2 on page 321 and Figure 29-3 on page 323. The analog global bus has eight routes on each side, AGL[7:0] on the left and AGR[7:0] on the right. Within each side, the bus is divided into two groups, AGR[3:0] and AGR[7:4] for the right side and AGL[3:0] and AGL[7:4] for the left side. The lower four globals on each side are routed to the GPIO in the lower half of the die and the upper four globals on each side are routed to the GPIO in the upper half of the die. All eight analog globals on each side get routed to ARBs on the same side. Analog globals can be used as single-ended or differential signal paths. The left and right half globals may operate independently or they may be joined through the switches that are shown at the top and bottom of Figure 29-3 on page 323. Each GPIO may be connected to an analog global through a switch in the following manner: ■ In the lower left half, Px[3:0] maps to AGL[3:0] and Px[7:4] maps to AGL[3:0] ■ In the upper left half, Px[3:0] maps to AGL[7:4] and Px[7:4] maps to AGL[7:4] 322 Analog Mux Bus (AMUXBUS) There are two AMUXBUS routes in PSoC 5LP devices. The device can be divided into two halves (left and right), with each half having one AMUXBUS (AMUXBUSR, AMUXBUSL). The left and right AMUXBUS may be shorted together with an analog switch. Every GPIO has the provision to connect to an AMUXBUS through an analog switch. CapSense applications use the AMUXBUS for their operation. See CapSense chapter on page 359 for details on using this bus for CapSense applications. PRT[x]_AMUX registers are used to configure the AMUXBUS routing for each GPIO port pin. See 29.6 Analog Routing Register Summary on page 333 for register details. 29.3.3 Liquid Crystal Display Bias Bus (LCDBUS) The LCD bias bus contains five routes that connect to every GPIO. These routes are continuous around the device periphery and are not separated by switches at the midline as are the analog globals and AMUXBUS. Each LCD route is individually configurable so that they are driven by the analog local bus or LCD bias voltage to the LCD driver buffer located in the GPIO. Connecting to an analog bus allows low frequency analog signals to drive off-chip through the LCD driver buffers. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Analog Routing The LCDBUS mux selections are given in the following table. See the LCD Direct Drive chapter on page 345 for LCD operation and biasing. See 29.6 Analog Routing Register Summary on page 333 for register details. Table 29-1. LCD Bias Bus Mux Selections Output Mux Selections LCD_BIAS_BUS[0] {0=LCDDAC_V0,1=abusr[0],2=abusl[0],3=NA} LCD_BIAS_BUS[1] {0=LCDDAC_V1,1=abusr[1],2=abusl[1],3=NA} LCD_BIAS_BUS[2] {0=LCDDAC_V2,1=abusr[2],2=abusl[2],3=NA} LCD_BIAS_BUS[3] {0=LCDDAC_V3,1=abusr[3],2=abusl[3],3=NA} LCD_BIAS_BUS[4] {0=LCDDAC_V4,1=AMUXBUSR,2=AMUXBUSL,3=NA} Figure 29-3. Analog Globals, AMUXBUS, and LCDBUS Routing AG[4]L AG[5]L AG[4]R AG[5]R AG[6]L AG[6]R AG[7]R AG[7]L 5 AMUXBUSL GPIO Px[0] GPIO Px[1] GPIO Px[2] GPIO Px[3] GPIO Px[4] GPIO Px[5] GPIO Px[6] GPIO Px[7] AMUXBUSR GPIO Px[0] GPIO Px[1] GPIO Px[2] GPIO Px[3] GPIO Px[4] GPIO Px[5] GPIO Px[6] GPIO Px[7] 7 6 5 43 2 1 0 01 2 3 4 56 7 Upper Left Quadrant Lower Left Quadrant Upper Right Quadrant Lower Right Quadrant AMUXBUSR AMUXBUSL GPIO Px[0] GPIO Px[1] GPIO Px[2] GPIO Px[3] GPIO Px[4] GPIO Px[5] GPIO Px[6] GPIO Px[7] LCD Bias Bus 01 2 3 4 56 7 76 5 43 2 1 0 AG[3]L AG[3]R AG[2]L AG[1]L AG[2]R AG[1]R AG[0]R AG[0]L GPIO Px[0] GPIO Px[1] GPIO Px[2] GPIO Px[3] GPIO Px[4] GPIO Px[5] GPIO Px[6] GPIO Px[7] Switch Connection PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 323 Analog Routing 29.3.5 Switches and Multiplexers Switches and multiplexers are used to establish connections using different analog routing buses. They are placed on the various buses to direct signals into and out of the GPIOs and ARBs. In a switch with ‘n’ inputs and one output, zero through ‘n’ switches may be on at a time, whereas in a multiplexer (mux) with ‘n’ inputs and one output, only one switch may be on at a time. Note that a group of eight analog switches requires eight bits for configuration, whereas, a mux with eight analog switches requires only three bits. Figure 29-4 illustrates the difference between switches and muxes, in switch and mux symbols. For example, in Figure 29-4, there are two muxes (ARB, LCD). In both these muxes, only one of the analog switches can be selected for routing. In the same figure, there are two switches (GPIO, ARB). For these switches, more than one analog switch can be selected for routing. Note that both muxes and switches are formed using analog switches. LCD MUX abus (1 of 8) AMUXBUS (1 of 2) There are eight analog local bus (abus) routes in PSoC 5LP devices, four in the left half (abusl[0:3]) and four in the right half (abusr[0:3]), as shown in Figure 29-2 on page 321. These are local routes located in the analog subsystem and are for interconnecting ARBs, which reduces the usage of AGs. They do not route directly out into GPIOs. It is possible to short the left and right abus’ together with four analog switches. ARBs may connect to each other through analog globals (AG) or the analog local bus (abus). For example, in Figure 29-2 on page 321, a DAC output (V1, for example) may be used as a reference for a comparator negative input (COMP1, for example). Using an analog switch, the DAC output can be placed on AGR0 and the comparator input switch can also be set to AGR0. Limited number of available analog globals (eight per side) and some block to block connections can be made through analog local bus for direct connections between blocks. For the above example, the DAC output (V1) can be routed directly to the analog local bus (abusr3) that goes to the negative input of the comparator (COMP1). This saves the GPIO routing resource from being used for interconnecting two ARBs. Figure 29-4. Difference Between Analog Switches and Muxes LCDBUS (1 of 5) Analog Local Bus (abus) Analog Global (1 of 16) 29.3.4 ARB Mux IN ARB1 GPIO GPIO Switch ARB2 OUT IN ARB3 ARB Switch OUT 29.3.5.1 Control of Analog Switches Analog globals (AGs), analog mux bus (AMUXBUS) and the analog local bus (abus) all use analog switches to establish connections. As stated earlier, analog switches can be grouped together to form multiplexers or switches. Each GPIO has two analog switches, one to connect the pin to the analog global and the other to connect the pin to the AMUXBUS. The open/close control signals for these analog switches can be generated by either of the following ways: 1. The registers corresponding to the GPIO pin, PRT[x]_AMUX and PRT[x]_AG, can be used to control the open/close state of the analog switches. This is the default option. 2. In addition, there is a provision to dynamically control these switches by means of the DSI control signal that is connected to the input of the port pin logic block. This option is enabled by setting the bit in the Port Bidirection Enable register (PRT[x]_BIE). For example, to control pin 3 of port 0, a value of 0x08 is written to PRT[0]_BIE. The switch control signal is the logical AND of the register setting, as in the first case, and the DSI control signal, as shown in Figure 29-5. Each GPIO is connected through two analog switches to an analog global and an AMUXBUS. The ARBs use ARB switches and ARB muxes for input/output routing options. 324 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Analog Routing Figure 29-5. GPIO Pin Input/Output Block Diagram Digital Input Path Naming Convention ‘x’ = Port Number ‘y’ = Pin Number PRT[x]CTL PRT[x]DBL_SYNC_IN PRT[x]PS Digital System Input PICU[x]INTTYPE[y] Input Buffer Disable PICU[x]INTSTAT Interrupt Logic Pin Interrupt Signal PICU[x]INTSTAT Digital Output Path PRT[x]SLW PRT[x]SYNC_OUT Vddio Vddio PRT[x]DR 0 In Digital System Output 1 Vddio PRT[x]BYP Drive Logic PRT[x]DM2 PRT[x]DM1 PRT[x]DM0 Bidirectional Control PRT[x]BIE Analog Slew Cntl PIN OE 1 0 1 Capsense Global Control 0 1 CAPS[x]CFG1 Switches PRT[x]AG Analog Global PRT[x]AMUX Analog Mux LCD Display Data PRT[x]LCD_COM_SEG Logic & MUX PRT[x]LCD_EN LCD Bias Bus 5 In addition, there are control signals that are dedicated for CapSense applications as shown in Figure 29-5. See the CapSense chapter on page 359 for the usage of these control signals. The analog switches corresponding to the analog resource blocks can be controlled only by the register settings of the respective ARBs. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F For example, to switch comparator input between two GPIOs that are connected to the same analog global, the register settings for the input select of the comparator are configured to select the analog global to which the GPIOs are connected. The DSI control signal can dynamically select between the two GPIOs after the corresponding PRT[x]_BIE register is configured. 325 Analog Routing 29.4 Analog Resource Blocks – Routing and Interface The analog interface (ANAIF) is the interface between the analog blocks and other PSoC systems (UDB, DSI, clock, and decimator). The analog interface has 2 kilobytes of memory, which stores the configuration settings of all analog resource blocks. The configuration space is written to and read by the PHUB. The analog interface also interfaces clock distribution to the various analog resource blocks. For ARBs that deal with both analog and digital signals, such as the ADC, DAC, and comparator, the analog interface connects the digital and analog portions. For example, the comparator output is routed to the digital systems interconnect (DSI) through the analog interface. The modulator output (digital) is routed to the decimator through the analog interface. Similarly, the strobe and other digital signals for the DAC are routed through the analog interface. More details about how the interfaces are provided by the ANAIF are given in the individual chapters in Section F: Analog System on page 303. The following figure shows the top level diagram of the analog interface. Figure 29-6. Analog Interface System Diagram Analog System Block Port Control All Clocks UDB Array 326 AHB CLK_A[3:0] CLK_A_DIG[3:0] ANAIF CLKDIST Decimator PHUB PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Analog Routing 29.4.1 Digital-to-Analog Converter (DAC) The DAC routing options and connections to other PSoC subsystems through the analog interface are shown in Figure 29-7. The output for each DAC is selected by control registers that are connected to multiplexer select lines. The DAC receives input data and control signals from the analog interface. The control signals include the strobe signal for the DAC, the reset signal, the DAC current-off signal, and output current direction. These control signals come from UDBs or control registers. See the Digital-to-Analog Converter chapter on page 369 to learn more about DAC control and operation. Figure 29-7. DAC Routing, Interface Reg DAC1.SW* Reg DAC0.SW* AGL0 AGL1 AMUXBUSL abusl1 abusl3 ANAIF dac_data V dac_data 8 P0[6] AMUXBUSL AGL0 AGL1 V 8 DAC0 DAC1 I 4 I dac1_cr dac0_cr 4 Reg DAC0.SW* Reg DAC3.SW* V 8 dac_data dac_data AGR4 AGR5 AMUXBUSR abusr0 abusr2 DAC3 I dac2_cr I dac3_cr 4 4 8 Reg DAC2.SW* V 8 DAC2 P0[7] AMUXBUSL AGL4 AGL5 P3[0] AMUXBUSR AGR0 AGR1 Reg DAC1.SW* Reg DAC2.SW* AGL4 AGL5 AMUXBUSL abusl0 abusl2 AGR0 AGR1 AMUXBUSR abusr1 abusr3 4 4 4 4 dac_data_udb dac0_cr_udb dac1_cr_udb dac2_cr_udb UDB PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F P3[1] AMUXBUSR AGR4 AGR5 dac3_cr_udb Reg DAC3.SW* dacn_cr includes strobe, reset, ioff, idir signals 327 Analog Routing 29.4.2 Comparator The comparator routing options and connections to other PSoC subsystems through the analog interface are shown in Figure 29-8. The input for each comparator is selected by control registers, which are connected to the multiplexer select lines. The outputs of the comparators are routed to the ANAIF for further processing. The analog interface contains lookup tables (LUTs) that are used to implement logic functions on comparator outputs. The LUT outputs (LUTN_OUT) are routed to the UDB block through the DSI. In addition, LUT outputs can generate interrupts (LUT_IRQ) to the device. See the Comparators chapter on page 335 to learn more about comparator control and operation. Figure 29-8. Comparator Routing, Interface AG L0 AG L1 AG L2 AG L3 AG L4 AG L5 AG L6 AG L7 AM UXBUSL abusl0 abusl1 refbufl Reg CM P 0.SW *, CM P 0.SW * AG L0 AG L2 AG L4 AG L6 AM UXBUSL abusl2 abusl3 VREF0 VREF1 Reg CM P 0.SW *, CM P0.SW * AG L0 AG L1 AG L2 AG L3 AG L4 AG L5 AG L6 AG L7 AM UXBUSL abusl0 abusl1 refbufl Reg CM P 2.SW *, CM P 2.SW * AG L1 AG L3 AG L5 AG L7 AM UXBUSL abusl2 abusl3 VREF0 VREF1 Reg CM P 2.SW *, CM P2.SW * AN AIF + com p0 _ + com p1 _ + _ com p3 + _ com p2 AG R0 AG R1 AG R2 AG R3 AG R4 AG R5 AG R6 AG R7 AM UXBUSR abusr0 abusr1 refbufr Reg CM P 1.SW *, CM P1.SW * AG R 0 AG R 2 AG R4 AG R 6 AM UXBUSR abusr2 abusr3 VREF0 VREF1 Reg CM P 1.SW *, CM P 1.SW * AG R0 AG R1 AG R2 AG R3 AG R4 AG R5 AG R6 AG R7 AM UXBUSR abusr0 abusr1 refbufr Reg CM P 3.SW *, CM P3.SW * 4 4 LUT 0 4 4 4 LU T1 4 LU T 2 4 4 LU T3 AG R 1 AG R 3 AG R5 AG R 7 AM UXBUSR abusr2 abusr3 VREF0 VREF1 Reg CM P 3.SW *, CM P 3.SW * UDB 328 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Analog Routing 29.4.3 Delta Sigma Modulator (DSM) The Delta Sigma modulator (DSM) is part of the Delta Sigma ADC and consists of various blocks that are mentioned in the Delta Sigma Converter chapter on page 377. The DSM can select its clock from any of the four analog clocks. The decimator block and the synchronization circuit in the ANAIF use the clock, CLK_DEC, which is selected from the corresponding digitally aligned analog clocks. The DSM output DSM0_DOUT and the overload detect status bits are routed to the ANAIF block for post processing. DSM also receives the reset signals and modulation signal from the analog interface. These control signals may originate from UDBs and or from control registers. See the Delta Sigma Converter chapter on page 377 to learn more about the control and operation of this block. Figure 29-9. DSM Routing, Interface AGL0 AGL1 AGL2 AGL3 AGL4 AGL5 AGL6 AGL7 dsm0_startup_reset_udb dsm0_startup_reset 8 AG1L AG3L AG5L AG7L AMUXBUSL abusl1 abusl3 VREF VSSA dsm0_dout_udb dsm0_dout 8 abusl0 abusl2 VSSA Reg DSM0.SW* dsm0_modbitin_udb dsm0_extclk_cp_udb dec_irq DSM dsm0_overload _one ANAIF dsm0_dout2scomp dsm0_overload _zero 4 Decimator UDB dec_start dec_clk dsm0_clk dsm0_reset_dec dsm0_modbitin dsm0_extclk_cp_udb Reg DSM0.SW* PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 329 Analog Routing 29.4.4 Switched Capacitor The switched capacitor block provides various analog functions. It has a modulator output SCN_MODOUT, which is routed to a register and is also routed to the UDB array as SCN_MODOUT_SYNC (Figure 29-10). The four analog clocks and the corresponding digitally aligned clocks, as well as the UDB generated clock, are selectable for each switched capacitor block instance. The interrupt signal corresponding to the switched capacitor blocks (SC_IRQ) is also routed to the UDB array. The polarity of the dynamic control input, SC_DYN_CNTRL, switches the amplifier between the inverting and non-inverting configuration. See the Switched Capacitor/Continuous Time chapter on page 305 to learn more about SC/CT. Figure 29-10. Switched Capacitor Routing, Interface AGR1 AGR3 AGR5 AGR7 abusr1 abusr3 AGL1 AGL3 AGL5 AGL7 abusl1 abusl3 SC1O Reg SC0.SW* AGL0 AGL2 AGL4 AGL6 VREF abusl0 SC1O Reg SC1.SW* SC0O AGR0 AGR2 AGR4 AGR6 VREF abusr0 SC0O ANAIF SC0 SC1 Reg SC1.SW* Reg SC0.SW* AGL0 AGL1 AGL2 AGL3 AGL4 AGL5 AGL6 AGL7 AMUXBUSL abusl0 abusl2 abusl3 VREF SC1O sc0_modout sc0_dyn_cntl sc1_dyn_cntl sc0_clk sc1_clk Reg SC0.SW* Reg SC1.SW* AGL0 AGL2 AGL4 AGL6 abusl0 abusl2 SC2O Reg SC2.SW* AGL1 AGL3 AGL5 AGL7 VREF abusl1 SC3O Reg SC2.SW* AG0L AG1L AG2L AG3L AG4L AG5L AG6L AG7L AMUXBUSL abusl1 abusl2 abusl3 VREF SC3O AGR0 AGR1 AGR2 AGR3 AGR4 AGR5 AGR6 AGR7 AMUXBUSR abusr0 abusr2 abusr3 VREF SC0O sc1_modout AGR0 AGR2 AGR4 AGR6 abusr0 abusr2 SC3O Reg SC3.SW* SC2 SC3 sc2_modout sc3_modout sc2_dyn_cntl sc3_dyn_cntl sc2_clk sc_dyn_cntl Reg SC3.SW* AG0R AG1R AG2R AG3R AG4R AG5R AG6R AG7R AMUXBUSR abusr1 abusr2 abusr3 VREF SC2O sc3_clk 4 Reg SC2.SW* AGR1 AGR3 AGR5 AGR7 VREF abusr1 SC2O sc_irq sc_clk_udb 4 Reg SC3.SW* sc_modout_sync UDB 330 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Analog Routing 29.4.5 Opamp The input and output routing options for the output buffer (opamp) are shown in Figure 29-11. See the Opamp chapter on page 341 for details on configuration and operation of this block. Figure 29-11. Opamp Input/Output Routing Reg ABUF0.SW* Reg ABUF1.SW* Reg ABUF0.SW* Reg ABUF1.SW* P0[3] P3[4] AGL4 AGL6 AGR4 AGR6 Reg ABUF0.MX* AGL4 AGL5 AGL6 AGL7 VREF abusl0 abusl1 abusl2 abusl3 Reg ABUF1.MX* Opamp0 Opamp1 P0[1] AGR4 AGR5 AGR6 AGR7 VREF abusr0 abusr1 abusr2 abusr3 P3[6] Reg ABUF0.MX* Reg ABUF1.MX* Reg ABUF0.SW* Reg ABUF1.SW* P0[2] P3[5] Reg ABUF2.SW* Reg ABUF3.SW* Reg ABUF2.SW* Reg ABUF3.SW* P0[5] P3[2] AG5L AG7L AGR5 AGR7 Reg ABUF2.MX* AGL4 AGL5 AGL6 AGL7 VREF abusl0 abusl1 abusl2 abusl3 Reg ABUF3.MX* Opamp2 Opamp3 P0[0] AGR4 AGR5 AGR6 AGR7 VREF abusr0 abusr1 abusr2 abusr3 P3[7] Reg ABUF2.MX* Reg ABUF2.SW* Reg ABUF3MX* Reg ABUF3.SW* P0[4] P3[3] = Analog Switch 29.4.6 Low-Pass Filter (LPF) Two tunable low-pass filter blocks are available. The inputs are selectable in a 2:1 mux for each LPF, as shown in Figure 29-12. On the left side, the LPF inputs are AMUXBUSL and AGL0. On the right side, the inputs are AMUXBUSR and AGR0. The outputs are connected through switches to abusL0 and abusR0, respectively. The tunability of the LPF allows the user to select an R of either 1 M or 200 k, and a C of either 5 pF or 10 pF. The LPF control registers are LPF0_CR0 and LPF1_CR0. Figure 29-12. LPF Routing AMUXBUSL ABUSL0 IN0 AGL0 OUT0 LPF0 LPF0.CR0 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F AMUXBUSR ABUSR0 OUT1 IN1 AGR0 LPF1 LPF1.CR0 331 Analog Routing 29.5 Low-Power Analog Routing Considerations For example, assume you want to connect P3.5 to P3.4 in a simple pass-through configuration. This is illustrated in Figure 29-13. This illustration is taken from the full chip diagram shown in Figure 29-2. P3.5 enters the chip on analog global AG5. P3.4 enters the chip on analog global AG4. To connect these two pins together, you need to track jump between AG4 and AG5. To do this, you can use the comparator ARBs comp1 positive input switches (assuming the rest of our project isn't using these switches). The switch for AG4 and AG5 on the comparator comp1 positive input is closed, while the rest of the switches remain open. The inputs to the comparator ARB itself are isolated from the switch group via a transmission gate. Figure 29-2 illustrates the analog global routing network, overlaid on top of the ARBs. Each ARB has a set of muxes and switches that it uses to connect to the global analog routing. By connecting to one of the analog routing channels virtually any ARB can be connected to any other ARB or pin on the chip. Not all pins or ARBs are connected to every analog global routing channel. To get a signal from a particular ARB out to a specific pin, the PSoC Creator analog routing algorithm implements a technique known as “Track Jumping”. Track jumping connects two analog globals together via one of the ARBs analog global switching structures, without connecting to that particular ARB resource. After this configuration is programmed into the device, any signal seen on P3.5 will show up on P3.4. Figure 29-13. Simplified Diagram of Routing P3.5 to P3.4 Using Track Jumping on the Positive Input of Comparator 1 P3.5 COMPARATOR P3.4 com p1 29.5.1 Mitigating Analog Routes with Degraded Low-power Signal Integrity The analog router in PSoC Creator uses track jumping to connect analog globals together. Track jumping is done on the muxes/switches of unused ARBs. The auto-router in PSoC Creator will always choose routes that ensure signal integrity in all power modes. If the auto-router is not sufficient and you need to resort to manual routing techniques to realize a design, then take special care. For performance reasons, the SC/CT ARB controls the availability of all of its associated analog switches. AG4 AG 5 AG 6 AG 7 input offs et cancelation routing components in PSoC Creator, make certain to avoid routing that uses the SC/CT block for track jumping purposes if the SC/CT is not enabled in Hibernate/Sleep modes. If a design uses SC/CT analog switches to realize a design in sleep/hibernate modes, but has this block powered down, significant degradation in signal integrity may be experienced. Explicitly Start() components derived from the SC/CT block and leave on in hibernate/sleep if it is necessary to still use these switches to route the design. Not starting the ST/ CT block is equivalent to stopping it. By starting this block its routing resources become available for routing. See the PSoC Creator datasheet associated with Manual Routing for more details on how to use the MARS tool. You can modify which analog routes are chosen by using the Manual Analog Routing (MARS) tool to force the routes. If you modify the analog routes chosen, by using the manual 332 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Analog Routing 29.6 Analog Routing Register Summary Table 29-2. Analog Routing Register Summary Name {PRT[0..11]_AMUX} {PRT15_AMUX} Brief Description These registers control the connection between the analog mux bus and the corresponding GPIO pin {PRT[0..11]_AG} {PRT12_AG} These registers control the connection between the analog global buses and the corresponding GPIO pin. Port 12 is the SIO port and the PRT12_AG register is for SIO reference selection. {PRT15_AG} {CMP[0..3]_SW0} {CMP[0..3]_SW2} {CMP[0..3]_SW3} {CMP[0..3]_SW4} Comparator positive input to analog globals 0-7 Comparator positive input to analog local bus Comparator positive input to AMUXBUS and reference buffer Comparator negative input to AMUXBUS and VREF Comparator negative input to analog globals 0-7 {CMP[0..3]_SW6} Comparator negative input to analog local bus {CMP[0..3]_CLK} Comparator sampling clock selection and clock control register {DSM0_SW0} Delta Sigma modulator positive input to analog globals 0-7 {DSM0_SW2} Delta Sigma modulator positive input to analog local bus {DSM0_SW3} Delta Sigma modulator positive input to AMUXBUS and VSSA Delta Sigma modulator negative input to AMUXBUS, VSSA, and VREF {DSM0_SW4} Delta Sigma modulator negative input to analog globals 0-7 {DSM0_SW6} Delta Sigma modulator negative input to analog local bus {DSM0_CLK} Delta Sigma modulator clock selection {DAC[0..3]_SW0} DAC voltage output to analog globals 0-7 {DAC[0..3]_SW2} DAC voltage output to analog local bus {DAC[0..3]_SW3} DAC voltage output to AMUXBUS DAC current output to AMUXBUS and direct to pad {DAC[0..3]_SW4} DAC current output to analog globals 0-7 {DAC[0..3]_SW6} DAC current to analog local bus {DAC[0..3]_STROBE} DAC strobe selection {SC[0..3]_SW0} Switched capacitor (SC) positive input to analog globals 0-7 {SC[0..3]_SW2} SC positive input to analog local bus {SC[0..3]_SW3} SC positive input to AMIUXBUS and VREF SC negative input to AMUXBUS and VREF {SC[0..3]_SW4} SC negative input to analog globals 0-7 {SC[0..3]_SW5} SC negative input to analog globals 8-15 {SC[0..3]_SW6} SC negative input to analog local bus {SC[0..3]_SW7} SC output to AMUXBUS, other SC negative and positive inputs {SC[0..3]_SW8} SC output to analog globals 0-7 {SC[0..3]_SW10} SC output to analog local bus {SC[0..3]_CLK} SC clock selection {ABUF[0..3]_MX} These registers select positive and negative inputs to the output buffer. {ABUF[0..3]_SW} These registers control the switch between the output and negative input, the switch between the output and GPIO, the switch between the negative input and GPIO, and the switch between the positive input and GPIO. {LUT[0..3]_CR} These registers select the signals to comparator LUT and also select the LUT function. {LCDDAC_SW[0:4]} These registers select the signals on the LCD bias bus. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 333 Analog Routing Table 29-2. Analog Routing Register Summary (continued) Name {BUS_SW0} Brief Description This register controls the switches that tie AGR[7:0] to AGL[7:0]. {BUS_SW2} This register controls the switches that tie abusL[7:0] to abusR[7:0] (left and right analog local bus) lines together. {BUS_SW3} This register controls the switch that ties AMUXBUSR to AMUXBUSL. {LPF0_CR0} {LPF1_CR0} 334 LPF registers PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 30. Comparators PSoC® 5LP devices have four analog comparator modules. The positive and negative inputs to the comparators come through muxes with inputs from analog globals (AGs), local analog bus (ABUS), analog mux bus (AMUXBUS), and precision reference. The output from each comparator is routed through a synchronization block to a two-input lookup table (LUT). The output of the LUT is routed to the UDB digital system interface (DSI). The comparator can also be used to wake the device from sleep. An ‘x’ used with a register name denotes the particular comparator number (x = 0 to 3). 30.1 Features PSoC® comparators have the following features: ■ Flexible input selection ■ Speed power tradeoff ■ Optional 10 mV input hysteresis ■ Low-input offset voltage (<1 mV) ■ Glitch filter for comparator output ■ Sleep and Hibernate wakeup 30.2 Block Diagram Figure 30-1 on page 336 is a block diagram of PSoC Comparators. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 335 Comparators Figure 30-1. Comparator Block Diagram AGL0 AGL1 AGL2 AGL3 AGL4 AGL5 AGL6 AGL7 AMUXBUSL abusl0 abusl1 refbufl Reg CMP0.SW*, CMP0.SW* AGL0 AGL2 AGL4 AGL6 AMUXBUSL abusl2 abusl3 VREF0 VREF1 Reg CMP0.SW*, CMP0.SW* AGL0 AGL1 AGL2 AGL3 AGL4 AGL5 AGL6 AGL7 AMUXBUSL abusl0 abusl1 refbufl Reg CMP2.SW*, CMP2.SW* AGL1 AGL3 AGL5 AGL7 AMUXBUSL abusl2 abusl3 VREF0 VREF1 Reg CMP2.SW*, CMP2.SW* AGR0 AGR1 AGR2 AGR3 AGR4 AGR5 AGR6 AGR7 AMUXBUSR abusr0 abusr1 refbufr ANAIF + comp0 _ + comp1 _ Reg CMP1.SW*, CMP1.SW* AGR0 AGR2 AGR4 AGR6 AMUXBUSR abusr2 abusr3 VREF0 VREF1 Reg CMP1.SW*, CMP1.SW* AGR0 AGR1 AGR2 AGR3 AGR4 AGR5 AGR6 AGR7 AMUXBUSR abusr0 abusr1 refbufr + _ comp3 + _ comp2 Reg CMP3.SW*, CMP3.SW* 4 4 LUT0 4 4 4 LUT1 4 LUT2 4 AGR1 AGR3 AGR5 AGR7 AMUXBUSR abusr2 abusr3 VREF0 VREF1 4 LUT3 Reg CMP3.SW*, CMP3.SW* UDB 30.3 How it Works The following sections describe the operation of PSoC comparators. 30.3.1 Input Configuration Inputs to the comparators are as follows: ■ Positive – from analog globals, analog locals, analog mux bus, and comparator reference buffer. See the CapSense chapter on page 359. ■ Negative – from analog globals, analog locals, analog mux bus, and voltage reference. All of the possible connections to the positive and negative inputs are shown in Figure 30-1. Inputs are configured using registers CMPx_SW0, CMPx_SW2, CMP_SW3, CMP_SW4, and CMP_SW6. 336 30.3.2 Power Configuration The comparator can operate in three power modes – fast, slow, and ultra low power. The power mode is configured using power mode select (SEL[1:0]) bits in the comparator control (CMPx_CR) register. The output of the comparators may glitch when the power mode is changed. Power modes differ in response time and power consumption; power consumption is maximum in fast mode and minimum in ultra low-power mode. Exact specifications for power consumption and response time are provided in the datasheet. 30.3.3 Output Configuration Comparator output can pass through an optional glitch filter. The glitch filter is enabled by setting the filter enable (FILT) bit in the control (CMPx_CR[6]) register. The output of the comparator is stored in the CMP_WRK register and can be read over the PHUB interface. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Comparators Four LUTs in the device allow logic functions to be applied to comparator outputs. LUT logic has two inputs: ■ Input A – selected using MX_A[1:0] bits in LUT control (LUTx_CR1:0) register ■ Input B – selected using MX_B[1:0] bits in LUT Control (LUTx_CR5:4) register The logic function implemented in the LUT is selected using control (Q[3:0]) bits in the LUT Control register (LUTx_CR) register. The bit settings for various logic functions are given in Table 30-1. Table 30-1. Control Words for LUT Functions Control Word (Binary) Output (A and B are LUT Inputs) 0000 FALSE(‘0’) 0001 A AND B 0010 A AND (NOT B) 0011 A 0100 (NOT A) AND B 0101 B 0110 A XOR B 0111 A OR B 1000 A NOR B 1001 A XNOR B 1010 NOT B 1011 A OR (NOT B) 1100 NOT A 1101 (NOT A) OR B 1110 A NAND B 1111 TRUE (‘1’) The output of the LUT is routed to the digital system interface of the UDB array. From the digital system interface of the UDB array, these signals can be connected to other blocks in the device or to an I/O pin. The state of the LUT output is indicated in the LUT output (LUTx_OUT) bit in the LUT clear-on-read sticky status (LUT_SR) register and can be read over PHUB interface. The LUT interrupt can be generated by all four LUTs and is enabled by setting the LUT mask (LUTx_MSK) bit in the LUT mask (LUT_MSK) register. 30.3.4 Hysteresis For applications that compare signals very close to each other, hysteresis helps to avoid excessive toggling of the comparator output when the signals are noisy. The 10 mV hysteresis level is enabled by setting the hysteresis enable (HYST) bit in the control (CMPx_CR5) register. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 30.3.5 Wake Up The comparator can run in sleep and hibernate modes and the output used to wake the device. Comparator operation in sleep and hibernate modes is enabled by setting the override (PD_OVERRIDE) bit in the control (CMPx_CR[2]) register. In low-power modes, the analog global pumps are disabled, increasing the resistance of analog routes for low-voltage applications that rely on the pumps in active mode. For best results, the device should be operated with VDDA at 2.7 V or higher. At lower voltages, the analog global routes will not meet their impedance specifications in low-power modes. 30.3.6 Comparator Clock Comparator output changes asynchronously and can be synchronized with a clock. The clock source can be one of the four digitally-aligned analog clocks or any UDB clock. Clock selection is done in mx_clk bits [2:0] of CMP_CLK register. The selected clock can be enabled or disabled by setting or clearing the clk_en (CMP_CLK[3]) bit. Comparator output synchronization is optional and can be bypassed by setting the bypass_sync (CMP_CLK[4]) bit. 30.3.7 Offset Trim Comparator offset is dependent on the common mode input voltage to the comparator. The offset is factory trimmed for common mode input voltages 0.1 V and Vdd - 0.1 V to less than 1 mV. If you know the common mode input range at which to operate the comparator, a custom trim can be done to reduce the offset voltage further. The comparator offset trim is performed in the CMPx_TR0 register. This register has two trim fields, trim1 (CMPx_TR0[3:0]) and trim2 (CMPx_TR0[7:4]). If shorting of the inputs is desired for offset calibration, the calibration enable field (cal_en) in the control register(CMP_CR[4]) helps to achieve it The method for a custom trim is as follows: 1. Set the two inputs ‘inn’ and ‘inp’ to the desired value. 2. Change the trim1 register settings: a. Depending on the polarity of the offset measured, set or clear trim1[3] bit. b. Increase the value of trim1[2:0] until offset measured is less than 1 mV. 3. If the polarity of the offset measured has changed but the offset is still greater than 1 mV, use trim2[3:0] to fine tune the offset value. This is valid only for the slow mode of comparator operation. 4. If trim1[2:0] is 07h, and the measured offset is still greater than 1 mV, set or clear trim2[3], depending on 337 Comparators offset polarity. Increase the value of trim2[2:0] until the offset measured is less than 1 mV. 338 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Comparators 30.3.8 Register Summary Table 30-2 is a summary listing of applicable registers. Table 30-2. Registers Register Function CMPx_SW0 Configures connection between positive input and analog globals 0-7 CMPx_SW2 Configures connection between positive input and analog locals 0-1 CMPx_SW3 Configures connection between analog mux bus to the two inputs and the voltage reference to negative input, CapSense® reference buffer to the positive input CMPx_SW4 Configures connection between negative input and analog globals 0-7 CMPx_SW5 CMPx_SW6 Configures connection between negative input and analog locals 0-1 CMPx_TR0 Trims the offset. Two groups of 4-bits for lower and higher end of common mode input ranges. CMP_WRK Stores the output state of the comparator CMPx_CLK These registers enable and disable synchronization of the output for comparators and the clock signal for synchronization CMPx_CR These registers are used to select the mode of operation of the comparator between the high speed and low speed modes and to enable/disable the comparator channel LUTx_CR Selects the input(s) and function for the LUT LUT_SR Stores the status of LUT outputs. It’s a clear on read register. LUT_MSK Enables interrupt request for a particular LUT output PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 339 Comparators 340 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 31. Opamp PSoC® 5LP devices have four operational amplifiers. An ‘x’ used with register name identifies the particular opamp number (x = 0 to 3). 31.1 Features PSoC operational amplifiers have the following features: ■ 25 mA current drive capability ■ 3-MHz gain bandwidth for 200-pF load ■ Offset trimmed to less than 0.5 mV ■ Low noise ■ Rail-to-rail to within 50 mV of Vss or Vdda for 1-mA load ■ Rail-to-rail to within 500 mV of Vss or Vdda for 25-mA load ■ Slew rate 3 V/µs for 200-pF load PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 341 Opamp 31.2 Block Diagram Figure 31-1 is the PSoC operational amplifiers block diagram. Figure 31-1. Operational Amplifiers Showing Available Connections OPAMP0.SW[0] OPAMP1.SW[0] OPAMP0.SW[1] OPAMP1.SW[1] P0[3] P3[4] AGL4 AGL6 AGR4 AGR6 OPAMP0.MX[4] OPAMP1.MX[4] OPAMP0 AGL4 AGL5 AGL6 AGL7 VREF abusl0 abusl1 abusl2 abusl3 P0[1] OPAMP1 P3[6] OPAMP1.MX[3:0] OPAMP0.MX[3:0] P0[2] P3[5] OPAMP0.SW[2] OPAMP1.SW[2] OPAMP2.SW[0] OPAMP3.SW[0] OPAMP2.SW[1] OPAMP3.SW[1] P0[5] P3[2] AGL5 AGL7 AGR5 AGR7 OPAMP3.MX[4] OPAMP2.MX[4] OPAMP2 AGL4 AGL5 AGL6 AGL7 VREF abusl0 abusl1 abusl2 abusl3 P0[0] OPAMP3 P3[7] P0[4] OPAMP2.SW[2] 31.3 How it Works Input and Output Configuration The positive and negative inputs to the operational amplifier can be selected through muxes and analog switches. A mux is used to connect an analog global, local analog bus, or reference voltage to an input, and an analog switch is used to connect a GPIO to an input. This is shown in Figure 31-1. Inputs are: Positive – The positive input analog switch, controlled by bit ABUFx_SW[2], is used to select an input from an external pin. The positive input mux (controlled by bits 342 P3[3] OPAMP3.SW[2] = analog switch PSoC 5LP devices have up to four operational amplifiers. The opamps are configurable as a unity gain buffer, to drive high current loads or as an uncommitted opamp. For example, a DAC output or voltage reference can be buffered using an opamp to drive a high current load. 31.3.1 AGR4 AGR5 AGR6 AGR7 VREF abusr0 abusr1 abusr2 abusr3 OPAMP3.MX[3:0] OPAMP2.MX[3:0] ■ AGR4 AGR5 AGR6 AGR7 VREF abusr0 abusr1 abusr2 abusr3 ABUFx_MX[3:0], is used to select an input from an internal signal. ■ Negative – The negative input analog switch, controlled by bit ABUFx_SW[1], selects an input from an external pin. The negative input mux, controller by bit ABUFx_MX[3:0], selects an input from an internal signal. The opamp output is connected directly to a fixed port pin. 31.3.2 Power Configuration The opamp can operate in three power modes – low, medium, and high. Power modes are configured using the (PWR_MODE[1:0]) power mode bits in the (OPAMPx_CR[1:0]) control register. The slew rate and gain bandwidth are maximum in high power mode and minimum in low-power mode. See the device datasheet for gain bandwidth and slew rate specifications in various power modes. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Opamp 31.3.3 Buffer Configuration The opamp is configured as a unity gain buffer by closing the feedback switch, using the OPAMPx_SW[0] bit. Setting the OPAMPx_SW[0] bit internally connects the output terminal to the negative opamp input. 31.3.4 Register Summary Table 31-1 summarizes applicable registers. Table 31-1. Registers Register OPAMPx_SW Function Controls positive input switch, negative input switch and feedback switch. OPAMPx_MX Selects the internal signal for positive and negative input. OPAMPx_CR Configures the power mode. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 343 Opamp 344 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 32. LCD Direct Drive The PSoC® liquid crystal display (LCD) drive system is a highly configurable peripheral that allows the PSoC device to directly drive a broad range of LCDs. The flexible power settings allow this peripheral to be used in applications where a battery is the power source. 32.1 Features Key features of the PSoC LCD system are: ■ LCD panel direct drive ■ Type A (standard) and Type B (low power) waveform support ■ Wide LCD bias range support (2 V to supply voltage) ■ Static, 1/3, 1/4, and 1/5 bias voltage levels ■ Internal bias voltage generation ■ Up to 62 total common and segment outputs ■ Supports up to 16 common glasses (16:1 mux) ■ Drives up to 736 total segments (16 backplane × 46 front plane) ■ 64 levels of software controlled contrast ■ Ability to move display data from memory buffer to LCD driver through direct memory access (DMA) without CPU intervention ■ Adjustable LCD refresh rate from 10 Hz to 150 Hz ■ Ability to invert LCD display for negative image ■ Various LCD driver drive modes, allowing power optimization 32.2 LCD System Operational Modes PSoC 5LP LCD architecture contains two operation modes. ■ LCD always active ■ LCD low power LCD always active mode is used when the device is not in low-power mode and when the LCD does not need to be operational in device low-power mode. LCD low-power mode is used when the LCD needs to be operational while the device is in low-power mode. This uses the same LCD always active system, but with some additional hardware. The LCD drive system does not work when the chip is placed in hibernate mode. The details of both modes are discussed in the following sections. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 345 LCD Direct Drive 32.3 LCD Always Active A complete functional LCD always active drive system is formed using the following major blocks: ■ ■ Dedicated LCD hardware ❐ LCD DAC ❐ LCD driver ❐ LCD bias generator System resources ❐ DMA ❐ Clocks: global ❐ RAM ❐ Universal digital block (UDB) Figure 32-1. LCD Always Active System Analog Global Bus Bias Select (LCDDAC.CR0[1:0]) System Resources Contrast Control (LCDDAC.CR1[5:0]) Continuous Drive (LCDDAC.CR0[3]) Dedicated LCD Component Hardware LCD DAC Drive/LCD CLK/Frame/ Mode[2:1] Clock UDB LCDDAC.SW 0/1/2/3/4 V0,V1,V2, V3,V4, GND LCD CLK Drive drq Frame Display Data LCD Driver Block Pin LCD Driver Block Pin Mode[2:0] DMA LCD Bias RAM LCD Bias Generator LCD Bias Port Data Registers LCD CLK Drive Frame Mode[2:0] LCD Bias Any LCD drive system requires the bias generating circuitry and system to interpret the data supplied, to display correctly on the LCD. PSoC 5LP contains dedicated LCD drive hardware, which works in conjunction with system resources. It contains a dedicated DAC that generates the five bias voltages, V0 to V4, along with ground. These bias voltages are distributed to all of the drivers of the LCD-capable pins. This DAC also helps to set contrast control. LCDs have two sets of pins: commons and segments. LCD functionality in PSoC 5LP GPIOs can be enabled by setting the appropriate bits of the PRT[0..11]_LCD_EN register. 346 These GPIOS can be configured to act as either common or segment drive pins by setting bits of PRT[0..11]_LCD_COM_SEG. The LCD driver blocks are the final interface to the pins. Each pin capable of driving an LCD contains driver logic. The function of this block is to select the bias level. It also drives the pin, depending on the LCD refresh state, whether the pin is configured as common or segment, and the display data. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F LCD Direct Drive The LCD display data resides in the system memory (SRAM). This display data needs to be transferred to the LCD driver logic. This is done using the direct memory access controller (DMAC). The DMAC takes the display data from the SRAM and loads it into the port data registers. The LCD driver latches this port data register value when a refresh action begins. The LCD DAC generates five voltages that are driven to LCD driver block. Important points regarding LCD DAC are: ■ All of the voltages V0 to V4 are generated using an internal resistor divider; V0 is the highest voltage and V4 the lowest voltage. By default, the five bias voltages, V0 to V4, from the LCD DAC are driven to each of the LCD driver blocks. Refreshing the LCD requires LCD state updates with accurate timing. This is done using a configurable clock, sourced from the internal main oscillator (IMO), which feeds the UDB block. The UDB is responsible for generating all of the control signals required by the rest of the blocks of the LCD system. ■ Analog mux bus and analog local bus can be selected to drive the LCD driver blocks, instead of the LCD DAC, by setting the appropriate bits of the LCDDAC_SW[0...4] registers. This is useful if you require external dividers to generate the drive voltages and optimize the power by switching off the internal DAC. In this mode, there is no software contrast control available. 32.3.1 ■ The LCD DAC can directly drive the LCD pixel, bypassing the LCD driver block. This is useful for driving the LCD even when the chip is put to sleep. You can do this by setting the LCDDAC.CR0[3] bit, which enables the continuous drive of the LCD DAC. Functional Description This section provides details of the LCD DAC, LCD driver, UDBs, clocking, DMA, CPU, and RAM, which all contribute to generating and sequencing the driving voltage for the LCD glass. 32.3.1.1 32.3.1.1.1 LCD DAC The LCD DAC is a 6-bit resistor ladder DAC. The LCD DAC is responsible for contrast control and bias voltage generation for the LCD drive system. When the device is put in lowpower mode, the LCD can remain operational. During this low-power mode, the DAC can directly drive the LCD pixel, bypassing the driver, thus compensating for the leakage. This is possible in LCD low-power mode, which is explained in section 32.4 LCD Low-Power Mode on page 351. Figure 32-2. LCD DAC (inputs and outputs) Contrast is controlled by varying the DAC output voltage, V0. This can be done by setting the LCD contrast control register (LCDDAC_CR1[5:0]), which sets the 6-bit DAC input (D[5:0], as shown in Figure 32-2). Thus, it provides 2 ^ 6 = 64 levels of contrast. Table 32-1 shows the V0 range and step size for 3.0-V and 5.5-V supply voltage. Table 32-1. LCD DAC V0 Range and Step Size 3.0 V Supply V0 5.5 V Supply V0 Range 2 V to 3 V 2 V to 5.5 V Step Size 27.3 mV 50 mV 32.3.1.1.2 D[5:0] pwrdn continuous drive enable hv holdb lcd bias select[1:0] Contrast Control Bias Ratio/Multiplex Ratio Selection Bias ratio/multiplex ratio is selected by setting the bias_sel field of the LCDDAC_CR0 register. This sets the DAC output voltages V1 to V4, as shown in Table 32-2 on page 348. V1 V2 V3 V4 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 347 LCD Direct Drive Table 32-2. LCD DAC Bias Select Bias Select Input: lcd_bias_select[1:0] Multiplex b1b0 Bias V0 V1 Ratio V2 V3 V4 Range in Volt 11 Invalid – default to 16:1 Default to 1/5 2.0 V to supply 0.800 × V0 0.600 × V0 0.400 × V0 0.200 × V0 10 16:1 1/5 2.0 V to supply 0.800 × V0 0.600 × V0 0.400 × V0 0.200 × V0 01 8:1 1/4 2.0 V to supply 0.750 × V0 0.500 × V0 0.500 × V0 0.250 × V0 00 4:1 1/3 2.0 V to supply 0.666 × V0 0.333 × V0 0.666 × V0 0.333 × V0 00 3:1 1/3 2.0 V to supply 0.666 × V0 0.333 × V0 0.666 × V0 0.333 × V0 00 2:1 1/3 2.0 V to supply 0.666 × V0 0.333 × V0 0.666 × V0 0.333 × V0 32.3.1.2 LCD Driver Block The LCD driver block is associated with each GPIO. The output of LCD DAC through MUX is provided to the LCD driver block to drive the LCD glass. Figure 32-3 shows the architecture of the LCD driver block. Figure 32-3. LCD Driver Block Inputs from LCD DAC pwrdn_n (global) V1 V2 V3 V4 GND od_h (global) 1 0 1 0 1 0 dispbInk (global) 1 0 hold_n_hw (global) V0 com_seg (individual) LCD Bias LCD Bias Generator pwrdn_n (individual) 00 01 10 11 hold_n_hv (global) Disp_data (individual) fr (global) drvr_in mode[2:0] (global) Buffer drive (global) bypass_en (global) pts (global) ESD Devices drvr_out 348 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F LCD Direct Drive The LCD driver contains three major blocks: ■ Buffer and associated control logic for power modes ■ 4:1 Output multiplexer ■ Common/Segment switches As shown in Figure 32-3 on page 348, the LCD driver block receives bias voltages V0 to V4 and GND voltage. It passes through a set of 2:1 muxes controlled by the COM-SEG bit of the PRT[x]_LCD_COM_SEG register. This register configures the pin as either a common or segment drive pin. If the bit is set, it configures the corresponding pin as common; otherwise, it is configured as a segment drive pin. As shown in Figure 32-3, V4 and GND voltages are forwarded to the next mux. If the pin is selected as a segment line, then V0, V2, V3, and GND are forwarded. These are the only voltages required at common and segment lines for any bias ratio, multiplex ratio, and LCD update state. Out of these four bias levels, only one level is selected by the 4:1 multiplexer. The select lines of the multiplexer are driven by display data and the frame signal. Frame is a global signal driven by the UDB control logic. This signal toggles every time the LCD waveform needs to be updated. Table 32-3 shows the 4:1 multiplexer output and driver input for different combinations of COM_SEG, DISP_DATA and the frame signal. Table 32-3. LCD DAC Output Selection com_seg disp_data fr drvr_in/out 0 0 0 V3 0 0 1 V2 0 1 0 GND 0 1 1 V0 1 0 0 V4 1 0 1 V1 1 1 0 V0 1 1 1 GND Note For proper functionality of the LCD driver, the port reset drive mode (PRTxRDM[1:0]) NVL bits of any port used for LCD drive must be configured as a high-Z input (00b). By default, these bits are set to the proper configuration. Ensure that the bits are not altered in an LCD application; otherwise, the drivers will fail to produce proper LCD drive signals. For more details on these registers and their configuration, see the Nonvolatile Latch chapter on page 85. 32.3.1.2.1 Buffer Modes The output of the 4:1 multiplexer is driven to the buffer, which drives the common or segment line of the LCD. The buffer in the LCD drive block has eight modes of operation, selectable from the Mode[2:0] bits. Mode[0] comes from PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F LCDDRV_CR[1]; the remaining two bits are driven from the UDB through the digital system interconnect (DSI). Each mode has a different power drive capability. Depending on the LCD, the appropriate one can be used to eliminate AC coupling between segment and common lines. Note that these buffer power modes are different than the I/O drive modes. Table 32-4. LCD Drive Modes Control Bits Mode Drive Strength Mode[2] Mode[1] Mode[0] 0 0 0 High Drive Seg = 1x, com = 1x 0 0 1 High Drive Seg = 1x, com = 2x 0 1 0 High Drive Seg = 1x, com = 4x 0 1 1 High Drive Seg = 2x, com = 2x 1 0 0 High Drive Seg = 2x, com = 4x 1 0 1 High Drive Seg = 4x, com = 4x 1 1 0 Low Drive Seg = 0.1x, com = 0.1x 1 1 1 Low Drive Seg = 0.2x, com = 0.2x The LCD display size and capacitance and the application power budget are two criteria for selecting buffer modes. The buffer is enabled only when the drive signal is high. Drive signal high time can be configured according to the application requirements. The drive current provided by the High Drive mode of the buffer (the mode that is normally used) is high, so it charges the pixel capacitance quickly. The disadvantage of this is higher power consumption. The time for which the buffer is kept on depends on the power budget and the LCD waveform's rise time requirements. The Low Drive mode of the buffer and the DAC are other options. It is possible to dynamically select the Low Drive mode by two mode control signals generated by the UDB. You do this in the case of extremely leaky glasses, when it is preferable to use the buffer to drive the LCD continuously throughout the refresh period. This is more effective than using the DAC, whose current drive ability is lower than that of the buffer Low Drive mode. Use the DAC when you have normal glasses and the charge leakage is small. If the leakage is small enough for the offset to be negligible, then the pin can be tristated by clearing the bypass_en bit, after charging the pixel using the High Drive buffer mode. In normal operation, the buffer in High Drive mode drives the LCD for a while, then a low-power source (either the DAC or the buffer in Low Drive mode) takes over and drives the LCD for the remaining time. ■ When using High Drive and DAC: Initially, for some period of time, the buffer quickly charges the LCD pixel capacitance near to the desired value. Later, when the drive signal goes low, the DAC directly drives the LCD for the remaining period (if the bypass_en bit is set) to sustain the voltage at the LCD 349 LCD Direct Drive pin. If the bypass_en bit is not set to 1, the pin is tristated and no source drives the LCD. This can lead to charge leakage from the pixel capacitance. ■ When using High Drive and Low Drive: The drive signal always remains high. This means that the buffer is always enabled. The UDB controls the time for which the buffer remains in High Drive and Low Drive modes. 32.3.1.2.2 LCD Driver Bias Generator The LCD bias generator block creates a bandgap-based voltage reference for the LCD driver block. The input to this block is a 2.5-µA bandgap current. The output is a bias voltage and the associated ground line. Figure 32-5 shows various control signals to the LCD driver block. Figure 32-4. The DAC Charging an LCD Segment Pin in Two Different States Buffer Driving LCD DAC Driving LCD Drive t Pixel Voltage t Slope depends on drive mode of buffer selected. Figure 32-5. Control Signals of LCD Driver Block Type A 1 Frame *1:4 Multiplex Ratio 1/3 Bias Type B (2x Type A Data Clock) 1 Frame 1 Frame LCD Drive Voltage Level Display Data Frame LCD CLK 350 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F LCD Direct Drive 32.3.1.3 UDB The UDB performs the following actions in the LCD system: ■ Triggers the DMA periodically to bring the display data from SRAM to the port data registers ■ Generates various control signals for the functioning of the LCD system hardware ❐ The drive signal, which is used to enable the driver buffer ❐ Two mode control signals for the buffer ❐ A synchronous LCD CLK, which is used to latch the port data register value for a particular pin ❐ The frame signal The clock for the UDB is derived from the IMO. The clock value changes with the refresh rate and the number of commons of LCD. 32.3.1.4 DMA DMA is used to transfer the display data into various port data registers. The display data is stored in SRAM. Data transfer is initiated by the UDB at the beginning of the LCD refresh cycle. Depending on which and how many ports are configured for the LCD drive, several transaction descriptors (TDs) associated with the DMA channel may need to be chained together. There is no separate display memory, as such, in PSoC. Display data resides in the SRAM connected to the peripheral hub (PHUB). The image/display buffer can be any block of available memory. To work more effectively with the DMA in transferring data to the LCD drivers, port data registers are aliased to a separate contiguous region in the memory map. These PRTx_DR_ALIAS registers are contiguous, to reduce the number of TDs required to move data. An additional set of registers (per port), the PRTx_BIT_MASK registers, mask off the write capability to the PRTx_DR_ALIAS registers on a bit level. This is an advantage if all of the pins on a given port are not being used for LCD; the unused pins can be masked off and used for other purposes. The port data register (PRTx_DR) can still be used to address pins masked off in the aliased data registers. 32.4 LCD Low-Power Mode This mode is useful when LCD is required to be functional while the device is in low-power mode. This requires special hardware and firmware logic to wake the system up at regular intervals, refresh the LCD, and put the device back to sleep. Periodic refresh should happen at the specified rate, even if there are other interrupts in the system. LCD low-power mode uses all the of components that are used for LCD always active mode. In addition to this, it also uses a programmable wakeup source and small dedicated digital logic to allow bug-free transitions to and from the lowpower mode. Figure 32-6 on page 351 shows the block diagram for the LCD low-power mode. To use the LCD in low-power modes, the LCD clock must be sourced by Digital Clock 0 (see clk_d0 in Figure 14-9). Figure 32-6. LCD Low-Power Mode Analog Global Bus System Resources Bias Select (LCDDAC.CR0[1:0]) Contrast Control (LCDDAC.CR1[5:0]) Continuous Drive (LCDDAC.CR0[3]) Dedicated LCD Component Hardware LCDTIMER_CFG[1] LCD DAC 1-kHz ILO CLK 8-kHz OPPS LCD Timer LCD INT Clock UDB Drive / LCD CLK / Frame / Mode[2:1] / LP_ACK LCDDAC.SW 0/1/2/3/4 V0,V1,V2, V3,V4, GND Frame Data 0x01 DMA drq LCD CLK Drive LCD CLK 0x00 Frame nrq LCD Driver Block Pin Mode[2:0] Display Data . . . RAM LCD Bias drq DMA LCD Bias Generator .. . . . . LCD CLK LCD Bias Port Data Registers Drive Frame LCD Driver Block Pin Mode[2:0] LCD Bias PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 351 LCD Direct Drive A complete functional LCD low-power system is formed using these major blocks: ■ ■ Dedicated LCD hardware ❐ LCD timer ❐ LCD DAC ❐ LCD driver ❐ LCD bias generator System resources ❐ Clocks: 1-kHz ILO and 8-kHz one pulse per second (OPPS) ❐ UDB implementation for sleep acknowledgement ❐ DMA for frame data transfer ❐ UDB implementation for control signal generation (frame, drive, LCD mode, LCD CLK) ❐ DMA for display data transfer The blocks in bold are unique to the LCD low-power system. The other blocks are same as the LCD always active system. What makes the LCD low-power system different from the LCD always active system? ■ It can wake the system ■ It can continuously drive the LCD even when the chip is put in low-power mode PSoC 5LP contains several clock sources that operate during device low-power mode. ILO and OPPS timer are examples. These clock sources are used to trigger periodic interrupts to the device to wake the system up. As shown in Figure 32-6, these two clock sources are selectable using a mux. The selected clock is fed to the 6-bit LCD timer. It is a continuously running timer; that is, when the timer overflows, the original period is reloaded in the timer register. The terminal count pulse from this timer triggers the interrupt to the chip. This restores the main clocks of the chip. When this happens, the interrupt signal from the LCD timer is intercepted by the UDB-implemented pulse generator. In response, the block generates a synchronous clock that causes several operations. See 32.4.1.2 UDB on page 352 for more details. Overall, the UDB's role is to provide control signals to various functional blocks of the LCD low-power system. nal (LP_ACK signal shown in Figure 32-6) is generated from the UDB, which keeps the LP command from the CPU on hold until the LCD refresh is completed. This control signal is driven to the power management controller of the device. There are two DMAs used in this architecture. One DMA is used for the transfer of display data to the port data register, which is the same as in an LCD always active system. the other DMA is used to update the frame information into the control register of the UDB each time the chip wakes up. 32.4.1 Functional Description This section gives details of the blocks and features used specifically in LCD low-power mode. 32.4.1.1 LCD Timer The LCD timer is a 6-bit timer dedicated only for the LCD drive application. Its period is set based on the required refresh rate of the LCD. The period of this timer can be configured by setting the period field of the LCDTMR_CFG register. There are two options for the LCD timer clock source: ■ 1-kHz ILO ■ 8-kHz OPPS. This requires that an external 32-kHz crystal be connected to the system. The source can be selected by setting the clk_sel field of the LCDTIMER_CFG register. The clock timer provides periodic interrupts to the system PM controller. The interrupt signal is also driven to the UDB to generate the LCD CLK signal. 32.4.1.2 UDB LCD low-power mode uses the UDB to generate various signals that control the functioning of the LCD system. These control signals are generated using the functional blocks listed below: ■ Pulse generator ■ BGREF timer ■ Drive pulse-width modulator (PWM) ■ Control register for frame data ■ Mode control signals to the LCD driver At this time, the system must be put back to sleep after the LCD refresh. In an LCD low-power system, the CPU issues a chip low-power (LP) mode command to the power management (PM) controller. (For this, firmware needs to be structured in a specific way, as explained in a later section.) Consent is given by the LCD hardware. This is because the LCD refresh happens in hardware and CPU does not know when it is completed. So, a control sig- 352 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F LCD Direct Drive Figure 32-7. LCD UDB Logic Pulse Generator LCD CLK DRQ DMA (Frame Data) NRQ BGREF Timer Drive PWM TC EN PWM Drive S SR Latch Q LP_ACK R Control Register Const Frame Mode [2:1] UDB The pulse generator samples the interrupt signal from the LCD timer; in response, it generates one synchronous clock pulse (LCD CLK), which is routed to the BGREF timer and DMA (for frame data). This synchronous clock triggers these operations: ■ Puts the sleep command issued by CPU, if any, on hold (using signal LP_ACK) until LCD refresh operation is completed. ■ Enables the BGREF timer. The BGREF timer is used to provide a 2.5-µs delay, which is necessary to stabilize the bandgap reference circuit. ■ Triggers the DMA to transfer the frame data into the UDB control register. Frame is a square wave signal that is used for proper sequencing of LCD refresh action. Each cycle of the frame signal represents one common update state. After the DMA transfer for frame data and the BGREF timeout are completed, Drive PWM is enabled. The Drive PWM output “Drive” signal is routed to all the LCD driver blocks associated with the GPIO. It enables the LCD buffer to drive PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F the LCD glass. The UDB also provides the two signals that set the drive mode of the LCD buffer. 32.4.1.3 DMA Two DMA channels are used by the LCD component for: ■ Transferring the frame information into the control register of the UDB from the system memory (RAM) ■ Transferring the display information from system memory (RAM) into the port register 32.4.1.4 LCD DAC and Driver: Low Power Feature The LCD DAC and driver have some features that are useful for LCD low-power mode functioning and help to achieve the lowest power consumption when the LCD system is shut down. The LCD DAC can remain active when the chip is put in sleep mode. In this mode, the DAC can continue to drive the inputs of LCD drivers. To enable this mode, set the continuous_drive bit in the LCDDAC.CR0 register to 1. The 353 LCD Direct Drive LCD DAC receives a pwrdn signal, which shuts the DAC off when it is HIGH. The LCD driver receives a display blank signal, dispbInk, controlled by the LCDDRV.CR register. This signal sets the output to be either tristated or grounded when the chip is in low-power mode. This function works when the power down signal (pwrdn_n) signal is low. The pwrdn_n signal is used when the LCD system needs to be shut down. Thus, for operation in sleep mode, for an LCD low-power system, continuous_drive, bypass_en, and pwrdn bit must be set to 1, and pwrdn_n must be set to 0. This causes the DAC to directly drive the LCD, bypassing the LCD driver section, which is shut down in chip low-power mode. The various operating modes of the LCD DAC and LCD driver are summarized in Table 32-5 and Table 32-6 on page 354. The buffer present in the LCD driver can be bypassed by setting the bypass_en bit of the LCDDRV.CR register to 1. Table 32-5. LCD DAC Operating Modes Chip Mode Block Mode pwrdn_n continuous_drive Description Active Active 1 X LCD DAC is active. It can drive I/Os or LCD drivers depending on the LCD driver mode. Sleep Sleep with bypass drive 0 1 LCD DAC is active and driving I/Os even though the chip is in sleep. LCD drivers are bypassed. Active/ Sleep/ Hibernate OFF 0 0 LCD DAC is powered down Table 32-6. LCD Driver Operating Modes Chip Mode Block Mode pwrdn_n dispbInk Drive bypass_en Active Active drive 1 X 1 X LCD driver is driving the pin in one of the High Drive or Low Drive modes. Active Active with bypass drive 1 X 0 1 LCD driver is bypassed. LCD DAC is driving the I/O. Active Active with tristate drive 1 X 0 0 LCD driver is active but the I/O is tristated. Sleep Off with bypass drive 0 0 X 1 LCD driver is powered down. LCD DAC is in sleep with bypass drive mode and driving the I/O. Active/ Sleep/ Hibernate Off with ground drive 0 1 X X LCD driver is powered down. Output is grounded. This is the power down mode for LCD applications. LCD DAC is off. Active/ Sleep/ Hibernate Off with tristate drive 0 0 X 0 LCD driver is powered down. Output is tri-stated. This is the power down mode for nonLCD applications. LCD DAC is off. 354 Description PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F LCD Direct Drive 32.4.2 Timing Diagram for LCD Low-Power Mode Figure 32-8 shows the timing in low-power mode. Figure 32-8. LCD Low-Power Mode Timing Diagram ILO LCD COUNT 1 P 0 P- 1 P- 2 LCD TC LCD INT BUS CLOCK ( Wakeup) ( Sleep) LCD CLK BGREF TIMER P 0 BGREF TIMER TC DMA Frame( current) Display Data( next) DMA TERMOUT( Frame) PWM( DRIVE_EN) “AND” PWM TC LP_ ACK A refresh timer overflow triggers an interrupt to the PM system and also drives the UDB pulse generator logic. After a few microseconds, system clocks are restored. This puts all of the resources on the chip in operation. The UDB-implemented pulse generator outputs an LCD CLK pulse, which: ■ Triggers the DMA to transfer frame information into the control register of the UDB ■ Enables the BGREF timer (implemented using UDB) ■ Copies the display data from the port data register into the driver for the present LCD state ■ Clears the refresh rate timer interrupt ■ Puts the sleep command from the CPU on hold After the frame information transfer, another DMA is triggered to transfer the display data into the port data register for the next LCD state. When the frame data transfer is completed and the BGREF timer overflows, the LCD drive buffer is enabled using the drive signal from the Drive PWM. This is when LCD glass refresh begins. The drive mode of the LCD drive buffer determines the current drive. After the drive time is set, the drive line goes low, disabling the buffer. This also releases the sleep command hold set by the LCD CLK. This causes PM to execute the sleep command issued by the CPU. During the rest of the period, the LCD is driven continuously from the LCD DAC, bypassing the driver buffer. Figure 32-9 on page 356 shows the sequence of operations and relative current consumption for low-power mode. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 355 LCD Direct Drive Figure 32-9. LCD Sequence of Operation NOT TO SCALE Chip Wake Up from LCD Source Chip Power Mode Active Sleep Sleep D E DMA Track Main/CPU Track B A End LCD Drive Pulse (LCD lp_ack Asserts) Turn On CPU and Bus Clock Relative Current Consumption Power Phase H G F Analog Track C Begin LCD Drive Pulse Chip Transition into Active Mode Turn On LCD DAC and Bias Generator 0 1 2 Chip Transition to Sleep Mode 3 Main/CPU Track A) Chip wake up process B) ‘Main’ execution: Check for interrupts, request sleep C) Power Manager (PM) asserts low power request (lp_req) to all subsystems and waits for all acknowledge signals (lp_ack) to assert H) PM Completes Active -> Sleep Mode transition DMA Track D) DMA TD: Update FR value (must complete before starting ‘G’) E) DMA TD: Setup Display Data for next LCD refresh cycle Analog Track F) LCD DAC and LCD bias generator power up (2.5 µs) (must complete before starting ‘G’) G) “Drive” pulse 356 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F LCD Direct Drive 32.5 LCD Usage Models The LCD can be used in these cases: ■ The chip is always maintained in active mode. The LCD driver buffer will drive in high drive mode for the specified time; later on, it will switch back to low drive mode. This mode can be used when the system is always on and a power saving feature is not needed. This uses LCD always active mode. ■ The chip enters low-power mode and the LCD does not need to function. Disable the entire LCD system before putting the device to low-power mode. This also uses LCD always active mode. ■ The chip enters low-power mode and the LCD must be functional. In this situation, the background LCD refresh timer allows the chip to be put to sleep and awakened at regular intervals to refresh the LCD glass. This system uses LCD low-power mode. There are restrictions in refresh rates due to the low frequency clock used for the LCD timer. Table 32-7 shows the allowed refresh rate values for this case: Table 32-7. Refresh Rate Limits ILO Commons Max ECO Min Max Min 2 125 21 128 32 4 125 21 128 32 8 63 21 128 20 16 31 - 128 20 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 357 LCD Direct Drive 358 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 33. CapSense PSoC® 5LP devices have a capacitive sensing feature called CapSense®. This feature allows users to take advantage of the capacitive properties of their fingers to toggle aesthetically superior buttons, sliders, and wheels. Touch pads and touchscreens are common examples of capacitive sensing interfaces. The underlying principle of these technologies is the measurement of capacitance between a plate (the sensor) and its environment. 33.1 Features Features of CapSense include: ■ Resources to support two capacitive sensors scanning simultaneously ■ Configurable low-pass filter to remove switching noise for accurate measurement ■ Reference buffer with High Drive mode for faster measurement 33.2 Block Diagram Figure 33-1 shows a block diagram of the overall capacitive sensing architecture. Figure 33-1. CapSense Module Block Diagram I/O Pins I/O Pins I/Os (Left Side) Reference Driver Reference Driver AGL<0> I/Os (Right side) AGR<0> UDBs AMUXBUSL AMUXBUSR MUX LPF V-I DAC Comparators MUX LPF V-I DAC Comparators System Bus LEFT SIDE PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F RIGHT SIDE 359 CapSense 33.3 How It Works connected to a large off-chip capacitor serving as integration or modulation capacitor. The PSoC device has configurable hardware for CapSense to optimize factors such as speed, power, sensitivity, noise immunity, and resource usage. It implements CapSense Sigma Delta (CSD) method of capacitive sensing. 33.3.4 33.3.1 Figure 33-2. GPIO Structure Reference Driver GPIO Configuration for CapSense The GPIO switching structure supporting CapSense is shown in Figure 33-2. This driver is used to quickly initialize nets to a voltage independent of the power supply. This ability speeds up capacitive scanning and improves power supply rejection ratio (PSRR). Two reference drivers operate independently; one drives to AMUXBUSL and one for AMUXBUSR. The driver is connected to the AMUXBUS by setting the out_en bit in the {CAPSx_CFG0}. AMUXBUSx Vdd The reference driver supports Normal and High drive modes; the drive mode is selected using the boost bit in the {CAPSx_CFG0} register. In Normal mode, capacitances up to 100 pF can be driven in less than 600 ns. In High mode, capacitances up to 30 nF can be driven in less than 15 µs. 33.3.2 Low-pass Filter Two tunable Low-pass Filter (LPF) blocks are available. The inputs are selectable in a 2:1 mux for each LPF. On the left side, the LPF inputs are AMUXBUSL and AGL[0]; on the right side, the inputs are AMUXBUSR and AGR[0]. LPF input is selected by using the swin[1:0] bits in the LPFx.CR0 register. The outputs are connected through switches to abusl[0] and abusr[0], respectively. The tunability of the LPF allows the user to select a (nominal) R of either 200 k or 1000 k, and a C of either 5 pF or 10 pF. The rsel and csel bits in the LPFx_CR0 register are used to select resistance and capacitance respectively. The LPF control registers are LPF0_CR0 and LPF1_CR0. 33.3.3 Analog Mux Bus All GPIO pins support CapSense operations except SIO and USB pins. The primary analog mux bus for CapSense is the AMUXBUS, which has two nets (AMUXBUSL and AMUXBUSR) for two simultaneous sensing operations. These can also be shorted to form a single net that connects to all GPIO. See the device datasheet for details about GPIOs available in each package and the Analog Routing chapter on page 319 for a diagram of AMUXBUS connectivity for the GPIO. AMUXBUSL and AMUXBUSR nets connect to all GPIO pins on their respective halves of the device. CapSense uses the AMUXBUS net, along with an analog global net (AGR[0] with AMUXBUSR, and AGL[0] with AMUXBUSL) to provide feedback to the reference driver. This feedback is from a pin 360 The port analog global mux register (PRT[x]_AMUX) is used to connect the port pin to the analog mux bus. The pull up or pull down is enabled using io_ctrl[1:0] bits in the CAPSx_CFG1 register. Sense capacitance is switched in two configurations, shown in Figure 33-3 on page 361 and Figure 33-4 on page 361, to convert the capacitance into equivalent resistance for measurement. The equivalent resistance can be calculated as: 1 R s = ------------- fs Cs Here: Cs=Sensor Capacitance 1 and 2 = Non-overlapping clocks, which may be configured in a pseudo random sequence (PRS). fs = Frequency of the clock Cmod = External Modulation Capacitance The CapSense methods can generally be done with either switching high or switching low at the GPIO pin. The rest of the hardware is configured with the appropriate polarity to match to the pull up or pull down choice. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F CapSense Vdd AMUXBUSx Vdd 1 RS AMUXBUSx Figure 33-3. Charging MUXBUS Through Sense Resistor 2 CS AMUXBUSx AMUXBUSx Figure 33-4. Discharging the MUXBUS Through Sense Resistor 2 CS 1 RS The CapSense clock is used for switching. Two alternatives are available to generate the CapSense clock (see also Figure 19-1 on page 152). ■ ■ The UDB generates two global clocks (caps_dsi_lft and caps_dsi_rt), and routes to GPIO logic of the I/O pins in the respective side. The PRT[x]_CAPS_SEL[y] registers (per port per pin basis) are set to select the global clock for switching the sensor during measurement. The DSI output to the I/O pin can be used to source the CapSense clock from the UDB. The PRTx_BIE[y] must be programmed for input (per port per pin basis) and PRT[x]_CAPS_SEL[y] is cleared to select the DSI output signal for the CapSense clock. With either of these paths, the nonoverlapping clock phases discussed above are automatically generated within the GPIO switching structure. GPIOs pins can be made as Shield Electrodes. The shield electrodes help in reliable operation in presence of water film or water droplets. The effect of these factors on shield electrode is measured and is removed from the CapSense buttons. The CapSense algorithms discussed below support the shield electrode. 33.3.5 Other Resources CSD CapSense techniques use many resources in PSoC 5LP devices. These include UDBs, Comparators, and V-I DAC. See the Universal Digital Blocks (UDBs) chapter on page 175, Comparators chapter on page 335, and Digital-to-Analog Converter chapter on page 369 for more details. Note that to connect an external integration capacitance (Cmod) statically (without switching), connect it to AMUXBUS using PRT[x]_AMUX register and then PRTx_CAPS_SEL[y] = 0 and PRTx_BIE[y] = 0. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 361 CapSense 33.4 CapSense Delta Sigma Algorithm 3. As the integration capacitor voltage moves back and forth across the comparator threshold, the comparator high outputs are counted in an interval to give a measure of the sense capacitor. The CapSense Delta Sigma (CSD) algorithm shown in Figure 33-5 and Figure 33-6 on page 363 measures capacitance with the hardware configured like a Delta Sigma modulator. Delta Sigma capacitive sensing operates by holding an integration capacitor voltage near a target threshold, and charging or discharging the capacitor, based on the present state of a comparator output. The sense capacitor is continuously switched between Vdd and the integration capacitor, which drives the integrated voltage up on each switching cycle. The CSD algorithm operates as follows: 4. The sense capacitance increases with touch, therefore equivalent resistance decreases. This decreased resistance causes an increase in the current flowing through switch CapSense resistor. 5. To maintain the voltage on Cmod near VREF during a touch, the IDAC sinks current for longer duration to compensate for the larger sense capacitance. This changes the count value accordingly. A PRS (pseudo random sequence) clock may be used instead of a fixed clock source to drive the precharge switches. The PRS clock produces less radiated noise on the sense capacitor, compared to a fixed clock source, hence improving EMI and interference performance. 1. When the integration voltage reaches the reference voltage, the comparator enables current DAC to discharge the capacitor. 2. When the capacitor voltage discharges below the reference voltage, the current DAC is disabled to allow the capacitor to continue charging. Figure 33-5. CSD Hardware Configuration VDD UDB 1 PRSCLK Prescale and PRS 2 CS IDAC En I LPF AMUXBUS C MOD Initialize Ref Driver Vin UDB D Vref Q Vmod UDB Counter C UDB CounterClock Prescale 362 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F CapSense Figure 33-6. CSD Waveform V CMOD Voltage Comp Out when No Touch Smaller count Comp Out when Finger is present Larger Count t The PSoC device also supports other variants of the CSD algorithm as follows: ■ Switched capacitor resistor (see Figure 33-3 on page 361) is used to charge the integration capacitor; an external bleeding resistor is used (instead of IDAC) to discharge the integration capacitor, based on comparator output. ■ Polarities are reversed so that the IDAC is used to charge up the integration capacitor and switched capacitor resistor (see Figure 33-4 on page 361) discharges the integration capacitor toward ground, based on comparator output. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 363 CapSense 364 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 34. Temperature Sensor The PSoC® 5LP devices have an on-chip temperature sensor that is used to measure the internal die temperature. The temperature sensor uses the Delta Vbe method for digital temperature measurement. The temperature sensor block has an auxiliary analog-to-digital converter (ADC) to measure the internal die temperature. The auxiliary ADC is a 10-bit accurate ADC in the system performance controller (SPC) primarily designed for measuring temperature sensor output. It is also possible to route the analog output of diode in temperature sensor block to analog globals to measure temperature using the higher resolution Delta-Sigma ADC in PSoC 5LP. 34.1 Features The temperature sensor offers the following features: ■ ± 5 degrees Celsius accuracy over commercial temperature range (–50ºC to +150ºC) ■ Ability to route temperature sensor output to analog global line, AGL3. 34.2 Block Diagram The block diagram for the temperature sensor is illustrated in Figure 34-1. Figure 34-1. Temperature Sensor Temperature Sensor Core Switch Network (Sequencer + Mode Select) Parallel Current Paths To AGL3 Digital Temperature Vbe Temperature Diode Curvature Compensation Circuit PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Auxiliary ADC 365 Temperature Sensor 34.3 How It Works The base-to-emitter voltage of a Bipolar Junction Transistor (BJT) device has a strong dependence on temperature at a constant collector current and zero collector-base voltage. The temperature sensor output (Vbe) is measured with two different drive currents: first with low bias current and second with high bias current. A current ratio of 1:29 is maintained between the conversions. By making the ratio between the two drive currents high, the voltage difference between the Vbe values is linearly proportional to temperature. The output voltage of the temperature sensor is either driven to the Delta Sigma ADC or other on-chip resources using analog global line (AGL3). To increase accuracy, the PSoC 5LP temperature sensors use the following techniques: ■ Dynamic Element Matching technique is implemented using a sequencer that cyclically selects among the eight current mirror paths during conversion (low current mode and high current mode). ■ Curvature compensation circuit to increase linearity when the temperature sensor output is routed to an external resource with a High Z buffer such as the onchip Delta Sigma ADC. ■ A two-point linear fit calibration routine for accurate temperature measurements using the auxiliary ADC. 34.4 Command and Status Interface The commands associated with the temperature sensor are executed through the simple command/status register interface. “Get Temp,” “Setup Temperature Sensor,” and “Disable Temperature Sensor” are commands associated with the temperature sensor. The command is sent as a series of bytes to either SPC_CPU_DATA or SPC_DMA_DATA, depending on the source of the command. Response data is read via the same register to which the command was sent. The status register, SPC_SR, indicates whether a new command can be accepted, when data is available for the most recent command, and success/failure response (status code) for the most recent command. Table 34-1. Command Registers Register Size (Bits) SPC_CPU_DATA 8 Data to or from CPU SPC_DMA_DATA 8 Data to or from DMAC SPC_SR 8 Status – ready, data available, status code 366 Description The command sequence consists of a 2-byte key, followed by command code and the parameters associated with the command. ■ Key byte #1 – always 0xB6 ■ Key byte #2 – 0xD3 plus the command code (ignore overflow) ■ Command code byte ■ Command parameter bytes ■ Command data bytes Before sending a command to the SPC_CPU_DATA or SPC_DMA_DATA register, the SPC_Idle bit in SPC_SR[1] must be ‘1’. SPC_Idle will go to ‘0’ when the first byte of a command (0xB6) is written to a DATA register, and then go back to ‘1’ when command execution is complete or an error is detected. Commands sent to either DATA register while SPC_Idle is ‘0’ are ignored. 34.4.1 Status Codes If the value of the 2-byte key is wrong or if any of the parameters passed are invalid, the command is ignored and the error condition is indicated by the status code in the Status register (SPC_SR). The Status_Code bits (7:2 in the Status register) are used to determine if the command operation is executed successfully or any error occurred. Table 34-2 lists the status code bit values. Table 34-2. Status Code Bit Values Status_Code Bit Values (Bits[7:2] in SPC_SR register) 0x00 34.4.2 34.4.2.1 Description Command successfully executed 0x02 Invalid key 0x0B Invalid command code 0x0D Invalid parameter 0x0E Temperature Sensor Vbe is currently driven to an external device Temperature Sensor Commands Get Temperature “Get Temperature” (command code: 0x0E). This command uses auxiliary ADC to measure the die temperature and the ADC output. It returns 2 bytes corresponding to a temperature value. The first byte is the sign of the temperature (0 = negative, 1 = positive). The second byte is the magnitude. These values are read from the SPC Data register. The command sequence is shown in Figure 34-2. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Temperature Sensor Figure 34-2. Get Temperature Command Sequence 2 -B y te K e y 0 x B 6 C o m m a n d 0 x E 1 0 x 0 E P a ra m e te r B y te s n u m S a m p C o m m a n d 0 x D 3 + C o m m a n d C o d e C o d e Command Parameters numSamp This parameter specifies the number of samples taken. The number of samples is equal to 2^numSamp. Valid values for this parameter are 0, 1, 2, 3, 4, or 5, thereby resulting in 1, 2, 4, 8, 16, or 32 samples, respectively. The ADC output is read after the averaging is done over all the samples as specified by this parameter. The averaging routine can be bypassed by selecting the numSamp value as 0. Reading Temperature Output After the command and its parameters are sent, the Temperature Sensor/ADC block is configured and starts the conversion. When the conversion is complete, the DATA READY bit in the Status register (SPC_SR) is set. The CPU must poll this bit to check if the ADC output is ready. When the bit is high, the first byte (Sign byte) of output is read from the Data register (SPC_CPU_DATA). The DATA READY bit is reset when a read operation is done. When the second byte (Magnitude byte) is ready to read, the DATA READY bit becomes high again and the second byte is read from the Data register (SPC_CPU_DATA). 34.4.2.2 Setup Temperature Sensor “Setup Temperature Sensor” (command code: 0x11). The purpose of this command is to connect the raw temperature sensor analog output onto AGL3 for measurement by the High Z buffer/Delta Sigma ADC(DSM) or other external resources. The auxiliary ADC cannot be operated at the same time when the sensor output is routed to AGL3. This command disables the functionality of the auxiliary ADC such that it does not load the sensor when the sensor output voltage is being driven into the DSM or other external ADCs. The “Setup Temperature Sensor” and “Disable Temperature Sensor” are the commands associated with this purpose and drive the temperature sensor output to AGL3. When temperature sensor output is routed to an analog global line, auxiliary ADC cannot be used to measure the temperature. Note that AGL3 should not be used by analog blocks other than the temperature sensor output when this command is executed. Even though PSoC Creator takes care of routing, ensure that there are no resource conflicts in using AGL3. The command sequence is shown in Figure 34-3. Figure 34-3. Setup Temperature Command Sequence Command Parameter Bytes 2-Byte Key 0xB6 0xE4 0x11 Sequence Select Sequence Freeze clkDivider Curvature Compensation Enable Command Code 0xD3 + Command Code PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 367 Temperature Sensor Command Parameters 34.4.2.3 Sequence Select. The temperature sensor output (Vbe) voltage is measured with low bias current and then with high bias current. A current ratio of 1:29 is established between the low bias and high bias current. This ratio is fixed and not configurable.The difference between the two output voltages is linearly proportional to temperature. “Disable Temperature Sensor” (Command code: 0x12). This command is used to disable the temperature sensor from driving its output voltage to the analog global line (AGL3). After calling this command, the “Get Temp” command can be executed, as well as commands using the erase portion of the Smart Write algorithm. This command has no parameters and does not return any value. The command sequence is shown in Figure 34-4. ■ 0 – Low bias current. The temperature sensor is driven with low bias current. ■ 1 – High bias current. The temperature sensor is driven with high bias current. Sequence Freeze. In low bias and high bias current modes, Dynamic Element Matching (DEM) is implemented by a sequencer that cyclically selects among the eight current mirror paths. ■ 0 – Sequencer is enabled. ■ 1 – Sequencer is disabled. No cycling of the current paths occurs. Disable Temperature Sensor Figure 34-4. Disable Temperature Command Sequence 2-byte Key 0xB6 0xE5 0xD3 + Command Code 0x12 Command Code clkDivider. This parameter sets the divider value for clock generation from the SPC clock (spcCLK, which is 36 MHz). This clock is used by the sequencer to cycle through the current mirrors. The clock frequency is equal to: spcCLK ----------------------------------------- clkDivider + 1 Equation 1 The clock divider value (clkDivider) is of 8 bits allowing clock to have 256 different frequencies ranging from spcCLK down to spcCLK/256 (spcCLK is 36 MHz). In general, the slower the clock, the better the linearity that will be achieved. Curvature Compensation Enable. The temperature sensor has a feature to correct for a curvature in its behavior and align it to a more linear path, thus giving it more accuracy when its output is routed to an external resource with a High Z buffer, such as the on-chip Delta Sigma ADC. A High Z buffer is required because the curvature compensation circuit needs to be buffered before driving an external ADC front end. ■ 0 – No curvature compensation is used. ■ 1 – Curvature compensation is enabled. 368 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 35. Digital-to-Analog Converter The 8-bit digital-to-analog converter (DAC) is configured to output either a voltage or a current. The 8-bit DAC supports CapSense®, power supply regulation, and waveform generation. 35.1 Features The DAC has the following features: ■ Adjustable voltage or current output in 255 steps ■ Programmable step size (range selection) ■ Eight bits of calibration to correct ± 25% of gain error ■ Source/sink option for current output ■ Output rate for current IDAC output: 8 Msps ■ Output rate for VDAC voltage output: 1 Msps ■ Monotonic in nature 35.2 Block Diagram A block diagram of the DAC is shown in Figure 35-1. Figure 35-1. DAC Block Diagram ISOURCE Range 1x, 8x, 64x Reference Source Scaler Vout Iout R 3R ISINK Range 1x, 8x, 64x PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 369 Digital-to-Analog Converter 35.3 How It Works This DAC generates either a voltage or a current output. It is built using current mirror architecture; current is mirrored from a reference source to a mirror DAC. Calibration and value current mirrors are responsible for the 8-bit calibration [DACx_TR] and the 8-bit DAC value. The current is then diverted into the scaler to generate the current corresponding to the DAC value. The DAC value can either be given from the register DACx_D or from 8 lines from the UDB. This selection is made using the DACx_CR1[5] bit. Using the UDB to write the DAC value uses the DAC bus. Because there is only one DAC bus available for each device, this bus must be shared by all the DACs in the device. The DAC is strobed to get its output to change for the input code. The strobe control is enabled by the DACx_STROBE[3] bit. The strobe sources for the DAC can be selected from the bus write strobe, analog clock strobe to any UDB signal strobe. This selection is done on the basis of setting in DACx_STROBE[2:0]. ■ Current (IDAC) Mode – The two mirrors for the current source and sink provide output as a current source or current sink, respectively. These mirrors also provide range options in the current mode. ■ Voltage (VDAC) Mode – The current is routed through resistors according to the range and voltage across it provided as output. The output from the DAC is single-ended in both IDAC and VDAC modes. 35.3.1 Current DAC When used as an IDAC, the output is an 8-bit digital-to-analog conversion current. This is done by setting the DACx_CR0[4] register. The reference source is a current reference from the analog reference called IREF(DAC). In this mode, there are three output ranges selected by register DACx_CR0[3:2]. ■ 0 to 2.048 mA, 8 µA/bit ■ 0 to 256 µA, 1 µA/bit ■ 0 to 32 µA, 0.125 µA/bit done using a UDB input. UDB control for the source-sink selection is enabled using the DACx_CR1[3] bit. 35.3.2 Voltage DAC When used as a VDAC, the output is an 8-bit digital-to-analog conversion voltage to support applications where reference voltages are needed. Here, the reference source is a voltage reference from the Analog reference block called VREF(DAC). The DAC can be configured to work in voltage mode by setting the DACx_CR0[4] register. In this mode, there are two output ranges selected by register DACx_CR0[3:2]. ■ 0 V to 1.024 V ■ 0 V to 4.096 V Both output ranges have 255 equal steps. The VDAC is implemented by driving the output of the current DAC through resistors and obtaining a voltage output. Because no buffer is used, any DC current drawn from the DAC affects the output level. Therefore, in this mode any load connected to the output should be capacitive. The VDAC is capable of converting up to 1 Msps. In addition, the DAC is slower in 4 V mode than 1 V mode, because the resistive load to Vssa is four times larger. In 4 V mode, the VDAC is capable of converting up to 250 ksps. 35.3.3 Output Routing Options Output routing options for the DAC are attained through two separate muxes for current and voltage modes. These muxes are controlled by the DACx_SWx registers, as shown in Figure 35-2 on page 371. For each level, there are 255 equal steps of M/256 where M = 2.048 mA, 256 µA, or 32 µA. In the 2.040 mA configuration, the block is intended to output a current into an external 600 load. The IDAC is capable of converting up to 8 Msps. You also have the option of selecting if the output is a current source or a sink. This is done by the DACx_CR1[2] register. The selection between source and sink for the IDAC can also be 370 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Digital-to-Analog Converter Figure 35-2. DAC Interconnect DAC Bus (data[7:0]) Register DAC0_SWV[4:0] AGL0 AGL1 AMUXBUS abusl1 L abusl3 P0[6] AMUXBUSL AGL0 AGL1 Register DAC0_SWI[3:0] V V DAC0 I DAC0_strobe DAC1_strobe DAC3_D[7:0] DAC2_D[7:0] V DAC3 DAC2 V I I DAC2_strobe You can route output as follows: ■ DAC1 I Register DAC2_SWV[4:0] AGL4 AGL5 AMUXBUSL abusl0 abusl2 P0[7] AMUXBUSL AGL5 AGL4 Register DAC2_SWI[3:0] ■ DAC1_D[7:0] DAC0_D[7:0] DAC3_strobe Register DAC1_SWV[4:0] AGR0 AGR1 AMUXBUSR abusr1 abusr3 P3[0] AMUXBUSR AGR0 AGR1 Register DAC1_SWI[3:0] Register DAC3_SWV[4:0] AGR4 AGR5 AMUXBUSR abusr0 abusr2 P3[1] AMUXBUSR AGR4 AGR5 Register DAC3_SWI[3:0] For example, the implementation of a 12-bit DAC using two 8-bit DACs require: Voltage Mode – to the analog globals, analog mux bus, or the analog local bus ■ Current Mode – to the analog globals, analog mux bus, or to a specific port One DAC scaled to the range 0 to 2.048 mA and the second one scaled to the range 0 to 32 µA. ■ The middle four bits of the lowest range DAC are used as inputs to the lower four bits. See Figure 35-4 on page 372. 35.3.4 Making a Higher Resolution DAC It is possible to achieve a higher resolution current output DAC by summing the outputs of two 8-bit current DACs, each one having a different segment of the input bus for input. The range of the two DACs used partially overlap. This architecture may have problems of mismatch in the two DACs and therefore might require adjustment and scaling. The last two bits of the LSB DAC are used for minor calibration requirements. Figure 35-3. Higher Resolution DAC Example 12-Bit DAC Input 11:4 8-Bit DAC 8 µA/Bit <0:7> GND 7:6 Input 3:0 8-Bit DAC 0.125 µA/Bit 5:2 Calibration Bits 1:0 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 371 Digital-to-Analog Converter Figure 35-4. 12-Bit DAC Using Two 8-Bit DACs Example DAC for MSB, 8-BIT MSB 8 µA/Bit 1024 512 256 128 64 32 16 8 1 µA/Bit 128 64 32 16 8 4 2 1 1/8 µA/Bit 16 8 4 2 1 1/2 1/4 1/8 DAC for LSB, 4-BIT MSB 35.4 Register List Table 35-1. DAC Register List Register Name Comments Features General Registers DACx_CR0 DAC Control register 0 Select DAC mode, range, and speed DACx_CR1 DAC Control register 1 Control DAC data source, reset, and direction DACx_SW0 DAC Analog routing register 0 Routing for the DAC voltage output to analog (global) bus DACx_SW2 DAC Analog routing register 2 Routing for the DAC voltage output to analog (local) bus DACx_SW3 DAC Analog routing register 3 Routing for the DAC current/voltage output to AMUXBUS DACx_STROBE DAC Strobe register DC strobe control DACx_D DAC Data register DACx_TR DAC Block Trim register 372 DAC trim values PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 36. Precision Reference A voltage/current reference with value independent of supply voltage and temperature is an essential building block of many analog circuits. For example, accurate biasing voltages are critical for many circuit schemes; in ADC, a reference voltage is required to quantify an input, while in V/I DAC, voltage/current reference is required to define the output full-scale range. 36.1 Block Diagram The PSoC® 5LP devices have a curvature compensated voltage bandgap along with a trim buffer to get absolute value accuracy. The trim buffer is a multiple reference generator. It takes the bandgap reference voltage as input and produces outputs ranging from 0.256 V to 1.2 V. The reference voltage is buffered by low-power 5 A, high accuracy buffers, and sent to multiple destinations. There is also a temperature corrected (to flat) current reference that is mirrored and sent to current DAC. The voltage reference block diagram is illustrated in Figure 36-1 on page 374. 36.2 How It Works The principle of the bandgap circuit relies on two groups of diode-connected bipolar junction transistors running at different emitter current densities. By canceling the negative temperature dependence of the PN junctions in one group of transistors with the positive temperature dependence from a PTAT (proportional-to-absolute-temperature) circuit (which includes the other group of transistors), a fixed DC voltage that does not change with temperature is generated. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 373 Precision Reference Figure 36-1. Voltage Reference Block Diagram 10 µA IREF(DAC) Vcca Vcca Trim Buffer (I) 1.2V Vcca Bandgap Generator (V) Buffered in DSM Block by 10 µA Buffers Buffers 5 µA VREF2 (DSM) +_ + 1.024V _ +_ Vssa Vssa Delta Sigma ADC VREF1 (DSM) Vdda BG_CR0[3] BG_CR0[2] +_ VREF0 (Comparator) +_ VREF (Opamp) +_ VREF (SC) 0 ABUSL0 1 0.9V +_ VREF (TEMP SENSOR) Resistor String 0.8V Vssa VREF1_CM (DSM) 0.7V VREF2_CM (DSM) BG_CR0[1:0] 1 0.256V VREF1 (Comparator) +_ 2 +_ VREF (DAC) Vssa Dynamic enabling or disabling of analog peripherals may disturb shared internal voltage references. This may parametrically or functionally affect "already-on" or "staying on" components. Because these interactions are at the system level, they cannot always be systematically addressed in firmware component design. Therefore, consider the implications of a disturbed reference on components in use when dynamically enabling or disabling analog components. This includes enabling or disabling analog components as part of system power mode transitions, whether hardware or firmware based. Note 1 Analog supply Vdda or Vdda/2 can be routed to the analog blocks through the analog local bus, ABUSL0. The voltage level is selected using the BG_CR0[3] bit and the switch is enabled using the BG_CR0[2] bit. Note 2 Reference voltage input (VREF1) to the comparator is selected using the BG_CR0[1:0] bits. It selects either bandgap reference voltage or the analog supply voltage. Note 3 IREF (DAC) is the reference current for the DAC during IDAC mode operation. 374 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Precision Reference Table 36-1. Reference Voltages and Blocks Voltage Block VREF0 (Comparator) Comparator VREF1 (Comparator) Comparator Value 1.024 V Description To Comparator negative inputs Vdda (or) Vdda/2 (or) To Comparator negative inputs 256 mV VREF (Opamp) VREF (SC/CT) Opamp 1.024 V To Opamp positive inputs SC/CT Block 1.024 V To SC/CT block positive and negative inputs Comparator Opamp ABUSL0 DAC SC/CT Vdda (or) Vdda/2 All blocks connected to the analog local bus ABUSL0 can get this voltage DSM VREF (DAC) DAC 256 mV Reference voltage for DAC during VDAC mode operation VREF2 (DSM) DSM 1.2 V Reference voltage to Delta Sigma Modulator. This voltage is buffered in the DSM block by a 10 A buffer. VREF1 (DSM) DSM 1.024 V Reference voltage to Delta Sigma Modulator. This voltage is buffered in the DSM block by a 10 A buffer. VREF1_CM (DSM) DSM 0.8 V Common mode reference voltage for Delta Sigma Modulator VREF2_CM (DSM) DSM 0.7 V Common mode reference voltage for Delta Sigma Modulator VREF (TEMP SENSOR) TEMP SENSOR 0.9 V Analog ground option to auxiliary ADC PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 375 Precision Reference 376 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 37. Delta Sigma Converter The PSoC® 5LP ADC is a high resolution ADC implemented in Delta Sigma technology. Delta Sigma converters are integrating converters that provide high SNR/resolution by oversampling, noise shaping, averaging, and decimation. A Delta Sigma analog-to-digital converter (ADC) has two main components: a modulator and a decimator. The modulator converts the analog input signal to a high data rate (oversampling), low resolution (usually 1 bit) bitstream, the average value of which gives the average of the input signal level. This bitstream is passed through a decimation filter to obtain the digital output at high resolution and lower data rate. The decimation filter is a combination of downsampler and a digital low-pass (averaging) filter that averages the bitstream to get the digital output. 37.1 Features ■ 8 - to 20-bit resolution ■ Configurable gain from 0.25 to 256 ■ Differential/single ended inputs ■ Optional input buffer with RC low-pass filter ■ Internal and external reference options ■ Reference filtering for low noise ■ Incremental/continuous mode ■ Gain and offset correction 37.2 Block Diagram Figure 37-1 is the converter block diagram. (From Analog Routing) Figure 37-1. Delta Sigma Block Diagram Positive Input Mux Input Buffer Delta Sigma Modulator Negative Input Mux High data rate (sampling rate) low resolution bitstream in thermometric format q[7:0] Analog Interface High data rate (sampling rate) low resolution bitstream in 2's complement (4-bit) Decimator 24-bit Output Register PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F High resolution (max 20 bits) low data rate (sampling rate/ decimation ratio) output 377 Delta Sigma Converter 37.3 How It Works Figure 37-2. Input Buffer Structure bypass_p The PSoC 5LP Delta Sigma converter has a third-order modulator, followed by a fourth-order decimation filter. The modulator has a high impedance front end buffer followed by a bypassable RC filter. ■ The output of the modulator is passed on to the analog interface that converts the thermometric output to two’s complement (4 bit) and passes it on to the decimation filter. ■ The decimation filter takes 4-bit two’s complement input and provides a higher resolution (user selectable) output at a lower data rate. A detailed description of the individual blocks and their configuration options is given in this section. 37.3.1 Input Buffer The input impedance of the modulator is very low and not suitable for many applications. For applications that require a higher input impedance, two buffers (one for each differential input) are provided. Figure 37-2 shows the buffer and the RC filter that follows it. The buffers are of very low noise, are independent of each other, and can be bypassed (DSM_BUF0[1], DSM_BUF1[1]) or powered down (DSM_BUF0[0], DSM_BUF1[0]) individually by setting the bits listed in the braces. The buffer can also be used to amplify the input signal; it can be configured to provide gain of 1, 2, 4 and 8 in DSM_BUF1[3:2] register bits. The buffer has two separate modes, selected in the DSM_BUF0[2] bit to support a 0 to Vdd - 0.2 V input common mode range. The modes are: ■ A INP The modulator sends out a high data rate bitstream in thermometric format (see 37.3.2.6 Quantizer on page 386). ■ ■ rc Level Shifted – Buffer output can be level shifted up from the input when the input is close to 0 V input common mode voltage range. The operating range is 0 – vdda- 600 mV. Rail-to-Rail – This is used when input is rail-to-rail. The operating voltage range is vssa+200 mV to vdda-200 mV. The input structure is illustrated in Figure 37-2. 378 outp Cfilt To Modulator rc A INN outn bypass_p An additional RC filtering option (DSM_BUF2[1]) is provided for lower noise contribution from the buffer, at the cost of the input voltage not settling completely. This incomplete settling causes a gain error that must be corrected later, as a part of the downstream filtering in the decimator. There is also an option to chop (DSM_BUF3[3]) the input and output stages of the buffer to keep the offset as low as 100 µV. The chopping frequency is user selectable (DSM_BUF3[2:0]) and can vary from 1/2 to 1/256 of the input sampling frequency. The buffer can also be operated in a low-power mode (DSM_BUF2[0]). The ADC (buffer) takes its inputs from analog globals, analog locals, analog mux bus, reference, and Vssa. Registers DSM_SW0, DSM_SW2, DSM_SW3, DSM_SW4, DSM_SW6 help configure the positive and negative inputs. Limit the maximum input signal amplitude to the modulator (after the buffer gain, if used) to the values in Table 37-1 for a proper operation. The values in Table 37-1 are for a 1.024 V reference. For other reference values, scale the maximum input amplitude accordingly. Table 37-1. Maximum Input Signal Levels (ADC Reference Vref -> 1.024 V) Gain Modulator Quantization - 9 Level 0.167 5.25 0.25 3.56 0.5 1.78 1 1.024 2 0.44 4 0.22 8 0.11 16 0.05 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter 37.3.2 Delta Sigma Modulator The Delta Sigma modulator does: ■ Sampling the input signal (oversampling) ■ Optional gain by adjusting the ration of Cin1 to Cref ■ Coarse quantization (2, 3, or 9 levels/1, 1.5, or 2.2 bits) ■ Overload detection and chopping PSoC 5LP Delta Sigma modulator implementation is shown in Figure 37-3. Figure 37-3. Delta Sigma Modulator Implementation VREF VCM Csumin inp Csum1 Cdac o1p Csum2 Cf1 INP Cf2 Cin2 Cin1 INT3 o2p Cin2 Cf1 o3p INT2 o1p Cin1 Csum3 o2p INT1 INN Cin3 o1p Csumfb o2p Cf3 Cin3 Cf2 Summer o3p Csum3 Cf3 o3n Csum2 Cdac Csumfb o2n Csum1 inn Csumin Quantizer b[8:0] QLEV[1:0] The Delta Sigma modulator consists of these subsystems: ■ Three active integrators ■ An active summer ■ A programmable quantizer ■ A switched capacitor feedback DAC PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 379 Delta Sigma Converter 37.3.2.2 A few points about the modulator: ■ ■ The three active integrators and the programmable quantizer form the third order modulator. The transfer function of the integrators and the quantizer together account for the high pass noise shaping. Higher the order of the modulator, better is the high pass filter response and lower is the noise in the signal frequency band. All of the capacitors shown in Figure 37-3 on page 379 have binary weighted programmability. The value of a capacitance can be configured by setting the following three fields: The three integrators and quantizer stages are followed by an active summer. The analog input and the output of all three opamp stages are summed here. ■ The summer output is quantized by a quantizer. The quantizer is programmable to output 2, 3, or 9 levels. ■ The DAC (Vref, Vgnd and Cref constitute the DAC) connects the quantizer output back to the first stage opamp input. It is this feedback DAC that ensures that the average of the quantizer output is equal to the average input signal level. 37.3.2.1 Capacitance Configuration ■ Offset Capacitance – single bit that enables or disables an offset capacitance ■ Cap Array[n:0] – n+1 binary weighted bits ■ LSB Enable – additional unit capacitance Capacitance configuration (configuration of the above fields) is done in registers DSM_CR4 through DSM_CR12. Capacitance value is described by following equation: Cap value = (offset × Coff) + (cap[n:0] × Cunit) + (EN × Cunit) Where: ■ Coff is the offset capacitor value. ■ Cunit is the unit capacitor value. ■ Offset is the binary value (single bit) programmed in the offset field. ■ EN (LSB enable) is the binary value (single bit) programmed in the EN field. ■ Cap[n:0] is the decimal equivalent of the binary value programmed in the cap array[n:0] field. Clock Selection Any one of the four analog clocks or a UDB-generated clock can be used as the input sampling clock. The clock input can also be disabled. The DSM0_CLK register helps in selecting the clock source and enabling or disabling it. The maximum clock that can be applied to the modulator is 6.144 MHz. Make certain that the clock to the decimator = fs/n, n = 2,3,4..., fs is the PHUB clock. The unit capacitance, offset capacitance, and default values for all of the capacitances are given inTable 37-2. Table 37-2. Capacitance Values Register Bit Description Value FCAP1OFFSET Offset cap for first stage feedback cap 3.4 pF FCAP1[6:0] Binary weighted first stage feedback cap Cunit = 100 fF FCAP1EN Enable for LSB CAP of FCAP1 100 fF - 12.8 pF in 100 fF steps Typical Value 0 IPCAP1OFFSET Offset cap for first stage input cap 4.8 pF IPCAP1[6:0] First stage Input CAP (binary) Cunit = 100 fF IPCAP1EN Enable for LSB cap of IPCAP1 100 fF - 12.8 pF in 100 fF steps DACCAP[5:0] DAC cap (each unit) - binary Cunit = 96 fF (2 LSBs) and 100 fF (4 MSBs) DACCAPEN Enable for LSB CAP of DAC 96 fF - 62898 fF in variable steps 1010000 8 pF 0 0 RESCAP[2:0] Resonator cap (binary) Cunit = 12 fF RESCAPEN Enable for LSB cap of RESCAP 12 fF - 96 fF in 12 fF steps FCAP2[3:0] Second stage Feedback cap - binary Cunit = 50 fF FCAP2EN Enable for LSB CAP of FCAP2 50-800 fF in 50 fF steps IPCAP2[2:0] Second stage input CAP - binary Cunit = 50 fF IPCAP2EN Enable for LSB Cap of IPCAP2 50-400 fF in 50 fF steps FACP3[3:0] Third stage feedback cap Cunit = 100 fF FCAP3EN Enable for LSB Cap of FCAP3 100 fF-1.6 pF in 100 fF steps 380 Default 0101100 4.4 pF 0 101100 4.4 pF 0 000 0 fF 0 1011 0.55 pF 0 101 0.25 pF 0 1110 1.4 pF 0 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter Table 37-2. Capacitance Values (continued) Register Bit Description Value IPCAP3[2:0] Third stage input cap Cunit = 50 fF IPCAP3EN Enable for LSB Cap of IPCAP3 50-400 fF in 50 fF steps SUMCAPIN[4:0] Summer cap for input path Cunit = 50 fF SUMCAPINEN Enable for LSB Cap of SUMCAPIN 50-1.6 pF in 50 fF steps SUMCAPFB[3:0] Summer cap for feedback path Cunit = 50 fF SUMCAPFBEN Enable for LSB Cap of SUMCAPFB 50-800 fF in 50 fF steps Default Typical Value 101 0.25 pF 0 00101 0.25 pF 0 1010 0.5 pF 0 SUMCAP1[2:0] Summer cap for first stage output 101 SUMCAP1EN Enable for LSB Cap of SUMCAP1 0 SUMCAP2[2:0] Summer cap for second stage output Cunit = 50 fF SUMCAP2EN Enable for LSB Cap of SUMCAP2 50-400 fF in 50 fF steps 0.25 pF 101 0.25 pF 0 SUMCAP3[2:0] Summer cap for third stage output 101 SUMCAP3EN Enable for LSB Cap of SUMCAP3 0 0.25 pF 37.3.2.3 Gain Configuration The modulator provides gain from 0.25 to 16 to the input signal. Gain is the ratio of input and DAC capacitances, as described in the following equation. Gain = C in C ref However, increasing only the input capacitance to increase gain disturbs the transfer characteristics of the modulator. Therefore, other capacitors also must be scaled to maintain the modulator transfer characteristics. Recommended values of capacitors for gains of 1, 2, 4, 8 are shown in Table 37-3, and those for 16, 0.25, and 0.5 are shown in Table 37-4. Equation 1 Table 37-3. Gains 1, 2, 4, and 8 Register Bit Gain = 1 Bit Setting Gain = 2 Typical Value Bit Setting Gain = 4 Typical Value 0 Bit Setting IPCAP1OFFSET 0 IPCAP1[6:0] 0101100 IPCAP1EN 0 0 1 DACCAP[5:0] 101100 101100 101100 DACCAPEN 0 SUMCAPIN[4:0] 00101 4.4 pF 1011000 4.4 pF SUMCAPINEN 1010 SUMCAPFBEN 0 SUMCAP1[2:0] 101 SUMCAP2[2:0] 101 SUMCAP2EN 0 SUMCAP3[2:0] 101 SUMCAP3EN 0 100 100 100 0.2 pF PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 0.2 pF 0 100 0.2 pF 0 0.2 pF 0 0 100 0.2 pF 0 0.2 pF 0.2 pF 0.25 pF 0100 0.4 pF 0 0 0.8 pF 0 100 100 0.25 pF 10000 1000 0 2.2 pF 0 0 0.2 pF 100 0.2 pF 0 17.6 pF 010110 0 100 1111111 0.8 pF 0.4 pF Typical Value 1 10000 1000 0.25 pF 17.6 pF 4.4 pF 0.4 pF 0 Bit Setting 1 0 0 0 1111111 4.4 pF 0.5 pF SUMCAP1EN 8.8 pF 01000 0 SUMCAPFB[3:0] 1 0 0.25 pF Gain = 8 Typical Value 0.2 pF 0 381 Delta Sigma Converter Table 37-4. Gains 16, 0.5, and 0.25 Register Bit Gain = 16 Bit Setting IPCAP1OFFSET Gain = 0.5 Typical Value 1 IPCAP1[6:0] 1111111 IPCAP1EN 1 DACCAP[5:0] 001011 DACCAPEN 0 Bit Setting 10000 SUMCAPINEN 0 0010 SUMCAPFBEN 0 17.6 pF 0010110 100 SUMCAP21EN 0 100 SUMCAP2EN 0 100 SUMCAP3EN 0 1.1 pF 101100 4.4 pF 4.4 pF 0 0 00010 00001 0.1 pF 0.05 pF 0 0 1000 1000 0.4 pF 0.4 pF 0 0 100 100 0.2 pF 0.2 pF 0 0 100 100 0.2 pF 0.2 pF 0 0 100 100 0.2 pF 37.3.2.4 0001011 101100 0.2 pF SUMCAP3[2:0] Typical Value 0 0.2 pF SUMCAP2[2:0] 2.2 pF 0 0.1 pF SUMCAP1[2:0] Bit Setting 0 0.8 pF SUMCAPFB[3:0] Typical Value 0 1.1 pF SUMCAPIN[4:0] Gain = 0.25 0.2 pF 0.2 pF 0 0 Power Configuration There are separate power settings for the first opamp stage, the summer, and the quantizer. The second and third stages share the same power settings. The power for all of these stages is configured in registers DSM_CR14 and DSM_CR16. The various configurable power settings are shown in Table 37-5. Table 37-5. Configurable Power Settings Register Bit Description Truth Table, Typical IDD 000 - LOW (42 µA) 001 - MEDIUM (114 µA) 010 - HIGH (430 µA) POWER1 011 - 1.5X (650 µA) Power control for first stage 100 - 2X (900 µA) 101 - C/2 at 3MSPS (254 µA) 110 = C/4 at 3MSPS (170 µA) 111 - 2.5X (1.35 mA) 000 - LOW (4 µA) 001 - MEDIUM (16 µA) POWER2_3[2:0] Power control for second stage/third stage 010 - HIGH (62 µA) 011 - 1.5X (100 µA) 100 - 2X (135 µA) 000 - LOW (4 µA) 001 - MEDIUM (16 µA) POWER_SUM[2:0] Power control for summer 010 - HIGH (62 µA) 011 - 1.5X (100 µA) 100 - 2X (135 µA) 00 - Very Low (2.2 µA) POWER_COMP[1:0] Comparator power control 01 - Normal (8.6 µA) 10 - 6 MHz (17 µA) 11 - 6 MHz (35 µA) 382 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter Table 37-5 indicates how to configure power for the individual blocks. Power dissipation, capacitances, clock frequency and quantization levels are interrelated to each other. Configuring power without varying the other parameters mentioned above affects the proper operation of the modulator. The tables below show a set of operational modes that indicate how to configure power based upon the other parameters or vice versa. Table 37-6. Power Configuration Based on Quantization Levels and Clock Frequency Register Bit Mode - 3 MHz 9 Level Bit Setting FCAP1OFFSET 1010000 FCAP1EN 0 IPCAP1OFFSET IPCAP1EN Mode - 6 MHz 9 Level Bit Setting 0 FCAP1[6:0] IPCAP1[6:0] Typical Value 0 8 pF 0010100 2 pF 0001011 Bit Setting 1111110 0101100 0 0 0 101100 001011 101100 DACCAPEN 0 RESCAP[2:0] 000 1.1 pF 0 RESCAPEN 0 FCAP2[3:0] 1011 FCAP2EN 0 IPCAP2[2:0] 0 FACP3[3:0] 1110 FCAP3EN 0 SUMCAPIN[4:0] 00101 0 SUMCAP1[2:0] 101 SUMCAP21EN 0 SUMCAP2[2:0] 101 0 SUMCAP3[2:0] 101 SUMCAP3EN 0 QLEVEL[1:0] 10 FCHOP[2:0] 0.25 pF 0 0010 1010 0.1 pF 0 0.5 pF 0 001 101 0.05 pF 0 0.25 pF 0 001 101 0.05 pF 0 0.25 pF 0 001 0.25 pF ODET_TH[4:0] 00101 0.05 pF 0 0.25 pF 0 0.25 pF 0 00001 0.25 pF SUMCAP2EN 101 0.05 pF 0.5 pF 0 1.4 pF 0 001 1010 SUMCAPFBEN 1110 0.3 pF 0 0.25 pF SUMCAPFB[3:0] 0.25 pF 0 0011 0 SUMCAPINEN 101 0.05 pF 0 101 0.55 pF 0 001 0.25 pF IPCAP3EN 1011 0.1 pF 0 1.4 pF IPCAP3[2:0] 0fF 0 0001 0.25 pF IPCAP2EN 000 0 101 4.4 pF 0fF 0.55 pF 101 0.05 pF 0 level=9 10 01100 12 001 Fclk/4 4.4 pF 0 000 0fF 16 pF 0 1.1 pF DACCAP[5:0] 4.4 pF Typical Value 0 0 4.4 pF Mode - 3 MHz 2 Level and 3 Level 1 0 0 0101100 Typical Value 0.25 pF 0 level=9 00 or 01 level=2 or 3 01100 12 01100 12 001 Fclk/4 001 Fclk/4 NONOV[1:0] 01 3.5 ns 00 1.5 ns 01 3.5 ns POWER1[2:0] 010 430 µA 010 430 µA 010 430 µA POWER2_3[2:0] 010 62 µA 010 62 µA 010 62 µA POWER_SUM[2:0] 010 62 µA 010 62 µA 010 62 µA POWER_COMP[1:0] 01 9 µA 10 18 µA 01 9 µA PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 383 Delta Sigma Converter Table 37-7. Power Configuration Based on Capacitances Mode - C/2 Register Bit Bit Setting FCAP1OFFSET Mode - C/4 Typical Value Bit Setting 4 pF 0010100 0 Mode - C/8 Typical Value Bit Setting 2 pF 0001010 0 0 FCAP1[6:0] 0101000 FCAP1EN 0 0 0 IPCAP1OFFSET 0 0 0 IPCAP1[6:0] 0010110 IPCAP1EN 0 DACCAP[5:0] 010110 DACCAPEN 0 RESCAP[2:0] 000 RESCAPEN 0 FCAP2[3:0] 0101 FCAP2EN 0 2.2 pF 0001011 010 0 FACP3[3:0] 0101 FCAP3EN 0 IPCAP3[2:0] 010 0 SUMCAPIN[4:0] 00010 0 0100 SUMCAPFBEN 0 SUMCAP1[2:0] 010 0 SUMCAP2[2:0] 010 0 SUMCAP3[2:0] 010 SUMCAP3EN 0 0.05 pF 0 0011 1110 0.3 pF 0 0.3 pF 0 001 101 0.05 pF 0 0.05 pF 0 00001 00101 0.05 pF 0 0.05 pF 0 0010 0010 0.1 pF 0 0.1 pF 0 001 101 0.05 pF 0 0.05 pF 0 001 0.1 pF 101 0.05 pF 0 0.05 pF 0 001 0.1 pF QLEVEL[1:0] 101 0.05 pF 0 0.1 pF SUMCAP2EN 0.1 pF 0 001 0.2 pF SUMCAP21EN 1011 0.1 pF 0 0.1 pF SUMCAPFB[3:0] 0fF 0 0001 0.1 pF SUMCAPINEN 000 0fF 0 0.5 pF IPCAP3EN 0.5 pF 0 000 0.1 pF 0.5 pF 000101 1.1 pF 0 0.25 pF IPCAP2[2:0] 0000101 1 pF 0 001011 0fF IPCAP2EN 1.1 pF 0 2.2 pF Typical Value 101 0.05 pF 0 0.05 pF 0 10 level=9 10 level=9 10 01100 12 01100 12 01100 12 FCHOP[2:0] 001 Fclk/4 001 Fclk/4 001 Fclk/4 NONOV[1:0] 01 3.5 ns 01 3.5 ns 01 3.5 ns POWER1[2:0] 101 254 µA 110 170 µA 000 114 µA POWER2_3[2:0] 001 16 µA 001 16 µA 001 16 µA POWER_SUM[2:0] 001 16 µA 001 16 µA 001 16 µA POWER_COMP[1:0] 10 18 µA 10 18 µA 10 18 µA ODET_TH[4:0] 384 level=9 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter Table 37-8. Configuration Based on Power Register Bit Mode - Medium Power Bit Setting FCAP1OFFSET Mode - Low Power Typical Value Bit Setting 8 pF 1010000 0 0 FCAP1[6:0] 1010000 FCAP1EN 0 0 IPCAP1OFFSET 0 0 IPCAP1[6:0] 0101100 IPCAP1EN 0 DACCAP[5:0] 101100 DACCAPEN 0 4.4 pF 0101100 RESCAP[2:0] 000 0 FCAP2[3:0] 1011 FCAP2EN 0 4.4 pF 0 000 0fF 0 1011 0.55 pF IPCAP2[2:0] 101 0 FACP3[3:0] 1110 FCAP3EN 0 0.55 pF 0 101 0.25 pF 0.25 pF 0 1110 1.4 pF IPCAP3[2:0] 101 IPCAP3EN 0 1.4 pF 0 101 0.25 pF SUMCAPIN[4:0] 00101 SUMCAPINEN 0 0.25 pF 0 00101 0.25 pF SUMCAPFB[3:0] 1010 SUMCAPFBEN 0 0.25 pF 0 1010 0.5 pF SUMCAP1[2:0] 101 SUMCAP21EN 0 0.5 pF 0 101 0.25 pF SUMCAP2[2:0] 101 SUMCAP2EN 0 SUMCAP3[2:0] 101 SUMCAP3EN 0 0.25 pF 0 101 0.25 pF 0.25 pF 0 101 0.25 pF QLEVEL[1:0] 4.4 pF 101100 0fF IPCAP2EN 8 pF 0 4.4 pF RESCAPEN Typical Value 0.25 pF 0 10 level=9 10 01100 12 01100 12 FCHOP[2:0] 001 Fclk/4 001 Fclk/4 NONOV[1:0] 01 3.5 ns 01 3.5 ns POWER1[2:0] 010 114 µA 010 42 µA POWER2_3[2:0] 010 16 µA 010 4 µA POWER_SUM[2:0] 010 16 µA 010 4 µA POWER_COMP[1:0] 01 9 µA 01 9 µA ODET_TH[4:0] PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F level=9 385 Delta Sigma Converter 37.3.2.5 Other Configuration Options The modulator can be chopped for a low offset of 100 µV. The chopping frequency can be set from fclk/2 to fclk/256, where fclk is the input sampling clock. Chopping enable and chopping frequency setting are done in the DSM_CR2 register The modulator can be configured for inverting the gain by setting the sign bit in DSM_CR3[7]. 37.3.2.7 Reference Options The Delta Sigma channel has selectable analog reference input (REFBUF0) options, as shown in Figure 37-4 on page 387. Also illustrated are the opamp output common mode (VCMBUF0) and the negative input buffer (REFBUF1) selection schemes. The various reference selections for the DSM ADC may be broadly classified into the following modes: ■ The modulator can be reset (all capacitances are reset) by the UDB or decimator, and the reset source is selected by the DSM_CR2[7] register. More details about reset are in the Reset chapter on page 143. Internal Reference (reference generated on-chip) that is buffered but unfiltered (Figure 37-5 on page 387) ■ Internal Reference that is buffered and filtered with an external capacitor tied between P0[3] and ground or P3[2] and ground (Figure 37-6 on page 388) 37.3.2.6 ■ External Reference source driving reference into the DSM (Figure 37-7 on page 388) Quantizer The quantizer can be configured for 2, 3, or 9 levels. A 9 level quantizer offers a better SNR and a 2 level quantizer offers better linearity. Depending on the application requirement, the user can choose quantization levels. The number of quantization levels is configured in DSM_CR0[1:0] register bits. The quantizer outputs data in thermometric format. The quantizer output is stored in the register DSM_OUT1. Thermometric format is explained by the pattern of output levels shown in the following table. In thermometric format, the number of ones increases from LSB to MSB as the quantization level increases. Table 37-9. Quantizer Output Data Level Quantizer Output Data 2 Level Quantizer Level 1 Level 2 00000000 11111111 3 Level Quantizer Level 1 00000000 Level 2 00001111 Level 3 11111111 9 Level Quantizer 386 Level 1 00000000 Level 2 00000001 Level 3 00000011 Level 4 00000111 Level 5 00001111 Level 6 00011111 Level 7 00111111 Level 8 01111111 Level 9 11111111 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter Figure 37-4. Delta Sigma Channel Analog Reference Selection Vcmx (unbuffered VCM) vpwra No Selection 00 0.8V 01 0.7V 10 NC vcm_res_div_en vpwra/2 VCMBUF0 VCM To the opamps in the DSM 11 vssd en_buf_vcm en_buf_vref_inn vcmsel<1:0> + vgnd To the INN Mux of the Channel S12 REFBUF1 To the Quantizer Resistor string vgnd Vpwra S13 No Selection S6 S5 000 VDAC0L vdda/4 S11 S3 010 vdda/3 Vref = 1.024 V from bandgap Vref = 1.2V from bandgap NA 011 REFBUF0 100 + 101 110 S4 S10 S0 - S2 S9 S7 en_buf_vref 111 Resd1 Resd3 S8 Resd3 refmux<2:0> Resd2 vdda Resd1 S1 Resd2 NA vref_res_div_en Reference of the DSM ADC (to the Cref branch in first integrator) 001 vdda/3 vdda/4 P0[3] P3[2] vgnd Figure 37-5. Connection Scenario: Internal Reference with No RC Filtering (using P0[3]) Vref for Quantizer S6 S5 Vref for ADC Internal Reference Buffered Unfiltered S3 S4 S0 Res d2 S1 S2 Res d3 + - Res d1 Vref From Mux No External Reference or External Filtering PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 387 Delta Sigma Converter Figure 37-6. Connection Scenario: Internal Reference with RC Filtering S6 Vref for Quantizer Resistor String S5 S3 Vref From Mux S4 + S0 Res d3 Res d2 S1 Vref for ADC Internal reference Buffered Filtered with External Capacitor S2 Res d1 - C (External) Figure 37-7. Connection Scenario: External Reference Only Vref for Quantizer is the Same as the External Reference S6 S5 S3 S4 + S0 REFBUF0 is Powered Down (Output Tristated) Vref is Driven by an External Source Resd3 S1 S2 Resd2 - Resd1 Vref from Mux External Reference Input 388 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter ence buffers can be configured in low, medium, high, and turbo power modes in the DSM_CR17 register. The common mode voltage buffer, internal reference voltage buffer, and the negative input buffer are powered down, using DSM_CR17[1]; DSM_CR17[0], and DSM_REF0[3] register bits, respectively. There are several selectable options for internal reference, based on refmux[2:0] programming in DSM_REF0 register. The places in the DSM block (Figure 37-3 on page 379) that require a reference value are: ■ DAC capacitor (Cref) sampling in the first integrator ■ Reference for the resistive ladder inside the quantizer block ■ 1. Power on the VCMBUF0 for the DSM to function. 2. Turn on the reference buffer REFBUF1 only when you want to drive the ADC reference to the negative input mux of the DSM channel. Common Mode Voltage (VCM) for the differential circuits. This voltage is typically 0.8 V with an option to go to 0.7 V for better head rooms. A provision for applying VDD / 2 is also provided. 37.3.2.8 3. Power down REFBUF0 only when you want to drive reference to the ADC from an off-chip source (See the external reference option in Table 37-10). Reference for DSM: Usage Guidelines To get low reference noise, the option to filter is provided with the special connections to pins P3[2] and P0[3], as shown in Figure 37-4 on page 387 and Figure 37-7 on page 388. Therefore, for low noise floor requirements, use the external capacitor filter. Only two pins, P3[2] and P0[3], are dedicated for this purpose in PSoC 5LP devices. The switches in Table 37-10 that are marked as ON mean that the switch is closed, and a path is created for reference to reach DSM. Empty cells indicate that the switches are open. The following table shows the state of various switches and the two reference buffers for certain selectable reference options. Not every possible combination of closing the switches marked S0-S13 is discussed in this section. The configuration of these switches (therefore the reference selection) is made in registers DSM_REF2 and DSM_REF3. The refer- Table 37-10. Analog Reference Modes for the Delta Sigma Channel SN Mode Switch States S0 1 2 S1 S2 Internal Reference (No Filtering) Internal Reference (Filter with P3[2]) 3 Internal Reference (Filter with P0[3]) 4 External Reference only (P3[2]) 5 External Reference only (P0[3]) 6 Vpwra is internal reference ON ON S3 S4 ON ON S5 ON S6 ON S9 S10 S11 S12 S13 ON ON ON ON ON ON ON ON ON ON REFBUF0 S8 ON ON ON S7 OFF ON ON ON OFF ON OFF Table 37-11. Guidelines for External Reference Bypass Capacitors Resolution Bypass Capacitor Value less than 16-bit 0.1 µF 16 bit or more 0.1 µF to 10 µF PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 389 Delta Sigma Converter 37.3.3 Analog Interface The analog interface connects the modulator to the other blocks including the decimator and the UDB. As shown in Figure 37-8 on page 390, the analog interface converts thermometric code sent by the modulator to two’s complement and allows for selection of modulation input, selecting and synchronizing clocks. Figure 37-8. Analog Interface DSMn vp vn ANALOG dsmn_clk dsmn_modbitin dsmn_dout[7:0] dsmn_startup_reset CLK SEL clk_a[3:0] mx_clk[2:0] bypass_sync ANAIF SYNC clk_en dout_sync[7:0] clk_a_dig[3:0] CLK SEL dec_clk dout_sat[7:0] modbitin_en mx_startup_reset 1 0 mx_modbitin[3:0] tempcode_in[7:0] TEMPCODE2SCOMP 8 lut_outputs[7:0] qlev[1:0] dout2scomp[3:0] out0[7:0] out1[7:0] mx_dout dsmn_startup_reset_udb dsmn_extclk_cp_udb 1 0 dsmn_clk_udb dec_clk dsmn_dout2scomp[3:0] dsmn_reset_dec DECIMATOR dsmn_modbitin_udb dec_irq dsmn_dout_udb[7:0] dec_start UDB 390 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter 37.3.3.1 Conversion of Thermometric Code to Two’s Complement The following table shows the conversion from thermometric format to two’s complement for 2, 3, and 9 level quantizations performed by the analog interface. This two’s complement input is fed to the decimator. Table 37-12. Two’s Complement Conversion Table Inputs Output qlev[1:0] dout[7:0] dout2scomp[3:0] 00 00000000 1111 -1 00 11111111 0001 +1 01 00000000 1111 01 00001111 0000 0 01 11111111 0001 +1 1x 00000000 1100 -4 1x 00000001 1101 -3 1x 00000011 1110 -2 1x 00000111 1111 -1 1x 00001111 0000 0 1x 00011111 0001 +1 1x 00111111 0010 +2 1x 01111111 0011 +3 1x 11111111 0100 +4 37.3.3.2 -1 Modulation Input As discussed in 37.3.2.5 Other Configuration Options on page 386, modulator gain can be inverted by the sign bit in DSM_CR3. The sign can also be changed by a direct digital input from LUTs or the UDB. The modulation input assists in this process. Depending on whether the modulation input is high or low, the gain is normal or inverted. The modulation input can be enabled by setting the DSM_CR3[4] register bit. Modulation input is selected by DSM_CR3[3:0] control bits. 37.3.3.3 Clock Selection and Synchronization The output of the modulator (quantizer) Q[7:0] can be synchronized with respect to the digitally aligned clock of the analog clock selected for the modulator. As mentioned in 37.3.2 Delta Sigma Modulator on page 379, clock selection is done by DSM_CLK[2:0] register bits. Clock synchronization is enabled by clearing the DSM_CLK[4] register bit. 37.3.4 Decimator The decimator takes the 4-bit input (low resolution) in two’s complement format and converts it into a high resolution output. The 4-bit two’s complement values coming into the decimator at the input sampling rate are averaged over a specified number of samples (decimation ratio), down sampled, and passed through an optional post-processing filter, achieving a higher resolution. The decimator in PSoC 5LP devices is a fourth order Cascaded Integrator Comb (CIC) filter. The decimator structure is shown in Figure 37-9 on page 392. 37.3.4.1 Shifters There are two shifters in the block — one in front of the CIC filter and another one in front of the post processor. The input shift values are programmed depending on the decimation ratio and quantization level to ensure that ADC results are available in the Q31 format. The shift values are programmed in register DEC_SHIFT1. The shift values to be programmed in DEC_SHIFT1 and DEC_SHIFT2 for various decimation ratios (DR1 and DR2) and quantization levels are shown in Table 37-13 and Table 37-14 on page 392. Table 37-13. Programmed Shifter1 Values for Various Decimation Ratios (Programmed in DR1) Decimation Ratio Quantization Levels Max Values in Range 8 2, 3 4095 to -4096 12 Left shift 20 8 9 16383 to -16384 14 Left shift 18 16 2,3 65535 to -65536 16 Left shift 16 16 9 262143 to -262144 18 Left shift 14 32 2, 3 1048575 to -1048576 20 Left shift 12 32 9 4194303 to -4194304 22 Left shift 10 64 2, 3 16777215 to -16777216 24 Left shift 8 64 9 67108863 to -67108864 26 Left shift 6 128 2, 3 268435455 to -268435456 28 Left shift 4 128 9 173741823 to -1073741824 30 Left shift 2 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Bit Width Shift Adjustment 391 Delta Sigma Converter Table 37-14. Programmed Shifter2 Values for Various Decimation Ratios (Programmed in DR2) Value of D2 Final Resolution Right Shift Value 1 No shift, bypass sync (boxcar) filter 16 4 32 5 64 6 128 7 256 8 512 9 1024 10 37.3.4.2 Table 37-16. Decimation Ratios for 9 Level Quantization Clock, Decimation Ratio 12-bit 6.144 MHz, 16 14-bit 6.144 MHz, 32 16-bit 3.072 MHz, 64 20-bit 3.072 MHz, 16384 (CIC + post processor) 37.3.4.3 Post Processing Filter The Post Processor receives 28-bit data from the output of the CIC Decimation filter for further convenience or post processing. Available functions are: CIC Filter The CIC filter has four cascaded integrator sections operating at the modulator sample rate, followed by four cascaded comb sections operating at a lower sample rate (determined by DR1). This combination implements a sinc4 Finite Impulse Response (FIR) filter. The CIC filter is controlled by a finite state machine that allows it to sequence events in the various modes of operation of the decimator. The decimation ratio is programmed in the DEC_DR1 register. The registers in CIC filter are 32-bits wide and, therefore, for proper operation, the decimation ratio should not exceed the values given in Table 37-15. Table 37-15. Maximum Decimation Ratio Values for CIC Level Bit Width Encoding (Decimal) Max Allowed 2 32 -1, 1 256 3 32 -1, 0, 1 215 9 32 -4, -3, -2, -1, 0, 1, 2, 3, 4 152 The decimation ratios to be configured for 12, 14, 16 and 20 bit resolutions for 9 level quantization are shown in Table 37-16. ■ Add a programmable offset coefficient to the CIC result ■ Multiply a programmable gain coefficient to the CIC result ■ Apply both offset and gain ■ Apply a sinc1 FIR filter ■ Apply both a sinc1 filter and offset correction ■ Apply both a sinc1 filter and gain correction ■ Apply all three When more than one of the three functions is enabled to operate concurrently on the data, they are always performed in the order: FIR > Offset > Gain. The decimator process is shown in Figure 37-9. The offset value to be added is programmed in registers DEC_OCOR, DEC_OCORM, and DEC_OCORH. The 24-bit offset is given in signed two’s complement format. The registers are coherency interlock protected (see 37.3.6 Coherency Protection on page 393). Sample Data From Modulator Shifter1 Figure 37-9. Decimator Sample Data Out of Decimator CIC - Decimation Shifter2 Post Processor FIR Offset Gain Decimation Ratio DR1 Post Processor Enabled Decode of Which PP Features are Enabled 392 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter The gain correction coefficient is programmed in registers DEC_GCOR, DEC_GCORH. The number of bits that are valid in the above register is programmed in the DEC_GVAL[3:0] register bits. This allows use of a part of the 16-bits for gain correction. The registers are coherency interlock protected. If the gain feature is used, the value programmed into the DR1 register (CIC decimation ratio) cannot be smaller than 2+2*GVAL, allowing time for the hardware to do a shift-add multiple during the decimation period. 37.3.6 The FIR filter is a summer that implements the sinc1 filter. It is used in cases where decimation ratios greater than 128 are desired. When the FIR function is enabled, the Post Processor sums samples from the CIC filter, DR2 at a time, where DR2 (10 bits) is the decimation ratio programmed in the DEC_DR2, DEC_DR2H[1:0] registers. ■ Gain and Gain Value (write protected) – really two fields, but they are checked for coherency as if they are a single field protected on writes so that the underlying hardware does not incorrectly use the field when it partially updated by system software. ■ Offset Value (write protected) – protected on writes so that the underlying hardware does not incorrectly use the field when it is partially updated by the system software. ■ Output Sample Value (read protected) – protected on reads so that the underlying hardware does not update it when partially read by the system software or DMA. Depending on the configuration of the block, not all bits of the output sample register are of interest. Gain correction, offset correction, and FIR filtering features can be enabled and disabled in the DEC_CR[6:4] register bits. The Post Processor implements saturation logic that prevents over- and under-flow wraparound in the accumulator. If the DEC_CR[7] bit is set, the ALU does not wrap when the most positive or negative number is exceeded. The output of the conversion is stored in registers OUTSAMP, OUTSAMPM, and OUTSAMPH. In some configurations of the block, output results of interest are placed in bits 23:8 of the output sample field. To allow reading such values in one bus cycle, an alignment feature is added to shift the result right by 8 bits. This feature is enabled by the OUTPUT_ALIGN bit of the DEC_SR register. 37.3.5 ADC Conversion Time The conversion time is the time taken from the SOC to the ADC interrupt. The conversion time provided in the table is expressed in ADC_clk cycles. Because the SOC is an asynchronous signal to ADC_clk, there can be a skew error on all the conversion times. Table 37-17. ADC Conversion Time ADC Mode Conversion Time in ADC_clk Cycles Single Sample, Fast Filter, Continuous (DR1 x 4)+3 Fast FIR DR1 x (DR2 +3) +6 Single Sample, Fast Filter, Continuous with offset correction (DR1 x 4)+6 Fast FIR with offset correction DR1 x (DR2 +3) +7 Single Sample, Fast Filter, Continuous with offset correction (DR1 x 4)+GVAL+5 Fast FIR with offset correction DR1 x (DR2 +3) +GVAL+7 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Coherency Protection Coherency refers to the hardware added to a block to protect against malfunctions of the block in cases where register fields are wider than the bus access, leaving intervals in time when fields are partially written or read (incoherent). Coherency checking is an option and is enabled in the DEC_COHER register. The hardware provides coherency checking on three register fields that are all up to three bytes wide: The coherency methodology allows for any size output field and handles it properly. In the COHER register, coherency is both enabled, and a Key Coherency Byte is selected. The Key Coherency Byte allows the user to tell the hardware which byte of the field will be written or read last when an update to the field is desired. Each for the three protected fields has a Coherency Interlock Flag (CIF). This flag signifies whether the field is coherent. The coherency hardware understands both 8-bit and 16-bit accesses and when tracking coherency, handles each appropriately. A hard or soft reset sets all CIF to coherent. 37.3.6.1 Protecting Writes (Gain/Offset) with Coherency Checking Starting from a coherent state (CIF is set), the software can write any of the other non-key bytes. This action flags the field incoherent (clears the CIF). When a field is incoherent, it is ignored by the underlying hardware, and a shadow register containing the last valid value is used. The field remains flagged incoherent until the Key Coherency Byte is written. At this time, the field is flagged coherent (CIF is again set), and the next time the hardware needs the field value, the new value is used, and the shadow register is updated with the new value. 393 Delta Sigma Converter 37.3.6.2 Protecting Reads (Output Sample) with Coherency Checking Starting from a coherent state (CIF is set), the software can read any of the other non-key bytes of the field. This action flags the field incoherent (clears the CIF). When a field is incoherent, it is protected against updates from the underlying hardware, and any new samples that may be generated while incoherent are dropped (without warning). The field remains flagged incoherent until the Key Coherency Byte is read. At this time, the field is flagged coherent (CIF is again set), and the next time the hardware generates a new output sample result, the field is updated. 37.3.7 ■ Generates one output result ■ Repeats this sequence until signaled to halt The decimator is set to one of the four modes by DEC_CR[3:2] bits. All four modes are started by either a write to the start bit in the DEC_CR[0] register or an assertion of the input signal ext_start. Set the DEC_CR[1] register bit when using the external start feature. When set, this bit ignores the DEC_CR[0] start bit. Figure 37-10 on page 395 shows the state diagram of various modes of operation of the decimator. Modes of Operation This block has four primary operating modes: ■ Single Sample ■ Fast Filter ■ Continuous ■ Fast FIR In Single Sample mode, the block sits in the standby state waiting for one of two start signals (START_CONV bit in CR register or ext_start). When a start is signaled, the block performs one sample conversion (four decimation periods where a decimation period is the count programmed in register DEC_DR1). It then captures the result, and signals the system by a polling or an interrupt that the process is complete and waits for the next signal as it reenters the standby state. The Fast Filter mode captures single samples back to back, resetting itself and the Modulator between each sample. Upon completion of a sample, the next sample is initiated continuously. Polling and interrupts mark result events. Fast Filter mode is simply a continuous string of Single Samples with channel resets between them. This mode should be used when multiplexing channels. If signaled to run Continuous, the filter resets the channel then runs continuously from that point forward, until signaled to stop, with no intervening resets of the channel. The hardware blocks the first three decimation periods but then provides a result every decimation cycle thereafter. Fast FIR mode is very much like Continuous mode, except that the ADC channel is reset and the filter restarted when the FIR decimation period (DR2) is reached. For example, if the DR2 register is set to 15 and this mode is selected, the filter: ■ Resets the channel ■ Blocks the first three decimation periods (DR1) ■ Produces 16 samples for the FIR function to operate on 394 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Delta Sigma Converter Figure 37-10. Decimator Modes Single Sample Fast Filter No Start No Start IDLE IDLE Start Start Reset Modulator Clear Start Reset Modulator Clear Start Wait 3 Cycles to Prime 4 Stage CIC Wait 3 Cycles to Prime 4 Stage CIC Runs Once Terminal Count not Runs until stopped by CSR write of Soft Reset Count out DR1 Terminal Count not Count out DR1 Terminal Count Terminal Count Result in Output Register or to Post Processor if in Use Result in Output Register or to Post Processor if in Use Continuous Fast FIR No Start No Start IDLE IDLE Start Start Reset Modulator Clear Start Reset Modulator Clear Start Wait 3 Cycles to Prime 4 Stage CIC Wait 3 Cycles to Prime 4 Stage CIC Terminal Count not Terminal Count not Count out DR1 Count out DR1 Runs until stopped by CSR write of Soft Reset Terminal Count Runs until stopped by CSR write of Soft Reset Result in Output Register or to Post Processor if in Use Terminal Count Pass Result to Post Processor to Process Terminal Count not Count out DR2 Terminal Count Result in Output Register PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 395 Delta Sigma Converter 396 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 38. Successive Approximation Register ADC The PSoC® 5LP architecture has two successive approximation register analog to digital convertors (SAR ADC) in addition to the delta sigma ADC. The SAR ADC is designed for applications that require medium resolution and high data rate. The SAR ADC takes its input from the analog globals, locals and the mux bus and the output can be taken from a register or be sent to the UDB for further processing. 38.1 Features ■ 12-bit resolution ■ Single ended, differential input ■ Rail-to-rail input (0V to Vdda) ■ 1 MSPS sample rate ■ Four power modes ■ Single shot or continuous running mode Figure 38-1. SAR ADC Block Diagram AGR 0 AGR 1 AGR 2 AGR 3 AGR 4 AGR 5 AGR 6 AGR 7 ABUS R0 ABUS R2 VSSA AMUXBUSR PHUB AGL 0 AGL 1 AGL 2 AGL 3 AGL 4 AGL 5 AGL 6 AGL 7 ABUS L0 ABUS L2 VSSA AMUXBUSL SOC SOC + + SAR 0 SAR 1 AGL 0 AGL 2 AGL 4 AGL 6 ABUS L1 ABUS L3 Vref VSSA AGR 0 AGR 2 AGR 4 AGR 6 ABUS R1 ABUS R3 Vref VSSA End of Conversion Vref External Reference Vref Bandgap Reference PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F External Reference 397 Successive Approximation Register ADC 38.2 How It Works 38.2.1 Input Selection The SAR ADC takes differential inputs which are well connected to the analog routing structure. The positive input connects to analog globals, analog locals and Vssa. The negative input connects to analog globals, analog locals, analog mux bus, voltage reference and Vssa. The input selection, both positive and negative, is made through the input selection mux, which can be controlled through either the SAR routing registers in the analog interface or through the UDB. Setting the SARx_CSR[4] bit takes the positive input through UDB and clearing the bit takes the positive input through registers. Similarly, setting the SARx_CSR[3] bit takes the negative input through UDB and clearing the bit takes the negative input through registers. If the positive and negative input selection is made through the registers, registers SARx_SW0, SARx_SW2, SARx_SW3, SARx_SW4 and SARx_SW6 help in making the selection. 38.2.2 power modes, namely maximum power, half of maximum power, 1/3 of maximum power or 1/4 of maximum power. There is a direct tradeoff between reducing the power with one of these modes and the SNR. The power selection is done in SARx_CSR0[7:6] register bits. 38.2.5 Reference Selection The SAR ADC requires a stable reference voltage between 1 V and Vdda (maximum 5.5 V). The SAR ADC can take either an internal or an external reference. The internal reference can be Vdda/2, 1.024V, 1,2V or DAC's output voltage. The reference selection is done in SAR_CSR1[7:5] register bits. For the vdda/2 reference selection be available, SAR_CSR3[6] register bit has to be set. Clock Selection The clock to the SAR can come from one of the four available analog clocks or a UDB generated clock. SAR ADC requires a low RMS jitter clock source. The maximum frequency of the clock source is 18 MHz. The maximum jitter on the sampling clock for 12-bit resolution is 32 ps RMS. The clock selection for the SAR is made in SARx_CLK[2:0] register bits. The clock can be enabled or disabled through the gate control bit SARx_CLK[3]. The maximum input clock that can be applied to the SAR is 18 MHz. The digital output will be synchronized with respect to the corresponding digitally aligned clock of the selected analog clock. This synchronization can be bypassed using SARx_CLK[4] register bit. 38.2.3 Input Sampling The input sampling time can be programmed from the 1 to 64 cycles in register SARx_CSR2[5:0] register bits. The user can also retain the earlier DAC value or clear it at the beginning of the new sampling clock. This is done in SARx_CSR0[3] register bit. The conversion time is 18 cycles for input sampling time up to four cycles. The maximum conversion time is 78 cycles for input sampling time of 64 cycles. The sampling time is chosen based on the source's input impedance so that the input settling time is lower than the sampling time. 38.2.4 Power Modes The SAR ADC can be operated in different power modes. The user can configure to operate the SAR ADC in four 398 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Successive Approximation Register ADC Figure 38-2. SAR ADC Reference Voltage Vdda S7 Vdda Reference to SAR S3 Reference Mux Vdda/2 Vdac 1.024V 1.2V S7 Reference buffer S4 + S4 SAR_CSR3[6] S2 SAR_CSR1[7:5] Resd2 Vref/2 S2 S2 Resd3 Resd4 HiZ S2 External reference/ capacitor Resd1 Ganged Table 38-1. SAR Analog Reference Modes Switch States Mode S2 S3 S4 S7 1 External Reference ON OFF OFF OFF 2 Internal Reference with External Capacitor ON OFF OFF OFF 3 Internal Reference without External Capacitor OFF ON ON OFF 4 Vdda as Reference Voltage OFF OFF OFF ON The SAR ADC has four main reference modes: external reference, internal reference with external bypass capacitor, internal reference without external bypass capacitor, and Vdda. The external reference and internal reference with bypass capacitor modes have similar performance and can run up to a 18-MHz clock. Internal reference mode without bypass capacitance cannot meet the 1-Msps rate due to reference buffer settling limitations. The maximum clock fre- quency for this mode is 3 MHz. The bypass capacitance requirement is a minimum of 1 µF for the 12-bit noise floor requirement. The reference buffer startup time is within 500 µs with the external capacitor value of 1 µF. The reference mux selects the internal references, as shown in Table 38-2. Table 38-2. Reference Mux Input Selection SAR_CSR1[7:5] Reference Mux Input Selected 010 Vdda/2 011 Vdac 100 1.024 V Bandgap reference 1 101 1.2 V Bandgap reference 2 000, 111 NA Not used, Mux output tristated 001, 110 NA Not used 38.2.6 Description Internal Vdda/2 reference Vdac input Operational Modes The SAR can be configured in two modes, single capture or continuous. In single capture, the SAR ADC completes one conversion on a trigger; in the continuous mode the SAR ADC performs continuous conversion. The trigger can be either software or hardware. The software trigger comes PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F from SARx_CSR0[0] register bit and the hardware trigger is from the UDB. The selection between software and UDB trigger is made in SARx_CSR0[2] register bit. As long the SOF stays high the conversion continues, the conversion stops when the SOF goes low. 399 Successive Approximation Register ADC The two modes, single capture and continuous, is realized in the way the SOF bit is configured, i.e., level or edge sensitive SOF. In the level sensitive mode, the SAR ADC performs the conversion as long as the SOF bit is asserted high. So, the level sensitive mode is used for continuous conversion. In the edge sensitive mode, the SAR performs a conversion on the edge and the bit is automatically reasserted low on the completion of the conversion (on the end of frame (EOF)). So, it has to be reasserted high for the next edge for the SAR ADC to start conversion. This mode helps in performing single sample conversions. In case of hardware enabled SOF, the user can sync the conversion to a PWM frequency by configuring it in the edge mode. The level or edge triggered function of the SOF signal is configured in the SARx_CSR0[1] register bit. The conversion time of the SAR is more than 18 cycles for input sampling time of four cycles when hardware trigger is used. This is because SOF and EOF are routed through DSI routing and these signals encounter a delay, resulting in longer conversion time. 38.2.7 SAR ADC Output The SAR ADC output includes: ■ End of Frame (EOF) bit ■ The output bits of user configured resolution ■ An optional interrupt on EOF The resolution can be configured to be 8, 10 or 12 bits in SARx_CSR2[7:6] register bits. After a conversion is complete the End of Frame (EOF) bit is asserted high in SARx_CSR1[0] register bit. This bit is a clear on read sticky status bit and is cleared automatically on a data read. The conversion result is stored in the registers SARx_WRK0 and SARx_WRK1 register. The SARx_WRK1 register bits [3:0] stores the higher four bits [11:8] of the output. Coherency protection can be applied to the SAR output by setting SARx_CSR0[4] register bit. It ensures that a new output is written only when both the registers are read. The EOF output can be used to generate interrupt to the CPU or DMA. The interrupt is enabled by setting the SARx_CSR1[1] register bit. The interrupt can be made edge/ level interrupt by setting/clearing SARx_CSR1[2] register bit. 400 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Section G: Program and Debug JTAG (4- or 5-wire) or serial wire debugger (SWD) (2-wire) interfaces are used to program and debug. The single wire viewer (SWV) can also be used for “printf” style debugging. By combining SWD and SWV, you can implement a full debugging interface with just three pins. These standard interfaces enables debugging or programming the PSoC® device with a variety of hardware solutions from Cypress or third-party vendors. This section includes the following chapters: ■ Test Controller chapter on page 403 ■ Cortex-M3 Debug and Trace chapter on page 415 ■ Nonvolatile Memory Programming chapter on page 423 Top Level Architecture Program and Debug Block Diagram PROGRAM AND DEBUG System Bus Program Debug and Trace Boundary Scan PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 401 Section G: Program and Debug 402 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 39. Test Controller The PSoC® 5LP architecture includes a test controller used for the following purposes: ■ Access to I/O pins for boundary scan testing. ■ Access to the device memory and registers (via the PHUB) through the PSoC 5LP Cortex-M3 Debug Access Port (DAP) for functional testing, device programming, and program debugging. The test controller connects to off-chip devices via the Joint Test Action Group (JTAG) interface or the serial wire debug (SWD) interface. These interfaces use I/O port pins; the exact number of pins depends on the type of interface used. 39.1 Features The test controller has the following features: ■ Supports JTAG or SWD interface to a debug host ■ SWD interface available on either GPIO or USB pins ■ Supports boundary scan in accordance with the JTAG IEEE Standard 1149.1-2001 “Test Access Port and BoundaryScan Architecture” ■ Supports additional JTAG instructions/registers beyond IEEE Standard 1149, for access to the rest of the device ■ Interfaces to PSoC 5LP debug modules for access to the rest of the device for program and debug operations 39.2 Block Diagram Figure 39-1.shows how the test controller (TC) acts as the interface between the external devices such as programmers, debuggers, and the internal debug port (DAP) in PSoC 5LP. Figure 39-1. PSoC 5LP Test Controller Block Diagram SWD IO TC PHUB bus (AHB) DAP JTAG Cortex-M3 AHB SRAM AHB PHUB DMA & AHB Bridge AHB SPC NVLs AHB EEPROM NV Write Bus JTAG IO SWD Flash AHB Program Program Cache Cache Program Interface PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Flash Interface 403 Test Controller to interact with the JTAG and SWD blocks in the test controller. The abbreviations used in the figure are: TC - Test Controller. DAP - Debug Access Port of Cortex-M3 CPU (ARM). 39.3.1 AHB - Advanced High-Performance Bus, def acto standard from ARM. In response to higher pin densities on ICs, the Joint Test Action Group (JTAG) proposed a method to test circuit boards by controlling the pins on the ICs (and reading their values) via a separate test interface. The solution, later formalized as IEEE Standard 1149.1-2001, is based on the concept of a serial shift register routed across all of the pins of the IC – hence the name “boundary scan.” The circuitry at each pin is supplemented with a multipurpose element called a boundary scan cell. In PSoC 5LP devices, most GPIO and SIO port pins have a boundary scan cell associated with them (see GPIO and SIO block diagrams in the I/O System chapter on page 151). PHUB - Peripheral HUB, advanced multi spoke bus controller which allows many different functional blocks to communicate without involving of CPU for setting up the bus transaction. DMA - Direct Memory Access controller. SRAM - Static Random Access Memory. SPC - System Performance controller implements R/W interface with non-volatile memory. NVL - non-volatile latch. 39.3 JTAG Interface The interface used to control the values in the boundary scan cells is called the Test Access Port (TAP) and is commonly known as the JTAG interface. It consists of three signals: Test Data In (TDI), Test Data Out (TDO), and Test Mode Select (TMS). Also included is a clock signal (TCK) that clocks the other signals. Background Information The following information helps to familiarize you with the JTAG interface and the IEEE 1149 Specification and the Serial Wire Debug (SWD) interface. An understanding of the state machine architecture in these interfaces is necessary TDI, TMS, and TCK are all inputs to the device and TDO is output from the device. This interface enables testing multiple ICs on a circuit board, in a daisy-chain fashion, as shown in Figure 39-2. Figure 39-2. JTAG Interface to Multiple ICs on a Circuit Board TMS TCK TMS TCK Device 1 TMS TCK Device 2 TMS TCK Device 3 TDI TDI TDO TDI TDO TDI TDO TDO 404 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Test Controller Within each device, the JTAG interface architecture is shown in Figure 39-3. Data at TDI is shifted in, through one of several available registers, and out to TDO. Figure 39-3. JTAG Interface Architecture Boundary Scan Path Boundary Scan Cells IO Pads Core Logic TDI Instruction Register BYPASS Register ID Register Other Register TCK TMS TRST TDO Test Access Port Controller The TMS signal controls a state machine in the TAP. The state machine controls which register (including the boundary scan path) is in the TDI-to-TDO shift path, as shown in Figure 39-4 on page 406. The following terms apply: ■ ir – the instruction register ■ dr – one of the other registers (including the boundary scan path), as determined by the contents of the instruction register ■ capture – transfer the contents of a dr to a shift register, to be shifted out on TDO (read the dr) ■ update – transfer the contents of a shift register, shifted in from TDI, to a dr (write the dr) PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 405 Test Controller Figure 39-4. TAP State Machine TMS = 1 test logic reset TMS = 0 TMS = 0 run test idle TMS = 1 TMS = 1 select dr scan TMS = 0 TMS = 0 TMS = 0 TMS = 1 capture dr TMS = 0 TMS = 0 TMS = 1 exit 1 dr exit 1 ir TMS = 0 TMS = 0 pause dr TMS = 1 TMS = 0 exit 2 dr TMS = 1 update dr update ir TMS = 1 TMS = 0 The registers in the TAP are: Instruction – Typically two to four bits wide, holds the current instruction that defines which data register is placed in the TDI-to-TDO shift path. ■ Bypass – one bit wide, directly connects TDI with TDO, causing the device to be bypassed for JTAG purposes. ■ ID – 32 bits wide, used to read the JTAG manufacturer/ part number ID of the device. ■ Boundary Scan Path (BSR) – Width equals the number of I/O pins that have boundary scan cells, used to set or read the states of those I/O pins. TMS = 0 The standard set of instructions (values that can be shifted into the instruction register), as specified in IEEE 1149, are: ■ EXTEST – Causes TDI and TDO to be connected to the boundary scan path (BSR). The device is changed from its normal operating mode to a test mode. Then, the device's pin states can be sampled using the capture dr JTAG state, and new values can be applied to the pins of the device using the update dr state. ■ SAMPLE – Causes TDI and TDO to be connected to the BSR, but the device is left in its normal operating mode During this instruction, the BSR can be read by the capture dr JTAG state to take a sample of the functional data entering and leaving the device. Other registers may be included in accordance with device manufacturer specifications. ■ 406 TMS = 0 exit 2 ir TMS = 1 ■ TMS = 0 pause ir TMS = 1 TMS = 1 TMS = 0 shift ir TMS = 1 TMS = 0 TMS = 1 capture ir shift dr TMS = 1 TMS = 1 select ir scan PRELOAD – Causes TDI and TDO to be connected to the BSR, but device is left in its normal operating mode. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Test Controller 39.3.2 The instruction is used to preload test data into the BSR before loading an EXTEST instruction. The SWD interface has two signals: data (SWDIO) and clock for data (SWDCK). The host programmer always drives the clock line, whereas either the programmer or PSoC 5LP device drives the data line. Host programmer and PSoC 5LP communicate in packet format through the SWD interface. Write packet refers to the SWD packet transaction in which the host writes data to PSoC 5LP. Read packet refers to the SWD packet transaction in which the host reads data from PSoC 5LP. The Write packet and Read packet formats are illustrated in Figure 39-5 and Figure 39-6, respectively Optional, but commonly available, instructions are: ■ IDCODE – Causes TDI and TDO to be connected to an IDCODE register. ■ INTEST – Causes TDI and TDO to be connected to the BSR. While the EXTEST instruction allows access to the device pins, INTEST enables similar access to the corelogic signals of a device. Serial Wire Debug Interface For more information, see the IEEE Standard, available at http://www.ieee.org. Figure 39-5. SWD ‘Write Packet’ Timing Diagram SWDCK (Driven by Host) SWDIO driven by: Parity wdata[31] 0 ACK[0:2] wdata[1] 0 wdata[0] 1 TrN z TrN A[2:3] Park (1) Stop (0) Parity APnDP RnW (0) z Start (1) SWDIO (Bidirectional) PSoC 5LP Host 0 0 0 Dummy Phase (3'b000) Host a.) Host Write Operation: Host sends data on the SWDIO line on the falling edge of SWDCK; PSoC 5LP reads that data on the next SWDCK rising edge (Example: 8-bit header data, Write data(wdata[31:0]), Dummy phase (3'b000)) b.) Host Read Operation: PSoC 5LP sends data on the SWDIO line on the rising edge of SWDCK; host reads that data on the next SWDCK falling edge (Example: ACK data (ACK[2:0]) c.) The host should not drive the SWDIO line during TrN phase. During first TrN phase (½ cycle duration) of SWD packet, PSoC 5LP drives the ACK data on the SWDIO line on the rising edge of SWDCK. The host should read the data on the subsequent falling edge of SWDCK. The second TrN phase is 1.5 SWDCK clock cycles. Both PSoC 5LP and the host will not drive the line during the entire second TrN phase (indicated as ‘z’). Host starts sending the Write data (wdata) on the next falling edge of SWDCK after second TrN phase. d.) “DUMMY” phase is three SWD clock cycles with SWDIO line low. This DUMMY phase is not part of SWD protocol. The three extra clocks with SWDIO low are required for the Test Controller in PSoC 5LP to complete the Read/Write operation when the SWDCK clock is not free-running. For a reliable implementation, include three IDLE clock cycles with SWDIO low for each packet. According to the SWD protocol, the host can generate any number of SWD clock cycles between two packets with SWDIO low. Figure 39-6. SWD ‘Read Packet’ Timing Diagram SWDIO driven by: Host TrN Parity rdata[31] 0 rdata[30] 0 ACK[0:2] rdata[1] 1 rdata[0] TrN Park (1) Stop (0) A[2:3] Parity RnW (1) APnDP SWDIO (Bidirectional) Start (1) SWDCK (Driven by Host) 0 0 0 Dummy Phase (3'b000) PSoC 5LP Host a.) Host Write Operation: Host sends data on the SWDIO line on the falling edge of SWDCK; PSoC 5LP reads that data on the next SWDCK rising edge (Example: 8-bit header data, Dummy phase (3'b000)) b.) Host Read Operation: PSoC 5LP sends data on the SWDIO line on the rising edge of SWDCK; the Host reads that data on the next SWDCK falling edge (Example: ACK data (ACK[2:0], Read data (rdata[31:0])) c.) The host should not drive the SWDIO line during TrN phase. During first TrN phase (½ cycle duration) of SWD packet, PSoC 5LP drives the ACK data on the SWDIO line on the rising edge of SWDCK. The host should read the data on the subsequent falling edge of SWDCK. The second TrN phase is 1.5 SWDCK clock cycles. Both PSoC 5LP and the host will not drive the line during the entire second TrN phase (indicated as ‘z’). Host starts sending the Dummy phase (3'b000) on the next falling edge of SWDCK after the second TrN phase. d.) “DUMMY” phase is three SWD clock cycles with SWDIO line low. This DUMMY phase is not part of SWD protocol. The three extra clocks with SWDIO low are required for the Test Controller in PSoC 5LP to complete the Read/Write operation when the SWDCK clock is not free-running. For a reliable implementation, include three IDLE clock cycles with SWDIO low for each packet. According to the SWD protocol, the host can generate any number of SWD clock cycles between two packets with SWDIO low. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 407 Test Controller A complete data transfer requires 46 clocks (not including the optional three dummy clock cycles in Figure 39-5 and Figure 39-6). Each data transfer consists of three phases: ■ Packet request – External host programmer issues a request to PSoC 5LP. ■ Acknowledge response – PSoC 5LP sends an acknowledgement to the host. ■ Data – This is valid only when a packet request is followed by a valid (OK) acknowledge response. The data transfer is either: ■ PSoC 5LP to host, following a read request – RDATA ■ Host to PSoC 5LP, following a write request – WDATA In Figure 39-5 and Figure 39-6, the following sequence occurs: 1. The start bit initiates a transfer; it is always logic ‘1’. 2. The APnDP bit determines whether the transfer is an AP access, ‘1’, or a DP access, ‘0’. 3. The next bit is RnW, which is ‘1’ for a read from PSoC 5LP, or ‘0’ for a write to PSoC 5LP. 4. The ADDR bits (A[3:2]) are register select bits for access port or debug port. See Table 39-8 for address bit definitions. 5. The parity bit has the parity of APnDP, RnW, and ADDR. This is even parity bit. If number of logical 1’s in these bits is odd, then parity must be ‘1’, otherwise it is ‘0’. If the parity bit is not correct, the header is ignored by the target device; there is no ACK response. For host implementation, the programming operation should be stopped and tried again by doing a device reset. 6. The stop bit is always logic ‘0’. 7. The park bit is always logic’1’ and should be driven high by the host. 8. The ACK bits are the device-to-host response. Possible values are shown in Table 39-1. Note that the ACK in the current SWD transfer reflects the status of the previous transfer. OK ACK means the previous packet was successful. WAIT response indicates that the previous packet transaction is not yet complete. For a Fault operation, the programming operation should be aborted immediately. Table 39-1. ACK Response for SWD Transfers ACK[2:0] JTAG SWD OK 010 001 WAIT 001 010 FAULT 100 100 a. For a WAIT response, if the transaction is a read, the host ignores the data read in the data phase. PSoC 5LP does not drive the line and the host must not check the parity bit as well. 408 b. For a WAIT response, if the transaction is a write, PSoC® 5LP ignores the data phase. However, the host must still send the data to be written from an implementation standpoint. The parity data corresponding to the data should also be sent by the host. c. A WAIT response indicates that the PSoC 5LP device is processing the previous transaction. The host can try for a maximum of four continuous WAIT responses to see if an OK response is received, failing which, it can abort the programming operation and retry. d. For a FAULT response, the programming operation should be aborted and retried by doing a device reset. 9. The data phase includes a parity bit (even parity, similar to the packet request phase). a. For a read data packet, if the host detects a parity error, then it must abort the programming operation and restart. b. For a write data packet, if the PSoC 5LP detects a parity error in the data packet sent by the host, it generates a FAULT ACK response in the next packet. 10. Turnaround (TrN) phase: According to the SWD protocol, the TrN phase is used both by the host and PSoC 5LP to change the Drive modes on their respective SWDIO line. During the first TrN phase after packet request, PSoC 5LP drives the ACK data on the SWDIO line on the rising edge of SWDCK in TrN phase. This ensures that the host can read the ACK data on the next falling edge. Thus, the first TrN cycle is only for half cycle duration. The second TrN phase is one-and-a-half cycle long. Neither the host nor PSoC 5LP should drive SWDIO line during both phases as indicated by ‘z’ in Figure 39-5 and Figure 39-6. 11. The address, ACK, and read and write data are always transmitted least significant bit (LSB) first. 12. At the end of each SWD packet in Figure 39-5 and Figure 39-6, there is a “DUMMY” phase, which is three SWD clock cycles with SWDIO line held low. The dummy phase is not part of the SWD protocol. The three extra clocks with SWDIO low are required for the Test Controller in PSoC 5LP to complete the Read/Write operation when the SWDCK clock is not free-running. For a reliable implementation, include three IDLE clock cycles with SWDIO low for each packet. According to the SWD protocol, the host can generate any number of SWD clock cycles between two packets with SWDIO low. Note The SWD interface can be reset anytime during programming by clocking 50 or more cycles with SWDIO high. To return to the idle state, SWDIO must be clocked low once. The host programmer can begin a new SWD packet transaction from the idle state. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Test Controller 39.4 How It Works Figure 39-7. Test Controller Internal Architecture CM3 DAP security logic SWD FSM usb port 1 swclktck / swditms dap tdi JTAG TAP Test Controller tck / tms tdi tc tdo JTAG TAP IO & Port Logic dap swdo dap tdo tc tdo SWD FSM PRT1ACQ TST_KEY USBACQ TST_CR1 TST_CR2 Test Logic TST_CR3 Control Signals TST_CR4 PM_CTRL tdo swdo tc swdo tc tdo dap tdo dap swdo Figure 39-7 illustrates the detailed architecture of the test controller. It shows how the test controller acts as the interface between external devices and the internal blocks in PSoC 5LP. An understanding of the working of the test controller is necessary to develop programmers and debuggers for PSoC 5LP. At the core of the TC are the JTAG TAP and SWD FSM. The JTAG Test Access Port (TAP) implements the state machine shown in Figure 39-4. There are two JTAG TAP's in PSoC 5LP – one inside the TC and the other inside the Cortex M3 Debug and Access Port (DAP). The SWD Finite State Machine (FSM) is used to decode the SWD packets shown in Figure 39-5 and Figure 39-6. There are two SWD FSMs – one inside the TC and the other inside the DAP; they are connected in parallel. The SWD interface is supported on two pairs of pins - Port 1 SWD and USB SWD pins. The acquisition logic inside the TC selects the appropriate SWD pair by monitoring the activity on these pins and also based on the debug port select (DPS), Debug Enable NVL settings explained in the Nonvolatile Latch chapter on page 85. Unlike SWD, JTAG interface is supported only on PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Port 1. The Port 1 SWD pins share functionality with the JTAG pins (TMS and SWDIO, TCK and SWDCK). The JTAG TAP, SWD FSM inside the test controller also provides access to the TC registers while the JTAG TAP, SWD FSM inside the CM3 DAP provides access to rest of the device registers and memory. 39.4.1 JTAG Interface Implementation In PSoC 5LP, there are two JTAG TAPs – one in the TC and another in the Cortex-M3's DAP. The two TAPs are connected in series, as shown in Figure 39-7, which means the TDO of the TC is connected to the TDI of the DAP. Each TAP has a 4-bit instruction register and a 35-bit data register. The TC TAP is used to access the TC registers and the boundary scan while the DAP's TAP is used for firmware debug and programming. Typically, only one of the TAP's will be used at a time and the other TAP will be in BYPASS mode of the JTAG. The instruction set for the JTAG TAP is listed in Table 39-2. 409 Test Controller Table 39-2. JTAG TAP Instruction Set Instruction Applicable Interface Instruction Code [3:0] PSoC 5LP Function EXTEST TC 0110 See IEEE 1149.1-2001 PRELOAD TC 0010 See IEEE 1149.1-2001 0101 Connects TDI and TDO to the BYPASS register, and sets the pins to the current contents of the boundary scan register CLAMP TC INTEST TC 0100 See IEEE 1149.1-2001 ABORT DAP 1000 Aborts the current AP access instruction. Connects TDI and TDO to the DP/AP Access register. DPACC TC and DAP 1010 Connects TDI and TDO to the DP/AP Access register, for accesses to the Debug Port registers. APACC TC and DAP 1011 Connects TDI and TDO to the DP/AP Access register, for accesses to the Access Port registers. IDCODE TC and DAP 1110 See IEEE 1149.1-2001 BYPASS TC and DAP 1111 See IEEE 1149.1-2001 EXTEST, IDCODE, BYPASS, and PRELOAD are all JTAG required instructions and information regarding these instructions can be found in the JTAG IEEE Test Access Port and Boundary-Scan Architecture Specification document (Std 1149.1). The 32-bit JTAG ID code is returned when the IDCODE instruction is executed. Note that the 32bit JTAG ID returned will be different for TC and the DAP. The TC returns the device JTAG ID code when an IDCODE instruction is executed. The device JTAG ID code is given in the respective device datasheets. The DAP has its own JTAG ID and it will always return the Cortex M3 JTAG ID irrespective of the P 5LP device manufacturing part number. From an end user perspective, only the TC JTAG ID is useful because it can be used to uniquely identify the different manufacturing part numbers. The 35-bit data register is used for the DPACC, APACC, and ABORT instructions. In the JTAG update dr state, the structure is as follows. ■ Bits 34 to 3 (32 bits of data) – If the Port register is less than 32 bits wide, only the N LS bits are transferred, where N is the width of the Port register. ■ Bits 2 to 1 – 2-bit address for Debug or Access Port register select (Table 39-3). ■ Bit 0 – RnW – 1 = read (from device to debug host); 0 = write (to device from debug host) In the JTAG capture dr state, the structure is as follows. The ABORT instruction is an ARM instruction used to abort the current AP access transaction. The DPACC instruction is used to access the Debug Port configuration registers. The APACC is the instruction used to access the JTAG Access Port registers. The address of the important APACC and DPACC access registers is given in Table 39-3. The complete register set can be found in the ARM Cortex-M3 Technical Reference Manual r1p0 (ARM TRM). ■ Bits 34 to 3 (32 bits of data) – If the Port register is less than 32 bits wide, only the N LS bits are transferred, where N is the width of the port register. ■ Bits 2 to 0 (ACK response code) – Depending on the interface, the ACK response is indicated in Table 39-4. Table 39-4. ACK Response for JTAG/SWD Transfers Table 39-3. JTAGP DP Access and AP Access Registers Addr[1:0] 2'b00 DPACC Reserved JTAG SWD 3'b010 3'b001 Control and Status Word WAIT 3'b001 3'b010 Transfer Address FAULT 3'b100 3'b100 APACC 2b01 DP Control/Status 2b10 AP Select Register Reserved 2'b11 Read Buffer Data Read/Write 410 ACK OK The IDCODE instruction only uses 32-bits of the data register and the BYPASS instruction only uses 1-bit of the data register. For all instructions, parallel data is entered into the data register during the capture data register state and parallel data is outputted and compared in the update data register state. When the BYPASS instruction is selected the 1bit data register is set to ‘0’ in the capture data register state. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Test Controller For instructions PRELOAD, EXTEST, and INTEST the boundary scan register is selected between TDI and TDO. The CLAMP instruction loads the 1-bit BYPASS register between TDI and TDO and asserts the TC's clamp output, which alerts the port logic to set the I/Os to the current contents of the boundary scan register. In most cases, the external device will see two TAPs connected in series and this determines if access is granted to the DAP. See Cortex M3 DAP Access Criteria on page 412 for details on DAP access. When configuring the two TAPs, the total instruction register length is 8-bits (4-bits per TAP). Figure 39-8. Instruction Register Length for both TAPs Instruction Regs. TDI TC TDO DAP TMS (swditms) is the input used to navigate between the JTAG states and is sampled on the rising edge TCK (clk_swtck). TDI is the input used to shift in the instruction and data and is also sampled on the rising edge of TCK. The instructions are loaded during the Shift-IR state of the JTAG FSM (lsb first) with the last bit of the instruction shifted in on the transition between Shift-IR and Exit-IR. TMS is then used to navigate to the Update-IR state which is the state that updates the newly loaded instruction. Similar to the instruction register, the data register is accessed by using TMS to navigate to the Shift-DR state. Data is shifted in (LSB first) with TDI, and TDO shifts out the data on the negative edge of TCK. As mentioned above, the data register changes size based on the selected instruction. The typical mode that will be used for debug and device programming is the TC's TAP in BYPASS with the DAP's TAP instruction set to APACC. APACC is the instruction used to access the Transfer Address register and the Data Read/ Write registers which are the two registers required to initiate data transactions over the system bus. Figure 39-9. TC in BYPASS while DAP in APACC Mode Data Regs. { bypass, apacc}, read_ data = data_reg[34:3] Data[34:3] 0 TDO TDI TC DAP ack[2:0] or {addr[2:1],RnW} Figure 39-10. TC in APACC while DAP in BYPASS Data Regs. { apacc, bypass}, read_ data = data_reg[35:4] Data[34:3] TDO TDI 0 DAP TC ack[2:0] or {addr[2:1],RnW} [35:0] A write transaction is triggered by first setting the Transfer Address register, then setting the Data Read/Write register with your write data. The write transaction is triggered at the Update-DR state. On a read request, bits [34:3] are ignored during the first read, and the returned read data will be loaded in the data register when the JTAG FSM goes through the Update-DR state to the Capture-DR state. When the contents of the data register are shifted out in the Shift-IR state the read data will be stored at bits [34:3], and bits[2:0] will contains the ACK response. So the data register has to be read twice and the data read during the second read is the actual register data. Note because the DAP is bypass the length of both data registers combined is 36-bits, so taking this into account the actual read data is stored in bits [35:4] and the ACK response will be stored in bits [3:1]. If the read is successful the "OK" response will be returned. If another read request is initiated while the last read request is in progress the "WAIT" response is returned. To avoid WAIT responses, read transfers require an additional five TCK clock cycles in the idle state in between the read request and the shift out of the data register which will return read data. Fault responses occur when attempting to access a non-TC register when access is not granted. 39.4.2 SWD Interface Implementation The serial wire access port uses the serial wire debug (SWD) protocol developed by ARM. SWD uses the same APACC and DPACC registers as JTAG to execute data transactions, but also allows SWD to read the DAPs SWD ID Code from the DPACC register space. See Table 39-5 Table 39-5. SWD DP Access and AP Access Registers Addr[1:0] DPACC APACC 2'b00 SWD ID Code Control and Status Word 2b01 DP Control/Status Transfer Address 2b10 AP Select Register Reserved 2'b11 Read Buffer Data Read/Write [35:0] TC registers can be accessed by putting the TC in APACC and the DAP in BYPASS. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F The SWD packet format for write and read transactions is given in Figure 39-5 and Figure 39-6 respectively. Unlike the JTAG TAPs, the SWD FSM's in TC and the DAP are connected in parallel as shown in Figure 39-7. Based on the address received from the host, the appropriate SWD FSM will respond. If the 32-bit address from the host corresponds 411 Test Controller to the test controller register address, the TC SWD FSM output is selected. For addresses corresponding to device memory (SRAM, flash) and other registers in PSoC 5LP, the output of the SWD FSM from the CM3 DAP will be selected. 39.4.3 Acquiring the Debug Port The first step in programming, debugging, or doing a boundary scan is to acquire the debug port. This gives the external device access to the physical interface of the selected protocol. There are two methods of acquiring the debug port. ■ Two nonvolatile latch (NVL) bits (CNVL_DPS[1:0]) determine the state of the Port 1 Debug interface pins after device reset is released. The meaning of these bits is shown in Table 39-6. Table 39-6. Port 1 Debug Port NVL Settings CNVL_DPS[1:0] Port Configuration 00 5 - Wire JTAG (nTRST is included) 01 (default) 4 - Wire JTAG (nTRST is not used) 10 Serial wire debug (SWD) 11 Debug Port Disabled (GPIO) If the NV latch bits are anything other than the Debug Port Disabled setting (GPIO setting), then the corresponding debug interface on Port 1 is accessible anytime to the external device. Boundary scan testing using standard boundary scan tools, device programming using standard JTAG programmers require that the CNVL_DPS[1:0] be set to one of the JTAG settings (4wire JTAG or 5-wire JTAG). Devices coming from the factory have the default DPS setting as 4-wire JTAG to enable support for standard third party JTAG tools. Note that these settings are applicable only for Port 1 debug pins. Programming, debugging over the USB SWD pins require the method given in next step to acquire the port. ■ If the Port 1 debug port is disabled (GPIO NVL setting) or to program/debug over the USB SWD pins, the only way to gain debug access to the part is to enter a valid port acquire key within a key window period of 8 µs after reset (8 µs is only the initial window, it extends to 400 µs if eight clocks are sampled in 8 µs). The port acquire key must be transmitted over one of the two SWD pin pairs, as indicated in the following table. Table 39-7. SWD Pin Pairs SWD Pin Pair SWDIO SWDCK Standard P1[0] P1[1] Alternate P15[6] (USB D+) P15[7] (USB D-) The SWD packet request phase consists of: ■ APnDP = 0 412 ■ RnW = 0 ■ ADDR = 11 ■ WDATA = 0x7B0C06DB with WDATA Parity = 0 The SWD frame should be transmitted until the ACK response is OK. If the debug port is disabled (GPIO setting), then the OK ACK will typically be received during the second try. The SWD interface will be in the idle state, ready for the next write. See PSoC 5LP Programming Specifications for detailed timing diagrams on the test controller acquisition. Note that irrespective of the CNVL_DPS settings, the debug port can always be acquired through one of the SWD pairs by sending the port acquire key. When acquired, the active debug port gets configured to the SWD port on which the acquisition key was sent. Further transactions can be done only on the selected SWD pin pair. If the Port 1 SWD pair was used for acquiring, there is the option of switching to JTAG interface by sending the JTAG to SWD switching sequence on the Port 1 pins. The active debug port will then be configured to 4-wire JTAG. 39.4.4 Cortex M3 DAP Access Criteria Cortex M3 DAP access is required for device programming and debugging. The DAP access criteria discussed in this section assumes that the debug port is already acquired as explained in the previous section. This is a necessary requirement for DAP access. The SWD FSM, JTAG TAP in the Cortex M3 DAP can be accessed only if the DAP access is enabled by the DEBUG_EN customer NVL bit or by test mode entry. The test mode entry details can be found in the PSoC 5LP Programming Specifications. Access will be completely blocked if the Write Once Latch (WOL) is set with the correct security key, as explained in Nonvolatile Latch chapter on page 85. When DAP access is granted, the two JTAG TAPs are in series and the bottom multiplexer in Figure 39-7 selects dap_tdo. When DAP access is not granted or the device is in a low-power mode, only the TC TAP will be seen by the external device and the bottom multiplexer selects tc_tdo. SWD mode follows the same security and low power rules as JTAG. When DAP access is not granted or the device is in a low-power mode, tc_swdo is selected by the bottom multiplexer in Figure 39-7. When DAP access is granted, the default selection of the bottom multiplexer is dap_swdo. The TC is always monitoring SWD transactions, and when either a port acquire attempt happens or TC register is being accessed, tc_swdo is selected. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Test Controller When the write once latch is set to enable device security, the debug port may still be acquired through the SWD interface but the TC functionality is limited: ■ Test Mode cannot be entered ■ Access to the DAP is denied ■ JTAG mode may not be entered (no boundary scan) ■ All reads of the Data Read/Write register return 0x5AFE5AFE 39.4.5 Device Programming Device programming refers to the programming of the nonvolatile memory in PSoC 5LP. The first two steps in device programming are the acquisition of the debug port and also enabling access to the CM3 DAP to perform the programming operations. The complete sequence for programming a PSoC 5LP device through the SWD, JTAG interface is given in the PSoC 5LP Programming Specifications. 39.4.6 Boundary Scan To perform a boundary scan: 1. At reset, assume that the pins state is unknown. 2. Optional: Do a port acquire within the key window, which enables the SWD interface. This step is not required if the DPS setting is 4-wire JTAG or 5-wire JTAG. 3. Optional: Shift to the JTAG interface from SWD. This step is required only if step 2 is applicable. 4. Reset the JTAG TAP. To perform a boundary scan in PSoC 5LP using standard boundary scan tools, the debug port NVL setting should only be either of 4-wire JTAG (default setting) or 5-wire JTAG. Also, the DEBUG_EN NVL setting determines whether the DAP TAP will be in series with the TC TAP (DEBUG_EN = 1) or the DAP TAP will not be accessible by the external 39.4.7 boundary scan tool ((DEBUG_EN = 0). Devices coming from the factory have the DEBUG_EN NVL setting as 1, thereby enabling the DAP TAP by default. In this case, the DAP TAP should be configured to be in BYPASS mode of JTAG and the TC TAP should be loaded with the boundary scan instructions. The data register access should also consider the 1-bit addition to its length due to the presence of 1bit BYPASS register along with the TC data register (similar to the example in Figure 6). If DEBUG_EN = 0, then only the TC TAP needs to be loaded with the boundary scan instructions and the data register length will only be that of the TC TAP according to the selected boundary scan instruction. 5. Start doing boundary scan operations. The instruction, data register accesses during boundary scan operation should also consider whether the DAP TAP is present (DEBUG_EN = 1) or not (DEBUG_EN = 0). Boundary Scan Pin Order For the 100-pin TQFP device, the boundary scan path (BSR) is connected to the I/O pins around the part from TDI (P1[4]) through TDO (P1[3]), in the order shown in the following table. Table 39-8. Boundary Scan Pin Order BSR# Pin BSR# Pin BSR# Pin BSR# Pin BSR# Pin BSR# Pin 1 P1[5] 13 P15[1] 25 P15[3] 37 P0[7] 49 P15[5] 61 P6[5] 2 P1[6] 14 P3[0] 26 P12[2] 38 P4[2] 50 P2[0] 62 P6[6] 3 P1[7] 15 P3[1] 27 P12[3] 39 P4[3] 51 P2[1] 63 P6[7] 4 P12[6] 16 P3[2] 28 P4[0] 40 P4[4] 52 P2[2] 64 XRES 5 P12[7] 17 P3[3] 29 P4[1] 41 P4[5] 53 P2[3] 65 P5[0] 6 P5[4] 18 P3[4] 30 P0[0] 42 P4[6] 54 P2[4] 66 P5[1] 7 P5[5] 19 P3[5] 31 P0[1] 43 P4[7] 55 P2[5] 67 P5[2] 8 P5[6] 20 P3[6] 32 P0[2] 44 P6[0] 56 P2[6] 68 P5[3] 9 P5[7] 21 P3[7] 33 P0[3] 45 P6[1] 57 P2[7] 69 P1[2] 10 P15[6] 22 P12[0] 34 P0[4] 46 P6[2] 58 P12[4] 11 P15[7] 23 P12[1] 35 P0[5] 47 P6[3] 59 P12[5] 12 P15[0] 24 P15[2] 36 P0[6] 48 P15[4] 60 P6[4] Similar boundary scan paths exist on the 68-pin QFN part. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 413 Test Controller 414 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 40. Cortex-M3 Debug and Trace The PSoC® platform provides extensive support for programming, testing, debugging, and tracing both hardware and firmware. PSoC 5LP supports four interfaces: JTAG, SWD, SWV, and TRACEPORT. Cortex-M3 debug and trace functionality enables full device debugging in the final system using the standard production device. Cortex-M3 debugging features are classified into two types: invasive debugging and noninvasive debugging. Invasive debugging includes program halting and stepping, breakpoints, data watchpoints, register value access, and ROM-based debugging. Noninvasive debugging includes memory access, instruction trace, data trace, software trace, and profiling. 40.1 Features ■ Debug access to all memory and registers in the system, including Cortex-M3 register bank when the core is running or halted ■ JTAG or SWD access ■ Flash Patch and Breakpoint (FPB) block for implementing breakpoints and code patches ■ Data Watchpoint and Trace (DWT) block for implementing watchpoints, trigger resources, and system profiling ■ Embedded Trace Macrocell (ETM) for instruction trace ■ Instrumentation Trace Macrocell (ITM) for support of printf style debugging ■ Support for six breakpoints and four watchpoints ■ Trace Port Interface Unit (TPIU) for bridging to a Trace Port Analyzer (TPA) PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 415 Cortex-M3 Debug and Trace Figure 40-1. Debug and Trace Block Diagram Debug control and data access occurs through the Advanced High-performance Bus-Access Port (AHB-AP) interface. This interface is driven by either the serial wire debug port (SW-DP) or serial wire JTAG debug port (SWJDP) components. Through internal PPB, the debugger can access: ■ Nested Vectored Interrupt Controller (NVIC) - for debug access to the processor core 40.2 How It Works The PSoC 5LP JTAG and SWD interfaces comply with standard specifications and offer extensions unique to PSoC 5LP architecture. 40.2.1 Test Controller (TC) The Test Controller is used for the following purposes: ■ DWT ■ Access to I/O pins for boundary scan testing. ■ FPB ■ ■ ITM Access to the device memory and registers (via the PHUB) through PSoC 5LP Cortex-M3 Debug Access Port (DAP) for functional testing, device programming, and program debugging. Through external PPB, the debugger can access: ■ ETM ■ Trace Port Interface Unit (TPIU) Through the DCode bus, the debugger can access memory located in the code space. The system bus provides access to bus, memory, and peripherals located in the system bus space. 416 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex-M3 Debug and Trace Figure 40-2. PSoC 5LP Test Controller interface TDI SWDITMS TDI_OUT SWCLKTCK nTDOEN TDO SWDOEN TMS_OUT Test Controller TCK_OUT DAP Cortex-M3 TDO_IN SWDO_IN SWDO In PSoC 5LP devices, under certain JTAG instructions, the JTAG or SWD signals are passed to the ARM Debug Access Port. For details, see the Test Controller chapter on page 403. 40.2.2 PSoC 5LP JTAG Instructions The PSoC 5LP JTAG interface complies with the IEEE 1149.1-2001 Specification and provides additional instructions. The instruction register is four bits wide. Instructions are listed in Table 39-2 on page 410. 40.2.2.1 Debug Port and Access Port Registers The registers are part of the ARM Cortex-M3 Debug Access Port (DAP). In the PSoC 5LP Cortex-M3, the DAP consists of the SWD/JTAG Debug Port (SWJ-DP) and the AHB Access Port (AHB-AP). The registers are listed in Table 39-3 on page 410. For more information on these ports and their registers, see the ARM Debug Interface Architecture Specification (for the SWJ-DP), and the ARM Cortex-M3 Technical Reference Manual (for the AHB-AP), available at http://www.arm.com. 40.2.2.2 Test Controller Interface Pins Two NV latch bits determine the state of the JTAG/SWD interface pins at reset. The settings of the bits are shown in Table 39-6 on page 412. The SWV interface consists of a single output signal (TRACESWO) that shares a pin with the JTAG TDO signal. When the pins are configured for SWD mode, then SWV is also routed to the TDO/TRACESWO pin. 40.2.3 TRACEPORT In PSoC 5LP devices, the TRACEPORT pins are a part of the TPIU and are used to provide the trace output. The TRACEPORT has five pins and is used for the fast transmission of large trace streams. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 40.3 Core Debug Core debug allows users to exercise features such as enabling debug, halting, stepping, and accessing the PSoC memory and registers. Core debug is accessed through the core debug registers. The main core debug registers are: ■ Debug Halting Control and Status Register (DHCSR) ■ Debug Exception and Monitor Control Register (DEMCR) ■ Debug Core Register Data Register (DCRDR) ■ Debug Core Register Selector Register (DCRSR) Among these the DHCSR allows enabling the core debug, providing status information about the state of the processor, and halting and stepping the processor. More details regarding these registers can be found in the ARM CortexM3 Technical Reference Manual, available at http:// www.arm.com. 40.3.1 Enabling the Debug The core debug can be enabled C_DEBUGEN bit of the DHCSR. by setting the Note that when the debug controller is enabled, it can read the entire flash memory regardless of the flash protection setting. Therefore, if flash protection is required, the debug controller also needs to be disabled. 40.3.2 Halting The debugger can halt the core by setting the C_DEBUGEN and C_HALT bits of the DHCSR. The core acknowledges when halted by setting the S_HALT bit of the DHCSR. 40.3.3 Stepping The core can be single stepped by halting the core, setting the C_STEP bit to ‘1’, and then clearing the C_HALT bit to ‘0’. The core acknowledges completion of the step and rehalts by setting the S_HALT bit of the Debug Halting Control and Status Register. The core can exit halting debug C_DEBUGEN bit in the DHCSR. 40.3.4 by clearing the Accessing PSoC Memory and Registers The DCRDR and DCRSR are used to access the PSoC memory and registers. The register and memory access are 32 bits wide. 417 Cortex-M3 Debug and Trace To use the registers to read the contents of a register, the perform the following steps: 40.4 1. Set the C_DEBUGEN and C_HALT bits of the DHCSR. This enables the debug and halts the core. The processor contains several system debug components that facilitate low cost debug, trace and profiling, breakpoints, watchpoints and code patching. 2. Wait for the S_HALT bit of the DHCSR to be set. This indicates that the core is halted. System Debug The system debug components are: 3. Write to the DCRSR with bit 16 set to ‘0’, indicating it is a read operation. ❐ Flash Patch and Breakpoint (FPB) unit to implement breakpoints and code patches. 4. Poll until the S_REGRDY bit in DHCSR is ‘1’. ❐ Data Watchpoint and Trace (DWT) unit to implement watchpoints, trigger resources, and system profiling. ❐ Instrumentation Trace Macrocell (ITM) for application-driven trace source that supports printf style debugging. ❐ Embedded Trace Macrocell (ETM) for instruction trace. The processor is supported in versions with and without the ETM. 5. Write the register number to be read into the DCRSR. 6. Read the value from the DCRDR. To write to a register, perform the following steps: 1. Make sure the processor is halted by following steps 1 and 2 mentioned above. 2. Write data value to the DCRDR. 3. Write to the DCRSR with bit 16 set to ‘1’, indicating it is a write operation. 4. Write the register number that you want to write to into the DCRSR. 5. Poll until the S_REGRDY bit in DHCSR is ‘1’. When the bit becomes ‘1’, the write operation is complete. The Memory Access Port (MEM-AP) provides access to the memory through the DAP. All accesses to a MEM-AP are made through the MEM-AP registers. All registers are 32 bits wide. The important registers required for memory access include: 40.4.1 Flash Patch and Breakpoint (FPB) Unit The main functions of the FPB are: ■ Implement hardware breakpoint (generates a breakpoint event to the processor to invoke debug modes such as halt or debug monitor). ■ Patch instruction or data from code memory space to SRAM. The FPB unit contains: ■ Control/Status Word Register (CSW) – The CSW Register configures and controls accesses through the MEM-AP to or from a connected memory system. ■ Two comparators for matching against literal loads from code space, and remapping to a corresponding area in system space. ■ Transfer Address Register (TAR) – The TAR holds the memory address to be accessed. ■ ■ Data Read/Write Register (DRW) – The DRW holds a 32-bit data value. In write mode, the DRW holds the value to write for the current transfer to the address specified in TAR[31:0]. In read mode, the DRW holds the value read in the current transfer from the address specified in TAR[31:0]. Six instruction comparators for matching against instruction fetches from code space, and remapping to a corresponding area in system memory space. Alternatively, it is possible to individually configure the comparators to return a Breakpoint Instruction (BKPT) to the processor core upon a match, providing hardware breakpoint capability. ■ Configuration Register (CFG) – The CFG Register provides information about the configuration of the MEM-AP implementation. It indicates whether memory accesses by the MEM-AP are big-endian or little-endian. ■ Debug Base Address Register (BASE) – The BASE Register provides an index into the connected memorymapped resource. This index value points to one of the following, the start of a set of debug registers or a ROM table that describes the connected debug components. The FPB has a flash patch control register that contains an enable bit to enable the FPB. In addition, each comparator comes with a separate enable bit in its comparator control register. Both the enable bits must be set to ‘1’ for a comparator to operate. If the comparison for an entry matches, the address is remapped to the address set in the remap register plus an offset corresponding to the comparator that matched, or is remapped to a BKPT instruction, if that feature is enabled. For more details on the Memory Access Port and registers, see the ARM Debug Interface Architecture Specification, available at http://www.arm.com. 418 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex-M3 Debug and Trace 40.4.2 Data Watchpoint and Trace (DWT) The DWT has a number of debugging functionalities. It has four comparators, each of which can be configured as follows: ■ Hardware watchpoint (generates a watchpoint event to processor to invoke debug modes such as halt or debug monitor) ■ ETM trigger (causes the ETM to emit a trigger packet in the instruction trace stream) ■ PC sampler event trigger ■ Data address sampler trigger ■ The first comparator can also be used to compare against the clock cycle counter instead of comparing to a data address The DWT also has counters for counting: ■ Clock cycles (CYCCNT) ■ Folded Instructions: A folded instruction is one that does not incur even one cycle to execute ■ Load Store Unit (LSU) Operations: LSU counts include all LSU costs after the initial cycle for the instruction ■ Sleep cycles ■ Cycles per instruction (CPI) ■ Interrupt overhead ■ PC sampling at regular intervals to count the number of core cycles ■ Applications and debuggers can use the counter to measure elapsed execution time ■ Interrupt events trace When used as a hardware watchpoint or ETM trigger, the comparator can be programmed to compare either data addresses or program counters. Otherwise, it compares the data addresses. 40.4.3 Instrumentation Trace Macrocell (ITM) The ITM is a an application driven trace source that supports printf style debugging to trace Operating System (OS) and application events, then emit diagnostic system information. The ITM emits trace information as packets. There are three sources that can generate packets. If multiple sources generate packets at the same time, the ITM arbitrates the order in which packets are output. The three sources in decreasing order of priority are: ■ Software Trace. Software can write directly to ITM stimulus registers. This emits packets. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F ■ Hardware Trace. The DWT generates these packets, and the ITM emits them. ■ Time Stamping. The ITM can generate timestamp packets that are inserted into a trace stream to help the debugger find out the timing of events. The ITM contains a 21-bit counter to generate the timestamp. The CortexM3 clock or the bit clock rate of the SWV output clocks the counter. One of the main uses of the ITM is to support printf style debugging. The ITM contains 32 stimulus ports, allowing different software processes to output to different ports, and messages that can be separated later at the debug host. Each port can be enabled or disabled by the Trace Enable Register (SWV_ITM_TER) and can be programmed (in groups of eight ports) to allow or disallow user processes to write to it. The output messages can be collected at the trace port interface or the Serial Wire Viewer (SWV) on the TPIU. The ITM is used in output of hardware trace packets. The packets are generated from the DWT and the ITM acts as a trace packet merging unit. To use DWT trace, you need to enable the DWTEn bit in the ITM Control Register (SWV_ITM_CR). ITM has a timestamp feature that allows trace capture tools to find out timing information by inserting delta timestamp packets into the traces when a new trace packet enters the FIFO inside the ITM. The timestamp packet is also generated when the timestamp counter overflows. The timestamp packets provide the time difference (delta) with previous events. Using the delta timestamp packets, the trace capture tools can then establish the timing of when each packet is generated and hence reconstruct the timing of various debug events. 40.4.4 Embedded Trace Macrocell (ETM) The ETM is an optional debug component that enables reconstruction of program execution. The ETM is designed as a high-speed, low-power debug tool that only supports instruction trace. This ensures that area is minimized, and that gate count is reduced. The Cortex-M3 system can perform low-bandwidth data tracing using the Data Watchpoint and Trace (DWT) and Instrumentation Trace Macrocell (ITM) components. To enable support of instruction trace with a low pin count, data trace is not included in the ETM. This considerably reduces gate count for the ETM because the triggering resources are simplified. Because the ETM does not generate data trace information, the lower bandwidth reduces the requirement for complex 419 Cortex-M3 Debug and Trace triggering capabilities. This means that the ETM does not include the following: internal comparators, counters, and sequencers. 40.5 For more details on system debug components and registers, see the Definitive Guide To ARM Cortex-M3 and ARM Cortex-M3 Technical Reference Manual, available at http://www.arm.com. Tracing Interface The Trace Port Interface Unit (TPIU) consists of SWV and TRACEPORT, which provides trace output from the DWT, ETM, and ITM. TRACEPORT is faster but uses more pins. SWV is slower but uses only one pin. The SWV and TRACEPORT interfaces provide trace data to a debug host via the Cypress MiniProg3 or an external trace port analyzer. The 5-pin TRACEPORT is used for rapid transmission of large trace streams. The single pin SWV mode is used to minimize the number of trace pins. SWV is shared with a JTAG pin. Figure 40-3. TPIU Block Diagram ETM ATB Slave Port ATB Interface TRACECLKIN Asynchronous FIFO TRACECLK Formatter ITM ATB Slave Port ATB Interface APB Slave Port ATB Interface Asynchronous FIFO The following functions are included in the TPIU: ■ ■ Asynchronous FIFO – The asynchronous FIFO enables trace data to be driven out at a speed that is not dependent on the speed of the core clock. Formatter – The formatter inserts source ID signals into the data packet stream so that trace data can be reassociated with its trace source. ■ Trace Out – The trace out block serializes formatted data before it goes off-chip. ■ ATB Interface – TPIU accepts trace data from the trace sources ETM or ITM. ■ APB Interface – The APB interface is the programming interface for the TPIU. ■ TRACECLKIN – Decoupled clock from ATB to enable easy control of the trace port speed. Typically, this is derived from a controllable clock source on-chip. Data changes on the rising edge only. 420 Trace Out (serializer) TRACEDATA [3:0] TRACESWO ■ TRACEPORT – It includes TRACEDATA[3:0] and TRACECLK. ■ TRACESWO – Trace output pin for SWV. 40.5.1 Single Wire Viewer Single Wire viewer (SWV) allows target resident code to communicate diagnostic information to the outside world through a single pin. The Serial Wire Viewer block is a combination of the Instrumentation Trace Macrocell (ITM) and the Serial Wire Output (SWO). ITM is a software application trace source. The SWV's trace output (TRACESWO) is channeled through the Test Controller, so that the Test Controller can output the trace data over the TDO pin when SWD is enabled. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Cortex-M3 Debug and Trace SWV can only be used when using the Serial Wire Debug (SWD) because its trace data is output over the same pin as JTAG's TDO. 40.5.1.1 Enabling SWV The Trace Enable Register (SWV_ITM_TER) is used to enable the stimulus ports so that trace data can be written into the stimulus port registers. Each bit in the Trace Enable Register is set to enable the corresponding stimulus port register. Also, the ITM should be enabled using the global enable bit, ITMEn, in the Control Register (SWV_ITM_CR). 40.5.1.2 Communicating with SWV Trace data is written into the stimulus port registers (SWV_ITM_SPR_DATA[0:31]). Each of the 32 stimulus ports has its own address. A write to one of these locations causes data to be written into the FIFO if the corresponding bit in the Trace Enable Register is set. Reading from any of the stimulus ports returns the FIFO status. A '0' is returned if the FIFO is full and a '0' is returned otherwise, only if the bit in the Trace Enable Register is set. 40.5.2 TRACEPORT TRACEPORT is used for rapid transmission of large trace streams. There are five TRACEPORT pins: four data pins, TRACEDATA[3:0] and one clock pin, TRACECLK. TRACEPORT supports synchronous mode of operation while TRACESWO does not. 40.5.2.1 40.5.2.2 Communicating with TRACEPORT As shown in Figure 40-3 on page 420, the trace data is passed onto the debug host via the TRACEDATA pins if the TRACEPORT mode is enabled. The output data on TRACEDATA pins changes on both edges of TRACECLK. 40.5.3 Using Multiple Interfaces Simultaneously If debugging and tracing are done at the same time, then SWD may be used with either SWV or TRACEPORT, or JTAG may be used with TRACEPORT, as shown in Table 40-1. Table 40-1. Debug Configuration Debug and Trace Configuration All Debug and Trace Disabled GPIO Pins Used 0 JTAG 4 or 5 SWD 2 SWV 1 TRACEPORT 5 JTAG plus TRACEPORT 9 or 10 SWD plus SWV 3 SWV plus TRACEPORT 7 Enabling TRACEPORT TRACEPORT mode can be enabled using the Select Pin Protocol Register (SWV_SWO_SPP). The format of the register is as shown in Figure 40-4. As shown in Figure 40-4, TRACEPORT can be enabled by writing 2'h10 to the SPP[1:0] bits. Figure 40-4. SPP Register 31 2 1 NA 0 SPP SPP DESCRIPTION 2'h00 Reset Value 2'h01 SWVMODE 2'h10 TRACEMODE PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 421 Cortex-M3 Debug and Trace 422 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 41. Nonvolatile Memory Programming PSoC® 5LP devices have three types of nonvolatile memory: flash, electronically erasable programmable read only memory (EEPROM), and nonvolatile latch (NVL). These can all be programmed by either the CPU running a boot loader program or by an external system via the JTAG/SWD interface. See PSoC 5LP Programming Specifications for details about device programming and programming specifications. 41.1 Features The nonvolatile memory programming system has the following features: ■ Simple command/status register interface ■ Flash can be programmed at the 288-byte row level ■ Each row of flash has 256 bytes of data plus an additional 32 bytes for ECC/configuration ■ EEPROM can be programmed at the 16-byte row level ■ All configuration NVL bytes can be programmed simultaneously ■ A single write once NVL byte can be programmed 41.2 Block Diagram Figure 41-1 is a block diagram of the flash programming system. Figure 41-1. Flash Block Diagram Test Controller (TC) Debug on-Chip (DOC) CPU PHUB EEPROM PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Flash Programming Interface NVL 423 Nonvolatile Memory Programming 41.3 How It Works less of CPU frequency. Disabled bus wait states may result in data corruption. All programming operations are done through a simple command/status register interface summarized in Table 41-1. Table 41-1. Command and Status Register Summary Register Size (Bits) SPC_CPU_DATA 8 Description Data to/from the CPU SPC_DMA_DATA 8 Data to/from the DMAC SPC_SR 8 Status – ready, data available, status code Commands and data are sent as a series of bytes to either SPC_CPU_DATA or SPC_DMA_DATA, depending on the source of the command. Response data is read via the same register to which the command was sent. The status register, SPC_SR, indicates whether a new command can be accepted, when data is available for the most recent command, and a success/failure response for the most recent command. The PANTHER_WAITPIPE register BYPASS bit must be set to ‘0’ before accessing the SPC register interface, regard- 41.3.1 Commands Before sending a command to the SPC_CPU_DATA or SPC_DMA_DATA register, the SPC_Idle bit in SPC_SR[1] must be ‘1’. SPC_Idle will go to ‘0’ when the first byte of a command (0xB6) is written to a data register, and go back to ‘1’ when command execution is complete or an error is detected. Commands sent to either data register while SPC_Idle is ‘0’ are ignored. All commands must adhere to the following format: ■ Key byte #1 – always 0xB6 ■ Key byte #2 – 0xD3 plus the command code (ignore overflow) ■ Command code byte ■ Command parameter bytes ■ Command data bytes The command codes are shown in Table 41-2. See 41.3.1.1 Command Code Descriptions on page 425 for details. Table 41-2. Command Codes Command Code Command Name Memory Type NVL Access Any Description 0x00 Load byte 0x01 Load multi bytes Flash, EEPROM Any Loads 1 to 32 bytes of data into the row latch 0x02 Load row Flash, EEPROM Any Loads a row of data 0x03 Read byte NVL Any Read a byte from NV memory 0x04 Read multi bytes Flash TC only 0x05 Write row Flash, EEPROM 0x06 Write NVL NVL 0x07 Program row Flash, EEPROM Any Programs a row with data in row latch 0x08 Erase sector Flash, EEPROM Any Erases a 64-row sector Any TC only Loads a single byte of data into the volatile latch Reads 1 – 256 data bytes, does not cross row boundaries Erases then programs a row with data in row latch Programs all of user NVL with data in the volatile latch 0x09 Erase all Flash TC only Erases all flash, including ECC and row protection bytes 0x0B Protect Flash TC only Program flash protection bits with data in row latch 0x0C Get Checksum Flash Any Some commands are available only when the device is being controlled by an external system via the JTAG/SWD interface and the test controller (see the Test Controller chapter on page 403). Some commands require an array ID as a parameter. Array ID codes are shown in Table 41-3. Table 41-3. Array ID Codes Array ID Code Memory Type 0x00 – 0x3E Single flash array 0x3F All flash arrays (used by the Erase All command) 0x40 Single EEPROM array 0x80 User NVL array 0xF8 Write Once NVL array 424 Computes 4 byte checksum for given memory locations A flash array has, at most, 64 KB plus ECC bytes. PSoC 5LP architecture has one or more arrays, where each array is 64K plus ECC bytes. For example, if a PSoC 5LP device has 256 KB Flash, there are four arrays, and the only valid array IDs are 0x00 – 0x03. An EEPROM array has, at most, 2 KB. PSoC 5LP devices have one EEPROM array, the size of which is 512 bytes, 1 KB, or 2 KB. PSoC 5LP devices have one user NVL array and one write once NVL array. For commands operating on flash or EEPROM, all array IDs within the number of flash and EEPROM arrays are valid. If a non-existent array is selected, the array ID wraps. For PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Nonvolatile Memory Programming example, if a device has two flash arrays (IDs = 0 and 1) and a command is sent with array ID = 3 then the upper bits of the ID are truncated and so array ID 1 is selected. are transferred from the nonvolatile cells to the volatile latch portion of the NVL. ■ Command Parameter Bytes – Array ID, Start address high, Start address mid, Start address low, Number of bytes (N) Some commands require an address as a parameter. As with array IDs, any address is valid for a flash or EEPROM array. Upper address bits are truncated to allow only addressing of valid locations. For example, if a device has 512 bytes EEPROM and address 0x202 (514) is passed as a parameter, the operation takes place on address 0x002. This command returns N + 1 data bytes from flash, starting at the given address. In flash arrays, two address spaces exist – data and ECC/configuration. Bit 7 of the address high parameter selects which of the two address spaces is addressed. If the bit is 0, then the data space is selected; otherwise, the ECC/configuration space is selected. For example, if the address is 0x80000B and N is 0x08, the command reads 9 ECC/configuration bytes starting at address 0x00000B. Array IDs and addresses do not wrap for NVL accesses. Some commands use the row latch size for flash and EEPROM. Row latch sizes are shown in the following table. Table 41-4. Row Latch Sizes Array Type Size (Bytes) Flash, with ECC Enabled 256 Flash, with ECC Disabled 41.3.1.1 The address plus N must not cross a row boundary – 256 for the flash data space, 32 for the flash ECC/configuration space. 288 (256 data bytes plus 32 configuration bytes) EEPROM Address wrapping applies; if the address is greater than the flash size, the upper bits are then ignored. For example, 16 bits of address are needed to access the data space in a 64 KB flash array, so the seven LS bits of the Address high parameter are ignored. Address 0x045A8B actually addresses 0x005A8B. 16 Command Code Descriptions The following are descriptions of the command codes listed in Table 41-2 on page 424. ■ Similarly, 13 address bits are needed to access the 8 KB ECC/configuration space associated with a 64 KB flash array. For example, for a 64 KB flash array (which also has 8 KB ECC/configuration bytes), valid address ranges are: Command 0x00 – Load Byte Command Parameter Bytes – Array ID, Address, Data This command loads the given data byte into the volatile latch for the selected NVL array (in accordance with the array ID) at the given address. Only addresses within the selected NVL array are valid. ■ Command 0x01 – Load Multiple Bytes Command Parameter Bytes – Array ID, Start address high, Start address low, Number of bytes (N), Data0, …, DataN ■ This command loads the given data bytes into a row latch for flash or EEPROM. The number of data bytes expected equals the row latch size. See Table 41-4. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Command 0x05 – Write Row The die temperature parameters can be acquired by sending the Get Temperature command (see the Temperature Sensor chapter on page 365). Command 0x03 – Read Byte This command returns a data byte from the selected NVL array (per the array ID), at the given address. Only addresses within the selected NVL array are valid. Note that when this command is executed all of the data bytes ECC/configuration space – 0x800000 – 0x801FFF (8 KB) For flash, data bytes and ECC/configuration bytes are both programmed. If ECC is enabled then the ECC syndrome bytes are automatically generated and loaded into the ECC/configuration bytes of the row latch before programming takes place. Command 0x02 – Load Row Command Parameter Bytes – Array ID, Address Data space – 0x000000 – 0x00FFFF (64 KB) ❐ This command erases the addressed flash/EEPROM row and then programs it with the data in the row latch. If the row ID is greater than the array size (in rows), then the row ID wraps (the upper bits are ignored). Command Parameter Bytes – Array ID, Data0, …, Data(row latch size -1) ■ ❐ Command Parameter Bytes – Array ID, Row ID high, Row ID low, Temperature sign, Temperature magnitude This command loads N + 1 given data bytes into a row latch for flash or EEPROM. N may range from 0 to 31 for flash or 0 to 15 for EEPROM. The given start address + N must be less than the array row latch size. See Table 41-4. ■ Command 0x04 – Read Multiple Bytes ■ Command 0x06 – Write User NVL Command Parameter Bytes – Array ID This command writes all of the bytes in the volatile latch for the selected NVL array (per the array ID) to that NVL 425 Nonvolatile Memory Programming array. All flash protection bits must be cleared (no flash protection) or the command fails. ■ Command 0x07 – Program Row Command Parameter Bytes – Array ID, Row ID high, Row ID low This command programs the addressed flash/EEPROM row with the data in the row latch. If the row ID is greater than the array size (in rows), the row ID wraps (the upper bits are ignored). The row must have been previously erased (commands 0x08 and 0x09). For flash, data bytes and ECC/configuration bytes are both programmed. If ECC is enabled, the ECC syndrome bytes are automatically generated and loaded into the ECC/configuration bytes of the row latch before programming takes place. For devices with multiple flash arrays, the All Flash array ID (0x3F) can be used with this command. This option causes each flash array to have its addressed row programmed with its row latch contents simultaneously with the other arrays, reducing the overall flash programming time. ■ ■ ■ Command 0x0C – Get Checksum Command Parameter Bytes – Array ID, Start row high, Start row low, Number of rows high, Number of rows low This command computes a 4-byte checksum for the given number of flash rows + 1, starting at the given row. The checksum is computed by a running simple addition of all values in the rows. If ECC is disabled, the computation includes all data from the user space and the ECC / configuration space. If ECC is enabled, the computation includes only data from the user space. If the array ID is All Flash, the checksum computed includes all flash data on all flash arrays on the device. The rest of the command parameters are ignored. The checksum value is returned MS byte first. 41.3.1.2 Command Failure Codes Command 0x08 – Erase Sector In response to commands, a success/failure code is returned in the SPC_SR register: These codes are described in Table 41-5. Command Parameter Bytes – Array ID, Sector ID Table 41-5. Command Failure Codes This command erases a sector of flash/EEPROM. A sector is a block of 64 contiguous rows that starts at a 64-row boundary. For flash arrays, all associated ECC/ configuration bytes are also erased. The sector ID wraps if it exceeds the number of sectors. ■ each flash array to have its protection row programmed with its row latch contents simultaneously with the other arrays, reducing the overall flash programming time. Success/Failure Code (Bits[7:2] in SPC_SR register) 0x00 Meaning Command successfully executed 0x01 Invalid array ID Command 0x09 – Erase All 0x02 Invalid key Command Parameter Bytes – None 0x03 Array is asleep This command erases all flash data and ECC/configuration bytes, all flash protection rows, and all row latches in all flash arrays on the device. 0x04 External access failure: command must be sent via test controller Command 0x0B – Protect 0x05 Invalid ‘N’ value 0x07, 0x08 Program/Erase failure Protection check failure: protection settings are in a state that prevents the command from executing Command Parameter Bytes – Array ID, Row Select 0x09 This command programs a flash protection row with data in the flash row latch (see 41.3.3 Flash Protection Settings on page 427). This command can be executed only if none of the protection bits are currently set – no flash protection. Any bytes of the protection row that are marked as unused space are programmed with 0x00. This occurs regardless of what values are loaded into the row latches prior to sending this command. 0x0A Invalid address 0x0B Invalid command code 0x0C Invalid row ID The Row Select parameter is used for flash arrays that have a row size less than 256 bytes. Because all flash arrays have 256-byte rows, this parameter should always be 0x00. 41.3.2 Register Summary All programming operations are done through a simple command/status register interface, shown in Table 41-1 on page 424. When the flash protection data is programmed, this command cannot be sent again until an Erase All command is sent first. For devices with multiple flash arrays, the All Flash array ID (0x3F) may be used with this command. This causes 426 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Nonvolatile Memory Programming 41.3.3 Flash Protection Settings Table 41-6. Protection Modes Each row of flash has its own protection settings. For each flash array, flash protection bits are stored in a “hidden” row in that array. A hidden row is one that is not readable by the CPU, and contains no CPU program or data bytes. In the hidden row, two bits per flash row are packed into a byte; therefore, each byte in the hidden row has protection settings for four flash rows. As shown in Figure 41-2, the flash rows are ordered so that the first two bits in the hidden row correspond to the protection settings of flash row 0. Protection is cumulative in that modes have successively higher protection levels and include the lower protection modes. The following table shows the protection modes. Mode Description Reada External Writeb Internal Writec Yes 00 Unprotected Yes Yes 01 Read Protect No Yes Yes 10 Disable External Write No No Yes 11 Disable Internal Write No No No a. Read – Applies to Test Controller and Read Commands, and cache data fetches. Cache code fetches are always allowed. b. External Write – Test Controller/third-party programmers. c. Internal Write – Boot loading or writes due to firmware execution. When a read/write/erase operation is to be done for a row, the corresponding protection bits are checked. The command is executed only if allowed under the current protection mode. If the command is not allowed, the command then fails. Figure 41-2. Flash Protection Bits Page 0 Bits [1:0] Page 1 Bits [1:0] Page 2 Bits [1:0] Page 3 Bits [1:0] ...... Page 255 Bits [1:0] Byte 0 in Flash Hidden Row 0: Contains protection data for Flash rows 0 through 3. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 427 Nonvolatile Memory Programming 428 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary The Glossary section explains the terminology used in this technical reference manual. Glossary terms are characterized in bold, italic font throughout the text of this manual. A accumulator In a CPU, a register in which intermediate results are stored. Without an accumulator, it is necessary to write the result of each calculation (addition, subtraction, shift, and so on.) to main memory and read them back. Access to main memory is slower than access to the accumulator, which usually has direct paths to and from the arithmetic and logic unit (ALU). active high 1. A logic signal having its asserted state as the logic 1 state. 2. A logic signal having the logic 1 state as the higher voltage of the two states. active low 1. A logic signal having its asserted state as the logic 0 state. 2. A logic signal having its logic 1 state as the lower voltage of the two states: inverted logic. address The label or number identifying the memory location (RAM, ROM, or register) where a unit of information is stored. algorithm A procedure for solving a mathematical problem in a finite number of steps that frequently involve repetition of an operation. ambient temperature The temperature of the air in a designated area, particularly the area surrounding the PSoC device. analog See analog signals. analog blocks The basic programmable opamp circuits. These are SC (switched capacitor) and CT (continuous time) blocks. These blocks can be interconnected to provide ADCs, DACs, multi-pole filters, gain stages, and much more. analog output An output that is capable of driving any voltage between the supply rails, instead of just a logic 1 or logic 0. analog signals A signal represented in a continuous form with respect to continuous times, as contrasted with a digital signal represented in a discrete (discontinuous) form in a sequence of time. analog-to-digital (ADC) A device that changes an analog signal to a digital signal of corresponding magnitude. Typically, an ADC converts a voltage to a digital number. The digital-to-analog (DAC) converter performs the reverse operation. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 429 Glossary AND See Boolean Algebra. API (Application Programming Interface) A series of software routines that comprise an interface between a computer application and lower-level services and functions (for example, user modules and libraries). APIs serve as building blocks for programmers that create software applications. array An array, also known as a vector or list, is one of the simplest data structures in computer programming. Arrays hold a fixed number of equally-sized data elements, generally of the same data type. Individual elements are accessed by index using a consecutive range of integers, as opposed to an associative array. Most high level programming languages have arrays as a builtin data type. Some arrays are multi-dimensional, meaning they are indexed by a fixed number of integers; for example, by a group of two integers. One- and two-dimensional arrays are the most common. Also, an array can be a group of capacitors or resistors connected in some common form. assembly A symbolic representation of the machine language of a specific processor. Assembly language is converted to machine code by an assembler. Usually, each line of assembly code produces one machine instruction, though the use of macros is common. Assembly languages are considered low level languages; where as C is considered a high level language. asynchronous A signal whose data is acknowledged or acted upon immediately, irrespective of any clock signal. attenuation The decrease in intensity of a signal as a result of absorption of energy and of scattering out of the path to the detector, but not including the reduction due to geometric spreading. Attenuation is usually expressed in dB. B bandgap reference A stable voltage reference design that matches the positive temperature coefficient of VT with the negative temperature coefficient of VBE, to produce a zero temperature coefficient (ideally) reference. bandwidth 1. The frequency range of a message or information processing system measured in hertz. 2. The width of the spectral region over which an amplifier (or absorber) has substantial gain (or loss); it is sometimes represented more specifically as, for example, full width at half maximum. bias 1. A systematic deviation of a value from a reference value. 2. The amount by which the average of a set of values departs from a reference value. 3. The electrical, mechanical, magnetic, or other force (field) applied to a device to establish a reference level to operate the device. bias current 430 The constant low level DC current that is used to produce a stable operation in amplifiers. This current can sometimes be changed to alter the bandwidth of an amplifier. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary binary The name for the base 2 numbering system. The most common numbering system is the base 10 numbering system. The base of a numbering system indicates the number of values that may exist for a particular positioning within a number for that system. For example, in base 2, binary, each position may have one of two values (0 or 1). In the base 10, decimal, numbering system, each position may have one of ten values (0, 1, 2, 3, 4, 5, 6, 7, 8, and 9). bit A single digit of a binary number. Therefore, a bit may only have a value of ‘0’ or ‘1’. A group of 8 bits is called a byte. Because the PSoC's M8CP is an 8-bit microcontroller, the PSoC devices's native data chunk size is a byte. bit rate (BR) The number of bits occurring per unit of time in a bit stream, usually expressed in bits per second (bps). block 1. A functional unit that performs a single function, such as an oscillator. 2. A functional unit that may be configured to perform one of several functions, such as a digital PSoC block or an analog PSoC block. Boolean Algebra In mathematics and computer science, Boolean algebras or Boolean lattices, are algebraic structures which "capture the essence" of the logical operations AND, OR and NOT as well as the set theoretic operations union, intersection, and complement. Boolean algebra also defines a set of theorems that describe how Boolean equations can be manipulated. For example, these theorems are used to simplify Boolean equations, which will reduce the number of logic elements needed to implement the equation. The operators of Boolean algebra may be represented in various ways. Often they are simply written as AND, OR, and NOT. In describing circuits, NAND (NOT AND), NOR (NOT OR), XNOR (exclusive NOT OR), and XOR (exclusive OR) may also be used. Mathematicians often use + (for example, A+B) for OR and for AND (for example, A*B) (in some ways those operations are analogous to addition and multiplication in other algebraic structures) and represent NOT by a line drawn above the expression being negated (for example, ~A, A_, !A). break-before-make The elements involved go through a disconnected state entering (‘break”) before the new connected state (“make”). broadcast net A signal that is routed throughout the microcontroller and is accessible by many blocks or systems. buffer 1. A storage area for data that is used to compensate for a speed difference, when transferring data from one device to another. Usually refers to an area reserved for I/O operations, into which data is read, or from which data is written. 2. A portion of memory set aside to store data, often before it is sent to an external device or as it is received from an external device. 3. An amplifier used to lower the output impedance of a system. bus 1. A named connection of nets. Bundling nets together in a bus makes it easier to route nets with similar routing patterns. 2. A set of signals performing a common function and carrying similar data. Typically represented using vector notation; for example, address[7:0]. 3. One or more conductors that serve as a common connection for a group of related devices. byte A digital storage unit consisting of 8 bits. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 431 Glossary C C A high level programming language. capacitance A measure of the ability of two adjacent conductors, separated by an insulator, to hold a charge when a voltage differential is applied between them. Capacitance is measured in units of Farads. capture To extract information automatically through the use of software or hardware, as opposed to hand-entering of data into a computer file. chaining Connecting two or more 8-bit digital blocks to form 16-, 24-, and even 32-bit functions. Chaining allows certain signals such as Compare, Carry, Enable, Capture, and Gate to be produced from one block to another. checksum The checksum of a set of data is generated by adding the value of each data word to a sum. The actual checksum can simply be the result sum or a value that must be added to the sum to generate a pre-determined value. clear To force a bit/register to a value of logic ‘0’. clock The device that generates a periodic signal with a fixed frequency and duty cycle. A clock is sometimes used to synchronize different logic blocks. clock generator A circuit that is used to generate a clock signal. CMOS The logic gates constructed using MOS transistors connected in a complementary manner. CMOS is an acronym for complementary metal-oxide semiconductor. comparator An electronic circuit that produces an output voltage or current whenever two input levels simultaneously satisfy predetermined amplitude requirements. compiler A program that translates a high level language, such as C, into machine language. configuration In a computer system, an arrangement of functional units according to their nature, number, and chief characteristics. Configuration pertains to hardware, software, firmware, and documentation. The configuration will affect system performance. configuration space In PSoC devices, the register space accessed when the XIO bit, in the CPU_F register, is set to ‘1’. crowbar A type of over-voltage protection that rapidly places a low resistance shunt (typically an SCR) from the signal to one of the power supply rails, when the output voltage exceeds a predetermined value. crystal oscillator An oscillator in which the frequency is controlled by a piezoelectric crystal. Typically a piezoelectric crystal is less sensitive to ambient temperature than other circuit components. cyclic redundancy check (CRC) A calculation used to detect errors in data communications, typically performed using a linear feedback shift register. Similar calculations may be used for a variety of other purposes such as data compression. 432 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary D data bus A bi-directional set of signals used by a computer to convey information from a memory location to the central processing unit and vice versa. More generally, a set of signals used to convey data between digital functions. data stream A sequence of digitally encoded signals used to represent information in transmission. data transmission The sending of data from one place to another by means of signals over a channel. debugger A hardware and software system that allows the user to analyze the operation of the system under development. A debugger usually allows the developer to step through the firmware one step at a time, set break points, and analyze memory. dead band A period of time when neither of two or more signals are in their active state or in transition. decimal A base-10 numbering system, which uses the symbols 0, 1, 2, 3, 4, 5, 6, 7, 8 and 9 (called digits) together with the decimal point and the sign symbols + (plus) and - (minus) to represent numbers. default value Pertaining to the pre-defined initial, original, or specific setting, condition, value, or action a system will assume, use, or take in the absence of instructions from the user. device The device referred to in this manual is the PSoC device, unless otherwise specified. die An non-packaged integrated circuit (IC), normally cut from a wafer. digital A signal or function, the amplitude of which is characterized by one of two discrete values: ‘0’ or ‘1’. digital blocks The 8-bit logic blocks that can act as a counter, timer, serial receiver, serial transmitter, CRC generator, pseudo-random number generator, or SPI. digital logic A methodology for dealing with expressions containing two-state variables that describe the behavior of a circuit or system. digital-to-analog (DAC) A device that changes a digital signal to an analog signal of corresponding magnitude. The analog-to-digital (ADC) converter performs the reverse operation. direct access The capability to obtain data from a storage device, or to enter data into a storage device, in a sequence independent of their relative positions by means of addresses that indicate the physical location of the data. duty cycle The relationship of a clock period high time to its low time, expressed as a percent. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 433 Glossary E emulator Duplicates (provides an emulation of) the functions of one system with a different system, so that the second system appears to behave like the first system. External Reset (XRES) An active high signal that is driven into the PSoC device. It causes all operation of the CPU and blocks to stop and return to a pre-defined state. F falling edge A transition from a logic 1 to a logic 0. Also known as a negative edge. feedback The return of a portion of the output, or processed portion of the output, of a (usually active) device to the input. filter A device or process by which certain frequency components of a signal are attenuated. firmware The software that is embedded in a hardware device and executed by the CPU. The software may be executed by the end user, but it may not be modified. flag Any of various types of indicators used for identification of a condition or event (for example, a character that signals the termination of a transmission). Flash An electrically programmable and erasable, volatile technology that provides users with the programmability and data storage of EPROMs, plus in-system erasability. Nonvolatile means that the data is retained when power is off. Flash bank A group of flash ROM blocks where flash block numbers always begin with ‘0’ in an individual flash bank. A flash bank also has its own block level protection information. Flash block The smallest amount of flash ROM space that may be programmed at one time and the smallest amount of flash space that may be protected. A flash block holds 64 bytes. flip-flop A device having two stable states and two input terminals (or types of input signals) each of which corresponds with one of the two states. The circuit remains in either state until it is made to change to the other state by application of the corresponding signal. frequency The number of cycles or events per unit of time, for a periodic function. G gain 434 The ratio of output current, voltage, or power to input current, voltage, or power, respectively. Gain is usually expressed in dB. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary gate 1. A device having one output channel and one or more input channels, such that the output channel state is completely determined by the input channel states, except during switching transients. 2. One of many types of combinational logic elements having at least two inputs (for example, AND, OR, NAND, and NOR (also see Boolean Algebra)). ground 1. The electrical neutral line having the same potential as the surrounding earth. 2. The negative side of DC power supply. 3. The reference point for an electrical system. 4. The conducting paths between an electric circuit or equipment and the earth, or some conducting body serving in place of the earth. H hardware A comprehensive term for all of the physical parts of a computer or embedded system, as distinguished from the data it contains or operates on, and the software that provides instructions for the hardware to accomplish tasks. hardware reset A reset that is caused by a circuit, such as a POR, watchdog reset, or external reset. A hardware reset restores the state of the device as it was when it was first powered up. Therefore, all registers are set to the POR value as indicated in register tables throughout this document. hexadecimal A base 16 numeral system (often abbreviated and called hex), usually written using the symbols 0-9 and A-F. It is a useful system in computers because there is an easy mapping from four bits to a single hex digit. Thus, one can represent every byte as two consecutive hexadecimal digits. Compare the binary, hex, and decimal representations: bin = hex = dec 0000b = 0x0 = 0 0001b = 0x1 = 1 0010b = 0x2 = 2 ... 1001b = 0x9 = 9 1010b = 0xA = 10 1011b = 0xB = 11 ... 1111b = 0xF = 15 So the decimal numeral 79 whose binary representation is 0100 1111b can be written as 4Fh in hexadecimal (0x4F). high time The amount of time the signal has a value of ‘1’ in one period, for a periodic digital signal. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 435 Glossary I I2C A two-wire serial computer bus by Phillips Semiconductors (now NXP Semiconductors). I2C is an Inter-Integrated Circuit. It is used to connect low-speed peripherals in an embedded system. The original system was created in the early 1980s as a battery control interface, but it was later used as a simple internal bus system for building control electronics. I2C uses only two bidirectional pins, clock and data, both running at +5 V and pulled high with resistors. The bus operates at 100 Kbps in standard mode and 400 Kbps in fast mode. ICE The in-circuit emulator that allows users to test the project in a hardware environment, while viewing the debugging device activity in a software environment (PSoC Designer™). idle state A condition that exists whenever user messages are not being transmitted, but the service is immediately available for use. impedance 1. The resistance to the flow of current caused by resistive, capacitive, or inductive devices in a circuit. 2. The total passive opposition offered to the flow of electric current. Note the impedance is determined by the particular combination of resistance, inductive reactance, and capacitive reactance in a given circuit. input A point that accepts data, in a device, process, or channel. input/output (I/O) A device that introduces data into or extracts data from a system. instruction An expression that specifies one operation and identifies its operands, if any, in a programming language such as C or assembly. instruction mnemonics A set of acronyms that represent the opcodes for each of the assembly-language instructions, for example, ADD, SUBB, MOV. integrated circuit (IC) A device in which components such as resistors, capacitors, diodes, and transistors are formed on the surface of a single piece of semiconductor. interface The means by which two systems or devices are connected and interact with each other. interrupt A suspension of a process, such as the execution of a computer program, caused by an event external to that process, and performed in such a way that the process can be resumed. interrupt service routine (ISR) A block of code that normal code execution is diverted to when the M8CP receives a hardware interrupt. Many interrupt sources may each exist with its own priority and individual ISR code block. Each ISR code block ends with the RETI instruction, returning the device to the point in the program where it left normal program execution. 436 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary J jitter 1. A misplacement of the timing of a transition from its ideal position. A typical form of corruption that occurs on serial data streams. 2. The abrupt and unwanted variations of one or more signal characteristics, such as the interval between successive pulses, the amplitude of successive cycles, or the frequency or phase of successive cycles. K keeper A circuit that holds a signal to the last driven value, even when the signal becomes un-driven. L latency The time or delay that it takes for a signal to pass through a given circuit or network. least significant bit (LSb) The binary digit, or bit, in a binary number that represents the least significant value (typically the right-hand bit). The bit versus byte distinction is made by using a lower case “b” for bit in LSb. least significant byte (LSB) The byte in a multi-byte word that represents the least significant values (typically the right-hand byte). The byte versus bit distinction is made by using an upper case “B” for byte in LSB. Linear Feedback Shift Register (LFSR) A shift register whose data input is generated as an XOR of two or more elements in the register chain. load The electrical demand of a process expressed as power (watts), current (amps), or resistance (ohms). logic function A mathematical function that performs a digital operation on digital data and returns a digital value. lookup table (LUT) A logic block that implements several logic functions. The logic function is selected by means of select lines and is applied to the inputs of the block. For example: A 2 input LUT with 4 select lines can be used to perform any one of 16 logic functions on the two inputs resulting in a single logic output. The LUT is a combinational device; therefore, the input/output relationship is continuous, that is, not sampled. low time The amount of time the signal has a value of ‘0’ in one period, for a periodic digital signal. low voltage detect (LVD) A circuit that senses Vddd and provides an interrupt to the system when Vddd falls below a selected threshold. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 437 Glossary M M8CP An 8-bit Harvard Architecture microprocessor. The microprocessor coordinates all activity inside a PSoC device by interfacing to the flash, SRAM, and register space. macro A programming language macro is an abstraction, whereby a certain textual pattern is replaced according to a defined set of rules. The interpreter or compiler automatically replaces the macro instance with the macro contents when an instance of the macro is encountered. Therefore, if a macro is used five times and the macro definition required 10 bytes of code space, 50 bytes of code space will be needed in total. mask 1. To obscure, hide, or otherwise prevent information from being derived from a signal. It is usually the result of interaction with another signal, such as noise, static, jamming, or other forms of interference. 2. A pattern of bits that can be used to retain or suppress segments of another pattern of bits, in computing and data processing systems. master device A device that controls the timing for data exchanges between two devices. Or when devices are cascaded in width, the master device is the one that controls the timing for data exchanges between the cascaded devices and an external interface. The controlled device is called the slave device. microcontroller An integrated circuit device that is designed primarily for control systems and products. In addition to a CPU, a microcontroller typically includes memory, timing circuits, and I/O circuitry. The reason for this is to permit the realization of a controller with a minimal quantity of devices, thus achieving maximal possible miniaturization. This in turn, will reduce the volume and the cost of the controller. The microcontroller is normally not used for general-purpose computation as is a microprocessor. mnemonic A tool intended to assist the memory. Mnemonics rely on not only repetition to remember facts, but also on creating associations between easy-to-remember constructs and lists of data. A two to four character string representing a microprocessor instruction. mode A distinct method of operation for software or hardware. For example, the Digital PSoC block may be in either counter mode or timer mode. modulation A range of techniques for encoding information on a carrier signal, typically a sine-wave signal. A device that performs modulation is known as a modulator. Modulator A device that imposes a signal on a carrier. MOS An acronym for metal-oxide semiconductor. most significant bit (MSb) The binary digit, or bit, in a binary number that represents the most significant value (typically the left-hand bit). The bit versus byte distinction is made by using a lower case “b” for bit in MSb. most significant byte (MSB) The byte in a multi-byte word that represents the most significant values (typically the left-hand byte). The byte versus bit distinction is made by using an upper case “B” for byte in MSB. 438 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary multiplexer (mux) 1. A logic function that uses a binary value, or address, to select between a number of inputs and conveys the data from the selected input to the output. 2. A technique which allows different input (or output) signals to use the same lines at different times, controlled by an external signal. Multiplexing is used to save on wiring and I/O ports. N NAND See Boolean Algebra. negative edge A transition from a logic 1 to a logic 0. Also known as a falling edge. net The routing between devices. nibble A group of four bits, which is one-half of a byte. noise 1. A disturbance that affects a signal and that may distort the information carried by the signal. 2. The random variations of one or more characteristics of any entity such as voltage, current, or data. NOR See Boolean Algebra. NOT See Boolean Algebra. O OR See Boolean Algebra. oscillator A circuit that may be crystal controlled and is used to generate a clock frequency. output The electrical signal or signals which are produced by an analog or digital block. P parallel The means of communication in which digital data is sent multiple bits at a time, with each simultaneous bit being sent over a separate line. parameter Characteristics for a given block that have either been characterized or may be defined by the designer. parameter block A location in memory where parameters for the SSC instruction are placed prior to execution. parity A technique for testing transmitting data. Typically, a binary digit is added to the data to make the sum of all the digits of the binary data either always even (even parity) or always odd (odd parity). path 1. The logical sequence of instructions executed by a computer. 2. The flow of an electrical signal through a circuit. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 439 Glossary pending interrupts An interrupt that is triggered but not serviced, either because the processor is busy servicing another interrupt or global interrupts are disabled. phase The relationship between two signals, usually the same frequency, that determines the delay between them. This delay between signals is either measured by time or angle (degrees). Phase-Locked Loop (PLL) An electronic circuit that controls an oscillator so that it maintains a constant phase angle relative to a reference signal. pin A terminal on a hardware component. Also called lead. pinouts The pin number assignment: the relation between the logical inputs and outputs of the PSoC device and their physical counterparts in the printed circuit board (PCB) package. Pinouts will involve pin numbers as a link between schematic and PCB design (both being computer generated files) and may also involve pin names. port A group of pins, usually eight. positive edge A transition from a logic 0 to a logic 1. Also known as a rising edge. posted interrupts An interrupt that is detected by the hardware but may or may not be enabled by its mask bit. Posted interrupts that are not masked become pending interrupts. Power On Reset (POR) A circuit that forces the PSoC device to reset when the voltage is below a pre-set level. This is one type of hardware reset. program counter The instruction pointer (also called the program counter) is a register in a computer processor that indicates where in memory the CPU is executing instructions. Depending on the details of the particular machine, it holds either the address of the instruction being executed, or the address of the next instruction to be executed. protocol A set of rules. Particularly the rules that govern networked communications. PSoC® Cypress’s Programmable System-on-Chip (PSoC®) devices. PSoC blocks See analog blocks and digital blocks. PSoC Creator™ The software for Cypress’s next generation Programmable System-on-Chip technology. PSoC Designer™ The software for Cypress’s Programmable System-on-Chip technology. pulse A rapid change in some characteristic of a signal (for example, phase or frequency), from a baseline value to a higher or lower value, followed by a rapid return to the baseline value. pulse-width modulator (PWM) An output in the form of duty cycle which varies as a function of the applied measure. 440 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary R RAM An acronym for random access memory. A data-storage device from which data can be read out and new data can be written in. register A storage device with a specific capacity, such as a bit or byte. reset A means of bringing a system back to a know state. See hardware reset and software reset. resistance The resistance to the flow of electric current measured in ohms for a conductor. revision ID A unique identifier of the PSoC device. ripple divider An asynchronous ripple counter constructed of flip-flops. The clock is fed to the first stage of the counter. An n-bit binary counter consisting of n flip-flops that can count in binary from 0 to 2n - 1. rising edge See positive edge. ROM An acronym for read only memory. A data-storage device from which data can be read out, but new data cannot be written in. routine A block of code, called by another block of code, that may have some general or frequent use. routing Physically connecting objects in a design according to design rules set in the reference library. runt pulses In digital circuits, narrow pulses that, due to non-zero rise and fall times of the signal, do not reach a valid high or low level. For example, a runt pulse may occur when switching between asynchronous clocks or as the result of a race condition in which a signal takes two separate paths through a circuit. These race conditions may have different delays and are then recombined to form a glitch or when the output of a flip-flop becomes metastable. S sampling The process of converting an analog signal into a series of digital values or reversed. schematic A diagram, drawing, or sketch that details the elements of a system, such as the elements of an electrical circuit or the elements of a logic diagram for a computer. seed value An initial value loaded into a linear feedback shift register or random number generator. serial 1. Pertaining to a process in which all events occur one after the other. 2. Pertaining to the sequential or consecutive occurrence of two or more related activities in a single device or channel. set To force a bit/register to a value of logic 1. settling time The time it takes for an output signal or value to stabilize after the input has changed from one value to another. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 441 Glossary shift The movement of each bit in a word one position to either the left or right. For example, if the hex value 0x24 is shifted one place to the left, it becomes 0x48. If the hex value 0x24 is shifted one place to the right, it becomes 0x12. shift register A memory storage device that sequentially shifts a word either left or right to output a stream of serial data. sign bit The most significant binary digit, or bit, of a signed binary number. If set to a logic 1, this bit represents a negative quantity. signal A detectable transmitted energy that can be used to carry information. As applied to electronics, any transmitted electrical impulse. silicon ID A unique identifier of the PSoC silicon. skew The difference in arrival time of bits transmitted at the same time, in parallel transmission. slave device A device that allows another device to control the timing for data exchanges between two devices. Or when devices are cascaded in width, the slave device is the one that allows another device to control the timing of data exchanges between the cascaded devices and an external interface. The controlling device is called the master device. software A set of computer programs, procedures, and associated documentation concerned with the operation of a data processing system (for example, compilers, library routines, manuals, and circuit diagrams). Software is often written first as source code, and then converted to a binary format that is specific to the device on which the code will be executed. software reset A partial reset executed by software to bring part of the system back to a known state. A software reset will restore the M8CP to a know state but not PSoC blocks, systems, peripherals, or registers. For a software reset, the CPU registers (CPU_A, CPU_F, CPU_PC, CPU_SP, and CPU_X) are set to 0x00. Therefore, code execution will begin at flash address 0x0000. SRAM An acronym for static random access memory. A memory device allowing users to store and retrieve data at a high rate of speed. The term static is used because, when a value is loaded into an SRAM cell, it will remain unchanged until it is explicitly altered or until power is removed from the device. SROM An acronym for supervisory read only memory. The SROM holds code that is used to boot the device, calibrate circuitry, and perform flash operations. The functions of the SROM may be accessed in normal user code, operating from flash. stack A stack is a data structure that works on the principle of Last In First Out (LIFO). This means that the last item put on the stack is the first item that can be taken off. stack pointer A stack may be represented in a computer’s inside blocks of memory cells, with the bottom at a fixed location and a variable stack pointer to the current top cell. state machine The actual implementation (in hardware or software) of a function that can be considered to consist of a set of states through which it sequences. sticky A bit in a register that maintains its value past the time of the event that caused its transition, has passed. 442 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Glossary stop bit A signal following a character or block that prepares the receiving device to receive the next character or block. switching The controlling or routing of signals in circuits to execute logical or arithmetic operations, or to transmit data between specific points in a network. switch phasing The clock that controls a given switch, PHI1 or PHI2, in respect to the switch capacitor (SC) blocks. The PSoC SC blocks have two groups of switches. One group of these switches is normally closed during PHI1 and open during PHI2. The other group is open during PHI1 and closed during PHI2. These switches can be controlled in the normal operation, or in reverse mode if the PHI1 and PHI2 clocks are reversed. synchronous 1. A signal whose data is not acknowledged or acted upon until the next active edge of a clock signal. 2. A system whose operation is synchronized by a clock signal. T tap The connection between two blocks of a device created by connecting several blocks/components in a series, such as a shift register or resistive voltage divider. terminal count The state at which a counter is counted down to zero. threshold The minimum value of a signal that can be detected by the system or sensor under consideration. Thumb-2 The Thumb-2 instruction set is a highly efficient and powerful instruction set that delivers significant benefits in terms of ease of use, code size, and performance. The Thumb-2 instruction set is a superset of the previous 16-bit Thumb instruction set, with additional 16-bit instructions alongside 32-bit instructions. transistors The transistor is a solid-state semiconductor device used for amplification and switching, and has three terminals: a small current or voltage applied to one terminal controls the current through the other two. It is the key component in all modern electronics. In digital circuits, transistors are used as very fast electrical switches, and arrangements of transistors can function as logic gates, RAM-type memory, and other devices. In analog circuits, transistors are essentially used as amplifiers. tristate A function whose output can adopt three states: 0, 1, and Z (high impedance). The function does not drive any value in the Z state and, in many respects, may be considered to be disconnected from the rest of the circuit, allowing another output to drive the same net. U UART A UART or universal asynchronous receiver-transmitter translates between parallel bits of data and serial bits. user The person using the PSoC device and reading this manual. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 443 Glossary user modules Pre-build, pre-tested hardware/firmware peripheral functions that take care of managing and configuring the lower level Analog and Digital PSoC Blocks. User Modules also provide high level API (Application Programming Interface) for the peripheral function. user space The bank 0 space of the register map. The registers in this bank are more likely to be modified during normal program execution and not just during initialization. Registers in bank 1 are most likely to be modified only during the initialization phase of the program. V Vddd A name for a power net meaning "voltage drain." The most positive power supply signal. Usually 5 or 3.3 volts. volatile Not guaranteed to stay the same value or level when not in scope. Vss A name for a power net meaning "voltage source." The most negative power supply signal. W watchdog timer A timer that must be serviced periodically. If it is not serviced, the CPU will reset after a specified period of time. waveform The representation of a signal as a plot of amplitude versus time. X XOR 444 See Boolean Algebra. PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Index Numerics A 24-bit data pointer registers 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 32.768 kHz crystal oscillator . . . . . . . . . . . . . . . . . . . . 129 50% duty cycle mode . . . . . . . . . . . . . . . . . . . . . . . . . 135 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 24-bit data pointer registers . . . . . . . . . . . . . . . . . . 64 active interrupt . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 arithmetic instructions . . . . . . . . . . . . . . . . . . . . . . 40 arithmetic logic unit . . . . . . . . . . . . . . . . . . . . . . . . 40 bit addressing mode . . . . . . . . . . . . . . . . . . . . . . . . 39 block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 boolean instructions . . . . . . . . . . . . . . . . . . . . . . . . 43 core enhancements . . . . . . . . . . . . . . . . . . . . . . . . 38 DPTR extension SFR . . . . . . . . . . . . . . . . . . . 38 Dual DPTR . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 vectored interrupt controller interface . . . . . . . 38 CPU halt mechanisms . . . . . . . . . . . . . . . . . . . . . . 66 CY wrapper logic . . . . . . . . . . . . . . . . . . . . . . . . . . 39 data transfer instructions . . . . . . . . . . . . . . . . . . . . 42 direct addressing mode . . . . . . . . . . . . . . . . . . . . . 39 dual data pointer registers . . . . . . . . . . . . . . . . . . . 63 external data memory space . . . . . . . . . . . . . . . . . 66 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . 38, 389 immediate constants mode . . . . . . . . . . . . . . . . . . 39 indirect addressing mode . . . . . . . . . . . . . . . . . . . . 39 instruction set . . . . . . . . . . . . . . . . . . . . . . . . . . 38, 39 instruction set details . . . . . . . . . . . . . . . . . . . . . . . 44 internal data space . . . . . . . . . . . . . . . . . . . . . . . . . 39 interrupt controller interface . . . . . . . . . . . . . . . . . . 38 interrupt enable register . . . . . . . . . . . . . . . . . . . . . 65 interrupt nesting . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 interrupt vector addresses . . . . . . . . . . . . . . . . . . . 97 IO port access registers . . . . . . . . . . . . . . . . . . . . . 65 jump instructions . . . . . . . . . . . . . . . . . . . . . . . . . . 43 logical instructions . . . . . . . . . . . . . . . . . . . . . . . . . 41 memory map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 program branching instructions . . . . . . . . . . . . . . . 43 program memory space . . . . . . . . . . . . . . . . . . . . . 66 PSoC 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 register addressing mode . . . . . . . . . . . . . . . . . . . . 39 register specific instructions mode . . . . . . . . . . . . . 39 special function registers . . . . . . . . . . . . . . . . . . . . 62 wrapper logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 ACC (SFR 0xE0) register . . . . . . . . . . . . . . . . . . . . . . . .40 active interrupt 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 active mode entering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 exiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 low power modes . . . . . . . . . . . . . . . . . . . . . . . . . .152 PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 alternative active mode . . . . . . . . . . . . . . . . . . . . . . . . .153 entering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .153 exiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .153 low power modes . . . . . . . . . . . . . . . . . . . . . . . . . .153 analog I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .176 analog subsystem PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 analog subsystem components PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 analog system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .319 architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .319 arithmetic instructions 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 arithmetic logic unit 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 asynchronous clocks . . . . . . . . . . . . . . . . . . . . . . . . . .138 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F B B (SFR 0xF0) register . . . . . . . . . . . . . . . . . . . . . . . . . . .40 bit addressing mode 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39 block diagram 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 cache controller . . . . . . . . . . . . . . . . . . . . . . . . . . . .67 clock distribution system . . . . . . . . . . . . . . . . . . . .132 clock divider Implementation . . . . . . . . . . . . . . . . .134 clock selection . . . . . . . . . . . . . . . . . . . . . . . . . . . .275 DMA controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73 external memory interface . . . . . . . . . . . . . . . . . . .113 Flash programming system . . . . . . . . . . . . . . . . . .109 GPIO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .168 I/O drive mode . . . . . . . . . . . . . . . . . . . . . . . . . . . .171 internal low speed oscillator . . . . . . . . . . . . . . . . .127 internal main oscillator . . . . . . . . . . . . . . . . . . . . . .127 interrupt controller . . . . . . . . . . . . . . . . . . . . . . . . . .91 master clock mux . . . . . . . . . . . . . . . . . . . . . . . . . .133 445 Index MHzECO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 phase shifter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 phase-locked loop . . . . . . . . . . . . . . . . . . . . . . . . . 130 PHUB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 port interrupt controller unit . . . . . . . . . . . . . . . . . . 179 RESET module . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 resync option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 SIO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 SRAM organization for the CY8C38 family . . . . . . 106 USB clock mux . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 USBIO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 voltage monitoring . . . . . . . . . . . . . . . . . . . . . . . . . 147 watchdog timer circuit . . . . . . . . . . . . . . . . . . . . . . 157 boolean instructions 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 boost converter PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 bypassed clock source . . . . . . . . . . . . . . . . . . . . . . . . . 136 C capture mode timing diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 capture signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 central processing unit cache controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 PHUB and DMA . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 clock system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 clock block components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 clock distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 clock distribution module . . . . . . . . . . . . . . . . . . . . . . . 134 clock divider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 50% duty cycle mode . . . . . . . . . . . . . . . . . . . . . . 135 single cycle pulse mode . . . . . . . . . . . . . . . . . . . . 135 clock doubler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 clock frequency USB mode operations . . . . . . . . . . . . . . . . . . . . . . 134 clock generator introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 clock selection block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 clock selection by timer block . . . . . . . . . . . . . . . . . . . . 274 clock signal naming . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 clock source selection by timer block . . . . . . . . . . . . . . . . . . . . . 274 clock sources distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 DSI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 internal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 phase-locked loop . . . . . . . . . . . . . . . . . . . . . . . . . 130 clock synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . 135 clock tree bypassed clock source output . . . . . . . . . . . . . . . . 136 phase delayed clk_sync output . . . . . . . . . . . . . . . 135 power gating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 446 resynchronized clock output . . . . . . . . . . . . . . . . . 135 types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 unsynchronized divided clock output . . . . . . . . . . 136 clocks asynchronous . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 high precision . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 low power mode operation . . . . . . . . . . . . . . . . . . 138 cock dividers main part of clock distribution module . . . . . . . . . 134 compare types in pulse width mode . . . . . . . . . . . . . . 284 core enhancements 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 counter mode register configuration . . . . . . . . . . . . . . . . . . . . . . 280 counters features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 CPU halt mechanisms 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 CPU system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 crystal oscillator low power operation . . . . . . . . . . . . . . . . . . . . . . . 129 CY wrapper logic 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 CY8C38 family Flash memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 PSoC 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 RAM implementation . . . . . . . . . . . . . . . . . . . . . . 107 D data transfer instructions 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 dead band feature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286 delay chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 development kits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 device configuration of nonvolatile latch . . . . . . . . . . . 101 digital I/O control by DSI . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 digital system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 digital-to-analog converter . . . . . . . . . . . . . . . . . . . . . . . 27 direct addressing mode 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 DMA controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 document glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443 overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 doubled clock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 DPTR extension SFR 8051 core enhancements . . . . . . . . . . . . . . . . . . . . 38 DSI control of digital I/O . . . . . . . . . . . . . . . . . . . . . . . . 174 DSI input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 DSI output to I/O port . . . . . . . . . . . . . . . . . . . . . . . . . . 174 dual data pointer registers PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Index 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Dual DPTR 8051 core enhancements . . . . . . . . . . . . . . . . . . . . 38 E EEPROM features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 EEPROM memory how erased . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 EMIF block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 external memory support . . . . . . . . . . . . . . . . . . . 114 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 memory mapped peripherals . . . . . . . . . . . . . . . . 120 registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 sleep mode behavior . . . . . . . . . . . . . . . . . . . . . . 117 timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 EMIF register list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 enable signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 enabling and disabling timer block . . . . . . . . . . . . . . . 275 enabling interrupts PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 erasing EEPROM memory . . . . . . . . . . . . . . . . . . . . . 112 external crystal oscillators . . . . . . . . . . . . . . . . . . . . . . 128 external data memory space 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 external memory interface block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 connection to peripheral devices . . . . . . . . . . . . . 120 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 register list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 sleep mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 external memory support in EMIF . . . . . . . . . . . . . . . . 114 F fast start IMO (FIMO) . . . . . . . . . . . . . . . . . . . . . . . . . . 127 features 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 counters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 EEPROM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 EMIF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Flash memory . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 I/O system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 interrupt controller . . . . . . . . . . . . . . . . . . . . . . . . . 91 low power modes . . . . . . . . . . . . . . . . . . . . . . . . . 151 nonvolatile latch . . . . . . . . . . . . . . . . . . . . . . . . . . 101 PHUB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 port interrupt controller unit . . . . . . . . . . . . . . . . . 179 power supply and monitoring . . . . . . . . . . . . . . . . 141 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F pulse-width modulator . . . . . . . . . . . . . . . . . . . . . .273 SRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .105 timers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .273 watchdog timer . . . . . . . . . . . . . . . . . . . . . . . . . . .157 flash interrupt attempted flash write . . . . . . . . . . . . . . . . . . . . . . .110 ECC – multi-bit . . . . . . . . . . . . . . . . . . . . . . . . . . .110 ECC – single bit . . . . . . . . . . . . . . . . . . . . . . . . . . .110 Flash memory CY8C38 family . . . . . . . . . . . . . . . . . . . . . . . . . . . .110 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110 PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109 regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110 free run mode timing diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . .279 free run timer mode . . . . . . . . . . . . . . . . . . . . . . . . . . .279 frequency generation . . . . . . . . . . . . . . . . . . . . . . . . . .125 G gated timer in pulse width timing diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . .281 general purpose input/output interface PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26 glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .443 GPIO block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . .168 GPIO pins in creation of buttons and sliders . . . . . . . .176 H help, getting development kits . . . . . . . . . . . . . . . . . . . . . . . . . . .29 hibernate mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .153 entering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .154 exiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .154 low power modes . . . . . . . . . . . . . . . . . . . . . . . . . .153 hibernate regulator . . . . . . . . . . . . . . . . . . . . . . . . . . . .143 high impedance analog drive mode . . . . . . . . . . . . . . .171 high impedance digital drive mode . . . . . . . . . . . . . . . .171 high voltage interrupt . . . . . . . . . . . . . . . . . . . . . 147, 148 hot swap SIO pin support . . . . . . . . . . . . . . . . . . . . . . . . . . .178 how it works 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38, 389 EEPROM memory . . . . . . . . . . . . . . . . . . . . . . . . .112 EMIF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .114 external memory interface . . . . . . . . . . . . . . . . . . .114 Flash memory . . . . . . . . . . . . . . . . . . . . . . . . . . . .110 I/O system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170 interrupt controller . . . . . . . . . . . . . . . . . . . . . . . . . .92 PHUB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .74 power regulators . . . . . . . . . . . . . . . . . . . . . . . . . .143 PSoC RAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 watchdog timer . . . . . . . . . . . . . . . . . . . . . . . . . . .158 447 Index I I/O drive mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 high impedance analog . . . . . . . . . . . . . . . . . . . . . 171 high impedance digital . . . . . . . . . . . . . . . . . . . . . 171 open drain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 resistive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 resistive pull up and pull down . . . . . . . . . . . . . . . 172 strong . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 I/O pins low power mode behavior . . . . . . . . . . . . . . . . . . . 179 overvoltage tolerance . . . . . . . . . . . . . . . . . . . . . . 178 I/O power supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 I/O system analog I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 CapSense . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 digital I/O controlled through DSI . . . . . . . . . . . . . 174 DSI input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 DSI output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170 I/O port reconfiguration . . . . . . . . . . . . . . . . . . . . . 178 I/O power supply . . . . . . . . . . . . . . . . . . . . . . . . . . 179 introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 LCD drive capabilities . . . . . . . . . . . . . . . . . . . . . . 176 low power mode effect on I/O pins . . . . . . . . . . . . 179 open drain modes . . . . . . . . . . . . . . . . . . . . . . . . . 172 overvoltage tolerance . . . . . . . . . . . . . . . . . . . . . . 178 port interrupt controller unit . . . . . . . . . . . . . . . . . . 179 port interrupt controller unit pin configuration . . . . 180 port register of digital I/O . . . . . . . . . . . . . . . . . . . . 172 power up I/O configuration . . . . . . . . . . . . . . . . . . 178 register summary . . . . . . . . . . . . . . . . . . . . . . . . . 181 resistive modes . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 resistive pull up and pull down mode . . . . . . . . . . 172 SFR to GPIO (PSoC3 only) . . . . . . . . . . . . . . . . . . 174 SIO functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 slew rate control . . . . . . . . . . . . . . . . . . . . . . . . . . 172 strong drive mode . . . . . . . . . . . . . . . . . . . . . . . . . 172 usage modes and configuration . . . . . . . . . . . . . . 170 identifying reset sources . . . . . . . . . . . . . . . . . . . . . . . . 160 immediate constants mode 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 indirect addressing mode 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 input signals of timer block . . . . . . . . . . . . . . . . . . . . . . 275 input voltage to boost converter . . . . . . . . . . . . . . . . . . 143 instruction set 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38, 39 instruction set details 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 internal data space 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 internal low speed oscillator . . . . . . . . . . . . . . . . .126, 127 internal main oscillator . . . . . . . . . . . . . . . . . . . . . . . . . 126 internal regulators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 interrupt controller block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 448 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 interrupt execution . . . . . . . . . . . . . . . . . . . . . . . . . 94 introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 level interrupt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 priority decoding unit . . . . . . . . . . . . . . . . . . . . . . . 94 pulse interrupt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 interrupt controller interface 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 interrupt enable register 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 interrupt execution interrupt controller . . . . . . . . . . . . . . . . . . . . . . . . . 94 interrupt nesting 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 interrupt service routine . . . . . . . . . . . . . . . . . . . . . . . . . 97 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 sleep mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 interrupt vector addresses 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 introduction clock generator . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 I/O system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 timer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 IO port access registers 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 J jump instructions 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 K kill signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 L LCD drive I/O system capabilities . . . . . . . . . . . . . . . . . . . . . 176 level interrupt interrupt controller . . . . . . . . . . . . . . . . . . . . . . . . . 94 logic diagram RESET module . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 logical instructions 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 low power modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 active mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 alternative active mode . . . . . . . . . . . . . . . . . . . . 153 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 hibernate mode . . . . . . . . . . . . . . . . . . . . . . . . . . 153 power mode transitions diagram . . . . . . . . . . . . . 151 sleep mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 timewheel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 low voltage detect interrupt processing . . . . . . . . . . . . 148 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Index low voltage interrupt . . . . . . . . . . . . . . . . . . . . . . . . . . 147 M mapping diagram digital system input to pad selection . . . . . . . . . . 174 master clock mux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 memory map 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 memory mapped peripherals in EMIF . . . . . . . . . . . . . 120 MHz crystal oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . 128 modes active mode in low power modes . . . . . . . . . . . . . 152 power consumption-reducing . . . . . . . . . . . . . . . . 152 N nonvolatile latch device configuration . . . . . . . . . . . . . . . . . . . . . . . device configuration register map . . . . . . . . . . . . features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . sleep mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . write once . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 101 101 103 103 102 O on the fly duty cycle update . . . . . . . . . . . . . . . . . . . . . 286 one shot mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 oscillators external crystal . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 internal PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 summary table . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 overview, document . . . . . . . . . . . . . . . . . . . . . . . . . 3, 21 getting started . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 overvoltage tolerance in I/O pins . . . . . . . . . . . . . . . . . 178 P peripheral device connection external memory interface . . . . . . . . . . . . . . . . . . 120 phase delayed clk_sync . . . . . . . . . . . . . . . . . . . . . . . 135 phase select and control . . . . . . . . . . . . . . . . . . . . . . . 136 phase shifter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 phase-locked loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 PHUB block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 port interrupt controller features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 port interrupt controller unit . . . . . . . . . . . . . . . . . . . . . 179 block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F pin configuration . . . . . . . . . . . . . . . . . . . . . . . . . .180 port register control of digital I/O . . . . . . . . . . . . . . . . . .172 power consumption-reducing modes . . . . . . . . . . . . . .152 power gating clock outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . .138 power mode transitions in low power modes . . . . . . . .151 power on reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .159 power regulators how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143 power supply and monitoring features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .141 priority decoding unit interrupt controller . . . . . . . . . . . . . . . . . . . . . . . . . .94 product upgrades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29 program and debug . . . . . . . . . . . . . . . . . . . . . . . . . . .413 architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .413 PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 program branching instructions 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 program memory space 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 Program Status Word SFR . . . . . . . . . . . . . . . . . . . . . . .40 PSoC active mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 analog subsystem . . . . . . . . . . . . . . . . . . . . . . . . . .27 analog subsystem components . . . . . . . . . . . . . . . .27 boost converter . . . . . . . . . . . . . . . . . . . . . . . . . . .143 DMA controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25 EEPROM memory . . . . . . . . . . . . . . . . . . . . . . . . .111 enabling interrupts . . . . . . . . . . . . . . . . . . . . . . . . . .92 Flash memory on-chip . . . . . . . . . . . . . . . . . . . . . .109 general purpose input/output interface . . . . . . . . . .26 program and debug . . . . . . . . . . . . . . . . . . . . . . . . .27 special input/output interface . . . . . . . . . . . . . . . . . .26 SRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .105 supply pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .141 timer blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .273 USB input/output interface . . . . . . . . . . . . . . . . . . . .26 PSoC 3 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 8051 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 CY8C38 family . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 major components . . . . . . . . . . . . . . . . . . . . . . . . . .23 PSoC Ram how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 PSoC3 SFR to GPIO . . . . . . . . . . . . . . . . . . . . . . . . . . . . .174 PSW0xD0 register . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40 pulse interrupt interrupt controller . . . . . . . . . . . . . . . . . . . . . . . . . .94 pulse width mode comparator mode . . . . . . . . . . . . . . . . . . . . . . . . .284 compare types . . . . . . . . . . . . . . . . . . . . . . . . . . . .284 pulse width modulator dead band feature . . . . . . . . . . . . . . . . . . . . . . . . .286 on the fly duty cycle update . . . . . . . . . . . . . . . . . .286 one shot mode . . . . . . . . . . . . . . . . . . . . . . . . . . . .287 register configuration . . . . . . . . . . . . . . . . . . . . . . .284 449 Index pulse-width modulator features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 R RAM implementation CY8C38 family . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 real time clock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 register addressing mode 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 register map nonvolatile latch . . . . . . . . . . . . . . . . . . . . . . . . . . 101 register specific instructions mode 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 register summary I/O system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 registers counter mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 EMIF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 pulse width modulator . . . . . . . . . . . . . . . . . . . . . . 284 timer block . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 regulator hibernate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 internal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 sleep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 identifying sources . . . . . . . . . . . . . . . . . . . . . . . . 160 introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 software initiated . . . . . . . . . . . . . . . . . . . . . . . . . . 160 watchdog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 RESET module logic diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 reset summary table . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 resynchronized clock . . . . . . . . . . . . . . . . . . . . . . . . . . 135 RETI instruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 S SFR – GPIO interface . . . . . . . . . . . . . . . . . . . . . . . . . . 39 single cycle pulse mode . . . . . . . . . . . . . . . . . . . . . . . . 135 SIO block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 SIO functions adjustable input level . . . . . . . . . . . . . . . . . . . . . . 177 hot swap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 regulated output level . . . . . . . . . . . . . . . . . . . . . . 177 sleep mode entering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 exiting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 external memory interface . . . . . . . . . . . . . . . . . . . 117 gating off of clock network outputs . . . . . . . . . . . . 138 interrupt service routine . . . . . . . . . . . . . . . . . . . . . . 98 low power mode . . . . . . . . . . . . . . . . . . . . . . . . . . 153 low power modes . . . . . . . . . . . . . . . . . . . . . . . . . 153 nonvolatile latch behavior . . . . . . . . . . . . . . . . . . . 103 timer block behavior . . . . . . . . . . . . . . . . . . . . . . . 288 450 sleep mode behavior in EMIF . . . . . . . . . . . . . . . . . . . 117 sleep regulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 slew rate control in I/O system . . . . . . . . . . . . . . . . . . 172 software initiated reset . . . . . . . . . . . . . . . . . . . . . . . . . 160 special function registers 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 special input/output interface PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 SRAM features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 supply pins PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 synchronization of clock . . . . . . . . . . . . . . . . . . . . . . . 135 system clock operation . . . . . . . . . . . . . . . . . . . . . . . . 138 system integration low power modes . . . . . . . . . . . . . . . . . . . . . . . . . 151 reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 system wide resources . . . . . . . . . . . . . . . . . . . . . . . . 123 architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 T timer introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . timer block enabling and disabling . . . . . . . . . . . . . . . . . . . . . input signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . selection of clock source . . . . . . . . . . . . . . . . . . . sleep mode behavior . . . . . . . . . . . . . . . . . . . . . . timer blocks PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . timer mode free run . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . period mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . stop on interrupt mode . . . . . . . . . . . . . . . . . . . . . timer reset signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . timing diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . timers features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . timewheel in low power modes . . . . . . . . . . . . . . . . . . timing EMIF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . timing diagram capture mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . free run mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . gated timer in IRQ mode . . . . . . . . . . . . . . . . . . . gated timer in period mode . . . . . . . . . . . . . . . . . . gated timer in pulse width mode . . . . . . . . . . . . . . timer reset signal . . . . . . . . . . . . . . . . . . . . . . . . . timing of EMIF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . timing of external memory interface . . . . . . . . . . . . . . 273 275 275 274 288 274 288 273 279 282 283 278 278 273 154 118 277 279 283 282 281 278 118 118 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F Index U unsynchronized divided clock . . . . . . . . . . . . . . . . . . . 136 update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 updating of clock divider . . . . . . . . . . . . . . . . . . . . . . . 137 upgrades . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 usage modes and configuration of I/O system . . . . . . 170 USB clock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 USB input/output interface PSoC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 USB mode operation clock frequency . . . . . . . . . . . . . . . . . . . . . . . . . . 134 USBIO block diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 V vectored interrupt controller interface 8051 core enhancements . . . . . . . . . . . . . . . . . . . . 38 voltage monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 W watchdog reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 watchdog timer clearing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 disabling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 enabling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 how it works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 operation in low power mode . . . . . . . . . . . . . . . . 158 protection settings . . . . . . . . . . . . . . . . . . . . . . . . 158 setting time period . . . . . . . . . . . . . . . . . . . . . . . . 158 wrapper logic 8051 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 write once latch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F 451 Index 452 PSoC 5LP Architecture TRM, Document No. 001-78426 Rev. *F
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.5 Linearized : Yes Author : shea Create Date : 2017:05:31 15:30:09Z Modify Date : 2017:05:31 15:30:43Z XMP Toolkit : Adobe XMP Core 5.2-c001 63.139439, 2010/09/27-13:37:26 Format : application/pdf Title : TRM_PSoC5.book Creator : shea Creator Tool : FrameMaker 10.0.2 Producer : Acrobat Distiller 10.1.16 (Windows) Document ID : uuid:3ec7a83d-1a2a-4c68-94ee-58341984c5e5 Instance ID : uuid:2d4e0ab2-0ac1-4067-b5d3-415804d7ace8 Page Mode : UseOutlines Page Count : 452EXIF Metadata provided by EXIF.tools