1991_Motorola_ECLin PS_Data 1991 Motorola ECLin PS Data
User Manual: 1991_Motorola_ECLinPS_Data
Open the PDF directly: View PDF .
Page Count: 180
Download | |
Open PDF In Browser | View PDF |
Numeric Index General Information ECIJ:PS Family Specifications & Device Data Sheets Design Guide This databook contains device specifications for Motorola's ECLinPS advanced ECL logic family. ECLinPS (ECL in picoseconds) was developed in response to the need for an even higher performance ECL family of standard logic functions, particularly in the Computer, Automated Test, Instrumentation and Communications industries. Family general features as well as specific functions were developed in close consultation with ECL systems design engineers. ECUnPS offers the user a single gate delay of 500 ps max., including package delay, and a flip-flop toggle frequency of 1100 MHz. ECUnPS is compatible with two different ECL standards. Each function is available with either MECL 10KH compatibility (MC10Exxx series) or 100K compatibility (MC100Exxx series). ECUnPS is offered in the 28-lead plastic leaded chip carrier (PLCC), a J-Iead surface mount IC package. This package was selected for high performance, reduced parasitics and good thermal handling in a low cost, standard package, and reflects an industry trend towards surface mount assembly. Suggested References: The user is referred to the following for general information on the MECL and 100K ECL families: Motorola MEeL Device Data Book, Motorola Inc., 1987. Stock code DL122/D. F100K EeL Data Book, Fairchild Camera and Instrument Corp. Motorola MEeL System Design Handbook, second edition. Motorola Inc., 1983. Stock code HB205RlID. Signetics EeL 10K/toOK Data Manual. ECLinPS MOSAIC, MECL 10K and MECL 10KH are trademarks of Motorola Inc. ©MOTOROLA INC .. 1991 Motorola reserves the right to make changes without further notice to any products herein to improve reliability, function or design. Motorola does not assume any liability arising out afthe application or use of any product or circuit described herein; neither does it convey any license under its patent rights nor the rights of others. Motorola products are not designed, intended, or authorized for use as components in systems intended for surgical implant into the body, or other applications intended to support or sustain life, or for any other application in which the failure afthe Motorola product could create a situation where personal injury or death may occur. Should Buyer purchase or use Motorola products for any such unintended or unauthorized application, Buyer shall indemnify and hold Motorola and its officers, employees, subsidiaries, affiliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim of personal injury or death associated with such unintended or unauthorized use, even if such claim alleges that Motorola was negligent regarding the design or manufacture of the part. Motorola and are registered trademarks of Motorola, Inc. Motorola, Inc. is an Equal Opportunity/Affirmative Action Employer. ® ECLinPS Numeric Index This section lists ECUnPS family functions in numericalorder. MC10E series devices are compatible with the MECL 10KH family. MC100E series are compatible with toOK ECL. ECLinPS 1-1 ECl:PS Numeric Index Numeric Index MC101 MC100 E016 El0l El04 El07 E111 El12 El16 E122 E131 E136 E137 E141 E142 El43 E150 E151 1:154 E155 E1!i6 E157 E158 E160 E163 E164 Function Page 8-Bit Synch. Binary Counter QUAD 4-lnput OR/NOR Gate QUINT 2-lnput AND/NAND Gate QUINT 2-lnput XOR/XNOR Gate 1:9 Differential Clock Driver QUAD Driver, Common Enable QUINT Diff. Line Receiver 9-Bit Buffer 4-Bit 0 Flip-Flop 6-Bit Universal Counter 8-Bit Ripple Counter 8-Bit Universal Shift Register 9~Bit Shift Register 9-Bit Hold Register 6-Bit 0 Latch 6-Bit 0 Register 5-Bit,.2: 1 Mux Latch . 6-Bit 2: 1 Mux Latch 3-Bit 4: 1 Mux Latch QUAD 2:1 Mux, Separate Selects 5-Bit ;1:1 Multiplexer 12-Bit Parity Gen/Checker 2-Bit 8: 1 Multiplexer 16:1 Multiplexer 3-3 3-9 3-11 3-13 3-15 3-18 3-20 3-22 3-24 3-26 3-27 3-28 3-30 3-32 3-34 3-36 3-38 3-40 3-42 3-44 3-46 3-48 3-50 3-52 MC101 MC100 Function Page E166 E167 E171 E175 E193 E195 E196 *E197 E212 E241 E256 E336 E337 E404 E416 E431 E445 E446 E451 E452 E457 *E1651 *E1652 9-Bit Magnitude Comparator 6-Bit2: 1 Mux Register 3-Bit 4:1 Multiplexer 9-Bit Latch w/Parity Gen/Checker 8-Bit EDAC/Parity Programmable Delay Chip Programmable Delay Chip High Speed Data Separator 3-Bit Scannable ECL Driver 8-Bit Scannable Register 3-Bit 4: 1 Mux Latch 3-Bit Registered Bus Xcvr 3-Bit Scannable Bus Xcvr QUAD High Freq. Diff. AND QUINT High Freq. Line Receiver 3-Bit Diff. Set/Reset Flip-Flop 1:4 Serial/Paraliel Converter 4: 1 Parallel/Serial Converter 6-Bit 0 Reg. Diff. 0 and Clk 5-Bit 0 Reg. Diff. 0, Clk and Q TRIPLE High Freq. Diff. 2: 1 Mux Dual Analog Comparator Dual Analog Comparator 3-54 3-56 3-58 3-60 3-62 3-64 3-68 3-73 3-87 3-89 3-91 3-93 3-95 3-98 3-100 3-102 3-104 3-105 3-106 3-108 3-110 3-112 3-114 *10E version only Nomenclature MC 10 E xxx FN MC 100 E xxx FN Motorola Standard Prefix • XC - non reliability-qualified • MC - fully qualified I IL _." ,mo, • FN - PLCC package 3-digit function identifier Compatibility Identifier ·10 - MECL 10KH compatible ·100 - lOOK compatible .....- - - - ECLinPS family identifier ECLinPS 1-2 Selection Guide Gates Quad 4-lnput OR/NOR Quint 2-lnput AND/NAND Quint 2-lnput XOR/XNOR Quad 2-lnput AND/NAND, Differential Counters 8-Bit Synchronous Binary Counter 6-Bit Synchronous Universal Counter 8-Bit Triple Counter El0l El04 El07 E404 E016 E136 E137 Shift Registers Buffers 9-Bit Buffer 1:9 Differential Clock Driver Quad Driver with Enable 3-Bit Scannable Driver 8-Bit 8-Bit 9-Bit 9-Bit 3-Bit E122 Ell1 El12 E212 Flip-Flops/Registers 4-Bit 6-Bit 6-Bit 9-Bit 3-Bit 5-Bit D (Async Set/Reset) D (Async Reset) D, Diff. Data & CLK Inputs Hold Register D, Edge Triggered Set & Reset Diff. DReg. E141 E241 E142 E143 E212 Parity Generator/Comparator 12-Bit Parity Generator/Checker 9-Bit Magnitude Comparator 8-Bit Error Detection/Correction (EDAC) 9-Bit Latch w/Parity Gen/Checker E131 E151 E451 E143 E431 E452 E160 E166 E193 E175 Line Receivers Quint Differential Line Receiver 1:9 Differential Clock Driver 6-Bit DReg., Diff. Data & CLK Inputs Quint High Freq. Differential Line Receiver 5-Bit Diff. DReg. Latches 6-Bit D (Async Reset) 9-Bit Latch w/Parity Gen/Checker Shift Register (bidirectional) Scannable Register (unidirectional) Shift Register (unidirectional) Hold Register Scannable Driver E150 E175 E116 Elll E451 E416 E452 Multiplexers Bus Transceivers 5-Bit 2:1 Multiplexer 3-Bit 4: 1 Multiplexer 2-Bit 8: 1 Multiplexer Single 16:1 Multiplexer Quad 2:1 Mux, Indiv. Select Triple 2: 1 Mux, Differential E158 E171 E163 E164 E157 E457 3-Bit Registered Bus XCVR 3-Bit Scannable Reg. Bus XCVR Miscellaneous Dual Analog Comparator w. Latch Dual Analog Comparator w. Latch & Hysteresis Programmable Delay Chip, Digital Programmable Delay Chip, Digital & Analog Hard Disk Data Separator 1:4 Serial/Parallel Converter 4:1 Parallel/Serial Converter Mux-Latches 5-Bit 6-Bit 3-Bit 3-Bit 2:1 Mux-Latch 2: 1 Mux-Latch 4:1 Mux-Latch 4: 1 Mux-Latch E154 E155 E156 E256 Mux-Registers 6-Bit 2:1 Mux-Register E336 E337 E167 ECLinPS 1-3 E1651 E1652 E195 E196 E197 E445 E446 ECLinPS 1-4 General Information CONTENTS This section contains a technical overview of the ECUnPS family as well as an outline of its electrical characteristics. In addition the section outlines the procedures and philosophies used to AC test the family. Family Overview . ....................... 2-2 Electrical Characteristics . ................ 2-7 ECLinPS 2-1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. GENERAL INFORMATION Family Overview SECTION 1 Family Overview Introduction system power and board space while maintaining exceptional system timing. Recent advances in bipolar processes have led to a proliferation of very high speed lSI and VlSI gate arrays in high end computer applications. The advent of these high speed arrays has created a need for a high speed logic family to tie or "glue" them together. Because arrays have a finite amount of circuitry and 1/0 pins, glue functions which are sensitive to either of these parameters may be better performed off of the array. In addition glue functions which require very tight skew control may be difficult to perform on an array due to the inherent skew of the large packages associated with large gate arrays. Therefore although the trend is to push more and more of the logic onto the array, there are design constraints which make performing some of the logic, such as clock distribution, multiplexing, decoding, latching, memory addressing and translating, in glue an attractive alternative. The high end computer segment is not the only market segment pushing for higher performance logic parts. ATE, instrumentation and communication designs can have data rate requirements ranging from 300MHz to as high as 2.5GHz. Because large high speed arrays do not always lend themselves to passing high frequency signals on and off chip, portions of the designs must be realized with discrete logic. The current bipolar logic families are not capable of operating at these high frequencies. To answer the call for a very high speed bipolar logic family Motorola has designed and produced the ECLinPS (ECl in Pica Seconds) logic family. The family was designed to meet the most stringent of system requirements in speed, skew and board density as well as maintaining compatibility to existing ECl families. Transmission line Drive Capability The low output impedance, high input impedance and high current drive cilpability of ECl makes it an ideal technology for driving transmission lines. Regardless of the technology, as system speeds increase, interconnect becomes more of a transmission line phenomenon. With ECl no special line driving devices are necessary as all ECl devices are line drivers. Constant Power Supply Current Drain Because of the differential amplifier design used for ECl circuits the current is not switched on and off but rather simply steered between two paths. Thus the current drain of an ECl device is independent of the logic state and the frequency of operation. This current stability greatly simplifies system power supply design. Input Pulldown Resistors ECl inputs have 50Kn - 75Kn internal pulldown resistors which pull the input to VEE (logic lOW) when left open. This allows unused inputs to be left open and greatly simplifies logic design. Differential Drive Capability Because of the presence of high current drive complimentary outputs, ECl circuits are ideally suited for driving twisted pair lines or cables over long distances. With common mode noise rejection of 1V or more ECl line receivers are less susceptible to common mode noise. In addition their differential inputs need only a few hundred millivolt voltage differences to correctly interpret the logic. Eel Design Benefits High Speed Design Philosophy The speed benefits of an ECl design over those of alternative logic technologies are well documented, however there are a number of other important features that make ECl an attractive technology for system designs. The ECLinPS logic family as with other ECl families afford the following advantages: Today a truly high speed logic family needs more than simply short propagation delays. The minimization of all types of skew as well as a level of logic density which affords a smaller amount of board space for an equivalent function are also necessities of a high speed family. The following summary will outline the steps taken by Motorola to achieve these goals in the development of the ECLinPS logic family. Complimentary Outputs Complimentary outputs are available on many functions with equal propagation delays between the two paths. This alleviates the need for external inverters and saves Fast Propagation Delays The ECLinPS family boasts 500ps maximum packaged ECLinPS 2-2 Family Overview ing situations a minimum of three single ended outputs per Veeo has been employed in the family. Optimum placement of these Vceo's also results in superior output to output skew. gate delays and typical flip flop toggle frequencies of 1 AGHz. Simple gate functions show typical propagation delays of 360ps at 25mW of power for a speed power product of only 9pJ. For higher density devices internal gates run at 100ps with 5mW of power for a speed power product of only .5pJ. Advanced Bipolar Processing Internal Differential Interconnect The propagation delay window size, skew between rising and falling inputs and susceptibility to noise are all phenomenon which are exacerbated by Vss switching reference variation. By extensively using differential interconnects internal to the chip the ECLinPS family has been able to achieve superior performance in these areas. The ECLinPS logic family is fabricated using Motorola's MOSAIC III process, a process which is two generations ahead of the process used in the development of the 10KH family. The small geometries and feature sizes ofthe MOSAIC III process enables the ECLinPS logic family to boast of a nearly 3 fold improvement in speed at less than half the power of existing ECl logic families. The MOSAIC III process is a double polysilicon process which uses a unique self-alignment scheme for device electrode and isolation definition. The process features self aligned submicron emitters as well as polysilicon base, collector and emitter electrodes. In addition polysilicon resistors, diodes and capacitors are available to minimize the parasitic capacitance of an ECl gate. Figure 1.1 shows a cross section for an NPN device using the MOSAIC III process. Propagation Delay Temperature Insensitivity The variation of propagation delay through an ECLinPS device across temperature is typically less than 50ps. This stability allows for faster designs due to tighter delay windows across temperature. Input impedance and Loading Capacitance The input structures of the ECLinPS family show a positive real impedance across the applicable input frequency range. This ensures that the system will remain stable and operate as designed over a wide range of input frequencies. The input loading capacitance typically measures only 1.5pF and is virtually independent of input fanout as the device capacitance is less than 5% of the total. Because the propagation delay of a signal down a transmission line is adversely affected by loading capacitance the overall system speed is enhanced. Input Buffers To minimize propagation delays in a system environment, inputs with a large internal fanout are buffered to minimize the loading capacitance on the transmission line. Figure 1.1 - MOSAIC III Cross Section High Level of Integration 28 pin designs allow for the design of 9 bit functions for implementation in byte plus parity applications. Full byte plus parity implementation reduces total package count and saves expensive board space. By incorporating the use of polysilicon contacts and resistors through the MOSAIC III process, the parasitic capacitances of an ECLinPS gate are minimized, thus minimizing the time constants which comprise the switching delays of the gate. The resultant gates show delays of 100ps for internal gates and 200ps for output gates capable of driving son loads. The small geometries of the process, nearly 350% reduction in device area compared to a 10KH device, allows these internal gate delays to be achieved at only 800flA of current. Space Efficient Package Surface mount PlCC package affords a high level of integration with a minimum amount of required board space. Quad layout of the package equalizes pin lengths thus minimizing the skew between similar internal paths. Universal Compatibility Flow Through Pin Assignment Input and output pins have been laid out in a flow through pattern with the inputs on one side of the package and the outputs on the other. This flow through pattern helps to simplify the PC board layout operation. Each member of the ECLinPS family is available in both of the existing ECl standards: 10E series devices are compatible with the MECl 1OKH family; 1OOE series devices are compatible with ECl 100K. In addition, to maintain compatability with temperature compensated, three level series gated gate arrays the 1OOE devices are guaranteed to Multiple VCCO pins To minimize the noise generated in simultaneous switch- ECLinPS 2-3 Family Overview operate without degradation to a VEE of -S.46V. The section below presents comparison between thE! two standards in the new context of. the ECLinPS family. The user is also referred to the Electrical Characteristics section of this book as well as appropriate family databooks and other literature for descriptive information on the earlier ECl families. Because no supplier previous to Motorola has offered both EClstandards on an identical process, comparison of existing 1OKH and 1OOK style devices has some limitations. Comparison of the two standards fabricated withtwd different processes has sometimes led to the errone'ous conclusion that there are inherent ACperformance differences between them. In reality this is not the case. The only inherent difference between the two standards is the difference in the behavior of the DC characteristics with temperature. and board space problems, the propagation delays through the DIP package were nearly twice as long as the delay through the silicon. The 28 pin PlCC package emerged as the clear favorite both internally and with the high speed market in general. The package offers a quad layout to minimize both lead lengths and lead length differences. As a result the parasitics and delays of the package are very well suited for a high speed logic family. In additionthe nearly matched lead lengths allow for tighter skew among similar paths through the Chip. The board density potential of the PlCC is also attractive in that it allows for a nearly 100% reduction in board space when compared to the DIP alternative. The package is approximately a half inch square with SO mil spaced Jbend leads. More detailed measurements can be found in the package section of this databook. The J bend leads provide a smaller footprint than a gull wing package and propose fewer temperature expansion coefficient mismatch problems than the lead less alternative. Thermally the standard PlCC exhibits a ElJA of 43.SoC per watt at SOO Ifpm air flow. With this thermal resistance most 28 pin functions can be implemented with the MOSAIC III process without encountering any severe thermal problems. For more details on thermal issues of the ECLinPS family refer to the thermal section of this databook. a AC Performance From an IC design standpoint the only differences between a 10E device and a100E device in the ECLinPS family is a small temperature compensation network in the 100E output gate, and very minimal differences in the two bias generator networks. Therefore one would expect that from an AC standpoint the performance of the two standards in the ECLinPS family should be nearly identical; measurements prove this to be the case. There is no significant measurable difference in the riselfall times, propagation delays or toggle frequencies when comparing a 10E and 1OOE device. The minor difference between previous 10KH and 100K designs is due to the fact that the two are fabricated on differentprocE!sses, and in some cases are designed for operation at different power levels. Abbreviation Definitions The following is a list of abbreviations found in this databook and a brief definition of each. Current Summary Total power supply current drawn from .the pOSitive supply by an ECLinPS unit under test. Summarizing the above information; in general the two ECl design standards, although differing somewhat in DC parameters, are nearly identical when one compares the AC performance for a given device. There may be very small differences in the AC measurements due to the slightly smaller output swing of the 100E device. However these difference are negligible when compared to the absolute value of the measurements. Therefore from an AC standpoint there is no real advantage in using one standard over the other, thus removing AC performance as a decision variable in high speed system design, Total power supply current drawn from an EClinPS device under test by the negative supply. Current drawn by the input of an ECLinPS device with a specified low level (Vll min) forced on the input. Current drawn by the input of an ECLinPS device with a specified high level (VIH max) forced on the input. Packaging The current sourced by an output under specified load.conditions. During the definition phase of the ECLinPS family much attention was placed on the identification of a suitable package for the family. The package h.ad to meet the criteria of minimum parasitics and propagation delays along with an attractive I/O vs board space relationship. Although the DIP package offered a level of familiarity and convenience the performance of the package with a very high speed logic family was inadequate. In addition to the obvious parasitics Voltage VBB The switching reference voltage Base-to-emitter voltage drop of a transistor at ECLinPS 2-4 Family Overview specified collector and base currents. Vpp Minimum peak to peak input voltage for differential input devices. VeMR The voltage range in which the logic HIGH voltage level of a differential input signal must fall for a differential input device. Power supply connection to the output emitter follower of an ECLinPS gate. For the ECLinPS logic family VCC and VCCO are common nodes. VeuT The logic LOW voltage level for ECL BUS outputs which attain cutoff of the output emitter follower. VEE The most negative supply voltage to an ECLinPS device. Vsup The maximum voltage difference between VEE and VCC for the E1651 comparator. V ,H Nominal input logic HIGH voltage level. Timing Parameters V ,H max Maximum (most positive) logic HIGH voltage level for which all parametric specifications hold. tR Waveform rise time of an output signal measured from the 20% to 80% levels of the signal. V ,H min Minimum (least positive) logic HIGH voltage level for which all parametric specifications hold. t, Waveform fall time of an output signal measured from the 20% to 80% levels of the signal. V'l Nominal input logic LOW voltage. TpD±± Propagation delay of a signal measured for a rising/falling input to a rising/falling output. V'l max Maximum (most positive) logic LOW voltage level for which all parametric specifications hold. xpt The crossing point of adifferential input or output signal. The reference point for which differential delays are measured. TplH The propagation delay for an output transitioning from a logic LOW level to a logic HIGH level. TpHl The propagation delay for an output transitioning from a logic HIGH level to a logic LOW level. f MAX Maximum input frequency for which an ECLinPS flip flop will function correctly. feouNT Maximum input frequency for which an ECLinPS counter will function properly. fSHIFT Maximum input frequency for which an ECLinPS shift register will function properly. tSKEW The maximum delay difference between similar paths on a single ECLinPS device. ts Setup time; the minimum amount oftime an input must transition before a clock transition to ensure proper function of the device. tH Hold time: the minimum amount of time an input must remain asserted after a clock transition to ensure proper operation of the device. tRR Release time or Reset Recovery Time; the mini Ves Collector-to-base voltage drop of a transistor at specified collector and base currents. Vee The most positive supply voltage to an ECLinPS device. Veeo V,lmin Minimum (least positive) logic HIGH voltage level for which all parametric specifications hold. V OH Output logic HIGH voltage level for the specified load condition. VOHA Output 10gic.HIGH voltage level with the inputs biased at VIH min or VOL max. VOH max Maximum (most positive) logic HIGH output volItage level. V OH min Minimum (least positive) logic HIGH output voltagelevel. Val Output logic LOW voltage level for the specified load condition. VOLA Output logic LOW voltage level with the inputs biased at VIH min or VOL max. Val max Maximum (most positive) logic LOW output voltage level. Val min Minimum (least positive) logic LOW output voltage level. VTT Output termination voltage for ECLinPS open emitter follower outputs. ECLinPS 2-5 Family Overview mum amount of time after a signal is de-asserted that a different input must wait before assertion to ensure proper functionality of the device. 1w min kage between the case and the ambient. Ifpm Linear feet per minute. Miscellaneous Minimum pulse width of a signal necessary to ensure proper functionality of a device. D.U.T. Device under test. G'N Input capacitance of a device. Z'N Input impedance of a device. GOUT Output capacitance of a device. Temperature The maximum temperature at which a device may be stored without damage or performance degradation. Junction (or die) temperature of an integrated circuit device. Output impedance of a device. Ambient (environment) temperature existing in the immediate vicinity of an integrated circuit package. The total dc power applied to a device, not including any power delivered from the device to the load. Thermal resistance of an integrated circuit package between the junction and the ambient. Load resistance RT Thermal resistance of an integrated circuit package between the junction and the case. Transmission line termination resistor. An input pull-down resistor. P.U.T. Thermal resistance of an integrated circuit pac- ECLinPS 2-6 Pin under test. MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. GENERAL INFORMATION Electrical Characteristics SECTION 2 Electrical Characteristics DC Characteristics The 10E devices are voltage compensated but not temperature compensated, therefore although the output voltage levels are insensitive to variations in VEE they do vary with temperature. The transfer curves in Figure 2.2 pictorially illustrate the behavior of the 10E outputs. In order to maintain noise margins overtemperature it is important that the VBB switching reference tracks with temperature in such a way as to remain centered between the VOH and VOL levels. As shown in Table 2.1 the temperature tracking rates of the V OH and VOL for a 10E device are not equal. Therefore it is ECLinPS Transfer Curves As mentioned in the previous section, except for the E16S1, E16S2 and E197 all ECLinPSdevices are offered in either 1OE or 1OOE versions to be compatible with 1OKH or 100K ECl logic respectively. The following information will overview the DC characteristics of the two versions of ECLinPS devices, for more detailed discussions the reader is referred to the MECl and F100K databooks. Both 10E and 100E devices produce =800mV output swings into a specified SOQ to -2.0V load. However because of the low output impedance (Figure 2.1) of both standards neither is limited to SOQ loads. larger load resistances can be used to reduce the system power without sacrificing the speed of the device. Of course the overall system speed will be reduced due to the increased delays of the interconnect traces. In addition, to better drive high capacitive lines, smaller resistances, down to 2SQ, can be used without violating the SOmA max output current specification. It is however recommended that for lines of less than 3SQ specialized 2SQ driver circuits or "ganged" output schemes should be used to ensure optimum long term reliability of the device. -0.8 I I~ -1.0 ~ " l!! '0 > '5 S::I 1\ I -1.2 ~5~ CI -1.4 II \ -1.6 i~ 0 ,1.B T -2.0 -1.8 .~ I~ JlT I VEE = -4.94V to -5.46V I -1.6 -1.4 -1.2 -1.0 -O.B Input Voltage (V) -0.8 25°C o·c -1.0 ~ " ~ \[\U, -1.2 ~A ~M CI 0 > '5 a. '5 7S"C '\ -1.4 -1.6 J\\ \ 0 -1.B ~ -2.0 -1.75 ·1.5 -1.25 -1.0 -0.75 -0.5 -2.0 -0.25 Output Voltage (V) -1.B [$ -1.6 ·1.4 -1.2 -1.0 -O.B Input Voltage (V) Figure 2.1 • Output Characteristics vs Load Figure 2.2 • ECLinPS 10E Transfer Curves ECLinPS 2-7 Electrical Characteristics necessary to design the VBB reference such that it tracks at a rate equal to the average rate of the difference between the high and low output level tracking rates. Table 2.1 also outlines the temperature tracking behavior of a 10E VBB switching reference. 10E min typ max IWoH/I1T (mV/oC) I1Vo/I1T (mV/oC) I1VBB/I1T (mV/oC) I1Vojl1VEE (mV/v) I1Vo/I1VEE (mV/v) I1VBB/I1VEE (mV/v) 1.1 0 0.6 0 0 0 1.2 0.4 0.8 5 10 5 1.4 0.6 1.0 20 30 20 100E min typ max I1V oH/I1T (mV/oC) I1VoL/I1T (mV/oC) I1VBB/I1T (mV/oC) I1VoH/I1VEE (mV/v) I1Vo/I1VEE (mV/v) I1VBB/I1VEE (mV/V) -.15 -.30 -.20 0 0 0 0 0 0 5 10 5 .15 .30 .20 20 30 20 Since the VBE'S of the current source transistor reduce with temperature, if the current source reference remains constant ,as is the case for 1DOE devices, the lEE of the device will vary with temperature. Careful scrutiny of the data sheets will reveal that the worst case lEE for a function is higher for the 100E version than the 10E version of that device. As a result from a power standpoint a 1DOE device operating at 85°C with a -4.5V VEE will be nearly identical to a 1OE device operating with a -5.2V VEE under identical temperature conditions. Although differing somewhat in many DC parameters, 1DE and 100E devices do share a couple of the same DC characteristics. Both designs show superior lEE vs VEE tracking rates due to the design of the voltage regulator. With a tracking rate of <3%/V this variation can effectively be ignored during system design. The output level and refer· ence level variation with VEE are also outstanding as can be seen in Table 2.1. Noise Margin The noise margin of a device is a measure of a device's resistance to undesirable switching. For ECLinPS as well as all ECl devices, noise margin is a DC specification. The noise margin is defined as the difference between the voltage level of an output of the sending device and the required voltage level of the input of the receiving device. Therefore a worst case noise margin can be calculated from the ECLinPS data sheets by simply subtracting the V,Lmax or V,Hmin from the VOL min or VOHmax respectively. Table 2.2 below illustrates the worst case and typical noise margins for both 1OE and 1DOE ECLinPS devices. Notice that the typical noise margins are approximately 100mV larger than the worst case. Table 2.1 - ECLinPS Voltage Level Tracking Rates The 1DOE devices, on the other hand, are temperature and voltage compensated, therefore the output levels remain fairly constant over variations in both VEE and temperature. Figure 2.3 shows the transfer characteristics for a 100E device. The associated tracking rates are illustrated in Table 2.1. Notice that in this case the VBB switching reference is designed to remain constant over temperature to maintain an optimum position within the output swing of the device. This flat temperature tracking of the internal reference levels leads to a phenomena particular to the 1OOE devices. 10E NM H1GH (mV) NM Low (mV) ·0.8 I ·1.0 -1.475, 1.035 ~ -1.2 '"" -1.4 l! g 1 $ I :; s- -1.6 -1.475, 1.610 ::I 0 1f Y '1 -1.165, -1.035 TI -2.0 -1.8 ~ lT -1.4 -1.2 -1.0 min typ 150 150 240 280 140 145 210 230 As mentioned above the noise margins of a device are a DC measurement and thus can lead to some false impressions of the noise immunity of a system. For instance from the chart the worst case noise margin is 140mV for a high level of a 100E device. This would suggest that an undershoot on this line of greater than 140mV could cause an error in the system. This however is not necessarily the case as the determination as to whether or not an AC noise signal is propagated is dependent on line impedances, output impedances and propagation delays as well as noise margins. A more thorough investigation of the noise immunity of a system can be found in Application Note AN-592. vEE"4.2ovlo,s.46v -1.6 typ Table 2.2 - DC Noise Margins -1.165, -1.610 JOOl -1.8 100E min -0.8 Input Voltage (V) Figure 2.3- ECLinPS 100E Transfer Characteristics ECLinPS 2-8 Electrical Characteristics AC Characteristics parameters such as skew, setup/hold times, release times, and maximum frequency. The following few paragraphs will outline the ways in which Motorola defines these parameters. Parameter Definitions The device data sheets in Section 3 contain specifications for the propagation delays and rise/fall times for each of the devices in the ECLinPS family. In addition, where applicable, skew, setup/hold, maximum toggle frequencies (fMAX ) , reset recovery and minimum pulse width specifications are included. The waveforms and terminologies used in describing the propagation delays and rise/fall times of the ECLinPS family are depicted in Figure 2.4 below. Skew Times In the design of high speed systems skew plays nearly as important a role as propagation delay. The majority of the devices in the ECLinPS family have the skew between outputs specified. This skew specification represents the typical difference between the delays of similar paths on a single chip. No maximum value for skew is specified due to the difficulty in the production testing of this parameter. The user is encouraged to contact an ECLinPS application engineer to obtain actual evaluation data if this parameter is critical in their designs. Set-Up and Hold Times Motorola defines the setup time of a device as the minimum time, prior to the transition of the clock, that an input must be stable to ensure that the device operates properly. The hold time, on the other hand, is defined as the minimum time that an input must remain stable after the transition of the clock to ensure that the device operates properly. Figure 2.5 illustrates the way in which Motorola defihes setup and hold times. Rise and Fall Times Data ts th 50% Single-Ended Propagation Delay Clock Tpd V out V out Vout Figure 2.5 - Set-Up and Hold Waveforms V out Release Times Release times are defined as the minimum amount of time an input must waillo be clocked after an enable, master reset or set Signal is deactivated to ensure proper operation. Because more times than not this specification is in reference to a master reset operation this parameter is often called reset recovery time. Figure 2.6 illustrates the definition of release time in the Motorola data sheets. Differential Propagation Delay Figure 2.4 - ECLinPS TPO Measurement Waveforms Propagation delays and rise/fall times are generally well understood parameters, howeverthere is sometimes confusion surrounding the definitions of more specialized AC ECLinPS 2-9 Electrical Characteristics Master Reset Y.Data Q Clock [> Clock Q Figure 2.6 - ECLinPS Release Time Waveforms Figure 2.7 - f MAX Measurement Measurement In general fMAX is measured in the manner shown in Figure 2.7 with the fail criterion being either a swing of 600mV or less, or a miscount. However in some cases the feedback method of testing can lead to a pessimistic value of f MAX because the feedback path delay is such thatthe setup times of the device are violated. If this is the Case it is necessary to have two free running signal generators to ensure that the setup times are observed. This parameter, along with fSHIFT and fCOUNT.represents the maximum frequency at which a particularflip flop, shift register or counter can be clocked with the divide, shift or count operation guaranteed. This number is generated from worst case operating conditions, thus under nominal operating conditions the maximum toggle frequency is higher. 'MAX AC Testing ECLinPS Devices The introduction of the ECLinPS family raised the performance of silicon t,o a new domain. As the propagation delays of logic devices become ever faster the task of correlating between test setups becomes increasingly challenging. To obtain test'results which correlate with Motorola, various testing techniques must be adhered to. A typical schematic for an ECLinPS test setup is illustrated in Figure 2.8. A solid ground plane is a must in the test setup, as the two power supplies are bypassed to this ground plane. A 20~F capacitor from the two power supplies to ground is Channel A son Coax son Coax vccd·--------~~ Oscilloscope (+2.0V) / ChannelS VCC (+2.0V) son Coax son 0.01~F -3.2V· • VEE = -3.2V for 10Exxx; -2.5V for 1OOExxx . •• MultipleVCCo's exist on most parts Figure 2.8 • Typical ECLinPS Test Setup ECLinPS 2-10 Electrical Characteristics phenomena which can lead to inaccuracies in AC measurements. To minimize degradation of the input and output edge rates a 50n coaxial cable with a teflon dielectric is recommended, however any other cable with a bandwidth of >5.0GHz is adequate. In addition, the cables from the device under test (OUT) to the inputs of the scope should be matched in length to prevent any errors due to different path lengths from the OUT to the scope. The interconnect fittings should be 50n SMA straight or SMA launchers to minimize impedance mismatches at the interface of the coax and test PC board. Although a teflon laminate board is preferable, an FR4 laminate board is acceptable as long as the signal traces are kept to five inches or less. Longer traces will result in significant edge rate degradation of the input and output Signals. To make the board useful for incoming inspection or other volume testing the board needs to be fitted with a socket. Although not suitable for AC testing due to different pin lengths and large parasitics, there are through hole sockets which are adequate for DC testing of ECLinPS devices. For AC testing purposes a 28 pin PLCC surface mount socket is recommended. At the publication of this databook there are two sockets available which the Motorola ECLinPS group recommends: AMP part # 822039-1 and Method Electronics part #'s 213-028-601 or 213-028-602. To ease the correlation issue Motorola has developed a universal AC test board which is now available to customers. The board is fitted with a PLCC socket and comes with instructions on how it can be configured for the different device types in the family. For ordering information see the description on the following page. Finally, to ensure correlation between Motorola and the customer, ·high performance state-of-the-art measuring equipment should be used. The pulse generator must be capable of producing the required input levels with rise and fall times of 500ps. In addition, if fMAX is going to be tested, a frequency of of up to 1.5GHz may be needed. The oscilloscope should also be of the utmost in performance with a minimum bandwidth of 5.0 GHz. 20!!F capacitor from the two power supplies to ground is used to dampen any supply variations. An RF quality .01!!F capacitor from each power pin to ground is used to decouple the fixture. These .01!!F capacitors should be located as close to the power pins of the package as possible. In addition, in order to minimize the inductance of the power pins, all of the power leads should be kept as short as possible. The power supplies are shifted by +2.0V so thatthe load comprises only the precision 50n input impedance of the oscilloscope. Use of this technique will assure that the customer and Motorola are terminating devices into equivalent loads and will improve test correlation. To further standardize testing any unused outputs should be loaded with 50n to ground. Because the power supplies are shifted, the input levels must also be shifted by an equal amount. Table 2.3 gives the typical input levels for the ECLinPS family and their corre- 10Exxx Typical Shifted VIL -1.75V +0.25V V,H -0.90V +1.10V 100Exxx Typical Shifted VIL -1.70V +0.30V V,H -0.95V +1.05V Table 2.3 - Eel levels after Translating by +2.0V sponding +2.0V shifted levels. The test fixture should be in a controlled impedance 50n environment, with any non-50n interconnects, or stubs, kept as short as possible «1/4"). This controlled impedance environment will help to minimize overshoot and ringing, two Figure 2.9 - ECLinPS AC Test Board ECLinPS 2-11 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. APPLICATIONS INFORMATION Engineering Evaluation Board for 28 Pin ECl Devices in the PlCC Package Part # EClPSBD28 DESCRIPTION This board is designed to provide a low cost characterization tool for evaluating ECl devices in the ECLinPS Product Family. The board provides a high bandwidth 50 ohm controlled impedance environment. The board is universal and can be configured by the user for any of the 28 pin PLCC devices in the family depending on the input, output, and power pinout layout of the device. The table below indicates common input/output/power devices. Group CONF1 CONF2 CONF3 CONF4 CONF5 CONF6 CONF7 CONF8 CONF9 CONF10 CONF11 CONF12 CONF13 CONF14 CONF15 Base Device Pin Qompatible Devices E196 E142 E337 E212 E156 E158 E154 E101 E112 E431 E111 E164 E451 E163 E193 E195 E016,E141,E143,E241 E336 E1 04,E1 07,E150,E151 E155,E167,E171,E256 E116,E122,E175,E416 E452 E131,E157,E404 E457 E160 E166 Table 1: Cross Reference of Board Configuations The board is designed to test devices using the fly-by (Kelvin contact) test method, therefore one input force trace and one input sense trace exists for each input pin. This allows termination of the input and output signals into the highly accurate 50 ohm impedance of an oscilloscope. The layout is engineered to have equal length traces from the device under test (OUT) socket to the sense outputs which simplifies the calibration requirements for accurate AC measurements. The kit provides a printed circuit board with an attached surface mount socket as well as assembly instructions. For superior impedance control from the cable to the board, Motorola recommends the use of SMA coaxial connectors. ECLinPS 2-12 ® MOTOROLA <:> ~ 0~ 10 uF capacitors Locations 0' o o A. Location of sense ring for SMAs B. Vias to the power planes Figure 1. Front View of ECLinPS Evaluation Board ECLinPS 2-13 # Group Part(s) m '}' g ... .j>. ::J iJ CJ) eONFl eONF2 eONF3 eONF4 CONFS eONF6 eONF7 eONF8 eONF9 eONF10 eONFll eONF12 eONF13 eONF14 eONF1S E196 E142 E337 E212 E1S6 E1SB E1S4 El01 E112 E431 El11 E164 E4S1 E163 E193 PI P2 P3 VEE VEE VEE VEE VEE VEE VEE VEE VEE VEE VEE VEE VEE VEE VEE I I I I I INB INB I I INB I I I I I I I I I I I I I I I VB I Ne I I PS P6 P7 VB Ne I I I I I I I I I I I I I I I Ne I I Ne 0 I I I I I I I I Ne I I I I I 0 I Vee I I 0 I 0 I I I I P4 Vee I I I I Vee I 0 I I I I P8 P9 Pl0 Pll P12 P13 I I 0 I 0 I 0 0 0 Ne 0 0 0 0 Vee 0 0 I 0 Vee Vee 0 0 0 0 0 Vee 0 0 0 Vee Vee 0 I 0 I I 0 INB Vee I I I I Vee I 0 I 0 I L I Vee Vee Vee 0 I 0 I 0 I Vee I 0 Vee 0 0 0 Vee Vee Vee 0 I 0 I 0 0 Ne 0 0 0 0 0 0 0 0 I 0 0 0 - _. P14 0 Vee Vee 0 0 0 0 0 0 0 0 0 Vee Vee Vee PIS P16 P17 P18 P19 P20 P21 P22 P23 P24 Ne 0 I 0 0 0 0 0 0 0 0 0 0 0 0 Ne 0 0 0 I Vee Vee I I I I I I I I I I I 0 I 0 I I I I Vee Vee 0 Vee 0 Vee 0 Vee 0 0 0 0 0 0 Vee 0 0 Ne 0 Vee Vee Vee Vee Vee Vee 0 Vee Vee Vee Vee I 0 Ne 0 0 0 0 0 0 0 0 0 0 0 0 Vee Vee 0 0 O· I 0 Vee Vee Vee 0 - I I I 0 Vee 0 I I I 0 0 Vee I 0 Vee I 0 Vee 0 Vee 0 I I INB 0 Vee 0 I I I I I I I I I I I Vee -L--- - I I I I I I I I 0 I 0 I I I I P2S P26 P27 P28 I I I I I I I I 0 INB 0 I I I I I I I I I I I I Vee I VEE I I I I I I I I I I I I I I I I VB I I L-.----~L...__ Table 2. Pin Cross Reference KEY: "I" "0" "VEE" "Vcc" "NC': "VB" designates an input designates an output designates the lower voltage rail designates the upper voltage rail designates a no connect designates VBB output which should not be terminated into 50 ohms I I I I I I I I I I I I I I I __ of eonnectors 35 37 37 33 40 32 38 40 26 44 25 44 37 42 38 --- ASSEMBLING THE ECLinPS EVALUATION BOARD The evaluation board is designed for characterizing devices in a laboratory environment using high bandwidth sampling oscilloscopes such as the Hewlett Packard 54120T, the Tektronix 11800 Series, or the Tektronix 7854. The board is designed using Kelvin contact (fly·by) techniques to present the input signals to the DUT. Each pin on the board has two traces, one force and one sense. Inputs pins use one force and one sense line, while outputs need only a sense line. This means that input signals are terminated through the sense line into the 50 ohm input of asampling oscilloscope instead of atthe inputto the DUT. Please refer to the AC Testing section of the ECLinPS Data Book for further information and a simplified figure of the test setup. The first step in building a board is determining which input/output/power configuration is necessary for the de· vice of interest. Table 1 on the first page of the Applications Information shows all the board configurations. For example, if the devices of interest were the E1 04 and the E151, then CONF2 would be selected. Table 2 is a pin cross reference for each configuration. I. Installing the SMA Connectors Table 2 indicates the number of SMA connectors needed to populate an evaluation board for a given configura· tion. Depending on the device and the parameters of interest, it may not be necessary to install the full comple· ment of SMA connectors. For example, some devices have two clock inputs or common clocks and individual clocks. Figure 1 is the front view of the ECLinPS evaluation board. Item A points to the inner ring which connects to the sense traces of the DUT. The outer ring connects to the force traces. An input requires one SMA connec· tor for the force and one SMA connector for the sense, while an output only requires a connection to the sense trace. Insert all the SMA connectors into the board and solder to the board. A simple assembly technique is to place a stiff piece of cardboard (8" x 7" or larger) on top of all the connectors and hold the board and cardboard together. Invert the board, place on a level surface, and all the connectors will be seated properly and can be soldered in place. II. Connecting Power Planes to OUT Socket There are four voltage planes on the ECLPSBD28. One is dedicated to ground and the other three: B1, B2, B3 are uncommitted. These planes are accessible through a power connection and sets of four vias that are adjacent to each sense trace. This is identified as Item B in Figure 1. For standard parts, B1 can be assigned VCC, B2 can be assigned to VEE, and B3 can be assigned to ground. Table 2 indicates which pins need to be connected to the various supply Voltages. On the front side of the board, solder ajumperwire from the closest VEE orVcc via to the sense trace for each VCC, VCCO, and VEE pin. Nearthe DUTthere are sets of ground/ bias plane vias that accommodate power supply decoupling capacitors. These are identified as Item C. On the front side of the board install 1a ~F capacitors and on the back side install a 0.01 ~F high frequency capacitor in parallel to decouple the VEE and VCC planes. III. Cutting Force Traces for Outputs Because of the design of the board all force traces for output pins will appear as transmission line stubs con· nected to the output pin. On the back side of the board, cut the force traces associated with the outputs using a razor blade knife. It is important to cut the trace very close to the DUT area to minimize the stub length. Also cut the force traces that are connected to VCC, Vcco, and VEE pins. IV. Installing the Chip Capacitors for the VCCIVCCO Pins In the kit are 0.01 ~F chip capacitors for use in decoupling the VCC and VCCO pins to the ground plane. This is critical because the power pins are not directly connected to the VCC plane as in an actual board layout. On the back side of the board beneath the DUT socket are pads for each pin which allow connection of chip capaci· tors to the center island (GND) for each VCC and VCCO pin. Stand the chip capacitors on edge when soldering them in place so that adjacent pins are not shorted together. ECLinPS 2·15 V. Final Assembly The board power plane interface is designed to accommodate a 15 pin right angle D connector. This can be used directly, or wires can be inserted into the vias to connect to the power planes that were connected to the DUT in part II. Attach standoffs into the four 0.25 inch holes at the corners of the board. This completes the assembly of the evaluation board and it should be ready to test. VI. SMA Connector Suppliers Below are two suppliers who manufacture PC Mount SMA connectors which interface to the evaluation board. Motorola has used these two connectors before, but there are other vendors who manufacture similar products. EF Johnson 299 Johnson Ave. P.O. Box 1249 Waseca, Minnesota 56093 (800)-247-8343 or (507)-835-6222 0.200" PC Mount SMA Jack Receptacle 142-0701-201 MACOM Omni Spectra 140 Fourth Avenue Waltham, Massachusetts 02254 (617)-890-4750 0.200" PC Mount SMA Straight Jack 2062-0000-00 ECLinPS 2-16 EC!:PS Family Specifications & Device Data Sheets This section contains AC & DC specifications for each ECUnPS device type. Specifications common to all device types can be found in the first part of this section. While specifications unique to a particular device can be found in the individual data sheets following the family specifications. Data Sheet Classification Advance Information - product in the sampling or pre-production stage at the time of publication. Product Preview- product in the design stage at the time of publication. ECLinPS 3-1 ECl::PS Family Specifications Absolute Maximum Ratings Beyond which device life may be impaired. 1 Characteristic Power Supply (Vee Input Voltage (Vee Symbol Rating Unit VEE -8 to 0 Vdc = 0 V) = 0 V) o to VI Output Current,- Continuous -Surge lout Operating Temperature Range 10E Series 100E Series TA Operating Range 2 -6 V Vdc 50 100 o to o to mA °e +75 +85 -5.7 to -4.2 VEE V 1. Unless specified otherwise on Individual data sheet. 2. Parametric values specified at: 100E series: - 4.2 V to - 5.46 V 10E series: -4.94 V to -5.46 V 10E Series DC Characteristics VEE = -5.2 V ± 5%; Vee = Veeo = GN01 Q'C 75°C 25'C 85°C Min Max Min Max Min Max Min Max Unit VOH Output HIGH Voltage -1020 -840 -980 -810 -920 -735 -910 -720 mV VOL Output LOW Voltage -1950 -1630 -1950 -1630 -1950 -1600 -1950 -1595 mV VIH Input HIGH Voltage -1170 -840 -1130 -810 -1070 -735 -1060 -720 mV VIL Input LOW Voltage -1950 -1480 -1950 -1480 -1950 -1450 -1950 -1445 mV IlL Input LOW Current 0.5 Symbol Characteristic 0.5 0.3 0.3 J,tA 1. 10E senes CircUits are deSigned to meet the de specifications shown In the table, after thermal equIlibrium has been established. The CirCUit IS In a test socket or mounted on a printed circuit board and transverse air flow greater than 500 Ifpm is maintained. Outputs are terminated through a 50 0 resistor to - 2.0 Yolts, except bus outputs which, where specified, are terminated into 25 n. 100E Series DC Characteristics VEE = -4.2Vto Symbol -5.46 V; Vee = Veeo = GNO;TA = oOeto Characteristic Min Typ +85°e Max Unit VOH Output HIGH Voltage -1025 -955 -880 mV VOL Output LOW Voltage -1810 -1705 -1620 mV VOHA Output HIGH Voltage -1035 VOLA Output LOW Voltage mV -1610 mV Conditions VIN = VIH(max) OrVIL(min) VIN = VIH(min) or VIL(max) Loading with 50 to -2.0 V n VIH Input HIGH Voltage -1165 -880 mV Guaranteed HIGH Signal for All Inputs VIL Input LOW Voltage -1810 -1475 mV Guaranteed LOW Signal for All Inputs IlL Input LOW Current 0.5 p.A VIN .. = VIL(min) ThiS table replaces the three tables at different supply voltages In the prevIous edition and In Eel 100K literature. The same DC parametnc values at VEe = -4.5 V now apply across the full VEE range of -4.2 to -5.46 V. ECLinPS 3-2 MOTOROLA SEMICONDUCTOR _ _ _ _ _ _ _ _ _ _ __ TECHNICAL DATA • • • • • • • • MC10E016 MC100E016 700 MHz Min. Count Frequency 1000 ps ClK to Q, TC Internal TC Feedback (Gated) 8-Bit Fully Synchronous Counting and TC Generation Asynchronous Master Reset Extended 100E VEE Range of - 4.2 V to - 5.46 V 75 k!l Input Pulldown Resistors 8-BIT SYNCHRONOUS BINARY UP COUNTER The MC10E/100E016 is a high-speed synchronous, presettable, cascadable 8-bit binary counter. Architecture and operation are the same as the MC10H016 in the MECl 10KH family, extended to 8-bits, as shown in the logic symbol. The counter features internal feedback of TC, gated by the TClD (terminal count load) pin. When TClD is lOW (or left open, in which case it is pulled lOW by the internal pull-downs), the TC feedback is disabled, and counting proceeds continuously, with TC going lOW to indicate an all-one state. When TClD is HIGH, the TC feedback causes the counter to automatically re-Ioad upon TC = lOW, thus functioning as a programmable counter. PINOUT: 28-lEAD PLCC (TOP VIEW) PE FUNCTION TABLE CE P7 P6 P5 Vcca 'fC 21 20 19 CE PE TClD MR ClK L L H l H H H X X X X X MR 07 ClK Os TClD Vcc VEE 05 Z ZZ NC VCCO Po 04 P1 03 = = X L H X X X l L L l L H Z Z Z Z ZZ Z Function Load Parallel (P n to On) Continuous Count Count; Load Parallel on fC = lOW Hold Masters Respond, Slaves Hold Reset (On: = LOW, TC : = HIGH) clock pulse (low to high); clock pulse (high to low) PIN NAMES P2 P3 P4 Vcca 00 10 11 01 02 Pin Po-P7 00- 0 7 CE PE MR ClK TC TCLD ECLinPS 3-3 Function Parallel Data (Preset) Inputs Data Outputs Count Enable Control Input Parallel Load Enable Control Input Master Reset Clock Terminal Count Output TC-load Control Input 8·BIT BINARY COUNTER LOGIC DIAGRAM 00 01 02"""00 07 i'E TClD~ II r---OOM r---1 m I~ IT (") W ~ r S· iJ Po (J) MR ---l ") I I I I I BIT 1 BIT 0 W I I I I I I I I I r I I -~ ~~Go 0 102 03 §TI5 BIT 2-BIT 6 I I I 06 P7 s:: I C') ...... I ------l I 0 BIT7 m ...... I 0 I I I I I O'l ~ s:: C') ...... 0 0 m 'I 0 ...... O'l ClK '" TC Note that this diagram is provided for understanding of logic operation only. It should not be used for propagation delays as many gate functions are achieved internally without incurring a full gate delay. MC10E016, MC100E016 DC Characteristics: VEE = V EE(min) to VEE(max); Vee = Vceo = GND O°C Symbol Characteristic I'H Input HIGH Current lEE Power Supply Current r-- min 25°C max min 'COUNT tplH tpHl ts Ih typ max Unit 150 ~A Characteristic min typ Max. Count Frequency 700 900 181 181 151 181 151 181 151 181 174 208 25°C ClKtoO 600 725 MRloO ClK to TC (O's loaded) ClK 10 TC (O's unloaded) 600 550 550 775 775 700 MRloTC 625 775 max 85°C min typ max min typ 700 900 1000 600 725 1000 1050 600 550 550 775 775 700 625 775 150 -30 150 600 600 max 700 900 1000 600 725 1000 1000 1050 600 550 775 1000 1050 900 1000 550 625 Unit Condition MHz Propagalion Delay 10 Oulput ps 900 1000 775 700 775 1 1 900 1000 Setup Time ps Pn CE 150 -30 600 400 600 400 PE TClD 600 500 400 600 500 400 300 Hold Time Pn 250 30 -400 250 250 0 30 -400 0 30 -400 100 -400 -300 0 100 -400 -300 0 100 -400 -300 700 900 700 900 700 300 500 -30 400 400 300 ps 0 0 IRA Reset Recovery Time 900 tpw Minimum Pulse Widlh ClK, MR 400 ps ps 400 400 Rise/Fall Times 20 -80% Condition VEE = VEE(min) to VEE(max); Vee = Veeo= GND TClD If min mA 151 151 CE PE I, 85°C max 150 O°C Symbol typ 150 10E 100E ACCharacteristics: typ ps 300 510 800 300 510 800 300 510 800 1. ClK to Tc propagation delay is dependent on the loading of the Q outputs. With all of the Q outputs loaded the noise generated in going from a 11111111 state to a 0000 0000 state causes the ClK to Tc+ delay to increase ECLinPS 3-5 ,MC1 OE016, MC100E016 FUNCTION TABLE FUNCTION PE CE MR TCLD CLK P7-P4 L H H H H L H H H H H H H H X X L L L L X H H L L L L L L X L L' L L L L L L L L L L L L H X L L L L X X X H H H H H H X Z Z Z Z Z Z Z Z Z Z Z Z Z Z H X X X X H X X H H H H H H X Load Count Load Hold Load On Terminal Count Reset X P3 P2 P1 PO H X X X X H X X L L L L L L X H L X X X X L X X L L L L L L X L X X X X L X X H H H H H H X X X X X H X X H H H H H H X Q7-Q4 Q3 Q2 Q1 QO H H H H L H H H H H H H H H L H H H H L H H H H H H L L H L H H H H L H H H H H H H H L L L L H H L L L L L H H H H L L L H L H L L L L H L H L H L L TC H H H L H H H H H H L H H H H Applications Information Cascading Multiple E016 Devices For applications which call for larger than 8-bit counters multiple E016's can be tied together to achieve very wide bit width counters. The active low terminal count (TC) output and count enable input (CE) greatly facilitate the cascading of E016 devices, Two E016's can be cascaded withoutthe need for external gating, however for counters wider than 16 bits external OR gates are necessary for cascade implementations. Figure 1 below pictorally illustrates the cascading of 4 E016's to build a 32-bit high frequency counter. Note the E1 01 gates used to OR the terminal count outputs olthe lower order E016's to control the counting operation of the higher order bits. When the terminal count of the preceeding device (or devices) goes low (the counter reaches an all 1's state) the more significant E016 is set in its count mode and will count one binary digit upon the next positive clock transistion. In addition, the preceeding devices will also count one bit thus sending their terminal count outputs back to a high state dis- abling the count operation of the more significant counters and placing them back into hold modes, Therefore, for an E016 in the chain to count all olthe lower order terminal count outputs must be in the low state. The bit width of the counter can be increased or decreased by simply adding or subtracting E016 devices from Figure 1 and maintaini'ng the logic pattern illustrated in the same figure. The maximum frequency of operation for the cascaded counter chain is set by the propagation delay of the TC output and the necessary setup time of the CE input and the propagation delay through the OR ga.m.controlling it (for 16-bit counters the limitation is only the TC propagation delay and the CE setup time). Figure 1 shows E101 gates used to control the count enable inputs, however if the frequency of operation is lower a slower ECL OR gate can be used. Using the worst case guarantees for these parameters from the ECLinPS data book the maximum count frequency for a greater than 16-bit counter is 475MHz and that for a 16-bit Loadcr----__________. -______--------,----------------------,--------------------, QO->Q7 "LO" CE QO·>Q7 QO-> Q7 CE PE E016 PE E016 E016 LSB QO->Q7 CE PE E016 MSB PO -> P7 Clocko-_t--------------.....--------------------_------------------~ Figure 1 - 32-Bit Cascaded E016 Counter ECLinPS 3-6 MC10E016, MC100E016 Applications Information counter is 625MHz. Note that this assumes the trace delay between the TC outputs and the CE inputs are negligible. If this is not the case estimates of these delays need to be added to the calculations. Divide Ratio Programmable Divider The E016 has been designed with a control pin which makes it ideal for use as an 8-bit programmable divider. The TCLD pin (load on terminal count) when asserted reloads the data present at the parallel input pin (Pn's) upon reaching terminal count (an all 1's state on the outputs). Because this feedback is built internal to the chip the programmable division operation will run at very nearly the same frequency as the maximum counting frequency of the device. Figure 2 below illustrates the input conditions necessary for utilizing the E016 as a programmable divider set up to divide by 113. Preset Data Inputs P7 P6 P5 P4 P3 P2 P1 PO 2 3 4 5 H H H H H H H H H H H H H H H H H H H H H H H L H L L H L H L H 112 113 114 H H H L L L L L L H L L L H H L H H L H H L H L 254 255 256 L L L L L L L L L L L L L L L L L L L H L . . . . ·· · · L ·· . · . · L H HLLLHHHH Table 1 - Preset Values for Various Divide Ratios P7 P6 P5 P4 P3 P2 P1 H PE L CE H TCLD PO full clock period. For even divide ratios, twice the desired divide ratio can be loaded into the E016 and the TC output can feed the clock input of a toggle flip flop to create a signal divided as desired with a 50% duty cycle. A single E016 can be used to divide by any ratio from 2 to 256 inclusive. If divide ratios of greater than 256 are needed multiple E016's can be cascaded in a manner similar to that already discussed. When E016's are cascaded to build larger dividers the TCLD pin will no longer provide a means for loading on terminal count. Because one does not want to reload the counters untill all of the devices in the chain have reached terminal count, external gating of the TC pins must be used for multiple E016 divider chains. Figure 4 on the following page shows a typical block diagram of a 32-bit divider chain. Once again to maiximizethe frequency of operation E1 01 OR gates were used. For lower frequency applications a slower OR gate could replace the E..! 0 1. Note that for a 16-bit divider the OR function feeding the PE (program enable) input CANNOT be replaced by a wire OR tie as the Tc output of the least significant E016 must also feed the CE input of the most significant E016. If the two TC outputs were OR tied the cascaded count operation would not operate properly. Because in the cascaded form the PE TC Figure 2 - Mod 2 to 256 Programmable Divider To determine what value to load into the device to accomplish the desired division the designer simply subtracts the binary equivalent ofthe desired divide ratio from the binary value for 256. As an example for a divide ratio of 113: Pn's = 256 - 113 = 8F,s = 1000 1111 where: PO = LSB and P7 = MSB Forcing this input condition as per the setup in Eigure 2 will result in the waveforms of Figure 3. Note that the TC output is used as the divide output and the pulse duration is equal to a Load '00' 0000 , 00' 000' "" 1100 11111101 11111110 Clock +--, P~E__________ TC Divide by 113 Divide by 113 E016 Programmable Divider Waveforms ECLinPS 3-7 11111111 Load MC10.E016, MC100E016 Applications Information "LD" E016 LSB PO -> P7 Clockir-_ir-------4_-----------il-------------l 32-81t Cascaded E016 Programmable Divider feedback is external and requires external gating the maximum frequency of operation will be significantly less than the same operation in a single device. Maximizing E016 Count Frequency The E016 device produces 9 fast transitioning single ended outputs, thus Vee noise can become significant in situations where all of the outputs switch simulataneously in the same direction. This Vee noise can negatively impact the maximum frequency of operation of the device. Since the device does not need to have the Q outputs terminated to count properly, it is recommended that if the outputs are not going to be used in the rest of the system they should be left unterminated. In addition if only a subset of the Q outputs are used in the system only those outputs should be terminated. Not terminating theunused outputs will not only cut down the Vee noise generated but will also save in total system power dissipation. Following these guidelines will allow designers to either be more aggressive in their designs or provide them with an extra margin to the published data book specifications. ECLinPS 3-8 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E101 MC100E101 • 500 ps Max. Propagation Delay • Extended 100E VEE Range of -4.2 V to -5.46 V • 75 kG Input Pulldown Resistors The MC10E/l00El0l is a quad 4-input ORINOR gate. QUAD 4-INPUT OR/NORGATE PINOUT: 28-lEAD PlCC (TOP VIEW) Caa D3b 0Jc D3d Veeo ii3 Q3 lOGIC SYMBOL DOa DOb Doc Dod Dla Dlb Ole Old 10 11 D2a D2b D2e D2d D3a PIN NAMES DOa-D3d 00-03 00-0 3 D3b Function Pin D3e Data Inputs True Outputs Inverting Outputs D3d ECLinPS 3-9 ~ 00 ao ~~ Ql ~~ 02 ~~ Q3 MC1 OE1 01, MC1 00E1 01 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo= GND 25°C O°C Symbol - Characteristic IIH Input HIGH Current lEE Power Supply Current 10E 100E ACCharacteristics: min typ max min 150 tpLH tpHL tSKEW tSKEW t, tf 85°C max min typ 150 max Unit 150 IlA Condition mA 30 30 36 36 30 30 36 36 30 35 36 42 VEE = VEE(min) to VEE(max); Vee = Veeo= GND 25°C O°C Symbo typ 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output DtoO 200 350 500 200 350 500 200 350 500 Condition ps Within-Device Skew Within-Gate Skew Rise I Fall Time 20-80% Unit 50 25 50 25 50 25 ps ps 300 380 575 300 380 575 300 380 575 1. Within-device skew is defined as identical transitions on similar paths through a device 2. Within-gate skew is defined as the variation in propagation delays of a gate when driven from its different inputs. ECLinPS 3-10 1 2 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • MC10E104 MC100E104 600 ps Max. Propagation Delay OR/NOR Function Outputs Extended 100E VEE Range of -4.2 V to -5.46 V 75 kn Input Pulldown Resistors The MC10E/l00El04 is a quint 2·input AND/NAND gate. The function output F is the OR of all five AND gate outputs, while F is the NOR. The Q outputs need not be terminated if only the F outputs are to be used. QUINT 2-INPUT AND/NAND GATE PINOUT: 28-LEAD PLCC (TOP VIEW) D3a D4b D4a Ne Veeo F LOGIC SYMBOL 18 D3b 04 D2a ~ D2b Vee VEE Q3 D1a 03 D1b 02 Doa DOa DOb D1a 01 D1b a, 02 DOb Veeo 00 GO 01 a, Veeo D2a 02 D2b 02 PIN NAMES Function Pin DOa-D4b Qo-C4 Qo-C4 F F Data Inputs AND Outputs NAND Outputs OR Output NOR Output FUNCTION OUTPUTS F = (DO a ' DOb) (D3a • D3b) + (Dla' Dlb) + (D2a' D2b) + + (D4a' D4b) ECLinPS 3-11 D3a 03 OJb Q3 D4a ~ D4b 04 MC10E104, MC100E104 DC Characteristics: VEE = VEE(min) to VEE(max); Vcc = V ceo = GND O°C Symbol Characteristic IIH Input HIGH Current lEE Power Supply Current min typ min 200 tpLH tSKEW If min typ max Unit 200 j.tA 46 38 46 38 46 100E 38 46 38 46 44 53 VEE = VEE(min) to VEE(max); Vee = Vcca= GND 25°C 85°C min typ max min typ max min typ max DloO 225 385 600 225 385 600 225 385 600 DtoF 500 725 1000 500 725 1000 500 725 1000 Characteristic Condition mA 38 Propagalion Delay 10 Oulpul Unit Condition ps Within-Device Skew ps DtoO I, max 200 O°C IpHL typ tOE ACCharacteristics: Symbol 85°C 25°C max 75 75 75 1 Rise I Fall Times ps 20-80% a 275 425 700 275 425 700 275 425 700 F 300 475 700 300 475 700 300 475 700 1" Within-device skew is defined as identical transitions on similar paths through a device ECLinPS 3-12 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E107 MC100E107 • 600 ps Max. Propagation Delay • ORINOR Function Outputs • Extended 100E VEE Range of -4.2 V to -5.46 V • 75 kO Input Pulldown Resistors The MC10EI100El07 is a quint 2-input XOR/XNOR gate. The function output F is the OR of all five XOR outputs, while F is the NOR. The Q outputs need not be terminated if only the F outputs are to be used. QUINT 2-INPUT XOR/XNOR GATE PINOUT: 28-LEAD PLCC (TOP VIEW) LOGIC SYMBOL ~-6Hi+--- 00 ~J-++++----ao DOb Veea 00 ao Q, 111 Veea .........---'iH---- Q, /o--ttil---- 111 .........-'H----Q2 PIN NAMES Pin Daa-D4b 00-0 4 00-0 4 F F ~)---++----02 Function Data Inputs XOR Outputs XNOR Outputs OR Output NOR Output ~-""'---Q3 ......-'I~-+---03 FUNCTION OUTPUTS F = (DO a EB Dab) + (Dla (D3a EB D3b) + (D4a ...........--4----<4 EB Dlb) + EB D4b) (D2a EB D2b) + ~J__---Cl4 ECLinPS 3-13 MC10E107, MC100E107 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo= GND 25°C O°C Symbol Characteristic IIH Input HIGH Current lEE Power Supply Current 10E 100E r-- AC Characteristics: min typ max min 200 tpLH tpHL tSKEW t, tf 85°C max min typ 200 max Unit 200 IlA Condition mA 42 42 50 50 42 42 42 48 50 50 50 58 VEE = VEE(min) to VEE(max); Vee = Veeo= GND 25°C O°C Symbol typ 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output DtoQ DtoF 250 500 410 600 1000 250 500 410 600 100 250 500 410 600 1000 Condition ps Within-Device Skew DtoQ Rise I Fall Times 20 -80% Q F Unit 725 725 725 ps 75 75 75 1 ps 275 300 450 475 700 700 275 300 450 475 700 700 1. Within-device skew is defined as identical transitions on similar paths through a device ECLinPS 3-14 275 300 450 475 700 700 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • • low Skew Guaranteed Skew Spec Differential Design Vas Output Enable Extended 100E VEE Range of - 4.2 V to - 5.46 V 75 kn Input Pulldown Resistors MC10E111 MC100E111 The MC10EI100E111 is a low skew 1-to-9 differential driver, designed with clock distribution in mind. It accepts one signal input, which can be either differential or else single-ended if the Vas output is used. The signal is fanned out to 9 identical differential outputs. An enable input is also provided. A HIGH disables the device by forcing all Q outputs lOW and all IT outputs HIGH. The device is specifically designed, modeled and produced with low skew as the key goal. Optimal design and layout serve to minimize gate to gate skew withindevice, and empirical modeling is used to determine process control limits that ensure consistent tpd distributions from lot to lot. The net result is a dependable, guaranteed low skew device. To ensure that the tight skew specification is met it is necessary that both sides of the differential output are terminated into 50 n, even if only one side is being used. In most applications, all nine differential pairs will be used and therefore terminated. In the case where fewer than nine pairs are used, it is necessary to terminate at least the output pairs on the same package side (i.e. sharing the same VCCO) as the pair(s) being used on that side, in order to maintain minimum skew. Failure to do this will result in small degradations of propagation delay (on the order of 10-20 psI of the output(s) being used which, while not being catastrophic to most designs, will mean a loss of skew margin. The Vas output is intended for use as a reference voltage for single-ended reception of ECl signals to that device only. When using for this purpose, it is recommended that Vas is decoupled to VCC via a 0.01 ILF capacitor. 00 Qo Ql Veeo a, Q2 LOGIC SYMBOL VEE Q3 EN 03 Q4 PINOUT: 28-LEAD PLCC (TOP VIEW) Vee Veco iN 04 Vaa Q5 Ne 05 08 Os 07 Veeo Q7 Os IN iN EN Q6 PIN NAMES Pin IN, IN EN QO, QO-QS' QS VSS Qo 00 Ql 02 IN 1:9 DIFFERENTIAL CLOCK DRIVER Function Differential Input Pair Enable Differential Outputs VSS Output Vaa - - - - ECLinPS 3-15 a, MC10E111, MC100E111 DC CHARACTERISTICS: VEE = VEE (min) to VEE (max); Vcc TA Symbol Vss Characteristic Min Output Reference Voltage 10E 100E IIH Input HIGH Current lEE Power Supply Current 10E 100E AC CHARACTERISTICS: VEE Typ VCCO TA Max Min = GND = 25'C Typ -1.38 -1.38 -1.27 -1.35 -1.26 -1.38 Min = 85°C Typ -1.25 -1.31 -1.26 -1.38 150 = TA Max Max Unit Conditions V -1.19 -1.26 150 150 pA mA 48 48 Characteristic Min = VCCO = DOC Typ TA Max 60 60 48 48 60 60 VEE (min) to VEE (max); VCC TA Symbol = = D'C Min = 48 55 60 69 GND = 25°C Typ TA Max Min tpLH tpHL Propagation Delay to Output IN (differential) IN (single-ended) Enable Disable tskew Within-Device Skew ts Setup Time EN to IN 200 0 200 0 200 tH Hold Time IN to EN 0 -200 0 -200 tR Release Time EN to IN 300 100 300 100 Vpp Minimum Input Swing 250 VCMR Common Mode Range -1.6 tr tf Rise/Fall Times 20-80% 275 = 85'C Typ Max Unit Note ps 430 330 450 450 630 730 850 850 25 430 330 450 450 50 630 730 850 850 25 50 375 -1.6 600 275 375 1 2 3 3 630 730 850 850 25 ps 4 0 ps 5 0 -200 ps 6 300 100 ps 7 50 250 250 -0.4 430 330 450 450 -0.4 -1.6 600 275 375 mV 8 -0.4 V 9 600 ps Notes: 1. The differential propagation delay is defined as the delay from the crossing points of the differential input signals to the crossing point of the differential output signals. (See Definitions and Testing of ECLinPS AC Parameters in this publication.) 2. The single-ended propagation delay is defined as the delay from the 50% pOint of the input signal to the 50% point of the output signal. _ (See Definitions and Testing of ECLinPS AC parameters in this publication.) 3. Enable is defined as the pro.l!agation delay from the 50% point of a negative transition on EN to the 50% point of a positive transition on Q (or a negative transition on Q). _ Disable is defined as the pr,Q.Pagation delay from the 50% point of a positive transition on EN to the 50% point of a negative transition on Q (or a positive transition on Q). 4. The within-device skew is defined as the worst case difference between any two similar delay paths within a single device. 5. The setup tim!..ls the minimum time that EN must be asserted prior to the next transition of IN/iN to prevent an output response greater than ± 75 mV to that INIIN transition (see Figure 1),_ 6. The hold time is the minimum time that EN must remain asserted after a negative going IN or a positive going IN to prevent an output response _ greater than :!:: 75 mV to that IN/iN' trnasitio~ee Figure 2). 7. The release time is the minimum time that EN must be deasserted prior to the next IN/IN transition to ensure an output response that meets the specified IN to Q propagation delay and output transition times (see Figure 3). 8. Vpp(min) is defined as the minimum input differential voltage which will cause no increase in the propagation delay. The Vpp(min) is AC limited for the E111 as a differential input as low as 50 mV will still produce full ECl levels at the output. 9. VeMR is defined as the range within which the VIH level may vary, with the device still meeting the propagation delay specification. The VIL level must be such that the peak to peak voltage is less than 1.0 V and greater than or equal to Vpp(min). ECLinPS 3-16 MC10E111, MC100E111 Setup Time Hold Time ~~ IN "1 ~ ~75mV1 o----v--l EN Q-f"\..-Tf IN "~ iN iN th EN 50% ~75 mV o--v--l 1 Q---A...-TT -~ ~ ~ EN~ " 50% a Q ><= ~75mV ~75mV Figure 1 Release Time Figure 2 ECLinPS 3-17 Figure 3 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • MC10E112 MC100E112 600 ps Max. Propagation Delay Common Enable Input Extended 100E VEE Range of -4.2 V to -5.46 V 75 k!l Input Pulldown Resistors The MC10E/l00El12 is a quad driver with two pairs of ORINOR outputs from each gate. and a common. buffered enable input. Using the data inputs the device can serve as an ECL memory address fan-out driver. Using just the enable input. the device serves as a clock driver. although the MC10E/l00Elll is designed specifically for this purpose. and offers lower skew than the El12. For memory address driver applications where scan capabilities are required. please refer to the E212 device. QUAD DRIVER LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) Qi 03a Q3b Q3a 22 Veeo Q2b 21 20 l'i2a: / " - - - - Q Oa 19 Veea Q2b OJ Q2a 02 Vee VEE Qlb 01 Qla DO Qlb DO -----4 " ' - - - - Oob ....01----- 00. '----Oob /"----Ql. EN °1---+--4 ....<:>-----Qla '-----Qlb Qla Ne Veea 00. QOb QOa 10 11 QOb Veea ~----Qlb /"----Q2a ~----Q2b A)-----l'i2a: '-----Q2b /"----Q3a PIN NAMES ~-----Q3a Function Pin 00-0 3 EN ana. onb ana. onb ..A:ll----- Q3b Data Inputs Enable Input True Outputs Inverting Outputs ~----Q3b ECLinPS 3-18 MC10E112, MC100E112 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo= GND O°C Symbol I'H Characteristic min typ Input HIGH Current D EN lEE tpLH tpHL tSKEW tr If max min typ 200 150 max Unit Condition 200 150 rnA 47 47 47 47 56 56 47 54 56 56 56 65 VEE = VEE(min) to VEE(max); Vec = Vcco= GND 25°C O°C Symbol typ IJA Power Supply Current 10E 100E ACCharacteristics: min 200 150 - - 85°C 25°C max 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output D EN 200 275 400 450 600 675 200 275 400 450 600 675 200 275 400 450 600 675 - Condition ps Within-Device Skew Dn toQn, Qn Qnato Qnb Rise / Fall Times 20- 80% Unit ps 80 40 80 40 80 40 1 2 ps 275 425 700 275 425 700 1. Within-device skew is defined as identical transitions on similar paths through a device 2. Skew defined between common OR or common NOR outputs of a single gate. ECLinPS 3-19 275 425 700 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TeCHNICAL DATA • • • • • 500 ps Max. Propagation Delay VSS Supply Output Dedicated VCCO Pin for Each Receiver Extended 100E VEE Range of -4.2 V to -5.46 V 75 kG Input Pulldown Resistors MC10E116 MC100E116 The MC10E/100E116 is a quint differential line receiver with emitter-follower outputs. An internally generated reference supply (VSS) is available for single-ended reception. Active current sources plus a deep collector feature of the MOSAIC III process provide the receivers with excellent common-mode noise rejection. Each receiver has a dedicated VCCO supply lead, providing optimum symmetry and stability. The receiver design features clamp circuitry to cause a defined state if both the inverting and non-inverting inputs are left open; in this case the Q output goes lOW, while the IT output goes HIGH. This feature makes the device ideal for twisted pair applications. If both inverting and non·inverting inputs are at an equal potential of >-2.5 V, the receiver does not go to a defined state, but rather current-shares in normal differential amplifier fashion, producing output voltage levels midway between HIGH and lOW, or the device may even oscillate. The device VSS output is intended for use as a reference voltage for singleended reception of ECl signals to that device only. When using for this purpose, it is recommended that VSS is decoupled to VCC via a 0.01 JLF capacitor. Please refer to the interface section of the design guide for information on using the E116 in specialized applications. The E116 features input pull-down resistors, as does the rest of the ECLinPS family. For application which require bandwidths greater than that of the E116, the E416 device may be of interest. QUINT DIFFERENTIAL LINE RECEIVER LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) OJ D4 D4 Veeo 04 <4 Veeo ii3 03 D2 03 D2 Vee VEE Q2 Ves 02 Do Veeo Do a;D, 0, Veeo 00 Do Veeo 0, PIN NAMES Pin DO, 00-04, 04 00, 00-04, 04 VBB Function Differential Input Pairs Differential Output Pairs Reference Voltage Output VBB 01----11 ECLinPS 3-20 MC10E116, MC100E116 DC Characteristics: VEE = V EE(min) to VEE(max); V cc = Vcco = GND DoC Symbol VBB Characteristic min typ 85°C max min typ max -1.27 -1.35 -1.25 -1.31 -1.19 100E -1.38 -1.26 -1.38 -1.26 -1.38 -1.26 Power Supply Current Unit Condition V -1.38 lEE VCMR 25°C max 10E Input HIGH Current Vpp(DC) typ Output Reference Voltage I'H - min 200 200 200 IJA mA 10E 29 35 29 35 29 35 100E 29 35 29 35 33 40 Input Sensitivity 150 Common Mode Range -2.0 150 -0.6 150 -2.0 -0.6 -2.0 -0.6 mV 1 V 2 1. VPP is the minimum differential input voltage required to assure full ECl levels are present at the outputs. 2. VCMR is referenced to the most positive side of the differential input signal. Normal operation is obtained when the "HIGH" input is within the VCMR range and the input swing is greater than VPPMIN and < 1V. AC Characteristics: VEE = V EE(min) to V EE(max); V cc = V cco = GND DOC Symbol tpLH tpHL Vpp(AC) tSKEW Characteristic max min typ max min typ max D 200 300 450 200 300 450 200 300 450 D(SE) 150 300 500 150 300 500 150 300 500 Propagation Delay to Output Minimum Input Swing t, Unit Condition ps 150 150 mV 150 Within-Device Skew 1 ps 50 50 50 ±10 ±10 ±10 2 Duty Cycle Skew tplH - tpHl tr 85°C typ Dn to an, an tSKEW 25°C min ps Rise / Fall Times 20- 80% ps 275 375 575 275 375 575 275 1 . M'nlmum ,nput sWing for which AC parameters are guaranteed. 2. Within-device skew is defined as identical transitions on similar paths through a device 3. Duty cycle skew defined only for differential operation when the delays are measured from the cross point of the inputs to the cross points of the outputs ECLinPS 3-21 375 575 3 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E122 MC100E122 • 500 ps Max. Propagation Delay • Extended 100E VEE Range of -4.2 V to -5.46 V • 75 kO Input Pulldown Resistors DESCRIPTION The MC10E/100E122 is a 9-bit buffer. The device contains nine non-inverting buffer gates. 9-BIT BUFFER PIN NAMES Pin DO-OS Qo-QS Function Data Inputs Data Outputs ECLinPS 3-22 MC10E122, MC100E122 DC Characteristics: VEE = VEE(min) to VEE(max); Vcc = Vcco= GND O°C Symbol Characteristic I'H Input HIGH Current lEE Power Supply Current 10E 100E I---- AC Characteristics: min typ min tpLH tpHL tSKEW max min typ 200 max Unit 200 ~ Condition mA 41 41 49 49 41 41 49 49 41 47 49 57 VEE = VEE(min).to VEE(max); Vcc = Vcco = GND 25°C 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output DtoO 150 350 500 150 350 500 150 350 500 Within-Device Skew Rise / Fall Times 20 -80% Unit Condition ps ps DtoO tr tf typ 200 O°C Symbol 85°C 25°C max 75 75 75 1 ps 300 425 800 300 .. 425 800 1. Within-device skew IS defined as Identical transitions on similar paths through a deVice ECLinPS 3-23 300 425 800 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • • 1100 MHz Min. Toggle Frequency Differential Outputs Individual and Common Clocks Individual Resets (asynchronous) Paired Sets (asynchronous) Extended 100E VEE Range of -4.2 V to -5.46 V 75 kll Input Pulldown Resistors MC10E131 MC100E131 4-BIT The MC10E/100E131 is a quad master-slave D-type flip-flop with differential outputs. Each flip-flop may be clocked separately by holding Common Clock (Cc) lOW and using the Clock Enable (CE) inputs for clocking. Common clocking is achieved by holding the CE inputs lOW and using Cc to clock all four flip-flops. In this case, the CE inputs perform the function of controlling the common clock, to each flipflop. Individual asynchronous resets are provided (R). Asynchronous set controls (5) are ganged together in pairs, with the pairing chosen to reflect physical chip symmetry. Data enters the master when both Cc and CE are lOW, and transfers to the slave when either Cc or CE (or both) go HIGH. lOGIC SYMBOL PINOUT: 28-lEAD PlCC (TOP VIEW) D2 R3 CE2 R2 VCCO 03 o FLlP·FlOP 03 S CE3 18 02 eE3- ~ I> D3 02 S12 VCC VEE a, D2 D 01 eE2 >--L-- > S03 00 R2 DO 00 CEO RO Dl CEl Rl 10 11 NC VCCO 0- D D3 R 00-- R3 S S03 S12 Cc - R 0- 00-- Rl PIN NAMES Pin Do-D3 CEo-CE3 Ro-R3 Cc S03,S12 00-0 3 00-0 3 eEl H-> Dl D - Ofr- 01 01--- 01 S Function Data Inputs Clock Enables (Individual) Resets Common Clock Sets (paired) True Outputs Inverting Outputs R RO eEODO ECLinPS 3·24 ~I> - IR D S Ofr01--- 00 MC10E131, MC100E131 DC Characteristics: VEE = V EE(min) to VEE(max); V cc = V cco = GND O°C Symbol IIH - lEE Characteristic min typ Inpul HIGH Currenl Cc min 300 300 CE 300 300 D 150 150 300 150 Condition mA 10E 58 70 58 70 58 70 100E 58 70 58 70 67 81 VEE = VEE(min) to VEE(max); Vcc = Vcco = GND IpLH Propagalion Delay 10 Output typ 25°C max min typ 85°C max 1100 1400 min typ max 1100 1400 360 CC 325 500 675 325 R 350 550 725 350 S 350 550 725 350 550 150 20 150 500 700 360 500 700 360 500 175 -20 400 150 700 500 675 325 500 675 550 725 350 550 725 725 350 550 725 20 150 20 175 -20 150 -20 400 150 400 150 ps 2 Hold Time D IAA Reset Recovery Time tpw Minimum Pulse Widlh Clk, S, R Condition MHz SelupTime D Unit ps CE ps 2 ps ps 400 Within-Device Skew 400 400 60 60 60 ps ps Rise / Fall Times 20 -80% Unit 450 Power Supply Current 1100 1400 If 350 450 Max. Toggle Frequency Ir max 300 f MAX I SKEW typ R O°C Ih min S min Is max fLA 350 Characteristic IpHL typ 350 450 ACCharacteristics: Symbol 85°C 25°C max 300 .. 480 675 300 480 675 1. Within-device skew IS defined as Identical transitions on similar paths through a device 2. Setup/hold times guaranteed for both Cc & CE ECLinPS 3-25 300 480 675 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E136 MC100E136 Product Preview 700MHz Min. Count Frequency Look-Ahead-Carry Input and Output Fully Synchronous Up and Down Counting Asynchronous Master Reset Internal 75kQ Input Pulldown Resistors Extended 1OOE VEE Range of -4.2V to -5.46V 6-81T UNIVERSAL COUNTER The MC 10E136/100E136 is an 6-bit synchronous, presettable, cascadable universal counter. The device offers a look-ahead-carry input and generates a look-ahead-carry output. These two features allow for the cascading of multiple E136's to wider bit widths that operate at nearly the same frequency as a stand alone counter. The CLOUT output will pulse low for one clock cycle one count before the counter reaches terminal count. The COUT pin will pulse low for one clock cycle when the counter reaches terminal count. The differential COUT output facilitates its use in programmable divider and self-stopping counter applications. PINOUT: 28-LEAD PLCC (TOP VIEW) 03 04 05 25 24 23 veea 05 22 21 Q4 veea 20 19 PIN NAMES 02 03 52 02 51 vee VEE veea elK eaUT elN eaUT eLlN CLOUT MR 01 DO veea 00 10 11 01 veea PIN FUNCTION DO-D5 00-05 S1, S2 MR CLK COUT,COUT CLOUT CIN CUN Preset Data Inputs Differential Data Outputs Mode Control Pins Master Reset Clock Input Carry Out Output (Active LOW) Look-Ahead-Carry Output Carry In Input (Active LOW) Loak-Ahead-Carry Input FUNCTION TABLE Sl S2 CIN MR ClK L L L H H H L H H L L H X Z Z Z Z Z Z X X L L L L L L H L H L H X X X Function Preset Parallel Data Inputs Increment ( Count Up ) Hold Count Decrement ( Count Down) Hold Count Hold Count Reset (On = LOW; COUT = HIGH) Z = low to High Transition This document contains information on a product under development. Motorola reserves the right to change or discontinue this product without notice. ECLinPS 3-26 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA ~- . • MC10E137 ~ Product Preview 1.2GHz Min_ Count Frequency Synchronous and Asynchronous Enable Pins Differential Clock Input and Data Output Pins Asynchronous Master Reset Internal 75k.Q Input Pulldown Resistors Extended IOOE VEE Range of ~4_2V to ~5_46V ... -.----------~ MC100E137 The MCI OE/I OOEI37 is a very high speed binary ripple counteL The two least significant bits were designed with very fast edge rates while the more significant bits maintain standard ECLinPS output edge rates. This allows the counter to operate at very high frequencies while maintaining a moderate power dissipation level. Soth synchronous and asynchronous enables are available to maximize the devices flexibility for various applications. The device is ideally suited for multiple frequency clock generation as well as a counter in a high performance ATE time measurement board. The asynchronous Master Reset resets the counter to an all zero state upon assertion. a-BIT RIPPLE COUNTER PIN NAMES PIN FUNCTION - CLK,C~ QO~Q7, QO~Q7 EN ESI, ES2 MR VSS a1 a1 Differential Clock Inputs Differential Q Outputs Asynchronous Enable Input Synchronous Enable Inputs Asynchronous Master Reset Switching Reference Output a6 a6 a7 a7 This document contains information on a product under development. Motorola reserves the right to change or discontinue this product without notice. ECLinPS 3-27 MOTOROLA SEMICONDUCTOR - - - - -_ _ _ _ _ __ TECHNICAL DATA • • • • • • • MC10E141 MC100E141 700 MHz Min. Shift Frequency 8-Bit Full-Function, Bi-Directional Asynchronous Master Reset Pin-Compatible with E241 Extended 100E VEE Range of -4.2 V to -5.46 V 75 kll Input Pulldown Resistors 8-BIT SHIFT REGISTER The MC10E/100E141 is an 8-bit full-function shift register. The E141 performs serial/parallel in and serial/parallel out, shifting in either direction. The eight inputs DO-D7 accept parallel input data, while DUDR accept serial input data for left/right shifting. The select pins, SELO and SEL 1, select one of four modes of operation: Load, Hold, Shift Left, Shift Right, according to the Function Table. Input data is accepted a set-up time before the positive clock edge. A HIGH on the Master Reset (MR) pin asynchronously resets all the registers to zero. PINOUT: 28-LEAD PLCC (TOP VIEW) SELO DL 25 24 I>] 06 05 Veeo Q7 FUNCTION TABLE SEll Os MR Vee SELl l l H H l H l H Function load Shift Right (On to On + 1) Shift left (On to On -1) Hold PIN NAMES Ne Pin Veeo DR SELO Function 00-07 DO DL, DR SElO, SEll ClK 01 00-0 7 MR LOGIC SYMBOL - , Parallel Data Inputs Serial Data Inputs Mode Select Inputs Clock Data Outputs Master Reset OL DR SEL1----~--~~--~--------~--+-~~----------~ SELO --------~~--~------------~~~------------~ eLK----------~_+--------------~~r---------------~ MR------------~----------------~~----------------~ ECLinPS 3-28 MC10E141, MC100E141 DC Characteristics: VEE = VEE(min) to VEE(max); Vcc = Vcco= GNO O°C Symbol - Characteristic min IIH Input HIGH Current lEE Power Supply Current 10E 100E AC Characteristics: typ 25°C max min typ 150 min 150 ~ Condition mA 131 157 131 157 131 157 151 181 VEE = V EE(min) to VEE{max); Vcc = Vcco = GND fSHIFT Max. Shift Frequency 700 900 tplH tpHl Propagation Oelay to Output Clk MR 625 600 750 725 25°C 175 max min typ 700 900 625 600 750 725 175 85°C max min typ 700 900 625 600 750 725 -125 175 350 300 +200 +150 350 300 +25 +200 +150 -25 -200 700 max Unit Condition MHz ps 975 975 975 975 Setup Time D SELO SEL1 350 300 25 200 150 Hold Time D SELO SEL1 200 100 100 -25 -200 -150 200 100 100 -25 -200 -150 200 100 100 tRR Reset Recovery Time 900 700 900 700 900 tpw Minimum Pulse Width Clk,MR 400 t, tf 150 157 typ tSKEW Unit 131 O°C th max 157 min ts typ 131 Characteristic Symbol 85°C max 975 975 ps ps -150 ps ps Within-Device Skew 400 400 60 Rise / Fall Times 20 -80% 60 60 ps 1 ps 300 525 800 300 525 800 300 525 800 1. Within-device skew IS defined as Identical transilions on similar paths through a device EXPANDED FUNCTION TABLE Function DL DR Load Shift Right X X X X Shift Left L H Hold Reset X X X L H X X X X X SElO SEll L L L H H H H L H H L L H H X X MR ClK 00 01 02 03 04 05 06 07 L L L L L L L H Z Z Z Z Z Z Z X DO L H L 00 00 00 L Dl 00 L 00 01 01 01 L 02 01 00 01 02 02 02 L 03 02 01 02 03 03 03 L 04 03 02 03 04 04 04 L 05 04 03 06 05 04 05 L L L L D7 06 05 L H H H L ECLinPS 3-29 Q4 05 05 05 L MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • MC10E142 MC100E142 700 MHz Min. Shift Frequency 9-Bit for Byte-Parity Applications Asynchronous Master Reset Dual Clocks Extended 100E VEE Range of -4.2 V to-5.46 V 75 kG Input Pulldown Resistors 9-BIT The MC10E/100E142 is a 9-bit shift register, designed with byte-parity applicaSHIFT REGISTER tions in mind. The E142 performs serial/parallel in and serial/parallel out, shifting in one direction. The nine inputs DO-DB accept parallel input data, while S-IN accepts serial input data. The SEl (Select) input pin is used to switch between the two modes of operation - SHIFT and lOAD. The shift direction is from bit 0 to bit B. Input data is accepted by the registers a set-up time before the positive going edge of ClK1 or ClK2; shifting is also accomplished on the positive clock edge. A HIGH on the Master Reset pin (MR) asynchronously resets all the registers to zero. lOGIC SYMBOL PINOUT: 28-lEAD PlCC (TOP VIEW) SEl 08 0] 06 05 Veea Os 25 24 23 22 21 20 19 MR 07 ClK1 06 ClK2 VCC VEE 05 S·IN Vcca 00 ~ a1 03 02 03 04 Vcca 00 01 02 FUNCTION TABLE SEL Mode L H LOAD SHIFT PIN NAMES Pin 00-08 S-IN SEL ClK1, ClK2 MR Qo-Os Function Parallel Data Inputs Serial Data Input Mode Select Input Clock Inputs Master Reset Data Outputs MR ECLinPS 3-30 MC10E142, MC100E142 DC Characteristics: VEE = VEE(min) to VEE(max); Vcc = Vcco= GND 25°C O°C Symbol - Characteristic I'H Input HIGH Current lEE Power Supply Current min typ max min tpLH tpHL 150 IlA Condition mA 145 120 145 100E 120 145 120 145 138 165 VEE = VEE(min) to VEE(max); Vcc = Vcco= GND Characteristic min typ Max. Shift Frequency 700 900 600 600 800 800 50 Propagation Delay to Output Clk SelupTime D 25°C min typ 700 900 600 600 800 800 max -100 50 300 150 300 75 85°C max min typ max 700 900 600 600 800 1000 800 1000 -100 50 -100 300 150 300 150 100 -150 300 75 100 -150 300 75 100 -150 700 900 700 900 700 Condition MHz 1000 1000 1000 1000 ps Hold Time ps Resel Recovery Time 900 Ipw Minimum Pulse Widlh Clk,MR 400 ps ps Wilhin-Device Skew Rise / Fall Times 20 -80% Unit ps IRR Ir I, Unit 120 D SEL I SKEW max 145 SEL Ih typ 120 MR ts min 150 150 O°C 'SHIFT 85°C max 10E AC Characteristics: Symbol typ 400 75 400 75 75 ps ps 300 .. 525 800 300 525 800 1. Within-deVice skew IS defined as Identical transitions on similar paths through a device ECLinPS 3-31 300 525 , 800 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • .• • • • • MC10E143 MC100E143 700 MHz Min. Operating Frequency 9-Bit for Byte-Parity Applications Asynchronous Master Reset Dual Clocks Extended 100E VEE Range of -4.2 V to -5.46 V 75 kO Input Pulldown Resistors The MC10E/100E143 is a 9-bit holding register, designed with byte-parity applications in mind. The E143 holds current data or loads new data, with the nine inputs 00-08 accepting parallel input data. The SEl (Select) input pin is used to switch between the two modes of operation - HOLD and lOAD. Input data is accepted by the registers a set-up time before the positive going edge of ClK1 or ClK2. A HIGH on the Master Reset pin (MR) asynchronously resets all the registers to zero. LOGIC SYMBOL PINOUT: 2a-lEAD PLCC (TOP VIEW) SEl 08 0] Os 05 VCCO Os MR 07 ClK1 Os ClK2 VCC VEE 05· Do C4 01 03 02 03 04 VCCO 00 01 9-BIT HOLD REGISTER 02 FUNCTION TABLE SEl Mode l H lOAD HOLD PIN NAMES Pin DO-OS SEl ClK1, ClK2 MR Oo-OS NC Function Parallel Data Inputs Mode Select Input Clock Inputs Master Reset Data Outputs No Connection SEl elK1 ClK2 MR---C>-----------~ ECLinPS 3-32 MC1QE143, MC100E143 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo= GND O°C Symbol Characteristic I'H Input HIGH Current lEE Power Supply Current 10E min 25°C max min 150 ~ 145 120 145 120 145 138 165 VEE = VEE(min) to VEE(max); Vee = Vcco= GND 700 900 tpLH Propagation Delay to Output Clk 600 600 800 800 50 300 -100 MR Setup Time D SEL 25°C max 85°C min typ 700 900 600 600 800 800 -100 150 50 300 300 75 100 -150 300 75 100 -150 900 700 900 700 min typ 700 900 600 600 800 800 -100 150 50 300 300 75 100 -150 700 max max D SEL Unit Condition MHz ps 1000 1000 1000 1000 1000 1000 ps 150 Hold Time ps tRR Reset Recovery Time 900 tpw Minimum Pulse Width Clk,MR 400 ps ps Within-Device Skew 400 75 400 75 75 ps Rise / Fall Times 20 -80% Condition mA 120 Max. Toggle Frequency tr tf Unit 145 f MAX tSKEW max 120 typ th typ 145 min ts min 120 Characteristic tpHL 85°C max 150 O°C Symbol typ 150 100E ACCharacteristics: typ ps 300 525 800 300 .. 525 800 1. Within-device skew IS defined as Identical transitions on similar paths through a deVice ECLinPS 3-33 300 525 800 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . .. . TECHNICAL DATA MC10E150 MC100E150 • 800 ps Max. Propagation Delay • Extended 100E VEE 'Range of ~4.2 V to -5.46 V • 75 kG Input Pulldown Resistors The MC10E/100E150 contains six D-type latches with differential outputs. When' both Latch Enables (LEN1, LEN2) are LOW, the latch is transparent and input data transitions propagate through to the output. A logic HIGH on either LEN1 or LEN2 (or both) latches the data. The Master Reset (MR) overrides all other controls to set the Q outputs low. PINOUT: 28-lEAD PLCC (TOP VIEW) LOGIC SYMBOL MR LEN2 LEN1 23 Ne Veeo 22 21 115 05 00 - - - I Os 1i4 04 Os VEE Oa Ci2 Do 02 NC VCCO qo 01 if," Veeo' PIN NAMES Pin DO-OS LEN1, LEN2 MR 00-0 5 00--05 Function Data Inputs Latch Enables Master Reset True Outputs Inverting Outputs LEN1 LEN2 MR ECLinPS 3-34 6-BIT D LATCH MC10E150, MC100E150 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo= GND 25°C O°C Symbol IIH lEE f------ Characteristic min typ max min 85°C max min typ max 200 150 200 150 Power Supply Current mA 10E 100E 52 52 62 62 52 52 52 62 62 60 62 72 VEE = VEE(min) to VEE(max); Vee = Veeo= GND O°C 85°C 25°C min typ max min typ max min typ max Propagation Delay to Output D 250 375 250 375 550 250 375 550 375 450 500 625 550 700 375 450 500 625 700 750 375 350 500 625 700 750 200 50 200 50 200 50 Hold Time D 200 -50 200 -50 200 -50 tRR Reset Recovery Time 750 650 750 650 750 650 tpw Minimum Pulse Width tpHL LEN MR ts MR tSKEW t, tf Condition 750 ps ps ps ps 400 Within-Device Skew Rise I Fall Times 20-80% Unit ps Setup Time D th Condition 200 150 Characteristic tpLH Unit IJ:A Input HIGH Current D LEN, MR ACCharacteristics: Symbol typ 400 400 50 50 50 ps ps 300 450 650 300 450 650 1. Within-device skew IS defined as identical transitions on similar paths through a device ECLinPS 3-35 300 450 650 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • MC10E151 MC100E151 1100 MHz Min. Toggle Frequency Differential Outputs Asynchronous Master Reset Dual Clocks Extended 100E VEE Range of -4.2 V to -5.46 V 75 kO Input Pulldown Resistors The MC10E/100E151 contains 6 D-type, edge-triggered, master-slave flip-flops with differential outputs. Data enters the master when both ClK1 and ClK2 are lOW, and is transferred to the slave when ClK1 or ClK2 (or both) go HIGH. The asynchronous Master Reset (MR) makes all Q outputs go lOW. 6-BIT D REGISTER PINOUT: 28-LEAD PLCC (TOP VIEW) MR ClK2 ClK1 25 24 23 NC VCCO 115 LOGIC SYMBOL 05 22 00 05 li4 04 D4 OJ VCC VEE 03 02 03 02 01 02 02 00 02 lio 01 01 NC VCCO 00 00 01 TIl 03 03 VCCO D4 fi4 PIN NAMES Pin 00-0 5 ClK1, ClK2 MR 00-0 5 00-0 5 05 Function Os Data Inputs Clock Inputs Master Reset True Outputs Inverted Outputs ECLinPS 3-36 MC1 OE151, MC100E151 DC Characteristics: VEE = VEE(min) to VEE(max); Vec = Vcco= GND 25'G O'G Symbol Characteristic IIH Input HIGH Current lEE Power Supply Current 10E 100E ~ ACCharacteristics: min typ max min Max. Toggle Frequency 1100 1400 Propagation Delay to Output Glk MR 475 475 650 650 0 Hold Time D tRR tpw Is th tSKEW tr tf typ max Unit 150 ~A Condition 78 78 65 65 65 75 78 78 78 90 VEE = VEE(min) to VEE(max); Vcc = Veco= GND min tplH tpHl min mA 65 65 Characteristic 'MAX 85'G max 150 150 O'G Symbol typ typ min typ min typ max 1100 1400 Unit Condition MHz ps 650 650 -175 0 350 175 Reset Recovery Time 750 550 Minimum Pulse Width ClK, MR 400 800 850 800 850 475 475 650 650 -175 0 -175 350 175 350 175 750 550 750 550 800 850 ps ps ps ps Within-Device Skew Rise / Fall Times 20-80% max 1100 1400 475 475 Setup Time D 85'G 25'C max 400 65 400 65 65 ps ps 300 450 700 300 450 700 1. Within-device skew is defined as identical transitions on similar paths through a device ECLinPS 3-37 300 450 700 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • • MC10E154 MC100E154 850 ps Max. LEN to Output 825 ps Max. 0 to Output Differential Outputs Asynchronous Master Reset Dual Latch-Enables Extended 100E VEE Range of -4.2 V to -5.46 V 75 Idllnput Pulldown Resistors The MC10E/l00E154 contains five 2:1 multiplexers followed by transparent latches with differential outputs. When both Latch Enables (LEN1, LEN2) are LOW, the latch is transparent, and output data is controlled by the multiplexer select control, SEl. A logic HIGH on either LEN1 or LEN2 (or both) latches the outputs. The Master Reset (MR) overrides all other controls to set the Q outputs LOW. LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) D4b D4a 25 24 OJb D3a Veeo 04 04 li3 SEL ,if.' il 5-BIT 2:1 MUX-LATCH Dna 00 Dab 00 LENl Dla LEN2 Vee Dlb 01 VEE 02 D2a 02 D2b 02 0Ja 03 OJb 03 D4a 04. D4b D.t 02 Daa 01 01 Dla Dlb D2a D2b Veeo 00 ao PIN NAMES Pin DOa-D4a DOb-D4b SEL LEN', LEN2 MR 00-14 00-14 01 Function SEL Input Data a Input Data b Data Select Input Latch Enables Master Reset True Outputs Inverted Outputs LENl LEN2 MR ECLinPS 3-38 MC10E154, MC100E154 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo = GND 25°C O°C Symbol Characteristic I'H Input HIGH Current lEE Power Supply Current r-- min typ max min 150 tpHL t, th Unit 150 IlA 91 76 91 100E 76 91 76 91 87 105 VEE = V EE(min) to V EE(max); V cc = Vcco = GND Characteristic 25°C 85°C min typ max min typ max min typ max 325 500 700 325 500 700 325 500 700 Propagation Delay to Output D 475 650 925 475 650 925 475 650 925 LEN 350 500 750 350 500 750 350 500 750 MR 450 600 800 450 600 800 450 600 800 D 300 100 300 100 300 100 SEL 500 250 500 250 500 250 300 200 -100 -100 -250 300 200 -100 -250 300 200 800 600 800 600 800 600 Setup Time Condition ps Hold Time ps Reset Recovery Time tpw Minimum Pulse Width -250 ps ps 400 MR Within-Device Skew 400 50 400 50 ps 50 Rise / Fall Times 20 - 80% IS Unit ps SEL 1. Within-deViCe skew Condition mA 76 tRA t, max 91 SEL tc typ 76 D tSKEW min 150 O°C tplH 85°C max 10E AC Characteristics: Symbol typ ps 300 475 800 300 475 800 defined as Identical transitions on similar paths through a device ECLinPS 3-39 300 475 800 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • • MC10E155 MC100E155 850 ps Max. LEN to Output .825 ps Max. D to Output Single-Ended Outputs Asynchronous Master Reset Dual Latch-Enables Extended 100E VEE Range of - 4.2 V to - 5.46 V 75 kn Input Pulldown Resistors The MC10E/100E155 contains six 2:1 mUltiplexers followed by transparent latches with single-ended outputs. When both Latch Enables (LEN1, LEN2) are LOW, the latch is transparent, and output data is controlled by the multiplexer select control, SEL. A logic HIGH on either LEN1 or LEN2 (or both) latches the outputs. The Master Reset (MR) overrides all other controls to set the Q outputs LOW. 6-BIT 2:1 MUX-LATCH LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) Dsa Doa----i D«b D4a D3b D3a Ne Veeo Dob----i 19 DSb Os LEN1 04 LEN2 Vee VEE 03 MR 02 SEL Veco Doa 10 DOb D1 a D1b D2a D2b Veco 11 00 Dsa---+-i PIN NAMES Pin DOa-D4a DOb-D4b SEL LEN1, LEN2 MR Qo--C4 Dsb---H Function Input Data a Input Data b Data Select Input Latch Enables Master Reset Outputs SEL LENl LEN2 MR ECLinPS 3-40 MC10E155, MC100E155 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo= GND 25°C O°C Symbol - Characteristic IIH Input HIGH Current lEE Power Supply Current 10E 100E AC Characteristics: min typ max min 85°C max min typ 150 150 max Unit 150 !lA 102 102 85 85 102 102 85 98 102 117 VEE = VEE(min) to VEE(max); Vee = Vceo = GND 85°C 25°C Characteristic min typ max min typ max min typ max Propagation Delay to Output D SEL LEN MR 325 475 350 450 500 675 500 600 700 925 750 850 325 475 350 450 500 675 500 600 700 925 750 850 325 475 350 450 500 675 500 600 700 925 750 850 SelupTime D 300 100 300 250 500 300 500 100 500 100 250 250 Hold Time D SEL 300 0 -100 -250 300 0 -100 -250 300 0 -100 -250 tRR Reset Recovery Time 800 650 800 650 800 650 tpw Minimum Pulse Width MR 400 tpLH tpHL Is SEL th tSKEW tr t, Condition mA 85 85 O°C Symbol typ Condition ps ps ps ps ps Within-Device Skew Rise / Fall Times 20-80% Unit 400 75 400 75 75 ps ps 300 450 800 300 450 800 " 1, Wllhln-devlce skew IS defined as IdentIcal transliions on similar paths through a deVice ECLinPS 3-41 300 450 800 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • • MC10E156 MC100E156 950 ps Max. D to Output 850 ps Max. LENto Output Differential Outputs' Asynchronous Master Reset Dual Latch-Enables Extended 100E VEE Range of -4.2 V to -.5.46 V 75 kO Input Pulldown Resistors The MC10EI100E156 contains three 4:1 multiplexers followed by transparent latches with differential outputs. When both Latch Enables (LEN1,LEN2)'are LOW, the latch is transparent, and output data is controlled by the multiplexer select controls (SELO, SELl). A logic HIGH on either LEN 1 or lEN2 (or both) latches the outputs. The Master Reset (MR) overrides all other controls to set the Q outputs LOW. LOGIC SYMBOL PINOUT: 28-lEAD PlCC (TOP VIEW) O, b O,a 02d 02e °2 b 25 24 23 22 21 02a Veeo 20 1---10 19 SELO 02 SELl Q2 MR Vee VEE a, LEN1 Q, LEN2 Veeo ao 10 O,d OOa Dob Doc Dod Veeo 11 4:1 MUX Go SElO SEll . PIN NAMES Pin Dox-DJX SELO, SEL' LEN', LEN2 MR 00-0 2 00-0 2 3-BIT 4: 1 MUX-LATCH' Function LENI LEN2 . Input Data Select Inputs Latch Enables Master Reset True Outputs Inverted Outputs MR ECLinPS 3-42 MC10E156, MC100E156 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veco= GND 25°C O°C Symbol - Characteristic I'H Input HIGH Current lEE Power Supply Current min typ max min 150 tpLH tpHL min typ max Unit 150 !!A 75 90 100E 75 90 75 75 90 75 90 86 25°C 90 103 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output D 400 600 900 400 600 900 400 600 900 550 450 775 1050 1050 1050 900 500 600 800 825 350 350 650 500 900 350 350 650 500 550 450 775 900 550 450 775 650 350 600 350 600 D SELO 400 275 400 275 400 275 700 600 300 400 700 600 300 400 700 SEL1 600 300 400 D 300 -275 -275 300 -275 SELO SEL1 100 -300 300 100 -400 200 -300 -400 100 200 200 -300 -400 800 600 800 600 800 600 SEL1 LEN MR 800 825 tRR Reset Recovery Time tpw Minimum Pulse Width MR tSKEW ps ps ps ps 400 Within-Device Skew 400 50 400 50 ps 50 Rise I Fall Times 20-80% . Condition 900 825 Hold Time th Unit ps Setup Time ts Condition mA 10E SELO t, tf . 85°C max 150 O°C Symbol typ ps 275 475 700 275 .. 475 700 1. Within-device skew IS defined as Identical transitions on similar paths through a device ECLinPS 3-43 275 475 700 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA. MC10E157 MC100E157 Individual Select Controls 550 ps Max. D to Output 800 ps Max. SEL to Output Extended I OOE VEE Range of -4.2V to -5.46V Internal 75kQ Input Pulldown Resistors The MC I OE/I 00E157 contains four 2: 1 multiplexers with differential outputs. The ouput data are controlled by the individual Select (SEL) inputs. The individual select control makes the devices well suited for random logic designs. PINOUT: 28·LEAD PLCC (TOP VIEW) SEL3 NC D3a 25 24 23 D3b vcco 22 OJ 03 20 19 21 D2b Q2 D2a 02 SEL2 VCC VEE 01 SEL1 01 D1a Qo D1b 00 SELO DOa DOb NC NC 10 11 NC Veca FUNCTION TABLE SEL Data H a L b PIN NAMES PIN FUNCTION DOa- D3a DOb- D3b SELO-SEL3 00-03 00-03 Input Data a Input Data b Select Inputs True Outputs Inverted Outputs ECLinPS 3-44 QUAD 2:1 MULTIPLEXER MC10E157, MC100E157 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo = GND 25°C O°C Symbol IIH lEE Characteristic min typ Input HIGH Current D SEL min tplH tpHl tSKEW tr tf 85°C max min typ max Unit Condition I1A 200 150 200 150 mA 32 32 38 38 32 32 32 37 38 38 38 44 VEE = VEE(min) to VEE(max); Vee = Veea = GND 25°C O°C Symbol typ 200 150 Power Supply Current 10E 100E AC Characteristics: max 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output D SEL 220 425 380 600 550 800 220 425 380 600 550 800 220 425 380 600 550 800 Condition ps Within-Device Skew Rise/Fall Times 20 - 80% Unit 70 70 70 ps ps 275 400 650 275 400 650 1. Within-device skew IS defined as Identical transitions on similar paths through a device ECLinPS 3-45 275 400 650 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . .. . TECHNICAL DATA • • • • • • MC10E158 MC100E158 600 ps Max. 0 to Output 800 ps Max. SEL to Output Differential Outputs One VCCO Pin Per Output Pair Extended 100E VEE Range of -4.2 V to -5.46 V 75 kG Input Pulldown Resistors The MC10E/l00E158 contains five 2:1 multiplexers with differential outpvts. The output data are controlled by the Select input (SEL). 5·BIT 2:1 MULTIPLEXER PINOUT: 28~LEAD PLCC (TOP VIEW) LOGIC SYMBOL FUNCTION TABLE SEL Data H L a Dab llo O,a 0, O,b 0, D2lI 02 02b 02 Daa 03 Dab 03 04a ~ 04bB Ao-As a: 0 ~ B>A VEE A=B. "- ::;; A.! VCCO B4 A>B 0 Bo-B8 NC B5 A6 B6 A7 B7 10 11 As B8 PIN NAMES Pin Ao-Aa Bo-Ba A>B B>A A=B Function A Data Inputs B Data Inputs A Greater than B Output B Greater than A Output A Equal to B Output (active-LOW) ECLinPS 3-54 u B>A MC10E166, MC100E166 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo = GND Characteristic I'H Input HIGH Current lEE Power Supply Current min typ max min tpHL t, tf max min typ 150 150 max Unit 150 ~A 136 113 136 113 136 100E 113 136 113 136 130 156 VEE = V EE(min) to VEE(max); Vee = Veeo = GND Characteristic 25°C 85°C min typ max min typ max min typ max DtoA=B 500 750 1100 500 750 1100 500 750 1100 DtoAB 500 850 1400 500 850 1400 500 850 1400 300 450 800 300 450 800 300 450 800 Propagation Delay to Output Unit ps Rise / Fall Time 20 - 80% Condition mA 113 O°C tpLH typ 10E AC Characteristics: Symbol 85°C 25°C O°C Symbol ps ECLinPS 3-55 Condition MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • • MC10E167 MC100E167 1000 MHz Min. Operating Frequency 800 ps Max. Clock to Output Single-Ended Outputs Asynchronous Master Resets Dual Clocks Extended 100E VEE Range of -4.2 V to -5.46 V 75 kG Input Pulldown Resistors The MC10E/100E167 contains six 2:1 multiplexers followed by D flip-flops with single-ended outputs. Input data are selected by the Select control, SEL. The selected data are transferred to the flip-flop outputs by a positive edge on ClK1 or ClK2 (or both). A HIGH on the Master Reset (MR) pin asynchronously forces all Q outputs lOW. lOGIC SYMBOL PINOUT: 2S-lEAD PlCC (TOP VIEW) Dsa D4b D4a D3b NC D3a 6-BIT 2:1 MUX-REGISTER Veeo Doa---f 1 - - - 00 Dob---f DSb 05 ClKl ClK2 Vce VEE 03 MR 02 SEl Vceo D, a --+-i 1 - - - 0, Dsa--+-i I---Os 0, Dob D, a D,b D2a D2b 10 11 VCCO 00 FUNCTION TABLE SEL Data H L a DSb _ _+-I b PIN NAMES Pin DOa-D5a DOb-D5b SEL CLK1,CLK2 MR Qo-0 5 Function SEl Input Data a Input Data b Select Input Clock Inputs Master Reset Data Outputs ClKl ClK2 MR ECLinPS 3-56 MC10E167, MC100E167 DC Characteristics: VEE = V EE(min) to VEE(max); Vee =Vceo =GND 25°C O°C Symbol - Characteristic min I'H Input HIGH Current lEE Power Supply Current typ max min 150 150 ~A 94 113 100E 94 113 94 113 108 130 VEE = V EE(min) to VEE(max); Vee 1000 1400 tpLH Propagation Delay to Output =Vceo = GND typ 85°C 25°C max min typ max 1000 1400 min typ max 1000 1400 650 800 450 650 800 450 650 800 MR 450 650 850 450 650 850 450 650 850 D 100 -50 100 -50 100 -50 SEL 275 125 275 125 275 125 D 300 SEL 75 50 -125 300 75 50 -125 300 75 50 -125 750 550 750 550 750 550 Setup Time ps Hold Time tpw Minimum Pulse Width Clk,MR ps ps ps 400 Within-Device Skew 400 400 75 75 75 ps ps Rise / Fall Times 20-80% Condition ps 450 Reset Recovery Time Unit MHz Clk tRR Condition mA 113 Max. Toggle Frequency tr tf Unit 94 f MAX tSKEW max 113 min th typ 94 Characteristic ts min 150 O°C tpHL 85°C max 10E AC Characteristics: Symbol typ 300 .. 450 800 300 450 800 1. Within-device skew IS defined as Identical transitions on similar paths through a device ECLinPS 3-57 300 450 800 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • MC10E171 MC100E171 725 ps Max. D to Output Split Select Differential Outputs Extended 100E VEE Range of -4.2 V to -5.46 V 75 k!l Input Pulldown Resistors The MC10E/100E171 contains three 4:1 multiplexers with differential outputs. Separate Select controls are provided for the leading 2: 1 mux pairs (see logic sym· bol). The three Select inputs control which one of the four data inputs in each case is propagated to the corresponding output. 3·BIT 4:1 MULTIPLEXER PINOUT: 28·LEAD PLCC (TOP VIEW) Dlb Dla D2d D2e D2b D2a Veeo LOGIC SYMBOL DOa 2:1 MUX DOb SEL SEL1A li2 SEL1B °2 SEL2 Vee VEE Q1 Ne °1 NC Veea 01a 00 01b Old Ooa Dob DOC Dod 10 11 Veea 00 00 00 DOC Dod Ole 2:1 MUX 01 a, Old FUNCTION TABLE Pin State SEL2 SELlA SEL1B H H H Operation 028 Output c/d data Input d data Input b data °2b PIN NAMES D2e Pin DOX-D2X SELlA, SEL1B SEL2 00-°2 '00-'02 Function 02d Data Inputs First-stage Select Inputs Second-stage Select Input True Output Inverted Output SEllA SEL1B SEL2 ECLinPS 3·58 02 li2 MC1 OE171, MC100E171 DC Characteristics: VEE = VEE(min) to VEE(max); 'V ee = Veeo= GND O°C Symbol Characteristic I'H Input HIGH Current lEE Power Supply Current - min typ 25°C max min 150 tpLH I SKEW tr tf min typ 150 max Unit 150 flA 67 56 67 56 67 100E 56 67 56 67 65 77 VEE = VEE(min) to VEE(max); Vee = Veeo= GND Characteristic 25°C 85°C min typ max min typ max min typ max D 275 480 650 275 480 650 275 480 650 SEll 450 650 850 450 650 850 450 650 850 SEL2 350 550 700 350 550 700 350 550 700 Propagation Delay to Output Unit Condition ps Within-Device Skew ps Dnm, Dnm to On 60 60 60 Da, Db, Dc, Dd 10 0 40 40 40 Rise / Fall Time 20-80% Condition mA 56 O°C tpHL 85°C max 10E ACCharacteristics: Symbol typ ps 300 475 650 300 475 650 300 475 650 1. Within-device skew IS defined as Identical transitions on similar paths through a device; n=O,l ,2 m=a,b,c,d. ECLinPS 3-59 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E175 MC100E175 9-bit Latch Parity Detection/Generation 800 ps Max. D to Output Reset Extended 1OOE VEE Range of -4.2V to -5.46V Internal 75kr.! Input Pulldown Resistors The Me 1OEl1 00E175 is a 9-bit latch. It also features a tenth latched output, ODDPAR, which is formed as the odd parity of the nine data inputs (ODDPAR is HIGH if an odd number of the inputs are HIGH). The E175 can also be used to generate byte parity by using D8 as the parity-type select ( L = even parity, H = odd parity), and using ODDPAR as the byte parity output. The LEN pin latches the data when asserted with a logical high and makes the latch transparent when placed at a logic low level. 9-81T LATCH WITH PARITY PINOUT: 28·LEAD PLCC (TOP VIEW) 06 07 08 vcca 08 Q7 25 24 23 22 21 20 LOGIC SYMBOL veea 05 06 04 05 03 vcc DO----~---------------4 D Q 00 Q 08 0 ODDPAR EN R bils VEE 04 LEN 03 MR vcca 02 02 '·7 D8~~~------------~~ D EN 01 00 veea 0 0 0 9 10 11 00 veea 01 R "0 D > :D R PIN NAMES LEN------------------~ PIN FUNCTION DO- D8 Data Inputs LEN Latch Enable Master Reset MR 00-08 ODDPAR MR----------------------~ Data Outputs Parity Output ECLinPS 3-60 MC10E175, MC100E175 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veea = GND O'C Symbol - Characteristic min IIH Input HIGH Current lEE Power Supply Current 10E 100E AC Characteristics: VEE typ min IplH tpHl ts Ih tRR tSKEW max Characteristic min 110 110 132 132 Vee Propagation Delay to Output DtoO DtoODDPAR LEN to 0 LEN to ODDPAR MR to 0 (lpHl ) MR to ODDPAR (tpHl ) 132 132 110 127 max Unit 150 flA Condilion 25'C 85°C min 450 850 525 525 525 525 600 800 1150 1450 700 900 700 900 700 900 700 900 450 850 525 525 525 525 Setup Time D(O) D (ODDPAR) 275 900 100 700 275 900 275 900 Hold Time D(O) D (ODDPAR) 175 -300 -100 -700 175 -300 175 -300 Reset Recovery Time 850 600 850 typ 132 152 = Veea = GND max max min 600 800 1150 1450 700 900 700 900 700 900 700 900 450 850 525 525 525 525 typ typ max Unit Condition ps 600 800 1150 1450 700 900 700 900 700 900 700 900 ps ps Within-Device Skew LEN, MR DtoO Dto ODD PAR 850 600 600 ps ps 75 75 200 75 75 200 75 75 200 ps 300 .. 500 800 300 500 800 FUNCTION TABLE D EN MR Q ODDPAR H L L L H L L L H H L H if odd no. of Dn HIGH H if odd no. of Dn HIGH 00 00 L L X typ 150 1. Wlthm-devlce skew IS defmed as Identical transitions on similar paths through a device X X min mA 110 110 = VEE(min) to VEE(max); Rise/Fall Times 20 - 80% t, tf Iyp 150 O'C Symbol 85'C 25'C max ECLinPS 3-61 300 500 800 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • MC10E193 MC100E193 Hamming Code Generation a-Bit Word, Expandable Provides Parity of Whole Word Scannable Parity Register Extended 100E VEE Range of -4.2 V to -5.46 V 75 kO Input Pulldown Resistors The MC10E/100E193 is an error detection and correction (EDAC) circuit. Modified Hamming parity codes are generated on an 8-bit word according to the pattern shown in the logic symbol. The P5 output gives the parity of the whole word. The word parity is also provided at the PGEN pin, after OddlEven parity control and gating with the BPAR input. This output also feeds to a 1-bit shiftable register, for use as part of a scan ring. The combinatorial part of the device generates the same code pattern as the MC10193, a member of the MECL 10K family. The user is referred to the 10193 data sheet in the MECL Device Data Book for a fuller description of pattern expansion to long words, along with check bit generation and decoding schemes. Used in conjunction with 12-bit parity generators such as the E160, a SECDED (single error correction; double error detection) error system can be designed for a multiple of an 8-bit word. ECLinPS 3-62 ERROR DETECTIONI CORRECTION CIRCUIT MC10E193, MC100E193 DC Characteristics: VEE = VEE(min) to VEE(max); Vee =Vceo = GND 25°C O°C Symbol IIH lEE f-- Characteristic min typ Input HIGH Current max min 112 112 134 134 VEE = V EE(min) to VEE(max); Vee Unit 150 ~ 112 112 134 134 112 129 155 =Vceo = GND 25°C 85°C typ max min typ max min typ max Propagation Delay to Output Bto P1, P2, P3, P4 BtoP5 350 400 700 1000 1150 350 400 700 775 1000 1150 350 400 700 775 1000 1150 350 650 850 1000 1450 550 850 350 650 850 1000 1450 350 600 300 650 850 1000 1450 400 150 400 300 50 350 250 300 750 775 600 300 550 850 600 300 550 850 ps EV/OD BPAR 750 500 1300 1300 B 1700 1100 EN - Hold Time SHIFT S-IN HOLD -EN - EV/OD BPAR B 850 400 150 50 300 50 350 250 850 750 500 1300 850 1700 1100 1300 350 250 850 850 500 1300 1300 850 150 1700 1100 ps 200 300 100 -150 -150 200 -150 -50 -350 300 100 -50 -350 100 -250 -200 -850 -200 -850 -300 -1100 100 -200 -250 300 300 200 300 100 -50 -350 100 -250 -200 -850 -200 -850 -300 -1100 -850 -200 -850 -300 -1100 Rise / Fall Times 20 -80% Unit ps Setup Time -HOLD Condition 134 min SHIFT S-IN t, tf max Characteristic ClK to PARERR th typ mA 10E 100E EV/OD, BPAR to PGEN Bto PGEN ts min 150 150 O°C tpLH tpHL 85°C max Power Supply Currenl AC Characteristics: Symbol typ ps 300 700 1100 ECLinPS 3-63 7000 1100 700 1100 Condition MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA 2ns Worst Case Delay Range ~20ps I Delay Step Resolution > 1GHz Bandwidth On Chip Cascade Circuitry Extended 100E VEE Range of -4.2V to -5.46V MC10E195 MC100E195 The MC 1OEl1 OOE 195 is a programmable delay chip (PDC) designed primarily forciockde-skewing and timing adjustment. It provides variable delay of a differential ECl input transition. The delay section consists of a chain of gates organized as shown in the logic symbol. The first two delay elements feature gates that have been modified to have delays 1.25 and 1.5 times the basic gate delay of approximately 80ps. These two elements provide the E195 with adigitally-selectable resolution of approximately 20ps. The required device delay is selected by the seven address inputs D[0:6], which are latched on chip by a high signal on the latch enable (lEN) control. Because the delay programmability of the E195 is achieved by purely differential ECl gate delays the device will operate at frequencies of >1GHz while maintaining over 600mV of output swing. The E195 thus offers very fine resolution, at very high frequencies, that is selectable entirely from a digital input allowing for very accurate system clock timing. PROGRAMMABLE DELAY CHIP An eighth latched input, D7, is provided for cascading multiple PDC's for increased programmable range. The cascade logic allows full control of multiple PDC's, at the expense of only a single added line to the data bus for each additional PDC, without the need for any external gating. 02 03 04 05 06 07 NC 25 24 23 22 21 20 19 01 NC 00 NC LEN vee PIN NAMES PINOUT: 28-LEAD PLCC (TOP VIEW) VEE Pin Function IN - INIIN EN D[0:7] Q/O lEN SET MIN SET MAX CASCADE Signal Input Input Enable Mux Select Inputs Signal Output latch Enable Min Delay Set Max Delay Set Cascade. Signal a VBB veeo 10 Ne Ne ~ w IN iN 03 z (fJ EN 02 EN f- VBB------1 01 • Q IN LOGIC DIAGRAM - SIMPLIFIED DO veeo D4 OS • delays are 25% or %50 longer than standard (standard",aOps) ECLinPS 3-64 06 07 x « :;; fW (fJ w 0 « a Ul « a 11 MC1 OE195, MC100E195 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo = GNO Symbol Characteristic min I'H Input HIGH Current lEE Power Supply Current typ max min 150 tpHl tRANGE 150 jlA 156 100E 130 156 130 156 150 179 VEE = VEE(min) to VEE(max); Characteristic Vee = Veeo = GNO min typ 25°C max min typ 85°C max min typ max Propagation Delay Unit Notes ps =0 = 127 EN to Q; Tap = 0 1210 1360 1510 1240 1390 1540 1440 1590 1765 3320 3570 3820 3380 3630 3880 3920 4270 4720 07 to CASCADE 300 IN to Q; Tap IN to Q; Tap Condition mA 130 1250 1450 1650 1275 1475 1675 1350 1650 1950 450 700 300 450 700 300 450 700 Programmable Range 2000 2175 2375 2580 2050 2240 ps Step Delay 03 High ps 17 17.5 68 105 136 180 55 115 42 84 120 180 65 140 168 205 70 140 105 04 High 250 272 325 250 280 325 305 336 380 05 High 505 544 620 515 560 620 620 672 740 06 High 1000 1088 1190 1030 1120 1220 1240 1344 1450 Linearity 01 DO 01 DO 6 21 35 34 55 115 01 DO 7 Duty Cycle Skew ±30 tpHL - tpLH ±30 ±30 ps 1 ps Setup Time o to LEN 200 OtolN 800 800 800 2 200 200 200 3 -EN to IN th Unit 156 02 High ts max 130 01 High tSKEW typ 156 DO High Lin min 130 tpD (max) - tpD (min) "'t max 150 O°C tplH typ 10E AC Characteristics: Symbol 85°C 25°C O°C 0 200 0 200 0 Hold Time ps LEN to 0 500 IN to EN 0 250 500 0 ECLinPS 3-65 250 500 0 250 4 Mel0E195,MC100E195 AC Characteristics (Cant): VEE = VEE(min) to VEE(max)' Vcc = Vcco = GND' QOC Symbol tR 4. 5. 6. 7. S. min Release Time EN to IN SET MAX to LEN SET MIN to LEN 300 800 800 tji! Jitter t, Rise/Fall Time 20 - 80% (Q) 20 - 80% (CASCADE) ~ 1. 2. 3. Characteristic typ 25°C max min typ 85°C max min typ max Unit Notes ps 300 800 800 <5 300 800 800 <5 5 ps <5 8 ps 125 300 225 450 325 650 125 300 225 450 325 650 125 300 225 450 325 650 Duty cycle skew guaranteed only for differential operation measured from the cross point pf the input to the cross point of the output. This setup time defines the amount of tim~rior to the input signal the delay tap of the device IJlllst be set. This setup time is the minimumJime that EN must be asserted prior to the next transition of IN/IN to prevent an output response greater than ±75mV to that IN/IN trimsit!2.0. _ . This hold time is the minimum time that EN mU'l!.remain asserted after a negative going IN or positive going IN to prevent an output response greater than ±75mV to that IN/IN transition. This release time is the minimum time that EN must be deasserted prior to the next INiiN transition to ensure an output response that meets the specified IN to Q propagation delay and transition times. Specifcation limits representthe amount of delay added with the assertion of each individual delay control pin. The various combinations of asserted delay control inputs will typically realize DO resolution steps across the specified programmable range. The linearity specification guarantees to which delay control input the programmable steps will be monotonic (ie. increasing delay steps for increasing binary counts on the control inputs Dn). Typically the device will be monotonic to the DO input, however under worst case conditions and process variation, delays could decrease slightly with increasing binary counts when the DO input is the LSB. With the D1 input as the LSB the device is guaranteed to be monotonic over all ~pecified environmental conditions and process variation. The jitter of the device is less than what can be measured without resorting to very tedious and specialized measurement techniques. ECLinPS 3-66 MC10E195, MC100E195 Applications Information A7--~----+-+-+-+-4-0 E195 E195 vee veca Chip #1 vee veca Chip #2 0 Input 0 a 0 z x oWIW :; ~ C3 {jvcco lifi >W >- Ul Ul z Ul lifi W '" '-' '" '-' Ul Output a {jvcco WIW x 00 :; ~ t3 >- >- Ul Ul W W '" '" Ul " '-' Ul Figure 1 - Cascading Interconnect Architecture Cascading Multiple E19S's To increase the programmable range of the E195 internal cascade circuitry has been included. This circuitry allows for the cascading of multiple E195's without the need for any external gating. Furthermore this capability requires only one more address line per added E195. Obviously cascading multiple poe's will result in a larger programmable range however this increase is at the expense of a longer minimum delay Figure 1 illustrates the interconnect scheme for cascading two E195's. As can be seen, this scheme can easily be expanded for larger E195 chains. The 07 input of the E195 is the cacade control pin. With the interconnect scheme of Figure 1 when 07 is asserted it signals the need for a larger programmable range than is acheivable with a single device. An expansion of the latch section of the block diagram is pictured below. Use of this diagram will simplify the explanation of how the cascade circuitry works. When 07 of chip #1 above is low the cascade output will also be low while the cascade bar output will be a logical high. In this condition the SET MIN pin of chip #2 will be asserted and thus all of the latches of chip #2 will be reset and the device will be set at its minimum delay. Since the RESET and SET inputs of the latches are overriding any changes on the AO-AS address bus will not affect the operation of chip #2. Chip #1 on the other hand will have both SET MIN and SET MAX de-asserted so that its delay will be controlled entirely by the address bus AO-AS. If the delay needed is greater than can be acheived with 31.75 gate delays (1111111 on the AO-AS address bus) 07 will be asserted to signal the need to cascade the delay to the next E 195 device. When 07 is asserted the SET MIN pin of chip #2 will be de-asserted and the delay will be controlled by the AO-AS address bus. Chip #1 on the other hand will have its SET MAX pin asserted resulting in the device delay to be independent of the AO-AS address bus. When the SET MAX pin of chip #1 is asserted the 00 and 01 latches will be reset while the rest of the latches will be set. In addition, to maintain monotonicity an additional gate delay is selected in the cascade Circuitry. As a result when 07 of chip #1 is asserted the delay increases from 31 .75 gates to 32 gates. A 32 gate delay is the maximum delay setting for the E195. To expand this cascading scheme to more devices one simply needs to connect the 07 input and CASCAOE outputs of the current most significant E 195 to the new most significant E195 in the same manner as pictured in figure 1. The only addition to the logic is the increase of one line to the address bus for cascade control of the second POCo [r= Bill 07 07 CASC'O' -CASCADE Co, SET MIN -+....l..------l:--.L.----+.....l______+--L____-+__L +-..l..____-l~ ____ .....J SET MAX_--'-______---'________....L________.l.....______.....l________...L______ Figure 2 - Expansion of the Latch Section of the E19S Block Diagram ECLinPS 3-67 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E196 MC100E196 2ns Worst Case Delay Range ~20ps Digital Step Resolution Linear Input for Tighter Resolution >IGHz Bandwidth On Chip Cascade Circuitry Extended 100E VEE range of -4.2V to -5.46V The MC1 OE/1 00E196 is a programmable delay chip (PDC) designed primarily for very accurate differential ECl input edge placement applications. The delay section consists of a chain of gates and a linear ramp delay adjust organized as shown in the logic symbol. The first two delay elements feature gates that have been modified to have delays 1.25 and 1.5 times the basic gate delay of approximately BOps. These two elements provide the E196 with a digitally-selectable resolution of approximately 20ps. The required device delay is selected by the seven address inputs 0[0:6], which are latched on chip by a high signal on the latch enable (lEN) control. The FTUNE input takes an analog voltage and applies it to' an internal linear ramp for reducing the 20ps resolution still further. The FTUNE input is what differentiates the E196 from the E195. PROGRAMMABLE DELAY CHIP An eighth latched input, 07, is provided for cascading multiple POC's forincreased programmable range. The cascade logic allows full control of multiple POC's, at the expense of only a single added line to the data bus for each additional PDC, without the need for any external gating. D2 D3 D4 D5 D6 D7 25 24 23 22 21 20 NC D1 FTUNE DO NC PIN NAMES VCC LEN Pin Function PINOUT: 28-LEAD PLCC (TOP VIEW) VEE - IN/IN EN O['?,:7] Q/Q lEN SET MIN SET MAX CASCADE FTUNE Signal Input Input Enable Mux Select Inputs Signal Output latch Enable Min Delay Set Max Delay Set Cascade Signal Linear Voltage Input vcca IN Q IN Q vcca VBB 10 NC NC EN z ~ fW C/l LOGIC DIAGRAM- SIMPLIFIED ~ Lll 0 ::;; < 0 f- < 0 w C/l rn 11 I~ ~ FTUNE VBB---I IN iN EN a a DO 01 02 03 D4 05 'delays are 25% or %50 longer than standard (slandard~80ps) ECLinPS 3-68 06 07 MC1 OE196, MC100E196 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo = GND 25°C O°C Symbol Characteristic min IIH Input HIGH Current lEE Power Supply Current typ max min 150 tpHL tRANGE th 150 j.lA 130 156 100E 130 156 130 156 150 179 VEE = VEE(min) to VEE(max); Characteristic Condition mA 156 min Vee = Veeo = GND typ 25°C max min typ 85°C max min typ max Propagation Delay Unit Notes ps IN to Q; Tap = 0 ~to Q; Tap = 127 EN to Q; Tap = 0 1210 1360 1510 1240 1390 1540 1440 1590 1765 3320 3570 3820 3380 3630 3880 3920 4270 4720 1250 1450 1650 1275 1475 1675 1350 1650 1950 D7 to CASCADE 300 450 700 300 450 700 300 450 700 Programmable Range 2000 2175 2375 2580 2050 2240 ps Step Delay ps 17.5 17 34 35 42 55 68 105 55 70 105 65 84 120 D3 High 115 136 180 115 140 180 140 168 205 D4 High 250 272 325 250 280 325 305 336 380 D5 High 505 544 620 515 560 620 620 672 740 D6 High 1000 1088 1190 1030 1120 1220 1240 1344 1450 Linearity D1 DO D1 DO 6 21 D2 High D1 DO 7 Duty Cycle Skew ±30 tpHl - tplH ts Unit 130 D1 High tSKEW max 156 DO High Lin typ 130 tpo (max) - tpo (min) L1t min 150 O°C tpLH 85°C max 10E AC Characteristics: Symbol typ ±30 ±30 Setup Time ps 1 ps Dto LEN 200 Dto IN 800 800 800 2 EN to IN 200 200 200 3 0 200 0 200 0 Hold Time ps LEN to D 500 INto EN 0 - 250 500 0 ECLinPS 3-69 250 500 0 250 4 MC1 OE196, MC100E196 AC Characteristics (Cant): VEE = VEE(min) 10 VEE(max); Vcc = Vcco = GND O°C Symbol IR Ijil Ir I, 1. 2. 3. 4. 5. 6. 7. S. Characteristic min Release Time ENlo IN SET MAX 10 LEN SET MIN 10 LEN 300 800 800 typ 25°C max min typ 85°C max min typ max Unit Notes ps Jitter 300 800 800 <5 Rise/Fall Time 20-80% (Q) 20 - 80% (CASCADE) 300 800 800 <5 5 ps <5 8 ps 125 300 225 450 325 125 650 300 225 450 325 650 125 300 225 450 325 650 Duty cycle skew guaranteed only for dlfferenltal operation measured from the cross pOint of the Input to the cross pomt of the output. This setup time defines the amount of tim~rior to the input signal the delay tap of the device IJl1lst be set. This setup time is the minimumjime that EN must be asserted prior to the next transition of IN/IN to prevent an output response greater than ±75mV to that IN/IN transition. This hold time is the minimum time that EN mus.lremain asserted after a negative going IN or positive going iN to prevent an output response greater than ±75mV to that IN/IN transition. This release time is the minimum time that EN must be deasserted prior to the next IN/iN transition to ensure an output response that meets the specified IN to Q propagation delay and transition times. Specifcation limits represent the amount of delay added with the assertion of each individual delay control pin. The various combinations of asserted delay control inputs will typically realize DO resolution steps across the specified programmable range. The linearity specification guarantees to which delay control input the programmable steps will be monotonic (ie. increasing delay steps for increasing binary counts on the control inputs On). Typically the device will be monotonic to the DO input, however under worst case conditions and process variation, delays could decrease slightly with increasing binary counts when the DO input is the LSB. With the 01 input as the LSB the device is guaranteed to be monotonic over all specified environmental conditions and process variation. The jitter of the, device is less than what can be measured without resorting to very tedious and specialized measurement techniques. Analog Input Characteristics: Flune = VCC 10 VEE Propagation Delay vs Ftune Voltage (100E196) 140 120 (ij' S >ca Qj I\.. 100 80 Q c 60 iCI 40 0 ca I\.. ...... Q. ... 0 D- 20 ...... t-h o -4.5 -3.5 -2.5 -1.5 Ftune Voltage (V) ECLinPS 3-70 -0.5 MC10E196, MC100E196 Analog Input Characteristics (cont): Ftune = VCC to VEE Propagation Delay vs Ftune Voltage (10E196) 'iii' B >ca 'iii c c: 0 iOl ca Q. 0 rr. 100 90 80 70 60 50 40 30 20 10 0 1\ r-- 1'\ '" '" r... r-5 -4 -3 -2 -1 o Ftune Voltage (V) USING THE FTUNE ANALOG INPUT The analog FTUNE pin on the E196 device is intended to enhance the 20ps resolution capabilities of the fully digital E195. The level of resolution obtained is dependent on the number of increments applied to the appropriate range on the FTUNE pin. To provide another level of resolution the FTUNE pin must be capable of adjusting the delay by greater than the 20ps digital resolution. From the provided graphs one sees that this requirement is easily achieved as over the entire FTUNE voltage range a lOOps delay can be achieved. This extra analog range ensures that the FTUNE pin will be capable even under worst case conditions of covering the digital resolution. Typically the analog input will be driven by an external DAC to provide a digital control with very fine analog output steps. The final resolution of the device will be dependent on the width of the DAC chosen. To determine the voltage range necessary for the FTUNE input, the graphs provided should be used. As an example if a range of 40ps is selected to cover worst case conditions and ensure coverage of the digital range, from the 1OOE196 graph a voltage range of -3.25V to -4V would be necessary on the FTUNE pin. Obviously there are numerous voltage ranges which can be used to cover a given delay range, users are given the flexibility to determine which one best fits their designs. ECLinPS 3-71 MC1 OE196, MC100E196 Applications Information ADDRESS BUS (AD - AS) " 8 " " " " 01 01 FTUNE DO DO E196 vee LEN Chip #1 veca VEE LEN VEE I:>-- IN Input I:>-- iN Q VBB liD IN ,. ''"" "wr" z IN x Q C3 (jVCCO en en ~ ~ en en VBB "'" "'" liD Figure 1 - Cascading Interconnect Architecture Cascading Multiple E196's To increase the programmable range of the E196 internal cascade circuitry has been included. This circuitry allows for the cascading of multiple E196's without the need for any external gating. Furthermore this capability requires only one more address line per added E196. Obviously cascading multiple POC's will result in a larger programmable range however this increase is at the expense of a longer minimum delay. Figure 1 illustrates the interconnect scheme for cascading two EI96's. As can be seen, this scheme can easily be expanded for larger E196 chains. The 07 input of the E196 is the cacade control pin. With the interconnect scheme of Figure 1 when 07 is asserted it signals the need for a larger programmable range than is acheivable with a single device. An expansion of the latch section of the block diagram is pictured below. Use of this diagram will simplify the explanation of how the cascade circuitry works. When 07 of chip #1 above is low the cascade output will also be low while the cascade bar output will be a logical high. In this condition the SET MIN pin of chip #2 will be asserted and thus all of the latches of chip #2 will be reset and the device will be set at its minimum delay. Since the RESET and SET inputs of the latches are overriding any changes on the AO-A6 address bus will not affect the operation of chip #2. Chip#1 on the other hand will have both SETMIN and SET MAX de-asserted so that its delay will be controlled entirely by the address bus AO-A6. If the delay needed is greater than can be acheived with 31.75 gate delays (1111111 on the AO-A6 address bus) 07 will be asserted to signal the need to cascade thedelaytothe next E196 device. When 07 is asserted the SET MIN pin of chip #2 will be de-asserted and the delay will be controlled by the AO-A6 address bus. Chip #1 on the other hand will have its SET MAX pin asserted resulting in the device delay to be independent of the AO-A6 address bus. When the SET MAX pin of chip #1 is asserted the DO and 01 latches will be reset while the rest of the latches will be set. In addition, to maintain monotonicity an additional gate delay is selected in the cascade circuitry. As a result when 07 of chip #1 is asserted the delay increases from 31.75gatesto 32 gates. A 32 gate delay is the maximum delay setting for the E196. When cacsading multiple POC's it will prove more cost effective to use a single E196 for the MSB of the chain while using E195 for the lower order bits. This is due to the fact that only one fine tune input is needed to further reduce the delay step resolution. To sel"1}""''' £F Bit7 07 07 CASC'D' -CASCADE Co" SET MIN -+.....l---+....l.---+-...l...---I-...L-_--l_L__-+_L__--1---.! SETMAX-....L----...L----.L.----L----1-----.l.----- Figure 2 - ExpanSion of the Latch Section of the E195 Block Diagram ECLinPS 3-72 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA Advanced Information MC10E197 2:7 and 1:7 Rll Format Compatible Fully Integrated VCO for 50Mb/s Operation External VCO Input for Higher Operating Frequency Anti-equivocation Circuitry to Ensure Pll lock The MCl OE197 is an integrated data separator designed for use in high speed hard disk drive applications. With data rate capabilities of up to 50Mb/s the device is ideally suited for today's and future state-of-the-art hard disk designs. The E197 is typically driven by a pulse detector which reads the magnetic information from the storage disk and changes it into ECl pulses. The device is capable of operating on both 2:7 and 1 :7 Rll coding schemes. Note that the E197 does not do any decoding but rather prepares the disk data for decoding by another device. For applications with higher data rate needs, such as tape drive systems, the device accepts an external VCO. The frequency capability of the integrated VCO is the factor which limits the device to 50Mb/s. A special anti-equivocation circuit has been employed to ensure timely lock-up when the arriving data and VCO edges are coincident. Unlike the majority of the devices in the ECLinPS family, the E197 is available in only 10KH compatible ECL. The device is available in the standard 28-lead PlCC. Since the E197 contains both analog and digital circuitry, separate supply and ground pins have been provided to minimize noise coupling inside the device. The device can operate on either standard negative ECl supplies or as is more common on positive voltage supplies. Data Separator PINOUT: 28-Lead PLCC (TOP VIEW) 0 z > 25 0 () > (5 () () () z 24 () > () N '" '" '">'" 22 23 21 20 8 () () > 19 TEST RDCLK EXTVeo RDCLK ENveo vee VEE RDATA ACQ RDATA PUMPUP TYPE RSETDN RDEN 10 :5 ~ '" II: LOGIC SYMBOL 0 I~ "'" II: z [i' ::; => D.. 11 6 () () > RDEN----------------------, REFCLK------------fp;;;~;;;;:;] PUMPUP VCOIN PUMPDN EXTVCO RSETUP ENVCO ASETDN RAWD RDATA ACQ TYPE RDCLK This document contains information on a new product. Specifications and information herein are subject to change without notice. ECLinPS 3-73 MC10E197 Pin Description REFCLK Reference clock equivalent to one clock cycle per decoding window. RDEN Enable data synchronizer when HIGH. When LOW enable the phase/ frequency detector steered by REFCLK. RAWD Data Input to Synchronizer logic. VCOIN VCO control voltage input CAPI/CAP2 VCO frequency controlling capacitor inputs --- ENVCO VCO select pin. LOW selects the internal VCO and HIGH selects the external veo input. Pin floats LOW when left open. EXTVCO External VCO pin selected when ENVCO is HIGH -- ACO Aquisistion circuitry select pin. This pin must be driven HIGH at the end of the data sync field for some sync field types. TYPE Selects between the two types of commonly used sync fields. When LOW it selects a sync field interspersed with 3 zeroes ( 2:7 RLL code). When HIGH it selects a sync field interspersed with 2 zeroes (1:7 RLL code). TEST Input included to initialize the clock flip-flop for test purposes only. Pin should be left open (LOW) in actual application. PUMPUP Open collector charge pump output for the signal pump PUMPDN Open collector charge pump output for the reference pump RSETUP Current setting resistor for the signal pump RSETDN Current setting resistor for the reference pump RDATA Synchronized data output RDCLK Synchronized clock output Vee. Veeo. Veevco Most positive supply rails. Digital and analog supplies are independent on chip VEE'VEEVeO Most negative supply rails. Digital and analog supplies are independent on chip ECLinPS 3-74 MC10E197 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = GND or Vee = 4.75V to 5.25V; VEE = GND 25°C O°C Symbol Characteristic min max typ min typ 150 85°C max min typ max Unit 150 J.lA 1 j.!A 1 Condition I'H Input HIGH Current I'L Input LOW Current 0.5 lEE Power Supply Current 90 180 90 180 90 180 mA ISET Charge Pump Bias Current 0.5 5 0.5 5 0.5 5 mA 2 lOUT Charge Pump Output 1 J.lA 3 Vee V max Unit 150 0.5 0.5 1 1 Leakage Current V AeT PUMPUP/PUMPDN V ee -2.5 Vee Vee -2.5 Vee Vee -2.5 Active Voltage Range 10 KH LOGIC LEVELS DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo =Veco,=Veeveo= GND OOG Symbol Characteristic min 25°C typ max min -980 typ 85°C max min typ -810 -910 -720 mV V OH Output HIGH Voltage -1020 -840 VOL Output LOW Voltage -1950 -1630 -1950 -1630 -1950 -1595 mV V ,H Input HIGH Voltage -1170 -840 -1130 -810 -1060 -720 mV V ,L Input LOW Voltage -1950 -1480 -1950 -1480 -1950 -1445 mV max Unit Condition POSITIVE EMITTER COUPLED LOGIC LEVELS DC Characteristics: VEE =V EEveo=GND; Vee = V ecoo=VCCo,=Veeveo=+5 volts' 25°C O°C min 85°C max Characteristic min VOH Output HIGH Voltage 3980 4160 4020 4190 4090 4280 mV VOL Output LOW Voltage 3050 3370 3050 3370 3050 3405 mV V ,H Input HIGH Voltage 3830 4160 3870 4190 3940 4280 mV V ,L Input LOW Voltage 3050 3520 3050 3050 3050 3555 mV Symbol typ max 'VOH and VOL levels will vary 1: 1 with Vee ECLinPS 3-75 typ min typ Condition MC10E197 AC Characteristics: VEE = VEE(min) to VEE(max); Vce = GND or Vee = 4.75V to 5.25V; VEE = GND O°C Symbo ts Time from RDATA Valid to Rising Edge of RDCLK tH Time from Rising Edge of RDCLK to RDATA invalid tSKEW fvco 1. 2. 3. 4. 5. 6. 7. Characteristic min 25°C max Tvco-500 150 Tuning Ratio 1.53 300 300 1.53 max Tvco 150 1.87 min T vco-500 Tvco 300 Frequency of the VCO 85°C max Tvco -500 Tvco Skew Between RDATA and RDATA min 150 1.87 1.53 Unit ps 4,7 ps 4,7 ps MHz 1.87 Condition 5 6 Applies to the Input current lor each Input except VeOIN For a nominal set current 01 3.72 mA, the resistor values lor RSETUP and RSETDN should be 130Q(0.1%). Assuming no variation between these two resistors the current match between the PUMPUP and PUMPDN output signals should be within ±3%. ISET is calculated as (VEE + 1.3v -V BE)/R; where R is RSETUP or RSETDN and a nominal value lor VBE is 0.85 volts. Output leakage current 01 the PUMPUP or PUMPDN output signals when at a LOW level. TVCD is the period 01 the yeo. The veo Irequency determined with VeOIN = VEE + 0.5volts and using a 1OpF tuning capacitor. The tuning ratio is delined as the ratio 01 I VCDMAX to IVCDMIN ,where IVCOMAX is measured at VeOIN = 1.3v + VEE and IVCDMAX is measured at VeOIN = 2.6v + VEE' Setup and hold timing diagrams: If.o"~--Is-If.o·~--IH-I Applications Information General Operation Operation The E197 is a phase-locked loop circuit consisting of an internal VCO, a Data Phase detector with associated acquisition circuitry, and a PhaselFrequency detector(Fig. 1). In addition, an enable pin(ENVCO) is provided to disable the internal VCO and enable the external VCO input. Hence, the user has the option of supplying the VCO signal. The E197 contains two phase detectors: a data phase detector for synchronizing to the non-periodic pulses in the read data stream during the data read mode of operation, and a phasel frequency detector for frequency (and phase) locking to an external reference clock during the "idle" mode of operation. The read enable (RDEN) pin muxes between these two detectors. ECUnPS 3-76 MC10E197 Data Read Mode The data pins (RAWD) are enabled when the RDEN pin is placed at a logic high level, thus enabling the Data Phase detector(Fig. 1) and initiating the data read mode. In this mode the loop is servoed by the timing information taken from the positive edges of the input data pulses. This phase detector samples positive edges from the RAWD signal and generates both a pump up and pump down pulse from any edge of the input data pulse. The leading edge of the pump up pulse is time modulated by the leading edge of the data signal, whereas the rising edge of the pump up pulse is generated synchronous to the veo clock. The falling edge of the pump down pulse is synchronous to the rising edge of the veo clock and the rising edge of the pump down signal is synchronous to the falling edge of the veo clock. Since both edges of the veo are used the internal clock a duty cycle of 50%. This pulse width modulation technique is used to generate the servoing signal which drives the yeo. The pump down signal is a reference pulse which is included to provide an evenly balanced differential system, thereby allowing the synthesis of a veo input control signal after appropriate signal processing by the loop filter. By using suitable external filter circuitry, a control signal for input into the veo can be generated by inverting the pump down signal, summing the inverted signal with the pump up signal and averaging the result. The polarity of this control signal is defined as zero when the data edges lead the clock by a half clock cycle. If the data edges are advanced with respect to the zero polarity data/yeO edge relationship, the control signal is defined to have a negative polarity; whereas if the veo is advanced with respect to the zero polarity datalVeO edge relationship, the control signal is defined to have a positive polarity. If there is no data edge present at the RAWD input, the corresponding pump up and pump down outputs are not generated and the resulting control output is zero. Acquisition Circuitry The acquisition circuitry is provided to assist the data phase detector in phase locking to the sync field that precedes the data. Forthe case in which lock-up is attempted when the data edges are coincident with the veo edges, the pump down signal may enter an indeterminate state for an unacceptably long period due to the violation of internal set up and hold times. After an initial pump down pulse, the circuit blocks successive pump down pulses, and inserts extra pump up pulses, during portions of the sync field that are known to contain zeros. Thus the data phase detector is forced to have a nonzero output during the lock-up period, and the restoring force ensures correction of the loop within an acceptable time. Hence, this circuitry provides a quasi-deterministic pump down output signal, under the condition of coincident data and veo edg~lIowing lock-up to occur without excessive delays. The AeO line is provided to disable(disable = HIGH) the acquisition circuit during the data portion of a sector block. Typically, this circuit is enabled at the beginning of the sync field by a one-shot timer to ensure a timely lock-up. The TYPE line allows the choice between two sync field preamble types: transitions interspersed with two zeros between transitions, or three zeros between transitions. These types of sync fields are used with the 1 :7 and 2:7 RLL coding schemes, respectively. Idle Mode In the absence of data or when the drive is writing to the disk, PLL servoing is accomplished by pulling the read enable line (RDEN) low and providing a reference clock via the REFCLK pins. The condition whereby RDEN is low selects the Phase/FrequenCy detector(Fig. 1) and the 1OE197 is said to be operating in the "idle mode." In order to function as a frequency detector the input waveform must be periodic. The pump up and pump down pulses from the Phase/Frequency detector will have the same frequency, phase and pulse width only when the two clocks that are being compared have their positive edges aligned and are of the same frequency. As with the data phase detector, by using suitable external filter circuitry, a VCO input control signal can be generated by inverting the pump down signal, summing the inverted signal with the pump up signal and averaging the result. The polarity of this control signal is defined as zero when all positive edges of both clocks are coincident. For the case in which the frequencies of the two clocks are the same but the clock edges of the reference clock are slightly advanced with respect to the veo clock, the control signal is defined to have a positive polarity. A control signal with negative polarity occurs when the edges of the reference clock are delayed with respect to those of the yeo. If the frequencies of the two clocks are different, the clock with the most edges per unit time will initiate the most pulses and the polarity of the detector will reflect the frequency error. Thus, when the reference clock is higher in frequency than the veo clock the polarity of the control signal is positive; whereas a control signal with negative polarity occurs when the frequency of the reference clock is lower than the veo clock. Phase Lock Loop Theory Introduction f---..-----ox"I'i Phase lock loop (PLL) circuits are fundamentally feedback systems used to synchronize the frequency of an oscillator to an incoming signal. In addition to frequency synchronization, the PLL circuitry is designed to minimize the phase difference between the system input and output signals. A block diagram of a feedback control system is shown in Figure 1. where: A(s) is the product of the feed-forward transfer functions. Figure 1 - Feedback System ECLinPS 3-77 MC10E197 B(s) is the product of the feedback transfer functions. NRZ Data Sequence The transfer function for this closed loop system is A(s) 1 + A(s)B(s) Typically, phase lock loops are modeled as feedback systems connected in a unity feedback configuration(B(s)=l) with a phase detector, a VCO(voltage controlled oscillator), and a loop filter in the feed-forward path, A(s). Figure 2 illustrates a phase lock loop as a feedback control system in block diagram form. Code Sequence 00 01 1000 0100 100 101 111 001000 100100 000100 1100 1101 00001000 00100100 Table 1 - 2:7 RLL Encoding Table NRZ Data Sequence Figure 2 - Phase Lock Loop Block Diagram XOl 010 XOO 1100 1101 1110 1111 The closed loop transfer function is: 1 + Kq, Code Sequence 00 01 10 K --f F(s) 010001 XOOOOO XOOOOl 010000 An X in the leading bit of a code sequence is assigned the complement of the bit where: K,= the phase detector gain. Ko= the VCO gain. Since the VCO introduces a pole at the origin of the s-plane, Ko is divided by s. F(s) = the transfer function of the loop filter. Table 2 - 1 :7 RLL Encoding Table Sync Pattern The 10E197 is designed to implementthe phase detector and VCO functions in a unity feedback loop, while allowing the user to select the desired filter function. Gain Constants As mentioned, each of the three sections in the phase lock loop block diagram has an associated open loop gain constant. Further, the gain constant of the filter circuitry is composed of the product of three gain constants, one for each filter subsection. The open loop gain constant of the feed-forward path is given by eqt. 1 Read Mode Idle Mode 2:7 121 mV/radian 484 mV/radian 1:7 161 mV/radian 483 mV/radian Table 3 - Phase Detector Gain Constants VCO Gain Constant The gain of the VCO is a function of the tuning capacitor. For a value of 10 pF a nominal value of the gain,Ko,is 20MHz per volt. and obtained by performing a root locus analysis. Filter Circuitry Gain Constant(s) Phase Detector Gain Constant The open loop gain constant of the filter circuitry is given by: The gain of the phase detector is a function of the operating mode and the data pattern. The 1OE197 provides data separation for signals encoded in 2:7 or 1:7 RLL encoding schemes; hence Tables 1 and 2 are coding tables for these schemes. Table 3 lists nominal phase detector gains for both 2:7 and 1:7 sync fields. eqt. 2 The individual gain constants are defined in the appropriate subsections of this document. ECLinPS 3-78 MC10E197 Loop Filter The two major functions of the loop filter are to remove any noise or high frequency components present in the phase detector output signal, and more importantly to control the characteristics which determine the dynamic response of the phase lock loop i.e. capture range, loop bandwidth, capture time, and transient response. Although a variety of loop filter. configurations exist, this section will only describe a filter capable of performing the signal processing as described in the Data Read Mode and the Idle Mode sections. The loop filter consists of a differential summing amplifier cascaded with an augmenting integrator which drives the VeOIN input to the 1OE197 through a resistor divider network(Fig. 3). The transfer function and the element valL\es for the loop filter are derived by dividing the filter into three cascaded subsections: filter input, augmenting integrator, and the voltage divider network(Fig.4). ~ (s) Figure 4 - Loop Filter Block Diagram A root locus analysis is performed on the open loop transfer function to determine the final pole-zero locations and the open loop gai n constant for the phase lock loop. Note that the open loop gain constant impacts the crossover frequency and that a lower frequency crossover point means a much more efficient filter. Once these positions and constants are determined the component values may be calculated. R, Loop Filter Transfer Function MC34182 The open loop transfer function of the phase lock loop is the product of each individual filter subsection, as well as the phase detector and veo. Thus, the open loop filter transfer function is: + 1 Ko Fo (s); K * -s- * F1 (s) * ~ (s) * Fd(s) R, where: Vccvco F1(S) K1 (s + P1) ~ (s) Fd(s) K, Kd [s2+ (21;0001 ) s + 0)~11 Filter Input (s + z) [s2+(21;00 )s+0)21 02 02 s 1 (s + P2 ) Figure 5 - Filter Input Sunsection The primary function of the filter input subsection is to convert the output of the phase detector into a single ended signal for subsequent processing by the integrator Circuitry. This subsection consists of the 1OE197 charge pump current sinks, two shunt capacitors, and a differential summing amplifier(Fig. 5). PUMPUPo---->r-vVV-~--~~---~-~~~-~~-~ Rv MC34182 + Ro R, PUMPDN o-----r~'V'v-_+ Vccvco R, VEEVCO Vccvco Figure 3 - Loop Filter Circuitry ECLinPS 3-79 MC10E197 RSETUP 0-___.------, 4640 ~--_--o 464n 4640: op-amp are obtained from the data sheets. Typically, op-amp manufacturers do not provide information on the location of the second open loop pole, however it can be approximated by measuring the roll off olthe op-amp in the open loop configuration; the second pole is located where the gain begins to decrease at a rate of 40dB per decade. The inclusion of both poles in the differential summing amplifier transfer function becomes important when closing the feedback path around the op-amp because the poles migrate; and this migration must be accounted for to accurately determine the phase lock loop transient performance. RSETDN 4640 Veevco Typically theop-amp poles can be approximated by a pole pair occurring as a complex conjugate pair making an angle of 45° to the real axis of the complex frequency plane. Two constraints on the selection of the op-amp pole pair are that the poles lie beyond the crossover frequency and they are pOSitioned for near unity gain operation. Performing a root locus analysis on the opamp open loop configuration and adhering to the two constraints yields the pole pOSitions contributed by the op-amp. Electronic Switch Figure 6 - Dual Bandwidth Current Source Implementation Hence, this portion of the filter circuit contributes a real pole and two complex poles to the overall loop transfer function F(s). Before these pole locations are selected, appropriate values for the current setting resistors(RSETUP and RSETDN) must be ascertained. The goal in choosing these resistor values is to maximize the gain of the filter input subsection while ensuring the charge pump output transistors operate in the active mode. The filter input gain is maximized for a charge pump current of 1.1 mA; a value of 4640 for both RSETUP and RSETDN yields a nominal charge pump current of 1.1 mAo It should be noted that a dual bandwidth implementation of the phase lock loop may be achieved by modifying the current setting resistors such that an electroniC switch enables one of two resistor configurations. Figure6 shows a circuit configuration capable of providing this dual bandwidth function. Analysis of the filter input circuitry yields the transfer function: Determination of Element Values Since the difference amplifier is configured to operate as a differential summer the resistor values associated with the amplifier are of equal value. Further, the typical input resistance to the summing amplifier is 1 kn; thus the op-amp resistors are set at 1 kO. Having set the input resistance to the op-amp and selected the position of the real pole, the value of the shunt capacitors is determined using the following relationship: Ip 1 I= -2-1t~~-C1 IN eql. 4 Augmenting Integrator The augmenting integrator consists of an active filter with a lag-lead network in the feedback path(Fig. 7). The gain constant is defined as: K =A __1_ 1 1 CIN CA RA RIA "'N eqt. 3 MC34182 where: + A 1 = op-amp gain constant for the selected pole positions. l 1 RIA CIN = phase detector shunt capacitor. VCCVCO The real pole is a function of the input resistance to the op-amp and the shunt capacitors connected to the phase detector output. For stability the real pole must be placed beyond the unity gain frequency; hence, this pole is typically placed midway between the unity crossover and phase detector sampling frequency, which should be about ten times greater. The second order pole set arises from the two pole model for an op-amp. The open loop gain and the first open loop pole for the V02 Figure 7 - Integrator Subsection Analysis of this portion of the filter circuit yields the transfer function: ECLinPS 3-80 * _1_ * s (s + z) [s2+(2~0))S + 0)2 02 02 1 MC10E197 Voltage Divider The gain constant is defined as: The input range to the VCOIN input is from 1.3v + VEE to 2.6v + VEE' hence the output from the augmenting amplifier section must be attenuated to meet the VCOIN constraints. A simple voltage divider network provides the necessary attenuation(Fig. eqt. 5 where: 8). A I ; op-amp gain constant for selected Rv ~No-------~~~--~------~~----~ pole positions. R A ; integrator feedback resistor. Ro RIA; integrator input resistor. The integrator circuit introduces a zero, a pole at the origin, and a second order pole set as described by the two pole model for an op-amp. As in the case of the differential summing amplifier, we assume the op-amp pole pair occur as a complex conjugate pair making an angle of 45° to the real axis of the complex frequency plane; are positioned for near unity gain operation; and are located beyond the crossover frequency. Since both the summing and integrating op-amps are realized by the same type of opamp(MC34182D), the open loop pole positions for both amplifiers will be the same. Further, the loop transfer function contains two poles located at the origin, one introduced by the integrator and the other by the VCO; hence a zero is necessary to compensate for the phase shift produced by these poles and ensure loop stability. The op-amp will be stable if the crossover pOint occurs before the transfer function phase angle becomes 180°. The zero should be positioned much less than one decade before the unity gain frequency. As in the case of the filter input circuitry, the poles and zero from this analysis will be used as open loop poles and a zero when performing the root locus analysis for the complete system. Figure 8 - Voltage Divider Subsection In addition, a shunt filter capacitor connected between the VCOIN input pin and VEE provides the voltage divider subsection with a single time constant transfer function that adds a pole to the overall loop filter. The transfer function for the voltage divider network is: The gain constant, Kd, is defined as: eqt. 9 Determination of Element Values The location of the zero is used to determine the element values for the augmenting integrator. The value of the capacitor, CA, is selected to provide adequate charge storage when the loop is not sampling data. A value of 0.1 ~F is sufficient for most applications; this value may be increased when the RDCLK frequency is much lower than 4 MHz. The value of RA is governed by: I I= z -2-1t-R"'-A--A C eqt. 6 The value of Kd is easily extracted by rearranging Equation 1 : Kol eqt. 10 Kd ; K * K * K * K <\l 0 1 I The gain constant Kd is set such that the output from the integrator circuit is within the range 1.3v +VEE to 2.6v +VEE. The pole for the voltage divider network should be positioned an octave beyond that for the filter input. Determination of Element Values For unity gain operation of the integrating op-amp the value of R'A is selected such that: eqt. 7 It should be noted that although the zero can be tuned by varying either RA or C A' caution must be exercised when adjusting the zero by varying CA because the integrator gain is also a function of CA. Further, the gain of the loop filter can be adjusted by changing the integrator input resistor R,A. Once the pole location and the gain constant Kd are established the resistor values for the voltage divider network are determined using the design guidelines mentioned above and from the following relationship: Having determined the resistor values, the filter capacitor is calculated by rearranging Equation 9: ECLinPS 3-81 MC10E197. Finally, a bias diode is included in the voltage divider network to provide temperature compensation. The finite resistance of this diode is neglected for these calculations. eqt. 9a Calculations For a 2:7 Coding Scheme Introduction positioned midway between the unity crossover point and the phase detector sampling frequency. Hence, the open loop filter input pole position is selected as: The circuit component values are calculated for a 2:7 coding scheme employing a data rate of 23 MbiVsec. Since the number of bits is doubled when the data is encoded, the data clock is at half the frequency of the RDCLK signal. Thus the operating frequency for these calculations is 46 MHz. Further, the pole and zero positions are a function of the data rate, hence the component values derived by these calculations must be scaled if a different operating frequency is used. Finally, it should be noted that the values are optimized for settling time. The analysis is divided into three parts: static pole positioning, dynamic pole positioning, and dynamic zero positioning. Dynamic poles and zeros are those which the designer may position, to yield the desired dynamic response, through the judicious choice of element values. Static poles are not directly controlled by the choice of component values. The voltage divider pole is set approximately one octave higher than the filter input pole. Thus the open loop voltage divider pole position is picked to be: P~ = -2.57 MHz Dynamic Zero Finally, the zero is positioned much less than one decade before the crossover frequency; for this design the zero is placed at: Static Poles z =-311 Hz Each op-amp introduces a pair of "static" complex conjugate poles which must lie beyond the crossover frequency. As obtained from the data sheets and laboratory measurements the two open loop poles for the MC34182D are: Once the dynamic pole and zero positions have been determined, the phase margin is determined using a Bode plot; if the phase margin is not sufficient, the dynamic poles may be moved to improve the phase margin. Finally, a root locus analysis is performed to obtain the optimum closed loop pole positions forthe dynamic characteristics of interest. P;a=-0.1 Hz P;b= -11.2 Hz Component Values Performing a root locus analysis and following the two guidelines previously stated, an acceptable pole set is: Having determined the closed loop pole and zero positions the component values are calculated. From the root locus analysis the dynamic pole and zero positions are: P1a = -5.65 + j5.65 MHz P1b = -5.65 - j5.65 MHz P 1 = -573 kHz Both op-amps introduce a set 01 static complex-conjugate poles at these positions for a total of four poles. Further, the loop gain for each op-amp associated with these pole positions is determined from the root locus analysis to be: P 2 = -3.06 MHz z =-311 Hz Filter Input Subsection Rearranging Equation 4: In addition to the op-amps, the integrator and the VCO each contribute a static pole at the origin. Thus, there are a total of six static poles. Dynamic Poles The filter input and the voltage divider sections each contribute a dynamic pole. As stated previously, the filter input pole should be and substituting 573 kHz for the pole position and 1 k.Q for the resistor value yields: ECLifiPS 3-82 MC10E197 Having determined the gain constant Kd, the value of R, is selected such that the constraints R, > Ro and: Augmenting Integrator Subsection f«J Rearranging Equation 6: _ 21tiP2iare fulfilled. The pole position P, is determined from the root locus analysis to be: and substituting 311 Hz for the zero position and 0.1 fl F for the capacitor value yields: RA = 5.11 P 2 : -3.06 MHz Hence, R, is selected to be: kn Rv = 2.15 kO From Equation 7 the value for the other resistors associated with the integrator op-amp are set equal to RA : and Ro is calculated to be: Ro RIA: R A : kn 5.11 = 700 0 Finally, using Equation 8a: Voltage Divider Subsection eqt. 8a The element values for the voltage divider network are calculated using the relationships presented in Equations 8, 9, and 10 with the constraintthat this divider network must produce a voltage that lies within the range 1.3v + VEE to 2.6v + VEE' Restating Equation 9, Note that the voltage divider section can be used to set the gain, but the designer is cautioned to be sure the input value to VCOIN is within the correct range. Kd : K • K • K' K a Ro and: and substituting 311 Hz for the zero position and 0.1 J.IF for the capacitor value yields: are fulfilled. The pole position P2 is determined from the root locus analysis to be: RA = 5.11 kQ P 2 = -2.73 MHz From Equation 7 the value for the other resistors associated with the integrator op-amp are set equal to RA : Hence, R, is selected to be: Rv = 2.15kQ and Ro is calculated to be: Voltage Divider Subsection Ro The element values for the voltage divider network are calculated using the relationships presented in Equations 8, 9, and 10 with the constraint that this divider network must produce a voltage that lies within the range 1.3v + VEE to 2.6v + VEE. Restating Equation 9, = 453 n Finally, using Equation 8a: eqt. 8a the capacitor value Cd is calculated to be: From the root locus analysis Ko' is determined to be: EGLinPS 3-85 MC10E197 Again, note the voltage divider section can be used to setthe gain, but the designer is cautioned to be sure the input value to VCOIN is within the correct range. Thus the element values for the filter are: Filter Input Subsection: = 588 pF Component Scaling C IN As mentioned, these design equations were developed for a data rate of 20 Mbitlsec. If the data rate is different from the nominal design value the reactive elements must be scaled accordingly. The following equations are provided are to facilitate scaling and were derived with the assumptions that a 1:7 coding scheme is used and that the RDCLK signal is twice the frequency of the data clock: R1 = 1 kQ Integrator Subsection: CA = 0.1 J.tf R A =5.11kQ CIN=294*~ f Cd = 156 * 3~ (pF) eqt. 13 (pF) eqt. 14 R IA = 5.11 kQ Voltage Divider Subsection: where f is the RDCLK frequency in MHz. Cd = 312 Example for an 10 Mbit/sec Data Rate Rv = 2.15 kQ As an example of scaling, assume the given filter and a 1:7 code are used but the data rate is 10 Mbitlsec. The dynamic pole positions, and therefore the bandwidth of the loop filter,are a function of the data rate. Thus a slower data rate will force the dynamic poles and the bandwidth to move to a lower frequency. From Equation 13 the value of C'N is: Ro = 453 Q C IN = 588 pF and from Equation 14 the value of Cd is: C d =312 pF pF Note, the poles P, and P2 are now located at: P 1 = -271 kHz P2 = -1.36 MHz And, the open loop filter unity crossover point is at 300 kHz. As in the case of the 2:7 coding scheme, the gain can be adjusted by changing the value of R'A and the value of Cd' Varying the gain by changing Cd is not recommended because this will also move the poles, hence affect the dynamic performance of the filter. ECLinPS 3-86 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E212 MC100E212 Scannable Version El12 Driver 1025ps Max. ClK to Output Dual Differential Outputs Master Reset Extended 100E VEE Range of -4.2V to -5.46V Internal 751<.0 Input Pulldown Resistors The MC 1OE/l OOE2l2 is a scannable registered ECl driver typically used as a fan-out memory address driver for ECl cache driving. In a VlSI array based CPU design, use of the E2l2 allows the user to conserve array output cell functionality and also output pins. The input shift register is designed with control logic which greatly facilitates its use in boundary scan applications. 3-BIT SCANNABLE REGISTERED ADDRESS DRIVER PINOUT: 28-LEAD PLCC (TOP VIEW) SHIFT MR 25 24 NC S·OUT VCCO Q2b Q2a 23 19 22 21 20 LOGIC SYMBOL lOAD Q2b ClK Q2a D2 VCC VEE Qlb Dl Q1a DO Qlb ,------S·OUT Q2b Q2a D2 Qla S·IN 10 NC VCCO QOa QOb QOa Qlb Qla 11 COb VCCO Dl PIN NAMES PIN FUNCTION 00-02 S-IN Scan Input -- lOAD SHIFT ClK MR S-OUT QOb QOa Data Inputs DO lOAD/HOLD Control Scan Control Clock Reset O[O:2]a, O[O:2]b Scan Output True Outputs O[O:2]a, O[O:2]b Inverting Outputs S·IN LOAD SHIFT ClK MR ECLinPS 3-87 QOa QOb MC10E212 MC100E212 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veco = GND 25°C O°C Symbol - Characteristic min typ max min 150 I'H Input HIGH-Current lEE Power Supply Current tpHL min typ 150 max Unit 150 J.LA 96 80 96 80 96 100E 80 96 80 96 92 110 VEE = VEE(min) to VEdmax); Characteristic Vcc = Vcco = GND 25°C 85°C min typ max' min typ max min typ max 575 800 1025 575 800 1025 575 800 1025 MR 575 800 1025 575 800 1025 575 800 1025 CLKto S-OUT 575 800 1025 575 800 1025 575 800 1025 D 175 25 175 25 175 25 SHIFT -LOAD 150 -50 150 -50 150 -50 225 225 150 150 50 -50 225 S-IN 50 -50 150 50 -50 D 250 25 250 25 250 25 100 Propagation Delay to Output CLK ps SHIFT 300 100 300 100 300 LOAD 225 0 225 0' 225 0 S-IN 300 100 300 100 300 100 600 350 600 350 600 350 ps Reset Recovery tAR Condition ps Hold Time th Unit ps Setup Time 's Condition mA 80 O°C tpLH 85°C max 10E AC Characteristics: Symbol typ tSKEW Within-Device Skew 100 100 100 ps 1 tSKEW Within-Gate Skew 50 50 50 ps 2 Rise/Fall Times tr ~ .. ps 20 - 80% 275 ., 425 650 275 425 650 275 425 650 1. Within-device skew IS defined as Identical tranSitions on similar paths through a device. 2. Within-gate skew is defined as the difference in delays between various outputs of a gate when driven from the same input. FUNCTION TABLE LOAD SHIFT MR MODE L L L Load H L L Hold X X H L Shift X H Reset ECLinPS 3-88 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • MC10E241 MC100E241 SHIFT overrides HOLD/lOAD Control 1000 ps Max. ClK to Q Asynchronous Master Reset Pin-Compatible with E141 Extended 100E VEE Range of - 4.2 V to - 5.46 V 75 kO Input Pulldown Resistors The MC10E/l00E241 is an 8-bit shiftable register. Unlike a standard universal shift register such as the E141, the E241 features internal data feedback organized so that the SHIFT control overrides the HOLD/lOAD control. This enables the normal operations of HOLD and lOAD to be toggled with a single control line without the need for external gating. It also enables switching to scan mode with the single SHIFT control line. The eight inputs 00-07 accept parallel input data, while S-IN accepts serial input data when in shift mode. Data is accepted a set-up time before the positive-going edge of ClK; shifting is also accomplished on the positive clock edge. A HIGH on the Master Reset pin (MR) asynchronously resets all the registers to zero. 8-BIT SCANNABLE REGISTER PINOUT: 28-lEAD PLCC (TOP VIEW) SElO Ne 0] Os 05 Veeo LOGIC SYMBOL 07 S-IN ---;::::;:=1~---==~ Os SEll 00---1 05 MR Vee VEE NC S·IN Veco Do (J.t 01 03 02 03 04 VCCO 00 01 02 HOLD/LOAD - _ _-' SHIFT _ _ _ _ _ _....J PIN NAMES Pin 00-0 7 S-IN SElO SEll ClK MR 00-0 7 ClK - -_ _ _ _ _ _---1 Function MR---------~ Parallel Data Inputs Serial Data Input SHIFT Control HOLD/LOAD Control Clock Master Reset Data Outputs ECLinPS 3-89 MC10E241, MC100E241 DC Characteristics: .. ,. VEE =V EE(min) to VEE(max); Vee . . =Vceo = GND O°C Symbol - min Characteristic liH . Input HIGH Currsnt lEE Power Supply Current max min tpLH tpHL ts Unit 150 IlA 125 150 100E 125 150 125 150 144. 173 VEE =VEE(min) to VEE(max); Characteristic min typ Max. Shift Frequency 700 900 Vee =Vceo = GND 25'C Clk 625 750 MR 600 725 175 max min typ 700 900 975 625 750 975 600 725 25 175 25 85'C max min typ max 700 900 975 625 750 975 975 600 725 975 175 25 ps SELO (SHIFT) 350 200 350 200 350 200 SEL 1 (HOLD/LOAD) 400 250 400 250 400 S-IN 125 -100 125 -100 125 250 ;100 200 -25 200 -25 200 -25 Hold Time ps 100 -200 100 -200 100 -200 SEL 1 (HOLD/LOAD) 50 -250 50 -250 50 -250 S-IN 300 100 300 100 300 100 900 600 900 600 900 600 Clk, MR ps ps 400 Within-Device Skew 400 60 400 60 60 ps ps Rise / Fall Times 20 - 800;:0 Condition ps Setup Time Minimum Pulse Width Unit MHz Propagation Delay to Output tpw Condition mA 150 Reset Recovery Time .. max 125 tRR tr tf typ 150 D SELO (SHIFT) tSKEW min 125 D th max 150 150 O°C fSHIFT typ 10E AC Characteristics: Symbol 85'C 25'C typ 300 525 800 300 .. 525 800 1. Within-device skew IS defined as Identical transitions on similar paths through a device ECLinPS 3-90 300 525 800 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • MC10E256 MC100E256 950 ps Max. D to Output 850 ps Max. LEN to Output Split Select Differential Outputs Extended 100E VEE Range of - 4.2 V to - 5.46 V 75 kD Input Pulldown Resistors The Me1 OE/1 00E256 contains three 4: 1 multiplexers followed by transparent latches with differential outputs. Separate Select controls are provided for the leading 2:1 mux pairs (see logic symbol). When the Latch Enable (LEN) is LOW, the latch is transparent, and output data is controlled by the multiplexer select controls. A logic HIGH on LEN latches the outputs. The Master Reset (MR) overrides all other controls to set the Q outputs LOW. LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) Dlb Dla D2d D2e D2b D2a Veeo 22 21 20 19 DOa DOb Q2 SELlA DOe SELlB DOd SEL2 Vee VEE 0, LEN 01 MR Veeo Ole 00 Dla Dlb Dle Old Old DOa DOb DOe DOd Veeo 00 D2a FUNCTION TABLE Pin State H H H SEL2 SELlA SEL1B D2b Operation Output cld Data Input d Data Input b Data D2e D2d PIN NAMES Pin DOx-D2x SEL lA. SEL 1B SEL2 LEN MR 00, 00-0 2, 02 3-BIT 4: 1 MUX-LATCH SEllA Function SELlB Data Inputs First-stage Select Inputs Second-stage Select Input Latch Enable Master Reset Data Outputs SEL2 LEN MR ECLinPS 3-91 MC10E256, MC100E256 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veeo= GND DoC Symbol Characteristic IIH Input HIGH Current lEE Power Supply Cu'rrent 10E 100E r--- ACCharacteristics: min typ 25°C max min 150 tpLH tpHL ts th typ , max Unit 150 IlA Condition mA 69 69 83 83 69 69 83 83 69 79 83 96 25°C 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output D SEL1 SEL2 LEN MR 400 550 450 350 350 600 775 900 1050 400 550 600 775 900 1050 400 550 600 775 900 1050 650 500 600 900 800 825 450 350 350 650 500 600 900 800 825 450 350 350 650 500 600 900 800 825 400 275 400 275 400 275 600 500 300 250 600 500 300 250 600 500 300 250 300 100 200 -275 -275 -300 -250 300 100 200 -300 -250 300 100 200 -275 -300 -250 600 700 600 700 600 Setup Time D SEL1 SEL2 Hold Time D Condition ps ps Reset Recovery Time 700 tpw Minimum Pulse Width MR 400 ps ps Within-Device Skew Rise I Fall Times 20-80% Unit ps tRR t, tf min VEE = VEE(min) to VEE(max); Vee = Veeo = GND SEL1 SEL2 tSKEW 85°C max 150 DoC Symbol typ 400 400 50 50 50 ps ps 275 475 700 275 475 700 1. Within-device skew is defined as identical transitions on similar paths through a device ECLinPS 3-92 275 475 700 1 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • • • • MC10E336 MC100E336 25 n Cutoff Bus Outputs 50 n Receiver Outputs Transmit and Receive Registers 1500 ps Max. Clock to Bus 1000 ps Max. Clock to 0 Bus Outputs Feature Internal Edge Slow-Down Capacitors Additional Package Ground Pins Extended 100E VEE Range of - 4.2 V to - 5.46 V 75 kn Input Pulldown Resistors 3-BIT REGISTERED BUS TRANSCEIVER The MC10E/100E336 contains three bus transceivers with both transmit and receive registers. The bus outputs (BUSD-BUS2) are specified for driving a 25 n bus; the receive outputs (00-02) are specified for 50 n. The bus outputs feature a normal HIGH level (VOH) and a cutoff lOW level - when lOW, the outputs go to - 2.0 V and the output emitter-follower is "off," presenting a high impedance to the bus. The bus outputs also feature edge slow-down capacitors. The Transmit Enable pins (TEN) control whether current data is held in the transmit register, or new data is loaded from the AlB inputs. A lOW on both of the Bus Enable inputs (BUSEN), when clocked through the register, disables the bus outputs to - 2.0 V. The receiver section clocks bus data into the receive registers, after gating with the Receive Enable (RXEN) input. All registers are clocked by a positive transition of ClK1 or ClK2 (or both). Additional leadframe grounding is provided through the Ground pins (GND) which should be connected to 0 V. The GND pins are not electrically connected to the chip. LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) TEN2 TEN1 B2 A2 NC VCCO 02 20 19 Ao BO 23 BUSEN1 GND BUSEN2 BUS2 VCC 01 VEE 01 ClK1 VCCO ClK2 BUS1 AO BUS2 GND 10 BO A1 B1 VCCO BUSO GND 11 00 TEN1 TEN2 RXEN D-----!--t----' BUSEN1 D---,...,[ BUSEN2 D---d.--1 ClK1 ClK2 ECLinPS 3-93 MC10E336, MC100E336 DC Characteristics: VEE = V EE(min) to VEE(max); Vee = Vceo = GND 25°C O°C Symbol VeuT IIH Characteristic min 'Cut-off Output Voltage -2.10 typ max min typ -2.03 -2.10 Input HIGH Current RXEN All Other Inputs 85°C max min typ -2.03 -2.10 max Unit -2.03 V I1A 225 150 Power Supply Current 10E 100E 'measured with VTT = -2.1 OV 225 150 225 150 mA lEE ACCharacteristics: 125 125 150 150 125 125 tpLH tpHL ts th tpw t, tf 150 150 125 144 150 173 VEE = VEE(min) to VEE(max); Vee = Veca= GND O°C Symbol Condition min typ max min Propagation Delay to Output ClktoO ClktoBUS 500 825 700 100 1250 1800 Setup Time BUS, RXEN BUSEN A, B Data TEN 150 100 300 450 Hold Time BUS, RXEN BUSEN A, BData TEN 450 500 350 200 Minimum Pulse Width Clk 400 Rise I Fall Times 20-80%(On) 20 - 80% ( BUSn Rise) 20 - 80% ( BUSn Fall ) 85°C 25°C Characteristic typ max min typ max 500 825 700 1000 1250 1800 500 825 700 1000 1250 1800 -150 -200 -50 150 150 100 300 450 -150 -200 -50 150 150 100 300 450 -150 -200 -50 150 150 200 50 -150 450 500 350 200 150 200 50 -150 450 500 350 200 150 200 50 -150 Unit ps ps ps ps 400 400 ps 300 500 300 450 800 500 700 1000 800 300 500 300 ECLinPS 3-94 450 800 500 700 1000 800 300 500 300 450 800 500 700 1000 800 Condition MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E337 MC100E337 Advance Information • • • • • • • • • • • • Scannable Version of E336 25 0 Cutoff Bus Outputs 50 0 Receiver Outputs Scannable Registers Sync. and Async. Bus Enables Non-Inverting Data Path 1500 ps Max. Clock to Bus (Data Transmit) 1000 ps Max. Clock to 0 (Data Receive) Bus Outputs Feature Internal Edge Slow-Down Capacitors Additional Package Ground Pins Extended 100E VEE Range of -4.2 V to -5.46 V 75 kO Input Pulldown Resistors 3-BIT SCANNABLE REGISTERED BUS TRANSCEIVER The MC10EI100E337 is a 3-bit registered bus transceiver with scan. The bus outputs (BUSO-BUS2) are specified for driving a 25 0 bus; the receive outputs (00-02) are specified for 50 O. The bus outputs feature a normal HIGH level (VOH) and a cutoff lOW level- when lOW, the outputs go to - 2.0 V and the output emitter-follower is "off," presenting a high impedance to the bus. The bus outputs also feature edge slow-down capacitors. Both drive and receive sides feature the same logic, including a loopback path to hold data. The HOLD/lOAD function is controlled by Transmit Enable (TEN) and Receive Enable (REN) on the. transmit and receive sides respectively, with a HIGH selecting lOAD. Note that the implementation of the E337 Receive Enable differs from that of the E336. A synchronous bus enable (SBUSEN) is provided for normal, non-scan operation. The asynchronous bus disable (ABUSDIS) disables the bus immediately for scan mode. The SYNCEi'ii input is provided for flexibility when re-enabling the bus after disabling with ABUSDIS, allowing either synchronous or asynchronous re-enabling. An alternative use is asynchronous-only operation with ABUSDIS, in which case SYNCEN is tied lOW, or left open. SYNCEN is implemented as an overriding SET control (active-lOW) to the enable flip-flop. Scan mode is selected by a HIGH at the SCAN input. Scan input data is shifted in through S_IN, and output data appears at the 02 output. All registers are clocked on the positive transition of ClK. Additional lead-frame grounding is provided through the Ground pins (GND) which should be connected to 0 V. The GND pins are not electrically connected to the chip. PIN NAMES Pin Ao--A2 Bo-B2 S-IN TEN,REN SCAN ABUSDIS SBUSEN SYNCEN ClK BUSO-BUS2 00-02 Function Data Inputs A Data Inputs B Serial (Scan) Data Input HOLD/lOAD Controls Scan Control Asynchronous Bus Disable Synchronous Bus Enable Synchronous Enable Control Clock 25 n Cutoff Bus Outputs Receive Data Outputs (02 serves as SCAN_OUT in scan mode) This document contains information on a new product. Specifications and information herein are subject to change without notice. ECLinPS 3-95 MC10E337, MC100E337 PINOUt: 28-LEAD PLCC (TOP VIEW) SBUSEN SYNCEN BO 25 24 AO ABUSDIS VCCO 23 20 00 19 GND SCAN S-IN BUSO TEN VCC VEE 01 VCCO BUS1 A1 GND B1 A2 B2 VCCO BUS2 GND 02 LOGIC SYMBOL _--------aBUS2 A2 B2 ,...4--1----0 BUS1 01 A1 B1 _-+--+----CI BUSO 00 AO BO S_INo---+...J TEN RENo----+--If---H-------l SCANo-----J'---+---l+------.....J ABUSDIS o - - - - - - I - - . J I SBUSEN 0 - - - - - - / SYNCEN 0 - - - - 4 CLKc>------l------------...J ECLinPS 3-96 MC10E337, MC100E337 DC Characteristics: VEE = VEE(min) to VEE(max); Vee = Veco = GND 25°C O°C Symbol VeuT IIH lEE Characteristic min 'Cut-off Output Voltage -2.10 typ max min VEE Characteristic rnA 145 174 145 174 145 174 145 174 145 174 167 200 min typ 25°C max min typ 85°C max min typ max Unit ps ClktoQ 450 1000 450 1000 450 Clk to BUS 800 1800 800 1800 800 1800 ABUSDIS 500 1500 500 1500 500 1500 800 1800 800 1800 800 1800 1000 Setup Time ps 350 350 350 SBUSEN 100 100 100 Data, S-IN 400 400 400 TEN,REN,SCAN 550 550 550 BUS 350 350 350 SBUSEN 500 500 500 Data, S-IN 350 350 350 TEN,REN,SCAN 200 200 200 400 400 400 Hold Time ps Minimum Pulse Width Clk ~ V Propagation Delay to Output BUS t, -2.03 Condition = VEE(min) to VEE(max); Vee = Veea = GND -SYNCEN tpw Unit 150 150 O°C th max = -2.10V AC Characteristics: ts typ ![A 150 100E tpLH min Power Supply Current "measured with VTT tpHL 85°C max -2.03 -2.10 -2.03 -2.10 Input HIGH Current All Other Inputs 10E Symbol typ ps Rise I Fall Times ps 20-80%(Qn) 300 800 20 - 80% ( BUSn Rise) 500 1000 20 - 80% ( BUSn Fall ) 300 800 300 300 800 500 800 1000 500 1000 300 800 300 800 EClinPS 3-97 Condition MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E404 MC100E404 Differential D and 0 700 ps Max. Propagation Delay High Frequency Outputs Extended lOOE VEE Range of -4.2V to -S.46V Internal7SkO Input Pulldown Resistors The MC 1OE404/1 00E404 is a 4-bit differential AND/NAND device. The differential . operation of the device makes it ideal for pulse shaping applications where duty cycle skew is critical. Special design techniques were incorporated to minimize the skew between the upper and lower level gate inputs. Because a negative 2-input NAND function is equivalent to a 2-input OR function. the differential inputs and outputs of the device also allow for its use as a fully differential 2 input OR/NOR function. The output RISE/FALL times of this device are significantly faster than most other standard ECLinPS devices resulting in an increased bandwidth. The differential inputs have clamp structures which will force the Q output of a gate in an open input condition to go to a LOW state. Thus inputs of unused gates can be left open and will not affect the operation of the rest of the device. LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) D3A D3A D3B 25 24 23 D3B veea 03 03 21 19 22 20 D o a G 00 DOa D2B 02 D2B 02 D2A vee VEE Qi D2A 01 D1B 00 D1B DOb DOb 00 D D1a 1 a G 01 D1b D1b 01 00 7 D1A D1A DOB 8 9 10 11 DOB DOA DOA veea D2aG02 D2a D2b 02 D2b PIN NAMES PIN FUNCTION - Differential Data Inputs D[0:4]. D[0:4] D 3aG03 D3a - Differential Data Outputs 0[0:4]. Q[0:4] D3b D3b FUNCTION TABLE Da Db Q L L H L H L H L L L H H QUAD DIFFERENTIAL AND/NAND - Da L L H H - Db L H L H - Q L H H H ECLinPS 3-98 . 53 MC10E404, MC100E404 DC Characteristics: VEE =VEE(min) to VEE(max); =Vcco = GND Vcc 25°C O°C Symbol Characteristic min I'H Input HIGH Current lEE Power Supply Current I--- Vpp(DC) VCMR typ max min typ 150 85°C max min typ 150 max Unit 150 ~A mA 10E 106 127 106 127 106 127 100E 106 127 106 127 122 146 Input Sensitivity 50 Common Mode Range Condition 50 -1.5 50 -1.5 0 0 -1.5 0 mV 1 V 2 1. Dlfferenlial Input voltage required to obtain a full ECl sWing on the outputs. 2. VCMR is referenced to the most positive side of the differential input signal. Normal operation is obtained when the input signals are within the VCMR range and the input swing is greater than V PP MIN and < 1V. AC Characteristics: VEE =V EE(min) to VEE(max); V cc =V cco = GND 25°C O°C Symbol tplH tpHl tSKEW Vpp(AC) tr tf Characteristic 85°C min typ max min typ max min typ max 350 475 650 350 475 650 350 475 650 700 Propagation Delay to Output Da(Diff) 300 475 700 300 475 700 300 475 Db (Dill) 375 500 675 375 500 675 375 500 675 Db (SE) 325 500 725 325 500 725 325 500 725 Within-Device Skew 50 50 150 50 150 150 ps 1 mV 2 Rise / Fall Time 20 - 80% Condition ps Da (SE) Minimum Input Swing Unit 150 400 150 400 1. Within-device skew is defined as identical transitions on similar paths through a device. 2. Minimum input swing for which AC parameters are guaranteed. ECLinPS 3-99 150 400 ps MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E416 MC100E416 Differential D and Q; V BB available 600 ps Max. Propagation Delay High Frequency Outputs 2 Stages of Gain Extended 1ODE VEE Range of -4.2V to -5.46V Internal 75kn Input Pulldown Resistors QUINT DIFFERENTIAL LINE RECEIVER The MC 1OE416/1 00E416 is a 5-bit differential line receiving device. The 2.0 GHz of bandwidth provided by the high frequency outputs makes the device ideal for buffering of very high speed oscillators. A V BB pin is available to AC couple an input signal to the device. More information on AC coupling can be found in the design handbook section of this data book. The design incorporates two stages of gain internal to the device making it an excellent choice for use in high bandwidth amplifier applications. LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) 03 04 04 veeo 04 04 veea 25 24 23 22 21 20 19 oo~oo DO 00 03 53 02 03 "~Q' vee 01 02 01 VEE Q2 VBB 02 ~~~ DO veea 02 DO 01 5 6 7 8 9 10 11 01 veea 00 00 veea D3~OO 01 01 03 04 FUNCTION - D[0:4). D[0:4) 03 "~~ PIN NAMES PIN 02 04 Differential Data Inputs - Q[0:4). Q[0:4) Differential Data Outputs VBB a ECLinPS 3-100 7' MC10E416, MC100E416 DC Characteristics: VEE =VEE(min) to VEE(max); =Veeo = GND Vee 25°C O°C Symbol Vee r--- Characteristic min Output Reference Voltage 10E 100E I'H Input HIGH Current lEE Power Supply Current 10E r--- VeMR Input Sensitivity max typ 85°C max min typ max Unit Condition V -1.19 -1.25 -1.31 -1.26 -1.38 150 -1.26 150 150 ~ mA 135 162 135 135 162 135 50 Common Mode Range min -1.27 -1.35 -1.26 -1.38 -1.38 -1.38 100E Vpp(DC) typ 162 162 50 -1.5 162 186 50 -1.5 0 135 155 0 -1.5 0 mV 1 V 2 1. Differential input voltage required to obtain a full ECl swing on the outputs. 2. VCMA is referenced to the most positive side of the differential input signal. Normal operation is obtained when the input signal are within the VCMA range and the input swing is greater than V PP MIN and < 1V AC Characteristics: VEE =VEE(min) to VEE(max); Vee =Vceo = GND O°C Symbol tplH tpHl typ max min typ max min typ max Propagation Delay to Output D(Diff) 250 350 350 500 550 250 350 350 500 550 250 350 500 200 350 550 tSKEW Within-Device Skew tSKEW Duty Cycle Skew ~ Minimum Input Swing Unit 200 200 50 50 50 ps 1 ±10 ±10 ±10 ps 2 mV 3 150 150 150 Rise / Fall Time 20 - 80% Condition ps tplH - tpHl tr 85°C min D(SE) Vpp(AC) 25°C Characteristic 100 200 350 100 200 350 100 200 350 1. Within-device skew is defined as identical transitions on similar paths through a device 2. Duty cycle skew defined only for differential operation when the delays are measured from the cross point of the inputs to the cross point of the outputs. 3. Minimum input swing for which AC parameters are guaranteed ECLinPS 3-101 ps MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA j;C10E431 ~100E431 Edge-Triggered Asynchronous Set and Reset Differential D. CLK and Q; VBB Reference Available 1100 MHz Min. Toggle Frequency Extended 1OOE VEE Range of -4.2V to -5.46V The MC 1OE/1 00E431 is a 3-bit flip-flop with differential clock. data input and data output. The asynchronous Set and Reset controls are edge-triggered rather than level controlled. This allows the user to rapidly set or reset the flip-flop and then continue clocking at the next clock edge. without the necessity of de-asserting the setlreset signal (as would be the case with a level controlled setlreset). The E431 is also designed with larger internal swings. an approach intended to minimize the time spent crossing the threshold region and thus reduce the metastability susceptibility window. LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) VBB CLK2 C.LK2 25 24 23 02 D2 R2 S2 22 21 20 19 3-BIT DIFFERENTIAL FLIP-FLOP so DO 00 DO CLKO 00 CLKO RO S1 01 5 6 CLKO CLKO 7 8 9 10 11 DO DO RO SO vcca 01 01 CLK1 01 CLK1 PIN NAMES PIN FUNCTION D[O:2]. D[O:2] CLK[O:2]. CLK[O:2] S[O:2] R[O:2] VBB Q[O:2]. Q[O:2] Differential Data Inputs R1 Differential Clock Inputs Edge Triggered Set Inputs Edge Triggered Reset Inputs VBB Reference Output Differential Data Outputs S2 02 02 02 CLK2 FUNCTION TABLE 02 CLK2 On CLKn Rn Sn Qn L H Z Z L L L L L L H Z L L Z L H X X L R2 VBB .. .. Z=Low to high transition ECLinPS 3-102 7 / MC1 OE431 , MC100E431 DC Characteristics: VEE =V EE(min) to VEdmax); Vcc = Vcco = GND O'C Symbol V BB Characteristic min 85'C 25'C max min typ max min typ max Output Reference Voltage -1.38 -1.27 -1.35 -1.25 -1.31 -1.19 100E -1.38 -1.26 -1.38 -1.26 -1.38 -1.26 Input HIGH Current lEE Power Supply Current 150 150 150 Condition !-LA mA 10E 110 132 110 132 110 132 100E 110 132 110 132 127 152 Common Mode Range Unit V 10E I'H V CMR typ -1.5 0 -1.5 0 -1.5 0 V 1 1. VeMR IS referenced to the most positive side of the differential Input signal. Normal specified operation IS obtained when the input signals are within the VeMR range and the input swing is greater than VPP MIN and < 1V. AC Characteristics: VEE = V EE(min) to V EE(max); V cc = V cco = GND 25'C O'C Symbol Characteristic min f MAX Max. Toggle Frequency 1100 1400 tpLH Propagation Delay to Output tpHL ts ClK(Diff) Vpp(AC) t, tf typ 85'C max 1100 1400 min typ max 1100 1400 Unit Condition MHz 600 750 450 600 750 450 600 750 400 600 800 400 600 800 400 600 800 R 550 725 925 550 725 925 550 725 925 S 550 725 925 550 725 925 550 725 925 200 0 200 0 200 0 R 1000 700 1000 700 1000 700 1 S 1000 700 1000 700 1000 700 1 200 0 200 0 200 0 Setup Time ps Hold Time ps Minimum Pulse Width ps 400 ClK tSKEW min ClK (SE) D tpw max ps 450 D th typ Within-Device Skew Minimum Input Swing 400 400 50 50 150 50 150 150 Rise/Fall Times ps 2 mV 3 ps 20 - 80% 275 450 650 275 .. 450 650 275 450 650 1. These setup times define the minimum time the ClK or SET/RESET Input must wall after the assertion of the RESET/SET input to assure the proper operation of the flip-flop. 2. Within-device skew is defined as identical transitions on similar paths through a device. 3. Minimum input swing for which AC parameters are guaranteed. ECLinPS 3·103 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. -·----·---- - - - :G TECHNICAL DATA Product Preview MC10E445 MC100E445 On Chip Clock +4 and +8 2.5Gb/s Data Rate Capability Differential Clock and Serial Inputs VBB Output for Single-ended Input Applications Asynchronous Data Synchronization Mode Select to Expand to 8 Bits Internal 75k!:! Input Pulldown Resistors Extended 1OOE VEE Range of -4.2V to -5.46V _ ... __ () SINA SINA 25 24 ~ PIN (J) w a: MODE NC VCCO 23 22 21 20 ----- PIN NAMES i;; (J) - 4-BIT SERIAL/PARALLEL CONVERTER The MC1 OE/1 00E445 is an integrated 4-bit serial to parallel data converter. The device is designed to operate for NRZ data rates of up to 2.5Gb/s. The Chip generates a divide by four and a _divide by 8 clock for both 4-bit conversion and a two chip 8-bit conversion function. The conversion sequence was chosen to convert the first serial bit to QO, the second to Q1 etc. Two selectable serial inputs provide a loopback capability for testing purposes when the device is used in conjunction with the E446 parallel to serial converter. The start bit for conversion can be moved using the SYNC input. A single pulse applied asynchronously for at least two cycles of the input clock signal shifts the start bit for conversion by one bit. For each additional shift required, an additional pulse must be supplied. The MODE input is used to select the conversion mode of the device. With the MODE input LOW, or open, the device will function as a 4-bit converter. When the MODE input is driven HIGH the data at the output will change on every eighth clock cycle thus allowing for an 8-bit conversion scheme using two E445's. PINOUT: 28-LEAD PLCC (TOP VIEW) ._" FUNCTION -- SINA, SINA SINS, SINS SEL SOUT, SOUT 19 lSJ SOUT 17 J 16 J vee 15 J 00 SOUT 00-03_ CLK,CLK ClI4, ClI4 ClI8, ClI8 MODE SYNCH 01 veeo Dill. Serial Data Input A Dill. Serial Data Input S Serial Input Select Pin Dill. Serial Data Output Parallel Data Outputs Dill. Clock Inputs Ditt. +4 Clock Output Ditt. +8 Clock Output Conversion Mode 4-biV8-bit Conversion Synchronizing Input 02 10 CUB CUB veeo eU4 11 FUNCTION TABLES eU4 veco 03 MODE Conversion L H 4 Sit 8 Sit SEL H L Serial Input A S This document contains information on a product under development. Motorola reserves the right to change or discontinue this product without notice. ECLinPS 3-104 MOTOROLA SEMICONDUCTOR - - - - - - - - - - - - TECHNICAL DATA MC10E446 MC100E446 Product Preview On Chip Clock +4 and +8 2.5Gb/s Data Rate Capability Differential Clock and Serial Inputs VBB Output for Single-ended Input Applications Mode Select to Expand to 8 Bits Internal 75kl:2 Input Pulldown Resistors Extended 1ODE VEE Range of -4.2V to -5.46V 4-BIT PARALLEL/SERIAL CONVERTER The MC1 OE/1 00E446 is an integrated 4-bit parallel to serial data converter. The device is designed to operate for NRZ data rates of up to 2.5Gb/s. The chip generates a divide by four and a divide by 8 clock for both 4-bit conversion and a two chip 8-bit conversion function. The conversion sequence was chosen to convert the parallel data into a serial stream from bit DO to 03. A serial input is provided to cascade two E446 devices for 8 bit conversion applications. The SYNC input can be used to reset the internal clock conversion unit to select the start of the conversion process. The MODE input is used to select the conversion mode of the device. With the MODE input LOW, or open, the device will function as a 4-bit converter. With the MODE input driven HIGH data at the serial input are read at one half of the +8 clock cycle thus allowing for an 8 bit conversion using two E446's. PIN NAMES PINOUT: 28-LEAD PLCC (TOP VIEW) DO D1 D2 D3 MODE NC NC 25 24 23 22 21 19 20 ClK 26 NC ClK 27 NC VBB 28 VCC VEE SOUT SIN SOUT SIN VCCO NC 10 VCCO CUB CUB vcco CU4 PIN FUNCTION SIN DO-D3 - SOUT, SOUT ClK,ClK Diff. Serial Data Input Parallel Data Input Diff. Serial Data Output Dill. Clock Input Dill. 4 Clock Output Dill. S Clock Output Conversion Mode, 4 billS bit Conversion Synchronizing Input CU4, CU4 CUS, CUS MODE SYNC FUNCTION TABLES 11 MODE Conversion l H 4 Bit S Bit CU4 VCCO This document contains information on a product under development. Motorola reserves the right to change or discontinue this product without notice. ECLinPS 3-105 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA • • • • • • MC10E451 MC100E451 Differential Inputs: Data and Clock Vss Output 1100 MHz Min. Toggle Frequency Asynchronous Master Reset Extended 100E VEE Range of -4.2 V to -5.46 V 75 kO Input Pulldown Resistors The MC10E/100E451 contains six D-type flip-flops with single-ended outputs and differential data inputs. The common clock input is also differential. The registers are triggered by a positive transition of the positive clock (ClK) input. A HIGH on the Master Reset (MR) input resets all Q outputs to lOW. The VSS output is intended for use as a reference voltage for single-ended reception of ECl signals to that device only. When using for this purpose, it is recommended that VSS is decoupled to VCC via a 0.Q1 JLf capacitor. PINOUT: 28-lEAD PlCC (TOP VIEW) liS DS 54 D4 53 D3 6-BIT D REGISTER, DIFF. DATA & CLOCK lOGIC SYMBOL Veeo DO 00 CLK Vee D, 0, CLK VCC VEE 03 MR VCCO NC 02 D2 02 D3 0, Do D, ii1 D2 52 03 " Vcca 00 D4 04 PIN NAMES Pin 00-0 5 Do-D5 CLK CLK MR Vee 00-0 5 DS Function OS +Data Input -Data Input + Clock Input - Clock Input Master Reset Input Vee Output Data Outputs CLK CLK MR Vee ECLinPS 3-106 I I MC1 OE451 , MC100E451 DC Characteristics: VEE = V EE(min) to V EE(max); V cc = V cco = GND 25°C O°C Symbol VBB Characteristic max min typ 85°C max min typ max -1.38 -1.27 -1.35 -1.25 -1.31 -1.19 100E -1.38 -1.26 -1.38 -1.26 -1.38 -1.26 lEE Power Supply Current 150 150 150 Condition IlA rnA 10E 84 101 84 101 84 101 100E 84 101 84 101 97 116 Common Mode Range Unit V 10E Input HIGH Current VCMR typ Output Reference Voltage I'H - min -2.0 -0.4 -2.0 -0.4 -2.0 -0.4 V 2 1. VCMR is referenced to the most positive side of the differential input signal. Normal operation is obtained when the "HIGH" input is within the VCMR range and the input swing is greater than VPPMIN and < 1V. AC Characteristics: VEE = V EE(min) to V EE(max); V cc = V cco = GND O°C Symbol Characteristic min f MAX Max. Toggle Frequency 1100 1400 tplH Propagation Delay to Output tpHl t. typ 85°C max 1100 1400 min typ max 1100 1400 800 475 650 ClK (SE) 425 650 850 425 MR 425 600 850 425 150 -100 150 250 100 250 800 475 650 800 650 850 425 650 850 600 850 425 600 850 -100 150 -100 100 250 100 ps Hold Time ps Minimum Input Swing 150 tRR Reset Recovery Time 750 tpw Minimum Pulse Width 159 150 600 750 600 750 mV 600 1 ps ps 400 Within-Device Skew 400 400 100 ps 100 100 Rise I Fall Times 20 - 80% Condition MHz Setup Time ClK, MR Unit ps 650 Vpp(AC) t, tf min 475 D tSKEW 25°C max ClK(Diff) D th typ ps 275 450 800 275 450 800 1. Minimum input voltage for which AC parameters are guaranteed 2. Within-device skew is defined as identical transitions on similar paths through a device ECLinPS 3-107 275 450 800 2 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E452 MC100E452 Differential 0, ClK and 0; VBB Reference Available 1100 MHz Min. Toggle Frequency Asynchronous Master Reset Extended lODE VEE Range of -4.2V to -5.46V The MC 1DEll 00E452 is a 5-bit differential register with differential data ( inputs and outputs) and clock. The registers are triggered by a positive transition of the positive clock (ClK) input. A high on the Master Reset (MR) asynchronously resets all registers so that the 0 outputs go lOW. The differential input structures are clamped so that the inputs of unused registers can be left ope.!! without ~tting the bias network of the device. The clamping action will assert the 0 and the ClK sides of the inputs. Because of the edge triggered flip flop nature of the device simultaneously opening both the clock and data inputs will result in an output which reaches an unidentified but valid state. The fully differential design of the device makes it ideal for very high frequency applications where a registered data path is necessary. 5-81T DIFFERENTIAL REGISTER LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) 0 0 - - - ' .... 03 OJ 04 i54 vcco Q4 Q4 25 24 23 22 21 20 19 50--",-,/ MR 03 ClK Q3 ClK vce VEE Q2 VBB 02 02 01 02 01 01 Di DO 50 vcco 10 11 QO Qli PIN NAMES PIN o 01--++-f' Di--++ 1.0 GHz bandwidth to meet the needs of the most demanding sYSitem clock. Both separate selects and a common select are provided to make the device well suited for both data path and random logic applications. LOGIC SYMBOL PINOUT: 28-LEAD PLCC (TOP VIEW) SEL2 D2A 25 24 D2A VBB 23 22 D2B 21 D2B eOMSEL 20 TRIPLE DIFFERENTIAL 2:1 MULTIPLEXER DOa 19 DOa 00 DOb 00 SELl 02 D1A 02 D1A vee VEE 01 VBB 01 D1B 00 Dla 01 D1B 00 Dlb 01 5 6 SELO DOA 7 8 DOA VBB 10 9 DOB DOb SELO Dla Dlb 11 SEL1 DOB veeo FUNCTION TABLE D2a SEL Data H L b D2a a D2b D2b SEL2-J....:>-.... PIN NAMES PIN FUNCTION - Dn[0:2], Dn[0:2] SEL COMSEL VBB 0[0:2], 0[0:2] eOMSEL Differential Data Inputs Individual Select Input Common Select Input VBB Reference Output Differential Data Outputs ECLinPS 3-110 02 MC1 OE457, MC100E457 DC Characteristics: VEE = VEE(min) to VEE(max); Vcc = Vcco= GND O°C Symbol V BB - - Characteristic 25°C max min typ 85°C max min typ max -1.38 -1.27 -1.35 -1.25 -1.31 -1.19 100E -1.38 -1.26 -1.38 -1.26 -1.38 -1.26 lEE Power Supply Current Unit Condition V 10E Input HIGH Current VCMR typ Output Reference Voltage IIH Vpp(DC) min 150 150 150 f!A mA 10E 92 110 92 110 92 110 100E 92 110 92 110 106 127 Input Sensitivity 50 Common Mode Range -1.5 50 50 -1.5 0 0 -1.5 0 mV 1 V 2 1. Differential input voltage required to obtain a full ECl swing on the outputs. 2. VeMR is referenced to the most positive side of the differential input signal. Normal operation is obtained when the input signal are within the VCMR range and the input swing is greater than VPP MIN and < 1V. AC Characteristics: VEE = V EE(min) to VEE(max); V cc = V cco = GND O°C Symbo tpLH tpHL Characteristic max min typ max min typ max D(Diff) 375 475 650 375 475 650 375 475 650 Propagation Delay to Output D 325 475 700 325 475 700 325 475 700 350 500 725 350 500 725 350 500 725 COMSEL 375 525 750 375 525 750 375 525 750 tSKEW Duty Cycle Skew tpLH - tpHL Minimum Input Swing Unit 40 40 40 ps 1 ±10 ±10 ±10 ps 2 mV 3 150 150 150 Rise / Fall Time 20-80% Condition ps SEL Within-Device Skew tr tf 85°C typ tSKEW Vpp(AC) 25°C min 150 275 450 150 275 450 150 275 1. Within-device skew is guaranteed for identical transitions on similar paths through a device 2. Duty cycle skew guarantee holds only for differential operation when the delays are measured from the cross point of the inputs to the cross point of the outputs. 3. Minimum Input Swing for which AC parameters are guaranteed. ECLinPS 3-111 450 ps MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA Advance Information • • • • • • MC10E1651 Typ. 3.0 dB Bandwidth> 1 GHz Typ. V to Q Propagation Delay ps Typ. Output Rise/Fall 300 ps Common Mode Range - 2.0 V to + 3.0 V Individual Latch Enables Differential Outputs DUAL ANALOG COMPARATOR W. LATCH The MC10E1651 is functionally and pin-for-pin compatible with the MC1651 in the MECL III family, but is fabricated on Motorola's advanced MOSAIC III process. The part has been designed with the goal of minimizing variations in propagation delay versus the amount of input overdrive. The output voltage levels are compatible with 10KH (and 10E) standard logic devices. PINOUT: 16-PIN CERAMIC DIP (TOP VIEW) GND Qii Qb LENb Vlb V2b Vcc NC VCC VEE LSUFFIX CERAMIC PACKAGE CASE 620 GND Qa 0. LENa V2a Vla LENb -----{;> VEE = -5.2V VCC = H.OV FUNCTION TABLE LEN V1, V2 H H L Vl > V2 Vl < V2 X Q 10E1651 20-LEAD PLCC PINOUTS H L Latched Qb LENb NC Vlb V2b 18 15 17 16 14 VCC Qb GND NC 11 Oa LENa NC V2a Vla This document contains information on a new product. Specifications and information are subject to change without notice. ECLinPS 3-112 NC MC10E1651 ABSOLUTE MAXIMUM RATINGS: Beyond which device life may be impaired Symbol VSUP VPP Characteristic min typ max Total Supply Voltage IVEEI + IVCCI 12.0 Differential Input Voltage IVl -V21 3.0 DC Characteristics: Unit V V VEE = -5.2 V ± 5%; Vee = +5.0V ± 5% O'C Symbol Characteristic min typ 25'C max min -980 V OH Ouput HIGH Voltage -1020 -840 VOL Output Low Voltage -1950 -1630 -1950 typ 85'C max min -810 typ max Unit -920 -735 mV -1630 -1950 -1600 mV IJA II Input Current (V1, V~ 65 65 65 I'H Input HIGH Current (LEN) 150 150 150 50 -55 mA 3.0 V Unit Icc Positive Supply Current 50 50 lEE Negative Supply Current -55 -55 VCMR Common Mode Range AC Characteristics: -2.0 3.0 -2.0 tplH tpHl tr tf -2.0 VEE = -5.2 V ± 5%; Vee = +5.0V ± 5% 25'C O'C Symbol 3.0 Characteristic 85'C min typ max min typ max min typ max VtoO 600 850 1150 600 850 1150 600 850 1150 LEN toO 500 750 950 500 750 950 500 750 950 200 300 700 200 300 700 200 300 700 Propagation Delay to Output - ps Rise / Fall Times 20 - 80% Condition ps Note: Contact factory for more complete data sheet ECLinPS 3-113 Condition MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. TECHNICAL DATA MC10E1652 Typ. 3.0 dB Bandwidth> 1.0GHz Typ. V to Q Propagation Delay of 850 ps Typ. Output Rise/Fall of 300 ps Common Mode Range -2.0 V to +3.0 V Individual Latch Enables Differential Outputs User Controlled Input Hysterisis DUAL ANALOG COMPARATOR wI LATCH The MC10E1652 is functionally compatible with the MC10E1651 and thus the MC1651 in the MECL III family. The hysterisis control pin HYS allows the user to define the amount of input hysterisis where as the MC1 OE1651 has a built in fixed level of input hysterisis. The device comes in a 10E version only, thus the outputs are only compatible with 10KH logic devices. The device is available in both a 16-pin DIP and a 20-pin PLCC surface mount package. PINOUT: 20-LEAD PLCC (TOP VIEW) Qb LENb Ne V1b LOGIC DIAGRAM V2b V1a 14 V2a Q 18 17 16 15 Qb vee GNO HYS 11 LENa Qa Qa NC HYS VEE vee V1b Q Qb Q Qb V2b Qa LENa Ne V2a V1a LENb PINOUT: 16-PIN CERAMIC DIP (TOP VIEW) GND Qb VEE = -5.2 V Qb LENb V1b V2b vce HYS vee = +5.0 V GND Qa 14 13 Qa LENa V2a V1 a vec VEE FUNCTION TABLE ECLinPS 3-114 LEN V1,V2 Q H H V1 > V2 V1 < V2 H L L X Latched MC10E1652 ABSOLUTE MAXIMUM RATINGS: Beyond which device life may be impaired Symbol VSUP VPP Characteristic min typ Total Supply Voltage IVEEI + Iveel 12.0 Differential Input Voltage IV1 ·V21 3.0 DC Characteristics: Unit max V V VEE =·5.2 V ± 5%; Vee =+5.0V ± 5% ooe Symbol Characteristic min typ 25°e min max typ 85°C max min V OH Ouput HIGH Voltage ·1020 ·840 ·980 ·810 ·920 VOL Output Low Voltage ·1950 ·1630 ·1950 ·1630 ·1950 typ max Unit ·735 mV ·1600 mV Input Current (V1. V~ Input HIGH Current (LEN) 65 150 65 150 65 150 IJA IIH Icc lEE Positive Supply Current Negative Supply Current 50 ·55 50 ·55 50 ·55 mA 3.0 V Unit II VCMR Common Mode Range AC Characteristics: ·2.0 3.0 ·2.0 tpLH tpHL t, tf ·2.0 VEE = ·5.2 V ± 5%; Vee = +5.0V ± 5% 25°C O°C Symbol 3.0 Condition 85°C Characteristic min typ max min typ max min typ max Propagation Delay to Output VtoQ LENtoQ 600 500 850 750 1150 950 600 500 850 750 1150 950 600 500 850 750 1150 950 Rise / Fall Times 20·80% 200 300 700 200 300 700 200 300 700 ps ps Note: Contact factory for more complete data sheet. ECLinPS 3-115 Condition ECLinPS 3-116 Design Guide This section contains a design guide written exclusively with the ECUnPS product family in mind. The design guide deals with system design aspects of using the family. This section is not meant to be a replacement for the MECL System Design Handbook but rather a supplement to the information contained in it. CONTENTS System Basics ........................ Transmission Line Theory ............... System Interconnect . ................... Interfacing with ECLinPS ................ Package and Thermal Information . ........ Quality and Reliability .................. ECLinPS 4-1 4-2 4-8 4-18 4-29 4-32 4-38 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. DESIGN GUIDE System Basics SECTION 1 System Basics Although both the 10E and 1OOE devices can tolerate variations in the VEE supply without any detrimental effects, it is recommended that the VEE supply also have a dedicated power plane. If this is not a feasible constraint care should be taken sothatthe IRdrops of the VEE busdo notcreatl') a VEE voltage outside of the specification range. To provide the switching currents resulting from stray capacitances and asymmetric loading, the VEE power supply in an ECl system . nel;!ds to be bypassed. It is recommended that the VEE supply be bypassed at each device with an RF quality .01 fJF capacitor to ground. In addition the supply should also be bypassed to ground with a 1J.lF -1 OJ.lF capacitor althe power inputs to a board. If a separate output termination plane is used the VEE supply will be of a static nature as the output switching current will return to ground via the VTT supply, thus the bypassing of every device may be on the conservative side. lithe design is going to include a liberal use of serial or Thevenin equivalent termination schemes a properly bypassed VEE plane is essential. Power Supply Considerations The following text gives a brief description of the requirements and recommendation for treatment of power supplies in an ECLinPS system design. A more thorough narration on the general subject of power supplies can be found in the Motorola System Design Handbook. Vee Supply As with all previous ECl families the ECLinPS logic family is designed to operate with negative power supplies, in other words with Vee connected to ground. However ECLinPS circuits will work fine with positive power supplies as long as special care has been taken to ensure a stable, quiet Vee supply. The output voltage levels for a positive supply system can be determined by simply subtracting the absolute value of the standard negative output levels from the desired Vee' To provide as small an AC impedance as possible, and minimize power bus IR drops, the Vee supply should have a dedicated power plane. By providing a full ground plane in the system the deSigner ensures that the current return path for the signal traveling down a transmission line does not encounter any major obstructions. It is imperative that'the noise and voltage drops be as small as possible on the Vec plane as the internal switching references and output levels are all derived off of the Vee power rail. Thus any perturbations on this rail could adversely affect the noise margins of a system. VTT Supply The output edge rates of the ECLinPS family necessitate an almost exclusive use of controlled impedance transmission lines for system interconnect (the details of this claim will be discussed in a latter section). Thus, unless Thevenin equivalent termination schemes are going to be used, a V TT supply is a must in ECLinPS designs. The choice of using only Thevenin equivalent termination schemes to save a power supply should not be made lightly as the Thevenin scheme consumes up to ten times more power than the equivalent parallel termination to a -2.0V VTT supply. As was the case for the VEE supply, a dedicated power plane, liberally bypassed as described above, should be used for the VTT supply. In designs which rely heavily on parallel termination schemes the VTT supply will be responsible for returning the switching current of the outputs to ground, therefore a low AC impedance is a must. For bypassing, many SIP resistor packs have bypass capaCitors integrated in their design to supply the necessary bypassing of the supply. The useof SIP resistors will be discussed more thoroughly in a latter chapter. VEE Supply To take advantage of increased logic density and temperature compensated outputs many designers are building array options with both temperature compensated output levels and a -5.2V VEE supply. To alleviate any problems with interfacing these arrays to ECLinPS 1OOE devices Motorola has specified the operation of 100E devices to include the standard 1OKH VEE voltage range. Moreover, because of the superior voltage compensation of the bias network, this guarantee comes without any changes in the DC or AC specification limits. With the availability of both 10KH and lOOK compatible devices in the ECLinPS family there is generally no need to run 1OE devices at lOOK voltage levels. If however this is desired, the 10E devices will function at1 OOE VEE levels with, at worst, a small degradation in AC performance for a few devices due to soft saturation of the current source device. Handling of Unused Inputs and Outputs Unused Inputs All ECLinPS devices have internal 50Kn-75Kn pulldown resistors connected to VEE' As a result an input which ECLinPS 4-2 System Basics is left open will be pulled to VEE and thus set at a logic lOW. These internal pulldowns provide more than enough noise margin to keep the input from turning on if noise is coupled to the input, therefore there is no need to tie the inputs to VEE external to the package. In addition by shorting the inputs to VEE external to the package one removes the current limiting effect of the pulldown resistor and under extreme VEE conditions the input transistor could be permanently damaged. If there are concerns about leaving sensitive inputs, such as clocks, open they should be tied low via an unused output or a quiet connection to VIT • Unless otherwise noted on the data sheets, the outputs to differential input devices will go to a defined state when the inputs are left open. This is accomplished via an internal clamp. Note that this clamp will only take over if the voltage at the inputs fall below = -2.SV. Therefore if equal voltages of greater than -2.SV are placed on the inputs the outputs will attain an undefined midswing state. Unlike saturating logic families the inputs to an ECLinPS, or any ECl device, cannot be tied directly to Vee to implement a logic HIGH level. Tying inputs to Vee will saturate the input transistor and the AC and DC performance will be seriously impaired. A logic HIGH on an ECLinPS input should be tied to a level no higher than 600mV below the Vee rail and more typically no higher than the specified V,Hmax limit. A resistor or diode tree can be used to generate a logic HIGH level or more commonly an output of an unused gate can be used. generation of crosstalk and other noise phenomena during simultaneous switching situations. Although the noise generated in EClsystems is minor compared toothertechnologies there are methods to even further minimize the problem. Figure 1.1 below illustrates the two output scenarios of an Eel device: differential outputs and single ended outputs. During switching the current in the output device will change by =17mA when loaded in the normal to -2.0V load. With differential outputs as one output switches from a low to a high state the other switches from a high to a low state simultaneously thus the resultant current change through the Vceo connection is zero. The current simply switches between the two outputs. However for the single ended output, the current change flows through the V ceo connection of the output device. This current change through the V ceo pin of the package causes a voltage spike due to the inductance of the pin. son 1 0UIPUI Current Flow OUT IN Unused Outputs The handling of unused outputs is guided by two criteria: power dissipation and noise generation. For single ended output devices it is highly recommended to leave unused outputs unterminated as there are no benefits in the alternative scheme. This not only saves the power associated with the output but also reduces the noise on the Vee line by reducing the current being switched through the inductance of the Vee pins. For the counters and shift registers of the family the count and shift frequencies will be maximized if the parallel outputs are left unterminated. Of course for applications where these parallel outputs are needed this is not a viable alternative. Forthe differential outputs, on the other hand, things are a little less cut and dry. If either of the outputs of a complimentary output pair is being used both outputs of the pair should be terminated. This termination scheme minimizes the current being switched through the Vee pin and thus minimizes the noise generated on Vee' If, however, neither of the outputs of a complimentary pair are being used it makes most sense to leave these unterminated to save power. Note that the E 111 device has special termination. rules; these rules are outlined on the data sheet for the device. Minimizing Simultaneous Switching Noise A common occurrence among ECl families is the Single-Ended Output OUT IN OUT Differential Outputs Figure 1.1 - Eel Output Structures Traditionally manufacturers of ECl products have attempted to combat this problem by providing a separate Vee pin for the output device (Vceo' V eeA etc.) and the internal circuitry. By doing this the noise generated on the Vceo of the output devices would see a high impedance internal to the ECLinPS 4-3 System Basics For ECLinPS devices the capacitive load produced by an input ranges from 1.2pF to 2.0pF. The majority (~95%) of this capacitance is contributed by the package with very little added by the internal input circuitry. For this reason the range is generally a result ofthe difference between a corner and a center pin for the PlCC package. A good typical capacitance value for a center pin is 1.4pf and for a corner pin 1. 7pf. The capacitances for the other pins can be deduced through a linear interpolation. chip and not couple onto the the Vee line which controls the output and internal bias levels. Unfortunately in practice the noise generated on the V ceo would couple into the chip Vee through the collector base capacitance ofthe output device, thus a large portion of the noise seen on the V ceo line would also be seen on the Vce line. Forthe ECLinPS family and its associated edge speeds it was decided that multiple V ceo pins would be necessary to minimize the inductance and the associated noise generation. A design rule was established so that there would be no more than three single ended outputs per V ceo pin. Initially the Vee and Vceo pins were kept isolated from one another. However it was discovered that in certain applications the parasitics of the package and the output device would combine to produce an instability which resulted in the outputs going into an OSCillatory state. To alleviate this oscillation problem it was necessary to make the V ce and V ceo metal common internal to the package. Subsequent evaluation showed that because of the liberal use of Vceo pins the noise generated is equal to or less than that of previous ECl families. To further reduce the noise generated there are some things that can be done at the system level. As mentioned above there should be adequate bypassing of the Vee line and the guidelines forthe handling of unused outputs should be followed. In addition for wide single ended output devices an increase in the characteristic impedance ofthe transmission line interconnect will result in a smaller time rate of change of current; thus reducing the voltage glitch caused by the inductance of the package. This noise improvement should, of course, be weighed againstthe potential slowing of the higher impedance trace to optimize the performance ofthe entire system. In addition the connection between the device Vee pins and the ground plane should be as small as possible to minimize the inductance of the Vee line. Note that a device mounted in a socket will exhibit a larger amount of Vee noise due to the added inductance of the socket pins. Wired-OR Connections The use of wired-or connections in ECl designs is a popular way to reduce total part count and optimize the speed performance of a system. The limitation of OR-tying ECl outputs has always been a combination of increased delay per OR-tie and the negative going disturbance seen at the output when one output switches from a high to a low while the rest of the outputs remain high. For high speed devices the latter problem is the primary limitation due to the increased sensitivity to this phenomena with decreasing outpultransition times. The following paragraph will attempt to describe the wire-OR glitch phenomena from a physical perspective. Figure 1.2 illustrates a typical wire-OR situation. For simplicity the discussion will deal with only two outputs, however the argument could easily be expanded to include any number of outputs. If both the A and the B outputs start in the high state they will both supply equal amounts of currenlto the load. If the B outpulthen transitions from a high to a low the line althe emitterof B will see a sudden decrease in the line Voltage. This negative going transition on the line will continue downward at the natural fall time of the output until the A output responds to the voltage change and supplies the needed current to the load. This lag in the time it takes for A to correct the load current and return the line to a quiescent high level is comprised of three elements: the natural response time of the A output, the delay associated with the trace length between the two outputs and the time it takes for a signal to propagate through the package. The trace delay can be effectively forced to zero by OR-tying adjacent pins. The resulting situation can then be considered "best case". In this best case situation if the delay through the package is not a significant portion of the transition time of the output the resulting negative going glitch will be relatively small (~1 OOmV). A disturbance of this size will not propagate through a system. As the trace length between OR-tied outputs increases, the magnitude of the negative going disturbance will increase. Older ECl families specified the maximum delay allowed between OR-tied outputs to prevent the creation of a glitch which would propagate through a system. As this glitch phenomena is a physical limitation, due to decreased edge rates, ECLinPS devices are susceptible to the problem to an even greater degree than previous slower ECl families. The package delay of even the 28-lead PlCC Effects of Capacitive Loads The issue of AC parametric shifts with load capacitance is a common concern especially with designers coming from the TTL and CMOS worlds. For ECLinPS type edge speeds wire interconnect starts acting like transmission lines for lengths greater than 1/2". Therefore in ECLinPS designs for the majority of cases the load on an output is seen by the transmission line and not the output of the driving device. The effects of load capacitance on transmission lines will be discussed in detail in the next section. If the load is close to the driving output «1/2"), the resulting degradation will be 15-25ps/pF for both propagation delays and edge rates. In general a capacitive load on an emitter follower has a greater impact on the falling edge than the rising edge. Therefore the upper end of the range given above represents the effect on fall times and the associated propagation delays while the lower end represents the effect on the riSing output parameters. ECLinPS 4-4 System Basics Package logic technologies are intimately tied to variations in the input thresholds. As illustrated in Figure 1.3 although the delays when measured from the threshold of the input to the 50% point of the output are equal; when measured from the specified 50% point olthe inputlo the 50% point of the output the delays will vary with any shift in the switching reference. Obviously the magnitude of the delay difference is also proportional to the edge rate of the input. In addition to increasing the size olthe delay windows, this reference shift will cause the duty cycle of the output of a device to be different than that of the input. Unfortunately these thresholds are perhaps the most difficult aspects of a logic device to control. As a result for the ultimate in low skew performance differential ECl devices are a must. A quick perusal of the ECLinPS databook will reveal a relatively large number of totally differential devices which will lend themselves nicely to very low skew applications such as clock distribu- -< ~ vn Figure 1.2 - Typical Wire-OR Configuration is a significant portion of the transition times for an ECLinPS device. Therefore even in the best case situation described above one can expect an ~200mV glitch on the OR-tied line. A glitch of this magnitude will not propagate through the system but it is significantly worse than the best case situation of earlier ECl families. In fact as long as the distance between OR-tied outputs is kept to less than 112" the resulting line disturbance will not be sufficientlo propagate through most systems. With this in mind the following recommendations are offered for OR-tying in ECLinPS designs. First OR-tying of clock lines should be avoided as even in the best case situation the disturbance on the line is significant and could cause false clocking in some situations. In addition wireORed outputs should be from the same package and preferably should be adjacent pins. Non adjacent outputs should be within 1/2" of each other with the load resistor connection situated near the midpoint of the trace (Figure 1.2). By following these guidelines the practice of wire-ORing ECl outputs can be expanded to the ECLinPS family without encountering problems in the system. A detailed discussion of wire-OR connections in the ECLinPS world of performance is beyond the scope of this text. For this reason a separate application note has been written which deals with this situation in a much more thorough manner. Anyone planning to use wire-OR connections in their ECLinPS design is encouraged to contact a Motorola representative to obtain this application note. Data I I Tponom Tpomax Tponom~ Tpomin L I . : ~~______________~~~~ ET'50% Output Figure 1.3 - Delay vs Switching Reference Offset Clock Distribution tion. Clock skew is a major contributor to the upper limit of operation of a high speed system, therefore any reduction in this parameter will enhance the overall performance of a system. Through the ECLinPS family and new offerings in the 10KH family Motorola is providing devices uniquely designed to meet the demands of low skew clock distribution. By far the largest contributor to system skew is the variation between different process lots of a given device. This variation is what defines the total delay window specified in the data sheets. This window can be minimized if the devices are fully differential due to the output level defined thresholds which ensure a "centered" input swing. The propagation delay windows of single ended ECl and other In addition to these generic differential devices there are several devices which were designed exclusively for clock distribution systems. With past ECl families designers were forced to build clock distribution trees with devices which were compromises at best. The ECLinPS family, however, was built around the Elll clock distribution chip; a fully differential 1:9 fanout device which boasts within part skews as well as part to part skews unequaled in today's market. Additionally, to further deskew clock lines the E195 programmable delay chip is available. Although static delay lines can remove built in path length difference skew they can not compensate for variations in the delays of the devices in the clock path. The E195 allows the user to delay a signal over a 2ns range in ~20ps steps. Through the use ECLinPS 4-5 System Basics The metastability behavior and measurement of a flip flop is a complicated subject and necessitates much more time than is available in this forum for a thorough explanation. As a result the following description is of an overview nature. Anyone interested in a more thorough narration on the subject is encouraged to contact a Motorola representative to acquire the application note which contains a more detailed discussion on the subject. In many designs occassions arise where an asynchronous signal needs to be synchronized to the system clock. Generally this task is accomplished with the use of a single or series ofD flip flops as pictured in Figure 1.4. Because the data signal and the clock signal are asynchronous the system designer cannot guarantee that the setup and hold specifications for the device will be met. This in and of itself would not cause a problem if it was not for the metastable behavior of a D flip flop. The metastable behavior of a flip flop is described by the outputs of a device attaining a nondefined logic level or perhaps going into an OScillatory state when the data and the clock inputs to the flip flop switch simultaneously. It has been shown that this metastable behavior occurs across technology boundaries as well as across performance levels within a technology. For Eel the characterisitic of a flip flop in a metastable state is a device whose outputs are in a non-defined state near the midpoint of a normal output swing. The output will return randomly to one of the two defined states some time later (Figure 1.5). The two parameters of importance when discussing metastability are the metastablity window; the of this device the designer can match skews between clocks to 20ps. Although these two devices satisfy the needs for many Eel designers they do overlook the needs of a special subset; the designer who mixes Eel and TTL technologies. When translating between Eel and TTL much of the skew performance gained through the Elll is lost when passed through the translator and distributed in TTL. To solve this problem a new set of translators have been introduced in the MECl 1OKH family. The H641 and H643 receive a differential Eel input and fan out nine TTL outputs with a guaranteed unparalleled skew between the TTL outputs. The H640 and H642 take differential EeL inputs and generate low skew TTL outputs which are ideal for driving clocks in 68030 and 68040 microprocessor systems. By using the EeL aspects of the Elll to distribute clock lines across the backplane to TTL cards and receiving and translating these signals with the H640, H641 , H642 or H643 a TTL clock distribution tree can be designed with a performance level unheard of with past logic families. Through the development of a library of differential devices, specialized low skew distribution chips and high resolution programmable delay chips Motorola has serviced the need for low skew clock distribution designs. These offerings open the door for even higher performance next generation machines. Metastability Behavior System 1 System 1 Clock System 1 Ouput .....J'"LSL 0 1 - - - - - - 1 System 2 Input Data D·Flip Flop System2~ - System 2 Clock Clock System 1 System Clock 1~ System 1 Ouput ...fLJ"L Data D-Flip Flop System Clock 2~ Clock 0 Data D-Flip Flop ,..- [>CIOCk TO Delay :! Figure 1.4 - Clock Synchronization Schemes ECLinPS 4-6 0 System 2 Input ,-- System 2 System Basics V The challenge then becomes, how to characterize metastability behavior given the above circumstances. The standard method in the industry is to use Stoll's' equation, combined with the standard MTBF equation, to develop the following relationship: / " where: Figure 1.5 - Metastable Behavior of an Eel Flip Flop fe: fo: T p: t: 't: Clock Frequency Data Frequency FF Propagation Delay Time Delay Between FF Clocks FF Resolution Time Constant Note thatthe clock frequency, data frequency and time delay between flip flops are user defined parameters, thus it is up to Motorola to provide only the propagation delays and the resolution time constants for the ECLinPS flip flops. The propagation delays are obviously already defined so this leaves only the resolution time constant yet to be determined. An evaluation fixture was fabricated and several ECLinPS flip flops were t;lvaluated for resolution time constants. The results of the evaluation showed thatthe time constant was somewhat dependent on the part type as all the flip flops in the ECLinPS family do not use the same general design. The time constants range from 125 - 225 ps depending on the part type. As an example for a system with a 1OOMHz clock and 75MHz data rate the required delay between clock edges of a cascaded flip flop chain for the E151 register, assuming a 't of 200ps, would be: window in time for which a transition on both the data and the clock will cause a metastable output, and the settling time; the time ittakes for a metastable outputto return to a defined state. For the single flip flop design of Figure 1.3, the data being fed into system 2 will be in an undefined state and thus unusable if the synchronizing flip flop enters a metastable state. Because of this a more popular design incorporates multiple flip flop chains with cascading data inputs and clock inputs which are delayed with respect to each other. This redundancy of flip flops helps to reduce the probability that the data entering system 2 will be at an undefined level which could wreak havoc on the logic olthat system. This reduction in probability relies on thefactthat even ilthe preceeding flip flop goes metastable it will settle to a defined state prior to the clocking ofthe following flip flop. Obviously once the first flip flop goes metastable there is an even chance that it will settle in the wrong state and thus information will be lost. However, there are error detection and correction methods to circumvent this problem. The larger the flip flop chain the lower the probability of metastable data being fed into system 2. Unfortunately for ECLinPS levels of performance both the window width and the settling time are difficult or impossible to measure directly. The metastable window for an ECLinPS flip flop is assuredly less than 5ps and most likely less than 1ps based on SPICE level simulation results. In either case, with todays measuring eqUipment it would be impossible to measure this window width directly. Although it is feasible to measure the settling time for a given occurence, this parameter is notfixed but rather is of a variable length which makes it impossible to provide an absolute guarantee. MTBF = 1/ (2*1 OOMHz*75MHz*800ps*1 0 'V200P' ) solving for an MTBF of 10 years yields: t = 3.1 ns therefore: So for an MTBF of 10 years for the above situation the second flip flop should be clocked 3.9ns afierthe first. Similar results can be found by applying the equation to different data and clock rates as well as different acceptable MTBF rates. 'Stoll, P. "How to Avoid Synchronization Problems," VlSI Design, November/December 1982. pp. 56-59. ECLinPS 4-7 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. DESIGN GUIDE Transmission Line Theory SECTION 2 Transmission Line Theory passing through a point on the line. Thus, 4, can be expressed in terms of the distributed inductance and capacitance of the line as shown by Equation 1. Introduction The ECLinPS family has pushed the world of ECl into the realm of picoseconds. Whim output transitions move into this picosecond region it becomes necessary to analyze system interconnects to determine if transmission line phenomena will occur. A handyrule of thumb to determine if an interconnect trace should be considered a transmission line is if the interconnect delay is greater than 118th of the signal transition time it should be considered a transmission line and afforded all of the attention required by a transmission line. If this rule is applied to the ECLinPS product line a typical PCB trace will attain transmission line behaviors for any length> 1/4" . Thus, a brief overview ofiransmission line theory is presented, including a discus.sion of distributed and lumped capacitance effects on transmission lines. For a more thorough discussion of transmission lines the reader is referred to Motorola's MECl Systems Design Handbook. (eqt. 1) where:· lo = Inductance per unit length (H) Co = Capacitance per unit length (F) Propagation Delay Propagation delay (Tpo) is also expressed as a function of both the inductance and capacitance per unit length of a transmission line. The propagation delay of the line is defined by the following equation: (eqt. 2) If lo is expressed as microHenry'.s per unit length and capacitance as picoFarad's per unit length, the units for delay are nanoseconds per unit length. The propagation velocity is defined as the reciprocal of the propagation delay: Background Exact transmission line analysis can be both tedious and time consuming, therefore assumptions for simplifying these types of calculations must be made. A reasonable assumption is that interconnect losses caused by factors such as bandwidth limitations, attenuation, and distortion are negligible for a typical PCB trace. The basis for this assumption is that losses due to the interconnect conductor are only a fraction of the total losses in the entire interface scheme. In addition, the conductivity of insulating material is very small, as a result the dielectric losses are minimal. A second and more fundamental assumption is that transmission line behavior can be described by two parameters: line characteristic impedance (4,), and propagation delay (Tpo)' lo and Co can be determined using the easily measured parameters of line delay (To)' line length (l), and the line characteristic impedance (4,) in conjunction with Equations 1 and 2. The propagation delay is defined as the ratio of line delay to line length: Tpo=Toll Combining equations 1 and 2 yields: (eqt. 3)· (eqt. 4) Characteristic Impedance An interconnect which consists of two conductors and a dielectric, characterized by distributed series resistances and inductances along with distributed parallel capacitances between them, is defined as a transmission line. These transmission lines exhibit a characteristic impedance over any length for which the distributed parameters are constant. Since the contribution of the distributed series resistance to the overall impedance is minimal, this term can be neglected when expressing the characteristic impedance of a line. The characteristic impedance is a dynamic quantity defined as the ratio of the transient voltage to the transient current Termination and Reflection Theory Figure 2.1 shows an ECLinPS gate driving a lossless transmission line of characteristic impedance and terminated with resistance 1\. Modifying the circuit of Figure 2.1 such that the driving gate is represented by its equivalent circuit gives the configuration shown in Figure 2.2. For a positive step function V,N , a voltage step Vs travels down the transmission line. The initial current in the transmission line is determined by the ratio Vs/Zo' When the 4" ECLinPS 4-8 Transmission Line Theory ==::J----r- ~ ~r:) Case 1: Rs < Zo; R T ", Zo The initial current in the transmission line is determined by the ratio VsIlo. However, the final steady state current is determined by the ratio VslRT; assuming ohmic losses in the transmission line are negligible. For case 1, an impedance discontinuity exists at the line-load interface which causes a reflected voltage and current to be generated at the instant the initial signal arrives at this interface. To determine the fraction of the traveling wave that will be reflected from the line-load interface, Kirchoff's current law is applied to node "a" in Figure 2.2. This results in the following: Zo Node "a" IRE Figure 2.1 - Typical Transmission Line Driving Scenario traveling wave arrives at the termination resistor R,-, Ohm's Law must be maintained. If the line characteristic impedance and the termination resistance match(i.e. lo=RT), the travel· ing wave does not encounter a discontinuity at the line-load interface; thus the total voltage across the termination resis· tance is the incident voltage Vs' However, if mismatches between the line characteristic impedance and the termination resistance occur, a reflected wave must be set up to ensure Ohm's Law is obeyed at the line-load interface. In addition, the reflected wave may also encounter a disconti· nuity at the interface between the transmission line and the source resistance, thereby sending a re-reflected wave back towards the load. When neither the source nor the load impedance match the line characteristic impedance multiple reflections occur with the reflected signals being attenuated with each passage over the transmission line. The output response of this configuration appears as a damped oscillation, asymptotically approaching the steady state value, a phenomenon often referred to as ringing. Using substitution: (eqt. 5) Since only one voltage can exist at node "a" at any instant in time: (eqt. 6) Combining Equations 5 and 6, and solving for VR yields: (eqt. 7) Therefore: The term PL' referred to as the load reflection coefficient, represents the fraction of the voltage wave arriving at the lineload interface that is reflected back toward the source. Similarly, a source reflection coefficient can be derived as: Figure 2.2 - Thevenin Equivalent Circuit of Figure 2.1 In performing transmission line analysis, designers may encounter one of three impedance situations: (eqt. 8) From equations 7 and 8 it is apparent that multiple reflections will occur when neither the source nor the load impedances match the characteristic impedance of the line. A general equation for the total line voltage as a function of time and distance is expressed by Equation 9. where: Rs = Source Resistance RT = Termination Resistance V(x,t) = ECLinPS 4-9 VA(t)*[U(t-TpD'x) + PL'U(t-TpD (2L-x) + PL'P s'U(t-TPD(2L+x)) + (P L**2)*(ps 'U(t-T PD(4L-x)) + Transmission Line Theory (P L**2)*ps "2)*U(t-Tpo(4Ltx» t ... ] t Voc (eqt. 9) ZO=50Q ) where: VA = Voltage Entering the Transmission Line T PO = Propagation Delay of the Line L = Total Line Length x = Distance to an Arbitrary Point on the Line V DC = Initial Quiescent Voltage of the Line Figure 2.4 " Transmission Line Model for ~ > Zo VTF = (65/71)*(-0.9) = -0.82V Finally, the output voltage, VI' can be derived from the reflection coefficient by combining Equations 6 and 7: The input voltage is a ramp from -1.75V to -0.9V. The initial voltage traveling down the line is: VT=(1 t(RT-ZO)/(RTtZO))*Vs Vs = (50/56)*0.85 = 0.76V VT=(2*R!(RTtZO)*Vs From Equations 7 and 8: The two possible configurations for the Case 1 conditions are RT>ZO and RT Zo Forthe case in which Rr>Za, PLis.positive, and the initial current at node "a" is greater than the final quiescent current From Equation 9, the output voltage VT after one line delay is: I'N'TIAL > IFINAL Likewise, after a time equal to three times the line delay, the output voltage VT is Hence: Thus a reflected current, IR' must flow toward the source in order to attain the final steady state current as shown in Figure 2.3. An example of a line mismatched at both ends, with the Additional iterations of Equation 9 can be performed to show that the ringing asymptotically approaches the final line voltage of -0.82V. Ringing is a characteristic response for transmission lines mismatched at both ends with Rr > ZOo A SPICE representation of configuration 1 is illustrated in Figure 2.5. m f vR ~I~ ~ ~ Vs -Is I -tIR I- I vT ~ I I Distance Figure 2.3 " Reflected Voltage Wave for ~ > Zo termination resistance greater than the load resistance is shown in Figure 2.4. The initial steady state output voltage is given by: ~ Time VTI = (65/71)*(-1.75) = -1.60V H=1000 psldiv V=200 mVldiv The final steady state output voltage is given by: Figure 2.5 SPICE Results for Circuit of Figure 2.4 ECLinPS 4-10 Transmission Line Theory VTF Configuration 2: RT < Zo For the case in which RT Line Delay Zo; 50n ) Figure 2.9 shows the line response for the same circuit as above, but for the case in which the input pulse width is less than the line delay. As in the previous example, the initial steady state voltage across the transmission line is -1.49 volts, and the reflection coefficients are -0.18 and -0.79 for the load and source respectively. However, the intermediate Figure 2.7· Transmission Line Model for R, < Zo ECLinPS 4·11 Transmission Line Theory ~~ f\ .... i I \ \ \ \ ~ ... '5 '5 o i\\... I LJ L L/\. "'\ - Time Time H= 1000 ps/div V=200 mV/div H= 1000 ps/div V=100 mV/div Figure 2.9 - SPICE Results for Circuit of Figure 2.7 with Input Pulse < Line Delay Figure 2.11 - SPICE Results for Shorted Line with the Input Pulse Width> Line Delay voltage across the transmission line is a series of positivegoing pulses of decreasing amplitude for each round trip of the reflected voltage, until the final steady state voltage of 1.49 volts is reached. source end the voltage is partially reflected back toward the shorted end in accordance with the source reflection coefficient. Thus, the voltage at the shorted end of the transmission line is always zero while at the source end the voltage is reduced for each round trip of the reflected voltage. The voltage at the source end tends toward the final steady state condition of zero volts across the transmission line. The values of the source and line characteristic impedances in this example are such that the amplitude decreases by 50% with each successive round trip across the transmission line. Figure 2.12 shows the line response for the same circuit as above, but for the case in which the input pulse width is less than the line delay. As in the previous example, the initial and final steady state voltages across the transmission line are zero, and the reflection coefficients are -1 and -0.5 for the load and source respectively. However, the intermediate voltage across the transmission line is a series of negative pulses with the amplitude of each pulse decreasing for each round trip of the reflected voltage until the final steady state voltage of zero volts is attained. Again, for this example the amplitude of the output response decreases by ShortedUne The shorted line is a special case of configuration 2 in which the load reflection coefficient is -1 , and the reflections tend toward the steady state condition of zero line voltage and a current defined by the source voltage and the source resistance. R - 16 7n Zo = 50n ~~):=J-)---r'l0 VT VIN CS? Figure 2.10 - Transmission Line Model for Shorted Line An example of a shorted line is shown in Figure 2.10. The transmission line response for the case in which the input pulse width is greater than the line delay is shown in Figure 2.11. The initial and final steady state voltages across the transmission line are zero. The source is a step function with a 0.85 volt amplitude. The initial voltage traveling down the line is: Vs f\ \ II \ = (50/66.7)*0.85 = 0.64V \ ! " From Equations 7 and 8, PL = (0-50)/(0+50) = -1 Time Ps = (16.7-50)/(16.7+50) = -0.5 Upon reaching the shorted end of the line, the initial voltage waveform is inverted and reflected toward the source. At the rv '- H=1000 ps/div V=200 mV/div Figure 2.12 - SPICE Results for Shorted Line with the Input Pulse Width < Line Delay ECLinPS 4-12 Transmission Line Theory 50% for each successive reflection due to the choice of source and transmission line characteristic impedances. Case 2: Rs $; Zo; RT = Zo As in Case 1, the initial current in the transmission line is determined by the ratio of VslZa' Similarly, since RT = Za the final steady state current is also determined by the ratio Vs/Za' Because a discontinuity does not exist at the line-load interface all the energy in the traveling step is absorbed by the termination resistance, in accordance with Ohm's Law. Therefore no reflections occur and the output response is merely a delayed version of the input waveform. / / v LJ / H=400ps/div V=200 mV/div Time ZO=50Q ) Figure 2.14· SPICE Results for Matched Termination resistance matches the line characteristic impedance. Ringing or stair-step output responses do not occur since the load reflection coefficient is zero. Figure 2.13 - Transmission Line Model for Matched Termination Case 3: Rs = Zo; RT", Zo When the termination resistance does not match the line characteristic impedance reflections arising from the load will occur. Fortunately, in case 3, the source resistance and the line characteristic impedance are equal, thus the reflection coefficient is zero and the energy in these reflections is completely absorbed at the source thus no further reflections occur. An example of a line mismatched at the source but matched at the load is shown in Figure 2.13. For an input pulse of -1.75V to -0.9V is given by: VTI = (50/56)*(-1.75) = -1.56V The final steady state output voltage is given by ZO=50Q ) VTF = (50/56)*(-0.9) = -0.80V The source is a step function with an 0.85 volt amplitude. The initial voltage traveling down the line is: Vs = (50/56)*0.85 = 0.76V Figure 2.15· Transmission Line Model for Vs =Zo From Equations 7 and 8, An example of a line mismatched at the load but matched at the source is shown in Figure 2.15. For an input pulse of -1.75V to -0.9V the initial steady state output voltage is given by: PL = (50-50)/(65+50) = 0 Ps = (6-50)/(6+50) = -0.79 From Equation 9, the output voltage VT after one line delay is: VTI = (65/115)*(-1.75) = -0.99V The final steady state output voltage is given by VT(L,Tpo) = VA(t)*[1 + pLl + Voc = -0.80V VTF = (65/115)*(-0.9) = -0.51 V Likewise, after a time equal to three times the line delay, the output voltage VT is: The source is a step function with a 0.85 volt amplitude. The initial voltage traveling down the line is: Vs = (50/100)*0.85 = 0.43V Thus the output response attains its final steady state value ( Figure 2.14) after only one line delay when the termination From Equations 7 and 8, ECLinPS 4·13 Transmission Line Theory PL = (65-50)/(65+50) =0.13 )-0 Ps = (50-50)/(50+50) = 0 From Equation 9, the output voltage VT after one line delay is: Figure 2.17 - Series Terminated Transmission Line Likewise, after a time equal to three times the line delay, the output voltage VT is: Thus the output response attains its final steady state value after one line delay when the source resistance matches the line characteristic impedance. Again, ringing or a stair-step output does not occur since the load reflection coefficient is zero ( Figure 2.16). / Ra I I vS = (50/100)*0.85 = 0.43V I V Time wave will be reflected toward the source. Since the source resistance matches the line characteristic impedance all the energy in the reflected wave is absorbed, and no further reflections occur. This "source absorption" feature reduces the effects of ringing, making series terminations particularly useful for transmitting signals through a backplane or other interconnects where discontinuities exist. As stated previously, the signal in the line is only at half amplitude and the reflection restores the signal to the full amplitude. It is important to ensure that all loads are located near the end of the transmission line so that a two step input signal is not seen by any of the loads. For the series terminated circuit of Figure 2.17 with + RST = Zo and an input pulse rising from -1. 75V to -0.9V, the initial line voltage VTI is -1.325V and the final line voltage, VTF , is -0.9V. The source is a step function with a 0.85 volt amplitude. The initial voltage traveling down t~e line is: From Equations 7 and 8, PL = (00-50)/(00+50) = 1 H=400ps/div V=100 mV/div Ps = (50-50)/(50+50) = 0 From Equation 9, the output voltage VT after one line delay is: Figure 2.16 - SPICE Results for Circuit of Figure 2.15 Series Termination Series termination represents a special subcategory of Case 3 in which the load reflection coefficient is + 1 and the source resistance is made equal to the line characteristic impedance by inserting a resistor, RST' between and in series with, the transmission line and the source resistance Ro' The reflections tend toward the steady state conditions of zero current in the transmission line, and an output voltage equal to the input voltage. This type of termination is illustrated by the circuit configuration of Figure 2.17. The initial voltage down the line will be only half the amplitude of the input signal due to the voltage division of the equal source and line impedances. Likewise, after a time equal to three times the line delay, the output voltage VT is: Since the load reflection coefficient is unity, the voltage at the output attains the full ECl swing, whereas the voltage at the beginning of the transmission line does not attain this level until the reflected voltage arrives back at the source termination (Figures 2.17 and 2.18). No other reflections occur because the source impedance and line characteristic impedance match. . . (eqt.10) Capacitive The load reflection coefficient tends to unity, thus a voltage wave arriving at the load will double in amplitude, and a reflected wave with the same amplitude as the incident Effect~ on Proppgation Delay Lumped Capacitive Loads ECLinPS 4-14 The effect of load capacitance on propagation delay Transmission line Theory r'-l ~./ II / I~ I / / f II II ) H=400 ps/div V=200 mV/div Time H= 1000 ps/div V=200 mV/div Time Figure 2.20 - Line Delay vs Lumped Capacitive Load Figure 2.18 - SPICE Results for Series Terminated Line in delay for load capacitances of 0, 1, 5, 10 and 20 picoFarads. The increase in propagation delay can be determined by using Thevenin's theorem to convert the transmission line into a single time constant network with a ramp input voltage. The analysis applies to both series and parallel terminations, since both configurations can be represented as a single time constant network with a time constant, t, and a Thevenin impedance Z'. Figure 2.21 shows the Thevenized versions for the series and parallel terminated configurations. The Thevenin impedance for the series configuration is approximately twice that for the parallel terminated case, thus the time constant will be two times greater for the series terminated configuration. Since t is proportional to the risetime, the risetime will also be two times greater;thus the reason for the larger impact of capacitive loading on the series terminated configuration. must be considered when using high performance inte· grated circuits such as the ECLinPS family. Although capacitive loading affects both series and parallel termination schemes, it is more pronounced for the series terminated case. Figure 2.19a illustrates a series terminated line with a capacitive load CL. Under the no load condition, CL=O, the delay between the 50% point of the input waveform to the Figure 2.19a - Lumped Load Transmission Line Model Z'=Zo 50% point of the output waveform is defined as the line delay To' A capacitive load placed at the end of the line increases the risetime of the output signal,thereby increasing To by an amount ~To (Figure 2.19b). Figure 2.20 shows the increase VIN~ z'c=ZOc L I-=- CL Thevenin Equivalent Series Termination ~ z=~~ VIN~ Line Input . -=- ICL -=- RT "'N/2~ - z'C=(Zo /2)C L Thevenin Equivalent Parallel Termination ~_ _ _ VHIGH ~Loaded /" Line Output Aol'---- t R= 0.6a Response a= 1.67t R VLOIW---_ _...J Unloaded Response t= a Figure 2.21 - Thevenin Equivalent Lumped Capacitance Circuits Figure 2.19b - ~TD Introduced by Capacitive Load ECLinPS 4-15 ICL -=- Transmission Line Theory gation delay. A circuit configuration for observing distributed capacitive loading effects is shown in Figure 2.23. 2.0 / 1.5 ,..- 1.0 0.5 1/ 0.0 0.0 / 0.5 / V Figure 2.23 - Transmission Line Model for Distributed Capacitive Load 1.0 1.5 2.0 2.5 Each capacitive load connected along a transmission line causes a reflection of opposite polarity to the incident wave. If the loads are spaced such that the risetime is greater than the time necessary for the incident wave to travel from one load to the next, the reflected waves from two adjacent loads will overlap. Figure 2.24 shows the output response for a transmission line with two distributed capacitive loads of 2pF separated by a line propagation time of 750ps. The upper trace, with a 20-80% input signal risetime of 400ps, shows two distinct reflections. The middle and lower traces with 20-80% risetimes of 750 ps and 950ps respectively, show that overlap occurs as the risetime becomes longer than the line propagation delay. Normalized Line-Load Time Constant (Z'C I t R ) Figure 2.22 - Normalized Delay Increase Due to Lumped Capacitive Load The relationship between the change in delay and the line-load time constant is shown in Figure 2.22. Both the delay change(j.T Dland the line-load time constant(Z'C) are normalized to the 20-80% risetime of the input signal. This chart provides a convenient graphical approach for approximating delay increases due to capacitive loads as illustrated by the following example. Given a 1000 series terminated line with a 5pF load at the end of the line and a no load rise time of 400ps the increase in delay, j.TD' can be determined using Figure 2.22. The normalized line-load time constant is: Z'C/IR = 1000*5pF/400ps = 1.25 Using this value and Figure 2.22: Therefore: j.T D = 0.9*400ps = 360ps Time Thus 360ps is added to the no load delay to arrive at the approximate delay for a 5pF load. For a 1000 line employing a matched parallel termination scheme, Z'=500, the added delay is only 240ps. This added delay is significantly less than the one encountered for the series terminated case. Thus when critical delay paths are being designed it is incumbent on the designer to give special consideration to the termination scheme, lumped loading capacitance and line impedance to optimize the delay performance of the system. H= 1000 ps/div V=100 mV/div Figure 2.24 - Reflections Due to Distributed Capacitance Increasing the number of distributed capacitive loads effectively decreases the line characteristic impedance as demonstrated by Figure 2.25. The upper trace shows that reflections occur for approximately 3.5 ns, during which time the characteristic impedance olthe line appearslower(= 760) than actual due to capacitive loading. After the reflections have ended, the transmission line appears as a short and the final steady state voltage is reached. The middle trace shows that decreasing the' termination resistance to match the effective line characteristic imper':mce produces a response typical of a properly terminated line. Finally, the lower trace shows that the original steady state output can be attained by changing the source resistance to match the load res is- Distributed Capacitive Loads In addition to lumped loading, capacitive loads may be distributed along transmission lines. There are three consequences of distributed capacitive loading of transmission lines: reflections, lower line impedance, and increased propa- EClinPS 4-16 Transmission Line Theory For the circuit used to obtain the traces in Figure 2.25, the distributed load capacitance is 4pF. From Equation 3, Co is calculated as / I~ Iii //1' ~ -1/ - ""- Co = 750 ps/930 = 8pF RS= RT = 93Q\ RS= 93!1, R T = 76Q Hence: r- Zo' I RS RT 76Qr 93Q/,t(1 + 4pF/8pF) = 760 Thus the effective line impedance is 170 lower than the actual impedance while reflections are occurring on the line. I Time = H=2000 psJdiv V=100 mVJdiv Line Delay Increase The increase in line delay caused by distributed loading is calculated by adding the distributed capacitance (CD) to the intrinsic line capacitance in Equation 2. Figure 2.25 • Characteristic Impedance Changes Due to Distributed Capacitive Loads tance and the effective characteristic capacitance. TpD Reduced Line Characteristic Impedance To a first order approximation the load capacitance(C L ) is represented as an increase in the intrinsic line capacitance along that portion of the transmission line for which the load capacitances are distributed. If the length over which the load capacitances are distributed is defined as "L" the distributed value of the load capacitance(CD) is given by TpD' = ,t(Lo '(Co+C o)) = ,t(Lo 'Co) TpD' = T pD ·,t(1 + CD/CO) (eqt.13) Once again for the circuit used to obtain the traces in Figure 2.25, the distributed load capacitance is 4pF. From the previous example the intrinsic line capacitance is 8 pF therefore, (eqt. 11) T po' = 750ps·,t(1 + 4pF/8pF) = 919ps The reduced line impedance is obtained by adding CD to Co in Equation 1. Thus, the effect of distributed load capacitance on line delay is to increase the delay by 169ps. From Equation 13 it is obvious that the larger the Co of the line the smaller will be the increase in delay due to a distributive capacitive load. Therefore, to obtain the minimum impedance change and lowest propagation delay as a function of gate loading, the lowest characteristic impedance line should be used as this results in a line with the largest intrinsic line capacitance. Zo = ,t(Lo/C o) Zo' = ,t(Lo/(Co + CD)) = ,t(Lo/(C o'(1+C D/C O ))) (eqt. 12) ECLinPS 4·17 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. System Interconnect DESIGN GUIDE SECTION 3 System Interconnect Introduction 3. Multilayer boards The most common printed circuit board material used for digital designs is a glass- epoxy laminate. These boards use a fiberglass dielectric with copper foils bonded to both sides of the dielectric material by an epoxy resin. Other substrate materials include a fiberglass dielectric with a polyimide resin and fiberglass dielectric with a teflon resin. For multilayer boards the inner layers are separated by sheets of prepreg which acts as both a dielectric material and a bonding agent between layers. The choice of substrate material depends on the function for which the board will be used, the environment in which the board is to operate, and costs. Table 3.1 lists several phYSical qualities which characterize several of the the available PCB types. Each available substrate material has its own properties which makes it ideally suited for particular applications. As mentioned earlier, edge rates of the ECLinPS family are such that most interconnects must be treated as transmission lines. Thus, a controlled impedance environment is necessary to produce predictable interconnect delays as well as limiting the reflection phenomena of undershoot and overshoot. The three most common techniques for circuit and/or system interconnect at high data rates are microstrip, stripline and coaxial cable; both microstrip and stripline are printed circuit board methods whereas coaxial cable is most often used for interconnecting different parts of a system which are separated by relatively large distances. For slower speed applications «300M Hz) a twisted pair scheme also works well. The scope of this writing will not include the twisted pair technique, however a detailed discussion of this topic can be found in the MECL System Design Handbook. Finally, wirewrap boards are not recommended for the ECLinPS family because the fast edge speeds exceed the capabilities of normal wirewrapped connections. Mismatches at the connections cause reflections which distort the fast signal, significantly reducing the noise immunity of the system and perhaps causing erroneous operation. Glass-Epoxy Possesses good moisture absorption, chemical, and heat resistance properties as well as mechanical strength over standard humidity and temperature ranges. The most widely used versions are G10 and FR4, the fire resistant version of G10. Printed Circuit Boards Glass-Polyimide Good for elevated temperature operation because of its tight tolerance of the coefficient of thermal expansion. Very hard material, so it may damage drilling eqUipment when being drilled. Printed circuits boards (PCB's) provide a reliable and economical means of interconnecting electrical signals between system components. Printed circuit boards consist of a dielectric substrate over which the conducting printed circuit material is placed. Three major categories of printed circuit boards exist: Glass-Teflon Good for use when a low dielectric material is required. Very soft material, so it may be difficult to build features 1. Single-sided boards 2. Double-sided boards Material Dielectric Constant Dissipation Factor Thermal Coefficient of Expansion Tensile Modulus Glass-Epoxy 4.8 (1MHz) 0.022 (1 MHz) 13 - 16 (1O. B/OC) 2.5 PTFE 2.1 (10GHz) 0.0004 (10GHz) 224 (1 O·B/oC) 0.05 Glass-Polyimide 4.5 (1MHz) 0.10 (1MHz) 12 - 14 (10· B/0 C) 2.8 Table 3.1 - Characterisitics of Common PCB Materials ECLinPS 4-18 System Interconnect requiring precise geometries. Relatively expensive material. 5 :c h =mils E= 4.8 " .5: Microstrip A microstrip line is the easiest printed circuit interconnect to fabricate because it consists simply of a ground plane and flat signal conductor separated by a dielectric ( Figure 3.1). 4 II. .eo 0 ~ 3 "c w h=60 S ... .. :::; 'u h = 100 2 a. t 0 T c 30 Ground Plane Figure 3.1 - Microstrip Line 6 The characteristic impedance, Zo, of a microstrip line is given by: 5 In [ 87 ,I (E, + 1.41) (5.98 • h) ] ( 0.8w+t) \ (eq!. 1) 3 where: = Relative Dielectric Constant of the Substrate w = Width of the Signal Trace = Thickness of Signal Trace t = Thickness of the Dielectric h 130 -f-~-+--+--+---IE= 4.8 t= 1.4mil 90 .§ 70 c :::; 50 85 105 125 Figure 3.2 is a plot of characteristic impedance as a function of trace width and dielectric thickness for a dielectric constant of 4.8 and a trace thickness of 1.4mils (1 ounce copper). Using the equation for Co developed in the previous chapter and Equation 1 above the capacitance per unit length can be calculated for various trace widths. Figure 3.3 plots Co vs trace width for several different dielectric thicknesses. In addition Figure 3.3 plots Co vs the characteristic impedance for a microstrip line for the dielectric constant and trace thickness given above. The propagation delay for a signal on a microstrip line is described by the following equation: h = mils a. 65 Figure 3.3 - Line Capacitance vs Line Impedance and Trace Width 150 110 45 r----. I---- Characteristic Impedance-Zo (n) To mitigate the effects of electric field fringing, an additional constraint is that the width of the ground plane be such that it extends past each edge of the signal line by at least the width of the signal line. c --........ 25 0.1 < w/h < 3.0 and 1 < E, < 15 ." ..... . ~ 1 Equation 1 is accurate to within ±5% when: ~ '\ 2 E, £0 110 E= 4.8 t = 1.4mils 1\ 4 Zo = 90 Trace Width (mils) h = 100 TPD = 1.016,1(0.475 • E, + 0.67) ns/foot (eqt2) h = 60 where: = Dielectric Constant of the Board Material E, 30 10 30 50 70 90 110 Note that the propagation delay is dependent only on the dielectric constant of the PCB substrate. Figure 3.4 plots the propagation delay of a microstrip line versus the dielectric constant of the PCB. Trace Width (mils) Figure 3.2 - Microstrip Impedance vs Trace Width ECLinPS 4-19 System Interconnect 200 ~ u . 180 c 160 .'" 140 !! 120 .5 .e ./' V Q Q. 11- 9:0 80 ~ c 70 N ./' ...... u V ./ ~ .!!! 90 / Q. 60 .§ / b=mils £=4.8 t= 1.4mil CD c ::i / 50 b = 20 40 100 2 3 4 5 6 7 8 9 5 10 11 8 14 17 20 Trace Width (mils) Dielectric Constant (e r ) Figure 3.6 - Stripline Impedance vs Trace Width Figure 3.4 - Propagation Delay vs Dielectric Constant As was the case with a microstrip line the capacitance per unit length of a stripline trace can be calculated using the Co equation from Chapter 2. The graphs of Figure 3.7 plot Stripline Stripline is a printed circuit board interconnect in which a signal conductor is placed in a dielectric medium which is "sandwiched" between two conducting layers (Figure 3.5). 4.0 ..,-_ _..,-_ _..,-_ _.,-_ _.,-_--, ~ u .5 IL .e 3.5 0 ===:0 VEE (eqt. 12) Figure 3.12 - Unterminated Transmission Line where: l = Line length tA = Rise time T po= Propagation Delay per unit length The function of RE is to provide the drive current for a high to low transition at the driver output. Since the reflection coefficient at the load is of opposite polarity to that at the source, the signal will be reflected back and forth over the transmission line with the polarity changing after each reflection from the source impedance. Thus, steps appear at the input tothe receiving gate. When RE is too large steps appear in the trailing edge of the propagating signal that slows the edge speed of the input to the receiving gate, subsequently causing an increase in the net propagation delay. A reasonable negative-going signal swing at the input of the receiving gate results when the value of RE is selected to produce an initial step of 600mV at the driving gate. Hence: 1* Zo > 0.6 (eqt. 8) 6.2Zo eo RE (1 DE), 4.9Zo eo RE (100E) (eqt. 9) Further, the propagation delay increases with gate loading, thus the actual delay per unit length (Tpo') is given as: Tpo' = TpD • ~(1 + CD I (l* Co)) Substitution of the modified delay per unit length into Equation 12 and rearranging yields Equation 13: (eqt. 13) Solving Equation 13 for the maximum line length produces: lma> = 0.5 '(~( (CD I Co) •• 2 + (tAl T pD )" 2) - Col Co (eqt. 14) Assuming a worst case capacitance of 2 pF and a rise time of 200 ps for the ECLinPS family gives a value of 0.3 inches for the maximum open line length. Table 3.3 shows maximum open line lengths derived from SPICE simulations for single and double gate loads, a maximum overshoot of 40% and undershoot of 20% was assumed. The simulation results indicate that for a son line a stub length of ~ 0.3 inches will limit the overshoot to less than 40%, and the undershoot to within 20% of the logic swing. Signal traces will most assuredly be larger than .3" for all but the simplest of interconnects, thus for most practical applications it will be necessary to use ECLinPS devices in a controlled impedance environment. Load resistors of less than 180n should not be used because the heavy load may cause a reduction in noise immunity when the output is in the high state due to an increased output emitter-follower VBE drop. When the driver gate delivers a full ECl swing, the signal propagates from point A arriving at pOint B a time To !ater. At point B the signal is reflected as a function of PL' The Input Impedance of the receiving gate is large relative to the line characteristic impedance, therefore: (eqt. 10) A large positive reflection occurs resulting in overshoot. The reflected signal reaches point A at time 2To' and a large ECLinPS 4-23 System Interconnect Microstrip Zo (0) 50 68 75 82 90 100 Fanout =1 Stripline Fanout =2 Fanout =1 Fanout =2 Lmax (in) Lmax (in) Lmax (in) Lm.. (in) 0.3 0.3 0.3 0.3 0.3 0.25 0.2 0.15 0.15 0.1 0.1 0.1 0.3 0.25 0.25 0.25 0.25 0.25 0.15 0.1 0.1 0.1 0.1 0.1 Table 3.3 - SPICE Derived Maximum Open Line Lengths for ECLinPS Designs additional supply voltage. An alternate approach to using a single power supply is to use a resistor divider network as shown in Figure 3.13b. The Thevenin equivalent of the two resistors is a single resistor equal to the characteristic impedance of the line, and terminated to VTr The values for resistors R, and R2 may be obtained from the following relationships: Parallel Termination When the fastest circuit performance or the ability to drive distributed loads is desired, parallel termination is the method of choice. An important feature of the parallel termination scheme is the undistorted waveform along the full length of the line. A parallel terminated line is one in which the receiving end is terminated to a voltage (VTT) through a resistor (RT ) with a value equal to the line characteristic impedance (Figure 3.13a). An advantage of this technique is that power consumption can be decreased by a judicious choice of VTr For 500 systems the typical value of VTT is negative two volts. Although the single resistor termination to VTT con· serves power, it offers the disadvantage of requiring an (eqt. 15) (eqt. 16) For a nominal 1OE supply voltage of -5.2V and VTT of -2V: (eqt. 17) (eqt. 18) R2 = 2.6' Zo R, = R2 /1.6 For a nominal 1OOE supply voltage of -4.5V and VTT of -2V R2 = 2.25' Zo R, = R2/ 1.25 (eqt. 19) (eqt. 20) Table 3.4 provides a reference of values for the resistor divider network of Figure 3.13b. Figure 3.13a Parallel Termination to VTT 10E lODE ZO(O) 50 70 75 80 90 100 120 150 R, (0) R2 (0) R,(O) R2 (0) 81 113 121 130 146 162 194 243 130 182 195 208 234 260 312 390 90 126 135 144 161 180 216 270 113 158 169 180 202 225 270 338 Figure 3.13b Thevenin Equivalent Parallel Termination Figure 3.13 - Parallel Termination Schemes Table 3.4 - Thevenin Termination Resistor Values ECLinPS 4-24 System Interconnect For both configurations, when the equivalent termination resistance matches the line impedance no reflection occurs because all the energy in the signal is absorbed by the termination. Hence, the primary tradeoff between the two types of termination schemes are power versus power supply requirements. As mentioned earlier the VTT scenario requires an extra power supply, however the Theveninization technique will consume 10 fold more DC power. Fortunately this extra power consumption will not be seen on the die, therefore both techniques will result in the same die junction temperatures. ECLinPS output drivers consists of emitter followers designed to drive a 500 load into a negative two volt supply (VTT)' Under these conditions the nominal 1OE output levels are -1. 75volts at 5mA for the low state and -0.9volts at 22m A for the high state. For the 1OOE devices the nominal output levels are -0.955 volts at 20.9mA for the high state and 1.705volts at 5.9mA for the low state. Figure 3.14 shows the nominal output characteristics for ECLinPS devices driving various load impedances returned to a negative two volt supply. This plot applies to both 10E and 1OOE versions of the ECLinPS family. The output resistances RH (high state output resistance) and RL (low state output resistance) are obtained from the reciprocal of the slope at the desired operating point. Many applications require loads other than 500, the resulting VOH and VOL levels can be estimated using the following technique. RH r------~----~-,l--oVOH 1 -770mV .,. VOL -1710mV Figure 3.15 • Equivalent Model for Calculating 1DE Output Levels where: IHOUT = (-770mV - VTT)/(60 + RT) and VOL = -171 OmV - S*ILQUT (eqt. 22) where: ILOUT = (-1710mV - VTT) I (SO + RT) lODE Devices The equivalent output circuit is shown in Figure 3.16. The output levels are estimated from Figure 3.16 as follows: VOH = -S30mV - 6*I HOUT (eqt. 23) where: IHOUT = (-S30mV - VTT) I (60 + RT) -2.0 -1.75 -1.5 -1.25 -1.0 -0.75 -0.5 -0.25 Output Voltage (V) Figure 3.14· ECLinPS Output Charactersitics -1660mV IDE Devices The equivalent output circuit is shown in Figure 3.15. The output levels are estimated from Figure 3.15 as follows: VOH = -770mV - 6*I HOUT Figure 3.16· Equivalent Circuit for Calculating 1DDE Output Levels (eqt. 21) ECLinPS 4·25 System Interconnect and VOL = -1660mV - S*I LOUT As mentioned in the Transmission line section, series termination techniques are useful when the interconnect lengths are long or impedance· discontinuities exist on the line. Additionally, the signal travels down the line at half amplitude minimizing. problems associated. with crosstalk. Unfortunately, a drawback with this technique is the possibility of a two step signal appearing when the driven inputs are far from the end of the transmission line. To avoid this problem the distance between the end of the transmission line and input gates should adhere to the guidelines specified in Table 3.3 from the section on unterminated lines. (eqt. 24) where: 'LOUT = (-1660mV - VTT)/(SQ + RT) SIP Resistors The choice of resistor type for use as the termination resistor has several a!ternatives. Although the use of a discrete, preferably chip resistor, offers the best isolation and lowest parasitic additions there are SIP resistor packs which will work fine for ECLinPS designs. SIP resistors offer a level of density which is impossible to obtain using their discrete counterparts. However, there are some guidelines which the user should follow when using SIP resistor packs. Always terminate complimentary outputs in the same pack to minimize inductance effects on the SIP power pin. Noise generated on this pin will couple directly into all of the resistors in the pack. In addition, the SIP package should incorporate bypass capacitors in the design (Figure 3.17). These capacitors are necessary to help maintain a solid VTT level within the package, again mitigating any potential crosstalk or feedthrough effects. A 10 pin SIP like the DALE CSRC-1 OB21-500J/1 03M, is suitable for providing 50Q terminations while maintaining a relatively noise free environment to non-switching inputs. Series Termination Theory When the output of the series terminated driver gate switches, a change in voltage(d V B) occurs at the input to the transmission line:. (eql. 25) where:. V,N = Internal Voltage Change Za = Line Characteristic Impedance Rs = Output Impedance of the Driver Gate RST = Termination Resistance Since Zo = RST + Rs substitution into Equation 25 yields: (eqt. 26) From Equation 26 an incident wave of half amplitude propagates down the transmission line. Since the transmission line is unterminated at the receiving end, the reflection coefficient at the load is approximately unity; therefore the reflection causes the voltage to double at the receiving end. When the reflected wave arrives at the source end its energy is absorbed by the series resistance producing no further rellections as the impedance is equal to the characteristic impedance of the line. A extension of the series termination technique using parallel fanout eliminates the problem of lumped loading at the expense of extra transmission lines (Figure 3.19). Figure 3.17 - Standard Eel 10 pin SIP Series Termination Technique Series Damping is a technique in which a termination resistance is placed between the driver and the transmission line with no termination resistance placed at the receiving end of the line (Figure 3.1S). Series Termination is a special case of series damping in which the sum of the termination resistor (RST ) and the output impedance of the driving gate (Ra) is equal to the line characteristic impedance. STn D- R ~"")---Z-O-~ n ( total number of lines) Figure 3.18 - Series Termination Figure 3.19 - Parallel Fanout using Series Termination ECLinPS 4-26 System Interconnect Calculation of RE For the 1DE series RE functions to establish VOH and VOL levels and to provide the negative going drive into RST and Zo when the driver output switches to the low state. The value of RE must be such that the required current is supplied to each transmission line while not allowing the output transistor to turn off when switching from a high to a low state. An appropriate model is to treatthe output emitter follower as a simple switch ( Figure 3.20 ). VOH = -0.9V, VSWING = 0.85V, VEE = -5.2V [-0.9 - (-5.2)] / (RST + RE + Zo) ~ 0.531 / Zo For the lODE series: VOH = -0.955V, VSWING = 0.75, VEE = -4.5V (eql. 30) Figure 3.19 showed a modification of the series termination scheme in which several series terminated lines are driven by a single Eel gate. The principle concern when applying this technique is to maintain the current in the output emitter follower below the maximum rated value. The value for RE can be calculated by viewing the circuit in terms of conductances. Figure 3.20 - Equivalent Circuit for RE Determination The worst case scenario occurs when the driver output emitter follower is cutoff during a negative going transition. When this happens the switch can be considered opened, and at the instant it opens the line characteristic impedance behaves as a linear resistor returned to VOH' The model becomes a simple series resistive network as shown in Figure 3.21 . GE > G, + G2 + ... + G, (eql. 31) For the 1DE series 1/ RE ~ 1/ (7.10"Z01 -Rsn ) + 1/( 7.10"Z02 - RST2 ) + 1 / ( 7.1 O"Zo, -RST' ) (eql. 32) where n is the number of parallel circuits. For the lODE series Figure 3.21 - Equivalent Circuit with Output Cutoff 1 / RE ~ 1 / ( 6.56"Z01 -RST1 ) + 1 / ( 6.56"Z02 - RST2 ) + 1 / ( 6.56"Zo, -RST' ) The maximum current occurs at the instant the switch opens and is given by Equation 27. (eql. 27) (eql. 33) The initial current must be sufficient to generate a transient voltage equal to half of the logic swing since the voltage at the receiving end ofthe line doubles forthe series terminated case. To insure the pull down current is large enough to handle reflections caused by discontinuities and load capacitances the transient voltage is increased by 25%. Therefore, I'N'T = ( 1.25"VSWING/2 ) / Zo where n is the number of parallel circuits. When a series terminated line is driving more than a single Eel load the issue of maximum number of loads must be addressed. The factor limiting the number of loads is the voltage drop across the termination resistor caused by the input currents to the Eel loads when the loads are in the quiescent high state. A good rule of thumb is to determine if the loss in high state noise margin is acceptable. The loss in noise margin is given by NMLOSS = IT " (R ST + Ro) (eql. 34) (eql. 28) To satisfy the initial constraints 1M AX> I'N'T ECLinPS 4-27 System Interconnect RO is 150j.tA. Thus for the circuit shown in Figure 3.22 in which three gate loads are present in a 50n environment the loss in high state noise margin is calculated as: RST RE Zo NM LOSS = 3 • 150j.tA • 50n = 22.5mV ECLinPS 1/0 SPICE Modeling Kit VEE Due to the heavier reliance on simulation tools for initial prototyping, Motorola has put together a SPICE modeling kit aimed at aiding the customer in modeling board interconnect behavior. The kit includes representative drivers and receivers as well as the necessary SPICE model parameters. In addition tips are provided for simulating a wide range of output conditions. The kit, in conjunction with todays CAD tools, can greatly simplify the design and characterization of critical nets in a design. Anyone interested in obtaining a copy is encouraged to contact an ECLinPS Application Engineer. Figure 3.22 - Noise Margin Loss Example where: IT = Sum of I'NH Currents For the majority of devices in the ECLinPS family the typical maximum value for quiescent high state input current ECLinPS 4-28 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. DESIGN GUIDE Interfacing SECTION 4 Interfacing with ECLinPS ±200ps/ns gate-gain delay (delay vs input edge rate) can be assumed or a more typical value of ±75ps/ns can be used. Interfacing to Existing ECl Families There currently exists two basic standards for high performance ECl logic devices: 10KH and 100K. To maximize system flexibility each member of the ECLinPS family is available in both of the existing ECl standards: 1OE series devices are compatible with the MECl 10KH family; 100E series devices are compatible with ECl lOOK. The difference in the DC behavior of the outputs of the two different standards necessitates caution when mixing the two technologies into a single ended input design. As illustrated in Figure 4.1 and Table 4.1, there is no problem when a 1OKH device is used to drive a lOOK device, however problems arise when the scenario is reversed. For the case of a lOOK device driving a 1OKH device the worst case noise margin is reduced to 35mV, a noise margin which is unacceptable for most designs. Since the problems of interfacing are an output tracking rate vs a Vee tracking rate problem if the system uses only differential interconnect between the two technologies there will be no loss of noise margin and the design will operate as desired. Fortunately the ECLinPS family, by offering devices in both standards, allows the user to integrate higher performance technology into his design without having to battle these interface problems. Another area of concern when interfacing to older slower logic families is the behavior of ECLinPS devices with slower input edge rates. Typically, other than clock inputs, the ECLinPS family will function properly for edge speeds of up to 20ns. For edges significantly slower than 20nS the Schmitt trigger circuit of Figure 4.2 can be used to sharpen the edge rates reliably. Obviously a very slow edge rate will amplify differences in delay paths due to any offset of the Vee switching reference. This extra delay should be included in speed calculations of a design. For calculation purposes a worst case lOOK Driving 10KH -0.8 -1.0 ~ -1.2 .l!! ~ -1.4 ~ -1.6 .5 -1.8 "S ~C. o 15 30 45 60 75 90 75 90 Temperature (OC) 10KH Driving lOOK -0.8 -1.0 -1.2 -1.4 -1.6 Drvr> Rcvr NM - High NM-low 10KH> 10KH 10KH > lOOK 100K> lOOK 100K> 10KH 150mV 145mV 130mV 35m V 150mV 125mV 135mV 130mV -1.8 o 15 30 45 60 Temperature (OC) Table 4.1 - Worst-Case Noise Margins of a Mixed 10KH and lOOK Design Figure 4.1 -Interfacing 10KH ECl and lOOK ECl ECLinPS 4-29 Interfacing Clock inputs on flip flop devices in the ECLinPS family are especially sensitive to slow edge rates. Flip flops have been successfully clocked in a noise free bench setup environment with edge rates of up to 20ns. However in ATE systems where more noise is present clocking problems arise with input edge rates of greater than 6 or 7ns. To ensure reliable operation in a system with input clock edges slower than 7ns it is recommended that the signals be buffered with an ECLinPS buffer circuit (E122, E116, E101 etc) or, for extreme conditions, the Schmitt trigger of Figure 4.2 to provide the gain necessary to sharpen the edges on the clock pulse. designer is using a GaAs device which produces these -.3V signals in an ECLinPS design he contact a Motorola Applications Engineer to determine if the ECLinPS device being driven will be susceptible to saturation. ACCoupling In some cases it may be necessary to interface an ECLinPS design with a signal which lacks any DC offset. The differential devices in the ECLinPS family are ideally suited for this application. As pictured in Figure 4.3 the signal can be AC coupled and biased around the VBB switching reference of the device. Note that this scheme only works for a data stream with no DC bias, for data streams such as RZ or unencoded NRZ DC restoration must be performed prior to AC coupling it to an ECLinPS device. IN IN o--J r------,:---j OUT .001I1F 4000 50Q 50Q 01~ 50n 50n Figure 4.2 - Schmitt Trigger w/100mV of Hysteresis Figure 4.3 - AC Couple Circuit Interfacing to TIL/CMOS Logic To interface ECLinPS devices to TTL or CMOS subsystems there exists several new product offerings, as well as several existing devices, in the MECl 1OKH family which are ideally suited to the task. These translation devices are specially suited for clock distribution, DRAM driving as well as general purpose translation in both single supply and dual supply environments. In mixed technology environments it is recommended that the noisy supplies of TTL and CMOS circuits be isolated from the ECl supplies. This can be done either through separate power planes in the board or a common plane with isolated ECl. and TTL power sub-planes. The planes of common voltages (ie. ECl Vee and TTL ground for split supply systems or common Vee and ground for a single supply system) should then be connected to a common system ground or power supply through an appropriate edge connector. The 50n resistor of Figure 4.3 provides the termination impedance while the VBB pin provides the DC offset. The capacitor used to couple the signal must have an impedance of «50n for all frequency components of the input signal. Because large capacitors appear somewhat inductive at high frequencies it may be necessary to use a small capacitor in parallel with a larger one to achieve satisfactory operation. In addition it is important to bypass the VBB line when used in this manner to minimize the noise coupled into the device. Because the AC signal is biased around VBB' the output of the ECLinPS device when AC coupled will have a duty cycle identical to the input. Thus this type of application is ideal for transforming high frequency sinusoidal waveforms from an oscillator into a square wave with a 50% duty cycle. The E416 device is a specialized line receiver with a much higher bandwidth than alternative ECLinPS devices, therefore for frequencies of >500MHz it is recommended that the designer use this device. The above mentioned scenario will work fine as long as the input signal is present, however if the the inputs to the AC coupled device are left open problems may occur. With no input signal both inputs will go to VBB and an undefined output, and perhaps an oscillating output, will result. If a defined output is necessary for an open input scenario, the InterfaCing to GaAs Logic In general GaAs logic is designed to interface directly with ECl, however in some instances the worst case VOH of a GaAs output can go as high as -0.3V. An ECLinPS device, depending on the input structure, may become saturated when driven with a -0.3V signal. It is recommended that if the ECLinPS 4-30 Interfacing to generate a VBB reference with the necessary current sinking capability. A single gate configured in this way will source or sink upto 1OmA without a significant shift in the generated V BB level. If more current is needed several gates can be connected in parallel to provide the extra drive capability . Note that the circuit pictured in Figure 4.4 will result in the Q outputs going high when the inputs are left open. If the opposite is desired the resistor to Vee can be tied to the inverting input and V BB to the non-inverting input. 2.SKU IN 0---1 f - - t - - - - j OUT .0011J.F son sou Figure 4.4 - AC Couple Circuit with DC Offset 500 circuit of Figure 4.4 can be used. The resistor tree between Vee and V BB creates an offset between the two inputs so that if the driving signal is lost a stable defined output will occur. Unfortunately this configuration will adversely affect the duty cycle of the output. Depending on the frequency of the output, the duty cycle will change due to the longer distance to threshold on a rising edge as opposed to a falling edge. With this in mind it becomes obvious that the smallest feasible offset would be the best solution. For stability a minimum of 25mV is recommended, however this will not produce full ECl levels at the outputs of an E116 and thus another differential gate should be used to further amplify the Signal. The gain of the E4l6 on the other hand is sufficient to produce acceptable levels at the outputs for DC input voltage differences of 25mV. If a 150mV offset is used full ECl levels will be seen at the outputs of the E 116 however the price in duty cycle skew will be high. Of course if the signal is divided after it is received the duty cycle will be restored. When using the circuit of Figure 4.4 care should be taken to limit the current sunk by the VBB pin to a maximum of O.5mA. To achieve an offset of greater than 25m V for the circuit of Figure 4.4 the DC current will necessarily need to be greater than O.5mA. To alleviate this dillema one of the gates olthe E116 can be configured as pictured in Figure 4.5 Single Gate VBB Generator Multiple Gate VBB Generator Vn Figure 4.5 - High Current V•• Generator ECLinPS 4-31 MOTOROLA SEMiCONDUCTOR . . . . . . . . . . . . . . . . . . . . . .. DESIGN GUIDE Package & Thermal SECTIONS Package and Thermal Information Package Choice and reel product. Therefore, to order the MC10Elll FN in tape and reel the part number would become MCl OElll FNR2. ECLinPS is offered in the 28-lead plastic leaded chip carrier (PLCC) package, a leaded surface mount IC package. The lead form is of the "J-Iead" type. For detailed dimensions of the 28-lead PLCC refer to the package description drawings at the end of this section. The PLCC was selected as the optimum combination of performance, physical size and thermal handling in a low cost standard package. While more exotic packages existto improve these qualities still further, the cost of these is prohibitive for many applications. The PLCC features considerably faster propagation delay and reduced parasitics compared to a DIP package of similar pin-count; two properties that make it eminently suitable for very high performance logic. The 28-lead PLCC for the ECLinPS family is available in tape and reel form to further facilitate automatic pick and place. The characteristics of the 28-lead PLCC reel are described in Figure 5.1 below. Reliability of Plastic Packages Although today's plastic packages are as reliable as ceramic packages under most environmental conditions, as the junction temperature increases a failure mode unique to plastic packages becomes a significant factor in the long term reliability of the device. Modern plastic package assembly utilizes gold wire bonded to aluminum bonding pads throughout the electronics industry. As the temperature of the silicon Ounction temperature) increases an intermetallic compound forms between the gold and aluminum interface. This intermetallic formation results in a significant increase in the impedance of the wire bond and can lead to performance failure of the affected pin. With this relationship between intermetallic formation and junction temperature established it is incumbent on the designer to ensure that the junction temperature for which a device will operate is consistent with the long term reliability goals of the system. Reliability studies were performed at elevated ambient temperatures (125°C) from which an arrhenius equation relating junction temperature to bond failure was established. The application of this equation yields the table of Figure 5.2. This table relates the junction temperature of a device in a plastic package to the continuous operating time before 0.1% bond failure (1 failure per 1000 bonds) ECLinPS devices are designed with chip power levels that permit acceptable reliability levels, in most systems, under the conventional 500lfpm (2.5m/s) airflow. General Information Reel Size: 13" (330mm) Tape Width: 24mm Units/Reel: 500 Mechanical Polarization @EBEBEBEBEBEB Bb3JB View From Tape Side Thermal Management As in any system, proper thermal management is essential to establish the appropriate trade-off between performance, denSity, reliability and cost. In particular, the designer should be aware of the reliability implication of continuously operating semiconductor devices at high junction temperatures. The increasing popularity of surface mount devices (SMD) is putting a greater emphasis on the need for better thermal management of a system. This is due to the fact that SMD packages generally require less board spacethantheir through hole counterparts so that designs incorporating Linear Direction of travel Figure 5.1 - 28-Lead PLCC Tape & Reel Information To order ECLinPS in tape and reel form a minimum 3000 piece order per device type is required. In addition orders must be full reels or multiples of full reels as no partial reels will be shipped. An R2 suffix has been established to add to the end olthe part number to signify the desire for tape ECLinPS 4-32 Package & Thermal T = 6.376 x 10"' e [~~;51i2:~J towards smaller denser designs makes it incumbent on the designer to provide for the removal of thermal energy from the system. Users should be aware that they control many of the variables which impact the junction temperatures and thus to some extent the long term reliability of their designs. Where: T = Time to .1 % bond failure Junction Temp·ee) Time (Hrs.) Time (Yrs.) 80 90 100 110 120 130 140 1,032,200 419,300 178,700 79,600 37,000 17,800 8,900 117.8 47.9 20.4 9.1 4.2 2.0 1.0 Calculating Junction Temperature The following equation can be used to estimate the junction temperature of a device in a given environment: TJ TA Po e JA Figure 5.2 - TJ vs Time to .1% Bond Failure = Junction Temperature = AmbientTemperature = Power Dissipation = Avg Pkg Thermal Resistance (Junction - Ambient) The power dissipation factor is made up oftwo elements: the internal gate power and the power associated with the output terminations. Essentially the two contributors can be calculated separately, then added to give the total power dissipation for a device. To calculate the power of the internal gates the user simply multiplies the lEE of the device times VEE' Since lEE in ECl is a constant parameter frequency need not be factored into the calculations. A worst case or typical number for chip power can be calculated by using either worst case or typical data book values for the lEE and VEE of a device. Next the power of the outputs needs to calculated so that the total power dissipation for a device can be determined. The output power is dependent on the termination resistance and the termination scheme used to pulldown the outputs. The most typical termination scheme for ECLinPS designs is a parallel termination into -2.0V. For this scheme the following equation describes the power for a single output of the device: SMD technologies have a higher thermal density. To optimize the thermal management of a system it is imperative that the user understand all of the variables which contribute to the junction temperature of the device. The variables involved in determining the junction temperature of a device are both supplier and user defined. The supplier through lead frame design, mold compounds, die size and die attach can positively impact the thermal resistance and thus the junction temperature of a device. Motorola continually experiments with new package designs and assembly techniques in an attempt to further enhance the thermal performance of its products. It can be argued that the user has the greatest control of the variables which commonly impact the thermal per· formance of a device. Ambient temperature, air flow and related cooling techniques are the obvious user controlled variables, however PCB substrate material, layout density, size of the air-gap between the board and the package, amount of exposed copper interconnect, use of thermallyconductive epoxies and number of boards in a box can all have significant impacts on the thermal performance of a system. PCB substrates all have different thermal characteristics, these characteristics should be considered when exploring the PCB alternatives. The user should also account for the different power dissipations of the different devices in his system and space them on the PCB accordingly. In this way the heat load is spread across a larger area and "hot spots" do not appear in the layout. Copper interconnect traces act as heat radiators therefore significant thermal dissipation can be achieved through the addition of interconnect traces on the top layer of the board. Finally the use of thermally conductive epoxies can accelerate the transfer of heat from the device to the PCB where it can more easily be passed to the ambient. The advent of SMD packaging and the industry push V OUT RT = V OH or VOL = Termination Resistance The power dissipated in the output of a device is dependent on the duty cycle of that output. For an output terminated to VTT the worst case situation would be if the output was in the high state all of the time, for an output terminated to VEE the low state will represent worst case. For single ended output devices typically the power is calculated with the outputs in the worst case and for a 50% duty cycle. For differential outputs the power for a differential pair is constant since they are always in complimentary states, therefore for a given output the power will simply be the average of the high and low state output powers. Figure 5.3 shows the various output power levels for the different output types and conditions. In addition the table includes power numbers for various other ECLinPS 4-33 Package & Thermal Output Power (mW) Termination Resistance Differential Output Single-ended Output (50% Duty Cycle) Single-ended Output (Worst Case) 10E 100E 10E lODE 10E 100E SO to-2.0V 68 to-2.0V 100 to -2.0V 14.3 10.5 7.1 15.0 11.1 7.5 14.3 10.5 7.1 15.0 11.1 7.5 19.8 14.6 9.9 20.0 14.7 10.0 510 to-5.2V 330 to-S.2V 180 to-5.2V 9.7 15.0 27.S 9.8 15.1 27.8 9.7 15.0 27.5 9.8 15.1 27.8 11.8 18.3 33.5 11.7 18.0 33.1 510 to-4.5V 330 to-4.5V 180 to-4.SV - 7.8 12.3 22.6 - 7.8 12.3 22.6 - 9.3 14.4 26.4 - Figure 5.3 - Output Power for Various Termination Schemes termination resistances and alternative termination schemes. These numbers can be derived by determining the loUT and V OUT for the different alternatives and applying the equation above. Now that the power dissipation of a device can be calculated one needs to determine which level of the parameter ( ie. typical, max etc .. ) to use to estimate the long term reliability of the system. Since this number is statistical in nature simply applying the worst case numbers will be overly pessimistic as these parameters vary statistically themselves. It is not very likely, for instance, that every device type will be operating at the maximum specified lEE level. Assuming all worst case conditions can have a significant impact on the resulting junction temperature estimates leading to erroneous conclusions about the reliability of the design. Another important parameter for calculating the junction temperature of a device is the junction-ambient thermal resistance,ElJA , of the package. ElJA is expressed in °C per watt (OCIW) and is used to determine the temperature elevation of the die Ounction) over the external package ambient temperature. Standard lab measurements of this parameter for the 28-lead PLCC yields the graph of figure 5.4. An alternative calculation scheme forTJsubstitutes the case temperature (Tc) and the junction-to-case (ElJc) thermal resistance for their ambient counterparts in the TJequation previously mentioned. The. ElJC for the 28-lead PLCC is 32°CIW. This parameter is measured by submerging the device in a liquid bath and measuring the temperature of the bath, therefore it represents an average case temperature. Thedifficulty in using this method arises in the determination 80 70 60 ...
Source Exif Data:File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.3 Linearized : No XMP Toolkit : Adobe XMP Core 4.2.1-c041 52.342996, 2008/05/07-21:37:19 Create Date : 2017:07:17 08:48:07-08:00 Modify Date : 2017:07:17 09:12:20-07:00 Metadata Date : 2017:07:17 09:12:20-07:00 Producer : Adobe Acrobat 9.0 Paper Capture Plug-in Format : application/pdf Document ID : uuid:83ac4898-151c-fd45-b3f4-ddcf3ab872b9 Instance ID : uuid:2ed4a1db-2a43-5f45-8c13-9c15259ba743 Page Layout : SinglePage Page Mode : UseNone Page Count : 180EXIF Metadata provided by EXIF.tools