60333300B_3234_Mass_Storage_Ctlrs_Jan73 60333300B 3234 Mass Storage Ctlrs Jan73
User Manual: Pdf 60333300B_3234_Mass_Storage_Ctlrs_Jan73
Open the PDF directly: View PDF .
Page Count: 91
Download | |
Open PDF In Browser | View PDF |
CONTROL DATA .--.- .. CONTROL DATA® 3234 MASS STORAGE CONTROLLERS HARDWARE REFERENCE MANUAL REVISION RECORD REVISION A (10-1-70) DESCRIPTION Manual released; this publication obsoletes the 3234-A Mass Storal!e Controller section of the 3000 Series Peripheral EQuipment Reference Manual Pub. No. 6-108800. Information contained herein is the same as that found in revision U of the obsolete manual. B (1-31-73) Manual revised; incoroorates En!lineerin!l Chan!le Order 32840 which incorporates Field Change Orders 32554 and 32555. This edition obsoletes all previous editions. Publication No. 60333300 Address comments concerning this manual to: © 1970, 1973 by Control Data Corporation Printed in the United States of America Control Data Corporation Technical Publications Department 4201 North Le}{ington Ave. Arden Hills. Minnesota 55112 or use Comment Sheet in the back of th is manual. PREFACE This publication contains reference information for CONTROL DATA® 3234-A/B/c/D/E Mass Storage Controllers which may be used with the CONTROL DATA ® 3000 Computer Data Channels. The user of this manual should be familiar with the Control Data 3000 Series Data Channel 110 specifications and operating characteristics. 60333300 B iii CONTENTS 1. GENERAL DESCRIPTION Strip File 2 ... 37 Disk Storage Drives 2-37 Introduction 1-1 Unsafe Conditions 2-38 Functional Description 1-2 Sector Timing 2-39 Subsystem Configuration 1-3 Buffer Timing 2-40 Controller 1-4 Peripheral Storage Devices 1-5 Data Format 1-11 Loading and Unloading Procedures 2-42 Sector Format 1-12 Ready State 2-46 1-13 1-14 Procedures for Writing Address Headers 2-48 Access Time 1-15 Procedures for Writing Data 2-49 1738 Compatibility 1-16 Checkword Data Transfer 2. Manual Operations Power Sequence Sample Program Routines Program Sequence OPERATION AND PROGRAMMING Codes 2-1 Connect Code (NODU) 2-3 Function Codes Status Codes 2-5 2-18 3. 2-41 2-41 2-50 2-50 Program A 2-51 Program B 2-55 SWITCHES AND INDICA TORS Maintenafl.ce Panel 3-1 2-24 MAINTENANCE PANEL 3-2 Connect 2-24 CONTROL A and CONTROL B 3-2 Simultaneous Connect Mode Restrictions 2-24 LOGICAL/PHYSICAL UNITS 3-2 2-24 DEVICE SELECT 3-3 REGISTER SELECT 3-3 Register Indicators/Switches 3-3 DEFECTIVE SECTOR 3-3 DA TA (WRITE) 3-3 DATA (READ) 3-3 HEADER 3-4 WRITE LOCKOUT 3-4 TRACK 3-4 CYL/STRIP 3-4 FILE/CELL 3-5 Programming Considerations Master Clear, Release and Disconnect 2-25 Restore (Strip File) 2-25 Interrupts Abnormal EOP Interrupt 2-25 2-26 Seek/ Sector Verification 2-26 Overlap Seek 2-27 Seek Interrupts IX and 4X F unction Codes 2-27 2-28 Records 2-28 System Errors and Performance 2-29 Addressing 2-33 60333300 B v REPEAT 3-5 SEEK ADRS. 3-~ MODE 3-7 MASTER CLEAR 3-5 MARGIN SELECT 3-7 3-6 OFFSET and STROBE 3-8 Repeat Seek 3-6 FAULT CLEAR 3-8 Repeat Seek Operate 3-6 Card Switches Switch Panel 3-7 FIGURES 1-1 Typical Configuration 1-2 2-4 1-2 1-3 Disk Pack and Data Cell Cylinder Concept of Data Recording 1-3 2-5 1-5 2-6 Disk File Address Format Disk Storage Drive Address Format 2-35 Strip File Addres s Format 2-35 2-34 813/814 Cabinet 1... 6 2-7 Cell Indicator and Elevating Door 2-44 1-5 Disk File Physical Layout 1-6 2-8 Cell Array 2-45 1-6 853 Disk Storage Drive 1-9 2-9 1-7 Strip File 1-10 Cell, Cell Cover~ and Protective Container 2-46 1-8 1-9 Array Subdivisions Sector Format 1-10 1 ... 12 1-4 2-10 Flow Chart, Program A 2-52 2-56 3-1 1-10 Disk Pack Storage Arrangement 1 ... 11 Block Diagram: Cyclic Encoder 1-13 2-11 Flow Chart, Program B 3-1 Maintenance Panel 1.. 13 3-2 Card Switches 3-6 2-1 2-2 Connect Code Format Function Code Format 2-4 2-5 3-3 Switc h Panel 3-7 2-3 Margin Selection 2 ... 12 Strip File Address Conversion Charts 2-36 WRITE LOCKOUT Switch and Keyswitch Functions 3-4 TABLES 1- 1 Controller Capabilities 1-2 Unit Storage Format and Capacity Connect, Function, and Status Codes 2 ... 1 vi 1-2 1 ... 11 2-2 3-1 2-1 60333300 B 1 GENERAL DESCRIPTION INTRODUCTION The 3234 Mass Storage Controller interfaces 3000 computer system data channels and one or more of the peripheral storage devices listed in Table 1-1. The controller" data channel" and one or more storage devices comprise a mass storage s.ubsystem. The controller synchronizes data transfer and controls the programmed selection of operating modes between the data channel and storage devices (refer to Figure 1-1). This manual provides descriptive" programming" and operating information for the mass storage subsystem. The following are the peripheral storage devices. • Special Option 60076 (adapter) t - The special adapter chassis (housed in the con- troller cabinet) interfaces the controller logic to the strip file. • Disk File - Either the 813 or 814 Disk File unit and cabinet in which the disks and associated electronics are housed • Disk Drive - Either the 853 or 854 Disk Storage Drive units" their associated disk pack" and the cabinet in which the pack and associated drive unit logic are housed • Disk Pack - The 850 Disk Pack assembly on a common vertical axis. co~sisting of 10 recording disks mounted Each disk is approximately 14 inches in diameter. The packs are portable and interchangeable among various disk drive units. • Strip File t t - The 2321 Data Cell Storage unit and the cabinet in which the cells" subcells" strips, associated logic" and electronics are housed • Data Cell t - The strip file storage assembly. Each cell consists of 20 sub-cells each containing 10 recording strips. The cells are portable and interchangeable between the various strip file units. A full complement of 10 cells (either data or ballast) is required to complete the array necessary for strip operations. • Ballast Cell t -A t dummy cell used to maintain the proper balance of the array when less than a full complement of data cells is used Figure 1-2 shows a disk pack and a data cell in their protective containers. t Applies to 3234-A/B/C models only t t Throughout the manual the IBM 2321 Data Cell will be referred to as the strip file. 60333300 B 1-1 TABLE 1-1. CONTROLLER APPLICATIONS Applicable Storage Devices and/or Error Recovery Capability Controller Model 813/814 Disk File 853/854 Disk Stor age Dri ve 2321 Strip File Sp Opt 60076 Data Gell Error Recovery Capability 3234-A Yes Yes Yes No 3234-B Yes Yes Yes No 3234-C Yes Yes Yes No 3234-D Yes Yes No Yes 3234 .. E Yes Yes No Yes - FUNCTIONAL DESCRIPTION The controller. in conjunction with the applicable peripherals (singly or in any combi- nation), operates as a mass storage SUQsystem having medium access time, nonvolatile, mass storage facilities. The subsystem provides large volume data storage with random access and interchangeable storage pack and data cell capabilities. 813/814 3234 DATA CHI\N STORAGE CONTROLLER , DATA CHAN DISK FILES MASS 853/854 - ~ DISK STORAGE DRIVES r--- - I SP I OPT - ~ : 60076 Figure 1-1. 1-2 . 2321 STRIP FILES Typical Configuration 60333300 B SUBSYSTEM CONFIGURATION The mass storage subsystem has expandable capabilities. The minimum subsystem consists of a single data channel, a controller, and one of the storage devices. The subsystem can be expanded with the addition of a second data channel and additional stor a ge devices. When two data channels are connected in the subsystem they operate on a time-shared basis. The maximum subsystem consists of two data channels, a controller, and up to eight peripheral storage devices. t To incorporate strip file units into the subsystem for the applic able controller models, the adapter chassis (special option 60076) must be added to the controller. The adapter can a ccommodate up to four strip file units. Figure 1-2. Disk Pack and Data Cell The types of storage units used in the subsystem may be intef mixed in any combination or may be all of one type. controller. No provision is made for more than one ad apter in the Therefore, no more than four strip files may be connected to the controller. The subsystem operates under program control from the computer. A 12-bit connect code selects the equipment (controller) a nd the unit (peripheral storage device). The 12-bit func- tion codes provide for selecting operating modes, interrupts, and loading and unlo a ding address information. t The 814 Disk File is considered as two storage units. 60333300 B 1-3 SUBSYSTEM CONFIGURATION The mass storage subsystem has expandable capabilities. The minimum subsystem consists of a single data channel, a controller, and one of the storage devices. The subsystem can be expanded with the addition of a second data channel and additional storage devices. When two data channels are connected in the subsystem they operate on a time-shared basis. The maximum subsystem consists of two data channels, a controller. and up to eight peripheral storage devices. t To incorporate strip file units into the subsystem for the applicable controller models, the adapter chassis (special option 60076) must be added to the controller. The adapter can accommodate up to four strip file units. Figure 1- 2. Disk Pack and Data Cell The types of storage units used in the subsystem may beintet-mixed in any combination or may be all of one type. controller. No provision is made for more than one adapter in the Therefore, no more than four strip files may be connected to the controller. The subsystem operates under program control from the computer. A 12-bit connect code selects the equipment (controller) and the unit (peripheral storage device). The 12 .. bit func- tion codes provide for selecting operating modes. interrupts. and loading and unloading address information. t The 814 Disk File is considered as two storage units. 60333300 B 1-3 CONTROLLER The standard 3000 series signals are exchanged between the controller and the data channels. The controller translates the connect code and the function codes issued by the computer to control peripheral operation. data channels. The controller buffers data between the peripherals and the It accepts data from the data channels in a parallel format, disassembles it, and transmits the data serially to the appropriate storage unit. In a similar manner, the controller accepts serial data from the selected storage unit, assembles it into 12-bit bytes, and transfers it in parallel to the appropriate data channel. The computer controls the subsystem operations through the 12-bit function codes and a 24-bit address word. Issuing a function code specifying the mode of operation prepares the controller and selected peripheral unit for an I/O operation. The issuance of a load address function code causes the controller to prepare for a seek operation t within the selected storage unit. Prior to the issuance of the load address code, the computer must do a write operation to issue output data containing address information to the controller. The data is loaded into the address register of the cont~oller and the seek operation starts immediately upon termination of the output operation. If the address loaded is a legal address, the selected storage unit seeks the position specified by the contents of the register in the controller. A sector verify operation automatically begins upon initiation of the I/O at the location specified by the address. If no load address operation is performed prior to receipt of a read or write Signal, the I/O operation begins at the address presently held in the address register (cylinder and sector plus 1; location at which the previous operation e~ded) t t except when a unit is at the last legal address. In the latter Case, a new seek must be initiated or an address error is generated. The 3234-D/E models contain error recovery procedures. A variety of functions are available for error identification and correction, depending on the type of error present. t An operation in which the R/W heads are physically pOSitioned to the addressed cylinder t tWhen an I/O operation is in progress (or ends) the controller automatically updates the address registers at the end of each sector except when an address error is detected, or when the abnormal EOP interrupt is selected and an abnormal condition occurs. 1-4 60333300 B PERIPHERAL STORAGE DEVICES The" peripheral storage devices use a ferrous oxide coating as the magnetic recording medium. The disk file and drive use the cylinder concept of recording to minimize positioner movement (Figure 1-3). The data is read or written consecutively from sector to sector within a track and from a track on one side of a disk to the same track on the opposite side of the same disk, then to the same track on the next sequential surface of the next disk, and so on. This method uses internal switching of the R/W heads to minimize positioner movement, which results in the data being recorded in a cylindrical drum -type pattern. Figure 1-3. Cylinder Concept of Data Recording DISK FILES The 813 and 814 Disk Files are contained in similar cabinets (Figure 1-4). cabinet contains two drive spindles. of the spindle (Figure 1- 5). A stack of 18 disks can be mounted on each half The 18 disks provide a total of 32 recording surfaces (the top and bottom s-urfaces are not used for recording). two groups. Each Each stack is subdivided into The spindles rotate (nonsynchronously) at approximately 1180 revolutions per minute less induction slip. 60333300 B 1-5 3G6 Figure 1- 4. Figure 1- 5. 1-6 813/8 14 Cabinet Disk File Physical Layout 60333300 B Each disk is approximately 26 inches in diameter; each stack is approximately 18 inches high. The 813 contains two stacks of disks mounted on the upper half. of the spindles. A single hydraulically positioned access assembly serves the two stacks with two separate horizontally opposed groups of R/W head arms. Mounted on the end of each head arm are two head pads~ each containing four R/W heads. for each of the 32 recording surfaces in a stack. There are four R/W heads The accesses (which move simulta- neously in opposite directions) can be positioned to any of 64 positions to provide a total of 128 recording cylinders (64 per stack). The 814 contains four stacks mounted on the two spindles with two independent hydraulically positioned access assemblies serving the four stacks. and positioner~ Each half (upper disks or lower disks and positioner) of the 814 Disk File is considered an independent unit and is addressed~ accessed~ and operated on independently. Thus. access and operation to either half of an 814 Disk File is identical to the access and operation in an 813 Disk File. DISK STORAGE DRIVES The disk storage drives (Figure 1-6) are random access storage devices. The access mechanism consists of ten arms mounted in pairs on a movable carriage. Each pair of arms is positioned between two disks. of each arm. A single R/W head is mounted on the extremity On an initial seek (performed by loading a disk pack~ closing the cover~ and pressing the START switch). the carriage moves horizontally from an initially retracted position (to facilitate disk pack insertion and removal) to the first track near the edge of the disk. The unit then moves the heads to the innermost track and then withdraws the heads back to track 000. During the latter process, the heads are loaded (put in a recording attitude near the surface of disks). The 853 and 854 Disk Storage Drives are contained in similar cabinets. The 853 has provlslons for positioning to 100 cylinders; the 854 for positioning to 203 cylinders. Both types utilize the standard 851 Disk Pack; however. because track spacing is different in the two types of devices, a pack recorded on one type (for example. 853) cannot be read on the other type (for example, 854). The packs rotate at approximately 2400 revolutions per minute. 60333300 B 1-7 STRIP FILE t The strip file (Figure 1-7) is a direct access storage device. The unit is capable of retrieving and restoring individual magnetic storage strips from an array of cells. data is written and read from the strips which are approximately 13 inches inches wide~ and O. 005 inch thick. long~ The 2-1/4 Each strip is coated on one side with an iron oxide recording medium and on the other side with an antistatic carbon coating. Each file contains 2000 strips set in an array consisting of 10 removable cells and arranged in a vertical cylinder (Figure 1- 8). with 10 recording strips per subcell. During Each cell is subdivided into 20 subcells operation~ the array is hydraulically positioned (in the direction of least travel) to any of the 200 subcells. linkage then selects the addressed it on a small drum. strip~ A mechanical pulls it from its storage location, and places The drum rotates the strip past a R/W head bar assembly. The assembly contains 20 R/W heads and can be positioned to any of five locations to provide access to the 100 recording tracks on the strip. When use of the strip is com- pleted, it is restored to its original storage location in the subcell. t Applies to 3234-A/ B/ C models only 60333300 B Figure 1-6. 60333300 B 853 Disk Storage Drive 1-9 Figure 1-7. Strip File SUBCEll 10 Strlpt per 5ube.1I 10 C.lh per Drive CEll 20 Subc.lI. per C.II Figure 1-8. 1-10 Array Subdivisions 60333300 B DATA FORMAT The. basic data word consists of a 12,..bit byte. Multiple bytes are recorded in each sector with multiple sectors recorded in each of the storage mediums. Physical arrangement of the tracks and cylinders in the storage devices varies dependent upon the device. Table 1-2 indicates the sector. track. and cylinder formats for the vari- ous storage devices. TABLE 1- 2. ~ 813 Division 814 Byte Sector Track Cylinder Access UNIT STORAGE FORMAT AND CAPACITY 853 2321 t t 854 12 bits per byte (all units) 128 bytes 128 bytes 128 bytes 128 bytes 1380 bytes (1536 bits) (1536 bits) (1536 bits) (1536 bits) (16,560 bits) 32 sectors 32 sectors (49,152 bits) 16 sectors 16 sectors (49,152 bits) (24" 576 bits) (24,576 bits) 1 sector (16,560 bits) 128 tracks (6,291,456 bits) 128 tracks (6,291,456 bits) 10 tracks (245,760 bits) 10 tracks (245,760 bits) 20 tracks (331,200 bits) 128 cylinders 128 cylinders NAt NA NA NA NA 5 cylinders . (805,306,568 bite4 (805, 306,568 bits) Strip t t NA NA (1,656,000 bits) Subcell t Cell t t t NA NA NA NA NA NA NA NA 10 strips 20 subcells (331,200,000 bits) Unit Total Capacity t t t 1 access 2 accesses 100 cylinders 200 cylinders 10 cells (805,306,568 bits) (1,610,612,936 bits) (24.576,000 bits) (49,152,000 bits) (3,312,000, 000 bits) Not applicable Applies to 3234-A/B/C models only 60333300 B 1-11 SECTOR FORMAT Figure 1-9 shows the basic sector format used in the peripheral storage devices. While only the datl;l. field in each sector can be changed by the program, the entire format is shown to aid in understanding the make-up of each sector in the track. Figure 1-10 shows disk, track, and sector organization for a disk pack (the disk file is similar). I BIT + ® ® ® f..-----j._-----I.--_-'------'------L-----.-t HEAD GAP SYNC PATTERN DATA FIELD r J tII SECTOR MARK I TOLERANCE GAP II END OF RECORD BIT Tolerance gap I ® Disk file Disk drive 60 hits Strip file t 360 bits Addres s field Upper portion Lower portion Checkword Head gap Disk file Disk drive 60 hits Strip file t 360 bits Disk file Disk drive 128 bytes + 1 (1536 bits) Strip file t 1380 hytes + 6 (16, 560 hits) Disk file Disk drive 95 hits Strip filet 830 bits Data field Tolerance gap II Provides initial head switching time Write lockout and defective track bits; sector address information and checkword 12 hits each Figure 1-9. R/W head, erase gap 12-bit checkword 12-hit checkwords Motor speed variation and R/W. erase gap Sector Format t Applies to 3234-A/B/ C models only 1-12 60333300 B Figure 1-10. Disk Pack Storage Arrangement CHECKWORD Address and data transfers are checked for accuracy by generation of a 12-bit dant checkword in the controller. During operations the controller generat~s redun~ and veri- fies checkwords to determine the correctness of addresses and data transferred be": tween the controller and the various peripheral storage devices. As data (or addresses) is read or written, each bit is also fed into the cyclic encoder. As the data (or addresses) enters the. encoder, it is shifted through the encoder and half adder stages (A, B, C, D, and E, data or address is entered. Figure 1-11) continuously until the last bit of The remainder is a cyclic code generated from the address or data being transferred. The checkword is obtained by dividing the address or data 12 3 (which is taken as a code polynomial) by the polynomial X + XlI + X + X2 + X + 1 (Figure 1-11) . • C D E • ATA AtX).I+X+X2+XJ+XIl+XI2 o 0 I I 'c O+ Figure 1-11. 60333300 B "HALF ADQEIt " - 0 o - 101 - I I 0 Block Diagram: Cyclic Encoder - The checkword is the 12-bit code left in the cyclic encoder after the last bit of address or data has entered the encoder. This is then written immediately following the ad- dress when the address is originally written or at the end of the sector for each write function (six checkwords are written per sector in the strip file). DUJ;ing subsequent address verification or read operations the address or data being read is again fed into the encoder I and if the original data was written and read correctly, the two check- words cancel each other. The encoder is then in a clear state. (If any stage of the encoder is set upon completion of the operation l an error has occurred in either the original writing on the storage unit or during the subsequent read operation, and a checkword error indication is generated.) The theory behind the operation of the cyclic encoder (cyclic code) involves a mathelnatical formula. t The error detection capability of the cyclic encoder falls within the following limits l where a burst equals the number of bits between any two bits in error. 1. If a burst is 12 bits or less l all errors are detected. 2. If a burst is 13 bits, an undetected error occurs approximately every 11 2(one out of every 2048 errors is undetected). 3. If a burst is 14 bits or greater, an ·undetected error occurs approximately every 2 -12 . (one out of every 4096 errors is undetected). DATA TRANSFER Data is addressed and written in a storage unit in discrete blocks (sectors); however I the data channel can read or record as little as one byte or as many bytes as necessary to reach the end of file or end of cell (in the case of the strip file). When reading or writing, the operation must commence at the start of a sector. When writing, if less than a full sector is writtenl the remainder of the sector is automatically filled with zeros. The nominal data transfer rate is: Disk file Disk drive Strip file (3234-A/B/C) 98 000 bytes per second 1 100 1 000 bytes per second 36 1 000 bytes per second t For a general description of cyclic code theory refer to "Error Correcting Codes!! by W. Wesley Peterson, MIT Press. For a detailed description of the type of cyclic encoder used in this controller refer to "A Study of Methods of Error Detection During Transmission of Binary Inform at ion II by Carl Superko, Control Data Corporation. 1-14 60333300 B Timing diagrams which indicate the relationship between the data signals and the reply signals during buffered operations are shown under Programming Considerations. ACCESS TIME Access time is equal to the cylinder-positioning time plus the rotational-latency time. The following maximum positioning times were achieved as an average over several maximum positioner moves. Maximum access time is the sum of maximum positioning time plus maximum latency time. Average access time is the sum of average positioning time plus average latency time. DISK FILE Maximum access time Average access time 162. 9 milliseconds 95. 7 milliseconds Maxim um positioning tim e (major move) Maximum positioning time (minor move) Average positioning time (1/3 of maximum move) 100.0 milliseconds 25. 0 milliseconds 70.0 milliseconds Maximum latency Average latency 51. 4 1. 5 milliseconds 25. 7 milliseconds ::!: DISK STORAGE DRIVE Maximum access time Average access time 190. 0 milliseconds 107.5 milliseconds Maximur.n positioning time Average positioning time (1/3 of maximum move) Cylinder to cylinder positioning time 165.0 milliseconds 95. 0 milliseconds 30. 0 milliseconds Maximum latency latency 25 STRIP FILE O. 5 milliseconds 12. 5 milliseconds :i: Av~rage t Access time is defined as the interval from the issuance of a seek command from the controller to the 2321 until the generation of a strip ready signal in the 2321. time to restore a previously addressed strip (if required). This includes Average access times under varying conditions are as follows: Move head bar (can be overlapped with other motions) Select head R/W element Maximum laten.cy Restore strip t 95. 00 O. 10 50 :i: 0.65 200. 00 milliseconds milliseconds milliseconds milliseconds Applies to 3234-A/B~C models only 60333300 B 1-15 Strip Seek Time Without Restore 175 milliseconds Number of Subcells Moved Strip Seek Time With Restore o 375 milliseconds 250 milliseconds 1 450 milliseconds 350 milliseconds 50 550 milliseconds 400 milliseconds 100 600 milliseconds 1738 CQMPATABILITY Certain design and operatin~ compatibility exists between the 1738 Disk Drive Subsystem and the 3234 Mass Storage Subsystem. For a complete explanation, refer to Disk Storage Drives under Programming Considerations. 60333300 B 2 OPERATION AND PROGRAMMING COCES Table 2-1 lists all codes applicable to the controller and the associated subsystem. A complete explanation of each code follows the table. TABLE 2-1. CONNECT, FUNCTION, AND STATUS CODES Connect NODU t Connect Code Function Release and disconnect 0000 Restore 0001 + 1 cylinder seek t t - 1 cylinder seek t t 0002 0003 Clear 0005 Load address 0010 Return address 0011 Read disk address t t 0013 Select interrupt on ready and not busy 0020 Release interrupt on ready and not busy 0021 Select interrupt on end of operation 0022 Release interrupt on end of operation 0023 t N = equipment number of controller tt 60333300 B D = U = unit number of storage device device type (1 = Disk Drive; 2 = Disk File; 3 = Strip File) 3234-D/E only 2-1 TABLE 2-1. CONNECT~ FUNCTION~ AND STATUS CODES (Cont'd) Function Select interrupt on abnormal end of operation 0024 Eelease interrupt on abnormal end of operation 0025 Select interrupt on oppos ite channel releas e 0026 Release interrupt on opposite channel release Select interrupt on end of seek Release interrupt on end of seek Disk file status t 0027 0030 0031 0032 Clear disk file status t 0033 Disable sector verify t 0034 'Write address normal t 0035 'Write address defective t 0036 Write address lockout t 0037 Read Write Search compare 0040 0041 0042 Masked search compare Checkword verify 0044 Read che ckword 004~ 0043 Magnitude search (record < buffer) Magnitude search (record ~ 0050 buffer) 0051 Equality search (record;:: buffer) 0052 Buffer mode 0053 End of record mode Margin selection t 0054 04XX Status Ready o XXXI Busy Abnormal / unavailable On sector I XXX2 2 XXX4 3. '2" XXIO Address error 3~ XX14 t 2-2 2 3234-D/E only 60333300 B TABLE 2-1. CONNECT. FUNCTION. AND STATUS CODES (Cont'd) Status No compare 4.2 XX20 Lost data 4.2 XX24 End of record 5.2 XX40 Checkword error 5.2 XX44 Write lockout on read (normal) 6.2 "Write lockout on write (abnormal) 6.2 Positioner ready X1XO X1X4 7 X2XX End of operation interrupt 8 X4XX Abnormal end of operation interrupt 9 1XXX Seek interrupt 10 2XXX Reserved 11.2 4XXO Defective track 11.2 4XX4 CONNECT CODE (NODU): The 12- bi~ connect code (Figure 2-1) designates the equipment (controller). the peripheral device type (disk drive. disk file. strip file). and the unit with which the computer is to communic ate. Reservation capability is set up on a data channel basis. Once the data channel is con- nected to the controller and unit. the controller and unit are reserved until specifically released by that channel via a master clear. channel clear. or release and discbnnect function code. Upon receipt of the connect code by the controller. a reply or reject is returned to the data channel. If the desired controller. device. and unit are available. a reply is returned immediately. a reject is returned. If the controller is unable to accept and perform the connect .. Upon receipt of a reject. the computer must request a status response and interrogate the status bits in order to determine whether the reject was a result of the controller being reserved. the storage type being nonexistent. or the unit being unavailable. Refer to explanation of status response bits unavailable (XXX4) and reserved (4XXX). 60333300 B 2-3 II 13 9 e 65 BITS~3 32 0 BITSl3 BITSI '---y-l '---y---1'-y--1 EQUIP DEV ICE UNIT NO. (PER IPHERAL TYPE) NO. Figure 2-1. Connect Code Format BITS 0 THROUGH 2 Bits 0 through 2 designate the logic unit number (0 through 7) with which the program communicates. The number designated is variable from 0 through 7 by means of the unit designation jumper wires located on the mainten~nce control panel (Figure 3-1). BITS 3 THROUGH 5 Bits 3 through 5 designate the type of peripheral device with which the program is to communicate. All other device codes are illegal and cause a reject to be XX1X Disk drive XX2X XX3X Disk file i~sued. Strip file BITS 6 THROUGH 8 Unused BITS 9 THROUGH 11 Bits 9 through 11 designate the equipment with which the progranl desires to communicate. The number designating the equipment (controller) is variable from 0 through 7 by means of the rotary equipment number switch located in the controller cabinet. 2-4 60333300 B FUNCTION CODES The function codes consist of control and address codes which affect the reservation of the controller and the positioning of the R/w heads. interrupt selections. The interrupt' codes set and remove The error recovery codes (3234-D/E) allow flexibility and control in the determination and correction of various error conditions. The 110 codes define the input and output operations and the mode select codes are used to set parameters of the search-compare operation. II o 65 6 BITS Figure 2- 2. I Function Code Format A reject is issued upon receipt of a function code whenever the controller is busy with an operation other than a seek. A reply is issued upon receipt of an unassigned function code, but the code does not produce action in the controller. CONTROL AND ADDRESS CODES The control codes provide the computer with a means of releasing the subsystem, removing all reserves, I/O selections, and interrupt signals, deselecting the data chan~el, and restoring the recording strip (in the case of the strip file), thus reducing the address time on future references to new strips. The address codes are used to address the selected storage unit or to return the current address. In either case the address is loaded into or taken from the controller address register. format varies for each type of storage unit. The address (Refer to addressing for complete address and format information.) RELEASE AND DISCONNECT (0000) This code logically releases the subsystem from the data channel. It causes all reserves to be removed, clears all interrupt signals, removes all 110 code selections, clears the read parity and write check error conditions, and drops the status response lines. 60333300 B 2- 5 RESTORE (0001) Dependent upon the type of device selectedl the restore code causes the following: Disk Drive Initiates a seek to positiori (cylinder) zero of the disk pack Disk File Causes the positioners to move out of the recording area Strip File t CaQses the file to restore the recording strip to its proper subcell location. If no strip is presently loaded, the code is ignored (no-op). Use of the restore code, in conjunction with the strip file, reduces strip wear as well as access time of possible future reqllests to a different strip. + 1 CYLINDER SEEK (0002)t t The +1 cylinder seek function code initiates a forward one-cylinder seek. A. seek for- ward and a difference count of 1 is sent to the disk unit without computing the difference or restoring to zero. Th,e automatic restore is cleared. the contents of the address register in the controller l This function does not affect the cylinder, head l and sector register in the drive. Interrupts resulting from this operation are the same as a seek operation, except for the seek interrupt. It is generated if the last operation was a seek but it is not generated if the last operation was a read or write. -1 CYLINDER SEEK (0003) t t The ..,1 cylinder seek function code is identical to the 0002 function code l except that a seek is initiated in the reverse direction. CLEAR (0005) The clear code clears all major components in the subsystem, but does not affect the connect, reserved, or unit select statu$ of the subsystem. t tt 2-6 Applies to 3234-A/B/C models only Applies to 3234-D/E models only 60333300 B LOAD ADDRESS (0010) The load address code causes 'the next output data to be loaded into the 24-bit controller address register. The data is continuously loaded into the address register (in the format indicated under addressing) until the output operation terminates. On completion, a seek operation is automatically initiated in the selected storage unit. All subsequent load address functions to the same storage unit are rejected until the seek is completed (or it is determined that the seek will be incomplete). RETURN ADDRESS (0011) The return address code, in conjunction with an input operation, causes the controller to return the content of the 24..,bit address register to the data channel. is continuously returned (in the format indicated under addressing) The address until the input oper ation terminates. READ DISK ADDRESS (0013) t The read disk address function code causes a 36-bit input buffer to return the address read from the disk unit. The address re,ad is that specified by the address register. An end 'of operation interrupt is generated after the third 12-bit byte of data. An end of record is sent on the fourth byte, if requested. INTERRUPT CODES The interrupt codes establish and remove the interrupt selections that determine which conditions send an' interrupt signal to the data channel. A manual master clear or channel clear removes all interrupt selections (except the interrupt on opposite channel release for the channel that is reserved). An interrupt signal may be dropped by a manual master clear, channel clear, or any function code including reselecting the same selection. t Applies to 3234-D/E models only 60333300 B 2-7 SELECT INTERRUPT ON READY AND NOT BUSY (0020) The select interrupt on ready and not busy code activates the interrupt line the next time the subsystem becomes ready and not busy (at the end of the next operation). RELEASE INTERRUPT ON READY AND NOT BUSY (0021) The release interrupt on ready and not busy code removes the associated interrupt selection set up by the 0020 code. No interrupt notification of ready and not busy is sent until the condition is res elected. SELECT INTERRUPT ON END OF OPERATION (0022) The select interrupt on end of operation code causes the 'interrupt line to be activated and the associated status bit to be set on completion of the next operation whether the end of operation is normal Or abnormal. The following conditions cause an interrupt on end of operation. 1. At the end of the sector in which a read or write operation ended 2. When the channel goes inactive at the end of a return address operation 3. After a seek command is sent to the storage unit on a load address operation RELEASE INTERRUPT ON END OF OPERATION (0023) The release interrupt on end of operation code removes the associated interrupt selection set up by the 0022 code. No interrupt indication of end of operation is sent until the condition is res elected. SELECT INTERRUPT ON ABNORMAL END OF OPERATION (0024) The select interrupt on abnormal end of operation code causes the interrupt line to be activated and the associated status bit to be set on the stopping of an operation because of any abnormal condition within the controller or selected storage unit. 2-8 60333300 B The following conditions are considered abnormal. 1. Detection of an address error 2. Any attempt to write at a location which is in a write lockout state 3. Occurrence of a checkword error (indication of an error in the data read from the selected storage unit) 4. Occurrence of a lost data error (data missed due to channel failure to maintain a fast enough transfer rate) 5. Any attempt to perform an operation on a defective track 6. Dropping ready during any operation RELEASE INTERRUPT ON ABNORMAL END OF OPERATION (0025) The release interrupt on abnormal end of operation code removes the associated interrupt selection set up by the 0024 code. No interrupt indication of abnormal end of operation is sant until the condition is reselected. SELECT INTERRUPT ON OPPOSITE CHANNEL RELEASE (0026) The select interrupt on opposite channel release code causes an interrupt signal to be sent whenever the opposite data channel (the channel presently maintaining a reserved state of the controller) releases its reservation of the controller and storage units. If only one data channel is connected to the controller, this code is not applicable and should not be used. NOTE The interrupt is conditioned upon the dropping of the reserve. Therefore, a master clear causes the interrupt only if the data channel executing the master clear has the subsystem reserved. RELEASE INTERRUPT ON OPPOSITE CHANNEL RELEASE (0027) • The release interrupt on opposite channel release code removes the associated interrupt selection set up by the 0026 code. No interrupt indication of a release by the opposite channel is sent until the condition is reselected. 60333300 B 2-9 SELECT INTERRUPT ON END OF SEEK (0030) The select interrupt on end of seek code causes the interrupt line to be activated and the associated status bit set at the end of the seek operation (in any storage unit) regardless of whether the seek was complete or incomplete. RELEASE INTERRUPT ON END OF SEEK (0031) The release interrupt on end of seek code removes the associated interrupt selection set up by the 0030 code. No interrupt indication of an end of seek operation is sent until the condition is res elected. ERROR RECOVERY CODES t Error recovery codes provide for positive error identification and recovery techniques. They allow, when needed, sampling of conditions in the disk file, disabling sector verify, or writing of address headers without certain restricttons. The margin selection function allows movement of the positioner in small increments to circumvent track defects, if necessary. It also provides for varying the read strobe to adjust for fluctu- ations in read timing. DISK FILE STATUS (0032) t The disk file status function brings up the read status select Hne to enable the file fault status onto the control bus. receive the file status. The controller then switches the channel status lines to The function is accepted by the controller and a reply generated whether the controller is ready or not, provided there is no transmission parity error. A reject is generated if the controller is busy. This function is cleared by a 0033, 0000, 0005 function, or a master clear. The file status Can be displayed on the maintenance panel by setting the REGISTER SELECT switch to FILE STATUS. t The following are the status bit assignments. Applies to 3234-D/E models only 2-10 60333300 B Bit 0 Vol tage error Bit 1 Logic error Bit 2 Write error Bit 3 Bit 4 Head select X error Head select Y error Bit 5 Read/ write error Bit 6 through Bit 11 } Unused CLEAR DISK FILE STATUS (0033) t The clear disk file status function code clears the disk file status function selection. The reply and reject conditions are the same as for the 0032 function code. DISABLE SECTOR VERIFY (0034) t The disable sector verify function code suppresses the need to verify the address before a read, write" or search operation. This function must precede the 004X functions and clears on the next end of operation. Multisector operation is possible. The defec- tive sector and write lockout status is suppressed during this operatiou. Selection of this function requires that the correct address of the pack or disk must have been specified in the last seek operation. A read disk address function (0013) should precede this function to ensure that the head is properly positioned. WRITE ADDRESS NORMAL (0035) t The write address normal function" when followed by a load address operation, writes a new header in the sector specified by the address register. ation initiates the write header operation. write header. The load address oper- No seek interrupt occurs on a programmed An end of operation is generated if previously selected. A read disk address function (0013) should precede this function to ensure that the read/write head is properly positioned. t Applies to 3234-D/E models only 60333300 B 2-11 WRITE ADDRESS DEFECTIVE (0036) t The write address defective function code is identical to the 0035 function except that the defective sector bit is added in the header. WRITE ADDRESS LOCKOUT (0037) t The write address lockout function code is identical to the 0035 function code except that the write lockout bit is added in the header. MARGIN SELECTION (04XX) t The margin selection function code sends eight bits to the disk which add small increments in cylinder movement or allow small variations in the read strobe time, depending on the bit assignments (Figure 2-3). A reply is sent to the channel if the controller is ready and not busy, and if no transmission parity error exists. A reject is generated if the controller is busy or not ready, with one exception (the clear fault function (0600) is accepted if the controller (or unit) is not ready). II 10 9 8 1 6 , NOT USED 5 4 o 2 T T READ cYLINDER OFFSET STROBE 1 CLEAR FAULT - - - MARGIN SELECTION FUNCTION Figure 2-3. t Margin Selection Applies to 3234-D/E models only 2-12 60333300 B CYLINDER OFFSET The cylinder offset portion of the function moves the positioner off cylinder in increments of 0.00075 inch in the 813 Disk File. minimum of 10 milliseconds. The disk file drops on cylinder for a The maximum time is less than one cylinder move. busy status goes up during the function and drops when the on cylinder comes up. read operations are allowed off cylinder. The Only A write operation in any offset position other than zero generates a fault and drops ready. No interrupts are generated during this operation. Function 0400 0401 0402 0403 0404 0405 0406 0407 0410 0411 0412 0413 0414 0415 0416 0417 Offset (inch) -.00000 +.00000 -.00075 +.00075 -.00150 +.00150 -.00225 +.00225 -.00300 +.00300 -.00375 +.00375 -.00450 +.00450 -.00525 +.00525 READ STROBE The read strobe portion of the function varies ·the read strobe in increments of 15 nanoseconds in the 813/814 Disk Files. No busy status or interrupts are generated other than the normal reply or reject signals. 60333300 B 2-13 Function Offset (nanoseconds) 0400 -00 0420 +00 0440 -15 0460 +15 0500 -30 0520 +30 0540 -45 0560 +45 CLEAR FAULT (0600) The clear fault bit, when set, clears all fault status conditions in the disk file except a solid fault condition. The following are examples of margin selection functions. 0407 Select offset +. 00225, and clear read strobe 0527 Select offset +. 00225, and select read strobe +30 nanoseconds 0537 Select offset +. 00525, and select read strobe +30 nanoseconds 0417 Select offset +.00525, and clear read strobe 0400 Clear offset and clear read strobe 0600 Clear fault, clear offset, and clear read strobe INPUT /OUTPUT CODES The following codes define the conditions for the various data handling operations which cause the transfer of data between the data channel andlor the controller and the selected peripheral storage device. With the exception of the checkword verify, the codes all require initiation of an input or output operation in order to activate the subsystem. Refer to Programming Considerations for timing diagrams of buffered 110 operations. READ (0040) The read, mode, in conjunction with an input operation, causes the controller to initiate a read operation from the selected storage unit at the address specified by the contents of the controller address register. The operation continues until halted by one of the conditions listed under interrupt on end of operation. When operating in EOR mode, an EOR signal is transmitted to the data channel on detection of a record mark at the end of a data block if the data channel tries to read past the end of record. 2-14 60333300 B WRITE (0041) The write mode". in conjunction with an output operation" causes the controller to initiate a write operation to the selected storage unit at the address specified by the contents of the controller address register" The operation continues until halted by one of the conditions listed under the interrupt on end of operation. If an output operation terminates before the end of the current data block .. the remainder of the block is automatically filled with zeros. The checkword is always written at the end of the block. When operating in EOR mode .. an EOR mark is recorded at the end of the last block written. SEARCH COMPARE (0042) The search compare mode, in conjunction with an output operation.. causes the controller to initiate a read operation from. the selected storage unit at the address specified by the contents of the controller address register. The output operation transmits data to the controller which is compared by bytes on a bit-by- bit basis with the data read from the storage unit. The compare is performed according to the previous settings of the mode select functions. Upon occurrence of a miscompare (unsuccessful comparison).. the no compare status bit is set. MASKED SEARCH COMPARE (0043) The masked search compare mode .. in conjunction with an output operation .. causes the controller to initiate a read operation from the selected storage unit at the address specified by the contents of the controller address register. The output operation transmits data to the controller which is compared by bytes on a bit- by- bit basis with the data read from the storage unit. If a byte in the output data contains all 1 's (7777 ), however .. no comparison is made on the associated byte read. 8 The compare operation is performed according to the previous settings of the mode 8 elect function. Upon occurrence of a miscompare (unsuccessful comparison)" the no compare status bit is set. 60333300 B 2-15 CHECKWORD VERIFY (0044) The checkword verify code is the only operating mode that does not require a buffer. On receipt of this function code by the controller, the data within the current cylinder (or strip in the strip file) is read and checked for an illegal checkword. Operation commences at the present address position of the read/write heads and continues to the end of that cylinder (or strip in the strip file). NOTE When operating in EOR mode, only one record is verified; if no record mark is present, the rest of the file/ cell is verified. READ CHECKWORD (0045) The read checkword operation, in conjunction with an input operation, causes the controller to initiate a read operation from the selected storage unit at the address specified by the contents of the controller address register. This operation is identical to the 0040 read function except that all checkword(s) read are returned to the data channel along with the data. There is one checkword per data block in the disk file and disk drive units. There are six checkwords per data block in the strip file (refer to Checkword Location). The operation continues until halted by one of the conditions listed under Select Interrupt o~ End of Operation. The following codes select and specify the conditions under which the 0042 and 0043 search compare functions operate, and determine and identify the conditions under which an end of operation occurs. MAGNITUDE SEARCH (RECORD < BUFFER) (0050) The magnitude search code modifies the search compare operation so the comparison is satisfied if the data read (searched) is equal to or less than the output data. The search is unsuccessful (no compare status bit sets) when the first (high order) unmasked output data bit is a 0 and the record bit is a 1. 2-16 60333300 B MAGNITUDE SEARCH (RECORD> BUFFER) (0051) The magnitude search code modifies the search compare operation so the comparison is satisfied if the data read (searched) is equal to or greater than the output data. The search is unsuccessful (no compare status bit sets) when the first (high order) unmasked output data bit is a 1 and the record bit is a O. EQUALITY SEARCH (RECORD = BUFFER) (0052) The equality search code modifies the search compare operation so the comparison is satisfied only if the data read (searched) is equal to the output data. Any master clear operation automatically selects this mode of operation. The search is unsuccessful (no compare status bit sets) when the first unmasked output data bit is different from the record bit read. BUFFER MODE (0053) The buffer mode code prepares the controller for a subsequent I/O operation wherein the end of operation is defined as the word count of the I/O operation. (Refer to Buffer Restrictions under Programming Considerations. ) END .OF RECORD MODE (0054) The end of record mode code prepares the controller for a subsequent I/O operation wherein the end of operation is defined as the limit of the record as follows: Write - a record mark is recorded at the end of the block in which the output operation ended. The original output operation may be from less than a single block to an entire file (cell) in length. Read - an end of record signal is sent to the data channel each time the record mark is detected if the data channel tries to read past the end of record. Any master clear automatically selects this mode of operation. 60333300 B 2-17 STATUS CODES The computer determines the state of the controller and storage units by the use of a 12-bit status response available to the data channel. The computer initiates a copy status instruction and samples the status response on the lines from the controller. The computer may sample a status response anytime it is connected, or after a connect attempt is rejected if the controller and/or peripheral units are under control or reservation by a different data channel. The status response may be a combination of any of the available response bits. The status response bits (Table 2-1) indicate the state of the controller and storage unit to which the data channel is connected or last attempted to connect. A 1 in the bit position indicates the condition is present (or has occurred); a 0 indicates the condition is not present (or has not occurred). An interrupt must previously have been selected or the associated interrupt status bit will be a 0 even though a condition that would normally set the interrupt has occurred (that is, a copy status does not indicate that an abnormal end of operation has occurred unless the abnormal end of operation interrupt is selected). If the abnormal end of operation interrupt is selected, the operation ends, the interrupt error status bits are set, and the interrupt is sent to the data channel immediately upon occurrence of the error condition. However, if the abnormal end of operation interrupt is not selected, the error status bits are set immediately upon occurrence of the error condition even though the operation may not end until the buffer is completed. READY (XX xl) - BIT 0 The presence of bit 0 indicates that the unit last connected is in an operable condition and ready for use. to operate. A storage unit is considered ready when it is available and ready The unit becomes not ready for the following conditions. DISK FILE AND DISK DRIVE UNI TS 1. Disk pack not loaded (applicable to disk drive unit only) 2. R/W heads not landed 3. Disk motor not up to speed 4. File or disk drive unsafe condition from selected unit (refer to Unsafe 5. No such unit 6. Wrong unit type designation Conditions) 2-18 60333300 B STRIP FILE 1. ENABLE/DISABLE (running time meter) switch in the disable position 2. Unit is in customer engineering mode (off-line) 3. Manual intervention 4. File unsafe condition (refer to Unsafe Conditions) 5. No such unit 6. Wrong unit type designation No such unit and wrong unit type designation conditions cause a reject or connect on all unit types. Any lX, 4X, or 5X function code is rejected when the controller is not ready; other function codes are accepted even though the controller is not ready. If a unit becomes not ready during an operation, the operation ceases immediately and (if selected) the abnormal end of operation interrupt is sent to the data channel. BUSY (XXX2) - BIT 1 The presence of bit 1 indicates that the controller and/ or peripheral unit specified by the connect code are currently performing an operation and are unable to initiate any new action at this time. The bit becomes a 0 at the end of operation. The busy status normally follows the channel busy signal. However, the busy status remaips until the checkword has been written or read, the address register updated, and in the case of the strip file, until the adapter busy signal drops. (This is the end of operation on a read or write function and if selected, the end of operation interrupt occurs at this time.) Any abnormal condition which causes an end of operation to occur causes the busy status to drop. In the case of a checkword search even though no buffer is initiated, the subsystem is busy until the search is finished. A storage unit is busy following a seek initiation until the seek is completed; however, the controller is available for operation to a different unit as soon as the positioner ready signal drops. 60333300 B 2-19 ABNORMAL/UNAVAILABLE (XXX4) - BIT 2 When the system is connected and reserved, the presence of bit 2 indicates that an abnormal condition exists in the controller or storage unit and assigns a different meaning to the status bits XXIX through XIXX and 4XXX. If a connect attempt is rejected, the presence of bit 2 indicates that the storage unit requested by the connect code was unavailable. When a unit is unavailable it is considered to be permanently unavailable (manual intervention is necessary to remove the cause of unavailability). ON SECTOR (XXlO) - BITS 3,2 DISK FILE AND DISK STORAGE DRIVE UNITS The on sector status bit comes up one sector prior to the addressed sector with or after the positioner ready status. This allows the computer a one-sector time slot in which to initiate an operation on the addressed sector. If no operation is initiated within the one-sector time slot (1 millesecond), the on sector status bit drops and comes up again one revolution later. STRIP FILE t When operating with the strip file unit there is only one sector per track; therefore, the on sector status bit comes up at the index (8 milliseconds before the sector starts) and stays up until the unit is operated on. Refer to Programming Considerations for timing diagrams showing the relationship between the on sector and positioner ready signals. ADDRESS ERROR (XX14) - BITS 3,2 ; The presence of the address error code indicates that an address error has been detected due to one of the following conditions. t Applies to 3234-A/ B/ C models only 2-20 60333300 B 1. The address register contains an illegal address. 2. The content of the address register does not compare with the address being read from the storage unit. 3. The controller has been requested to operate beyond the storage unit address limits. 4. The seek operation initiated by the load address was incomplete. 5. A defective track bit is present. r 6. A checkword error occurs during address verification. 7. A seek error occurs from the disk unit. If an error occurs during an operation, the operation ends immediately, and if selected, an abnormal end of operation interrupt is transmitted. A new load address function should be performed before any new I/O is attempted. For the 3234-D/E modelo, additional bits further define the address error status as follows. Other bit combinations are possible. (025.5) Address error with checkword error - This results from a read error which causes a header checkword error. (4355) Address error with checkword error, defective sector, and write-lockout detected. This condition is the same as 0255 except that the incorrect sync bit c aus es other bits to be detected. (4205) Defective sector status - The defective sector bit without address error indicates the header was tagged defective by the computer or the maintenance panel. NO COMPARE (XX20) - BITS 4,2 The presence of the no compare status code indicates that a miscomparison was detected during the preceding search compare operation. LOST DATA (XX24) - BITS 4,2 The presence of lost data status code indicates that data has been lost because of the data channel (computer) delay. Specifically, in the case of an output operation, no data was ready when the storage unit was ready to write, or in the case of an inp'ut operation, the data channel had not yet accepted the last byte read when the storage unit had another byte ready. t Applies to 3234-A/B/C models only 60333300 B 2-21 Upon detection of lost data, operation ends immediately if the abnormal EOP interrupt is selected and the associated status bits set. If the abnormal end of operation inter- rupt is not selected, operation continues in a normal manner; however, the lost data status bit sets. END OF RECORD (XX40) BITS 5,2 The presence of the end of record bit indicates that an end of record bit has been detected at the end of the last s ector when operating in end of record mode. If the sub- system is not operating in the end of record mode, this bit is a 0 even though an end of record bit is present at the end of a sector. CHECKWORD ERROR (XX44) - BITS 5,2 The presence of the checkword error status code indicates that an incorrect checkword has been detected during a read. search compare, or checkword verify operation. WRITE LOCKOUT ON READ (X1XO) - BITS 6,2 . On an input (read) operation it is permissible to operate in a write protected area and other than setting this status bit, operation proceeds in a normal manner. WRITE LOCKOUT ON WRITE (XI'X4) - BITS 6,2 For an output (write) operation the write lockout on write is an abnormal condition, . and if the abnormal end of operation interrupt is selected, operation ends immediately upon initiation of the operation. The abnormal end of operation interrupt sets. If the abnormal end of operation 'interrupt is not selected, the output data is accepted by the controller; however, the data accepted is ignored and no data is written on the selected stor age unit. 2-22 60333300 B POSITIONER READY (X2XX) - BIT 7 The presence of bit 7 indicates that a positioner ready signal has been received from the selected storage unit. In the disk file and disk drive units this signal comes up as soon as the positioner is settled on the cylinder. The positioner ready must come up before on sector can come up. The positioner ready status stays up until a new load address operation is initiated. selected" an end of operation interrupt sets when the positioner ready is dropped. If When the positioner ready comes back up" a ready not busy interrupt sets if selected. Refer to Programming Considerations for timing diagrams showing the relationship between the on sector and positioner ready signals. END OF OPERATION INTERRUPT (X4XX) - BIT 8 t The presence of bit 8 indicates the interrupt was caused by an end of operation. ABNORMAL END OF OPERATION INTERRUPT (lXXX) - BIT 9 t The presence of bit 9 indicates the interrupt was caused by an abnormal end of operation. SEEK INTERRUPT (2XXX) - BIT 10 The presence of bit 10 indicates the interrupt was caused by the end of a seek operation (in any storage unit) regardless' of whether the seek was complete or incomplete. RESERVED (4XXX)' - BITS 11,2 The presence of the reserved status code indicates that the last connect attempted to the subsystem was rejected because the subsystem was reserved by the opposite data channel. t For an explanation of end of operation and an abnormal end of operation, refer to interrupt codes 0022 and 0024. 60333300 B 2-23 DEFECTIVE TRACK (4XX4) - BITS 11,2 t The presence of the defective track status code indicates a defective track bit has been detected at the address referenced by the content of the controller address register. Once a track is marked defective any I/O operation attempted on that track hangs up and an address error occurs (refer to abnormal EOP interrupt). (The defective track bit is written manually into the header via the maintenance panel.) PROGRAMMING CONSIDERATIONS The following information concerns problems that may arise in programming mass storage subsystems. Most of the procedures are common to all computers and storage devices; those that are unique to a particular computer or unit are designated. CONNECT I Device type codes other than nect to be rejected. XXIX~ XX2X~ and XX3X are illegal and cause the con- The unavailable and not ready status bits set. SIMULTANEOUS CONNECT If the two data channels attached to the same controller simultaneously attempt to con- ?ect to the controller~ neither channel is given preference. The controller connects to the data channel recognized first. Initiation of the operation need not take place immediately after the connect is made; once the connect is made~ the channel has the controller and subsystem reserved until specifically released by that channel. MODE RESTRICTIONS Operations which are inconsistent with the mode selected cause the computer to hang up (that is~ initiation of an output (write) when read mode is selected). t Applies to 3234-A/B/C models. generated. 2-24 On the 3234-D/E models~ no address error is Individual sectors can be labeled defective by software. 60333300 B MASTER CLEAR, RELEASE.. AND DISCONNECT: A manual or program (channel) master clear~ clear function~ or restore function cause the address register to be cleared to zero. A release and disconnect has no effect on the content of the address register. The master clear, clear function~ and release and disconnect remove all interrupt selections except the interrupt on opposite channel release. A master clear automatically selects end of record and magnitude search (record = buffer) mode. STRIP FILE·t If the strip file is busy when a master clear is initiated, the master clear may cause the file to go inoperative and require manual intervention to remove the inoperative condition. DISK UNIT Any master clear to a unit which is busy with a write operation causes the operation to cease immediately; however. the previously recorded data and checkword in the remainder of the sector are unaltered (and now useless). Therefore. a subsequent read operation causes a checkword error to be generated from the inconsistent data recorded in the sector. Any l!-laster clear to a disk unit busy with a read operation causes operation to cease immediately~ but has no other effect on the present or subsequent operations. RESTORE (STRIP FILE) 1 When the equipment is disconnected (strip file released), the strip is automatically restored if the same strip is not readdressed (unit reconnected) within 16 revolutions of the drum (approximately 800 milliseconds). INTERRUPTS An interrupt signal may be dropped by any master clear~ clear function, release and disconnect function. or by any new function code including reselecting the same interrupt selection. t Applies to 3234-A/ B/ C models only 60333300 B 2-25 Only the interrupt on opposite channel release interrupt selection remains when a release and disconnect is performed. All other interrupt selections are removed when- ever a master clear, clear function, or release and disconnect are performed. NOTE The interrupt on opposite channel release can be cleared by a master clear or clear function only when the opposite channel is not res erved (that is, the channel B release interrupt selection can be cleared only if channel B is not reserved.) Any interrupt selection may be removed by its associated release interrupt code. A data channel which has the subsystem reserved receives all interrupts selected. The channel not in control (reservation) of the subsystem can be interrupted only by a previously selected interrupt on opposite channel release. All other interrupts from the controller to the data channel not in control are inhibited even though the interrupt is selected. ABNORMAL EOP INTERRUPT· The abnormal end of operation interrupt should always be used when performing any I/O operations (4X codes) other than the load and return address functions. If the interrupt is not used and either an address error or not ready condition occurs, the channel hangs up in a busy condition. With the interrupt selected, an interrupt lockout override is sent back to the data channel to allow the program (via a clear channel instruction) to recover from the hung condition. SEEK/SECTOR VERIFICATION: When any I/O code (4X) operation is initiated an automatic seek is performed to the address specified by the content of the controller address register. (Normally no posi- tioner or strip loading movement takes place since this seek operation would have been performed subsequent to the last load address operation.) A header verification takes place at the sector following the one in which the last operation ended unless the new I/O is initiated in a different unit than the one last referenced. t In the latter case, if the positioner is not in the proper cylinder, both a seek and header verify sequence are automatic ally initiated. tWhen any I/O is in progress (or ends), the controller automatically updates the address registers at the end of each sector except when an address error is detected or the abnormal EOP interrupt is selected and an abnormal condition occurs. 2-26 60333300 B OVERLAP SEEK Overlap seek capability is incorporated within the units and can be performed on any storage unit in the subsystem as follows: 1. Select unit, select seek interrupt, and initiate a seek operation (by performing a load addres s function). 2. Repeat step 1 for other desired storage units. 3. On detection of an interrupt, t search the selected storage units for an on sector status bit to determine which unit is ready for operation. 4. Initiate and perform I/O operation. 5. Continue from step 3. The on sector status is returned from a storage unit in which a seek operation (load address operation) is completed (positioner ready) until a read/write (or master clear) is performed on that unit. Therefore, if several overlapped seeks are initiated, but with no read/write operations, a seek interrupt is generated each time the seek interrupt is selected until all previously addressed units have had a read/write (or master clear) operation performed on them. NOTE The seek interrupt is generated from a line common to all units where an interrupt can be generated by a unit in which a seek was initiated even though the controller is no longer connected to that particular stor age device. SEEK INTERRUPTS The on sector status is returned from a storage unit after a seek operation has been completed. The on sector status line remains active until a read, write, or new seek is performed on that unit, or if a drive release, controller release, or channel clear is issued. If more than one unit is given a seek, the unit which completes its seek first will generate a seek interrupt (this unit can be determined by selecting each unit and checking on sector status). This interrupt can be cleared by issuing any function. However, even though another unit sends back an on sector signal, another interrupt will not occur. Another seek interrupt can occur only if an operation is performed. The end of operation enables the next on sector to set the seek interrupt if it is selected. t An alternate method is to omit the interrupt and immediately (on completion of seek initiations) commence a continuous search of the on sector status bits. 60333300 B 2-27 Any function sent to the controller clears an interrupt, and if the controller is in the condition of expecting a seek interrupt (that is, seek interrupt selected and one or more units with seeks in progress), a function issued during this period could occur at the same time as an on sector signal from a unit. If this occurred, the seek interrupt could be cleared before the computer had time to respond, and since the interrupt cannot occur again until an operation is performed, the result is a lost interrupt. It should also be noted that if more than one unit has a seek in progress and the computer receives a seek interrupt from one unit and performs an operation on it, the possibility exists of a lost interrupt at the time the end of operation occurs. A pos- sible solution is to deselect the seek interrupt before the operation is performed, and reselect after the operation is complete. On a seek operation the busy status remains up until the seek is completed, but the controller accepts a new connect or function as soon as the controller has sent the seek pulse to the unit. If the end of operation interrupt is set, the interrupt is genera- ted when the controller is ready to accept a new connect or function. lX AND 4X FUNCTION CODES When preparing the subsystem for operation via the issuance of the various function codes, the IX (address) or 4X (I/O codes) must be issued last or they will be cleared out by the next function code issued. With the exception of the checkword verify code, 0044, the IX and 4X function codes are designed to be followed by a buffered operation. RECORDS Records consist of full sectors only. If an output from the data channel consists of less than a full sector when operating in end of record mode, the remainder of the sector is automatically filled with zeros and the record mark is written at the end of the sector. It is possible to count and locate record marks in the various storage units by performing a checkword verify in end of record mode, and then performing a return address. The sector address returned is one greater than the address in which the record mark was detected. The difference between the starting and ending addres- ses is one greater than the number of sectors in the record. 2-28 60333300 B SYSTEM ERRORS AND PERFORMANCE The controller recognizes the following error conditions. 1. Transmission parity error 2. Lost data 3. Write lockout 4. Checkword error 5. Address error 6. Defective sector With the exception of the transmission parity error, the presence of the errors may be detected through the use of interrupts and/ or a status response. Other internal error conditions (refer to Unsafe Conditions) from the selected storage unit cause the subsystem to go to the not ready state. (The not ready condition may be detected via the status response.) TRANSMISSION PARITY ERROR The transmis sion (XMSN) parity circuits examine each byte transmitted to the controller from the data channel, generate a new parity bit for that byte, and co:mpare the parity bit generated with the parity bit accompanying the byte. If the bits do not agree, the PARITY ERROR indicator lights. Transmission of the error indication to the data channel is dependent upon the code or data causing the error. XMSN PARITY ERROR ON CONNECT If the error is detected on a connect code, the controller does not connect; if connected, it disconnects (including its status lines). The transmission PARITY ERROR indicator lights, but no reply, reject, or transmission parity error signals are sent to the data channel. XMSN PARITY ERROR ON FUNCTION If an error occurs on the function code, the transmission parity error signal is sent to the data channel and the error indicator lights; however, the function code is ignored by the controller (no reply or reject is sent). 60333300 B 2-29 XMSN PARITY ERROR ON DATA TRANSFER If an error is detected on a data byte received from the data channel, error signal is sent and the error indicator lights. the XMSN parity The controller returns a reply and uses the data in the normal manner. LOST DATA, A lost data error occurs when the data channel does not transmit or receive data at the fixed rate required by the subsystem. This condition can occur when the data channel serving the mass storage subsystem must compete with several other devices for access to the computer storage. Specifically, an error occurs when the computer does not have data ready to be written when the storage unit is ready to write (output operation), or the computer has not yet accepted the data read when the storage unit has another byte ready (input operation). WRITE LOCKOUT To prevent accidental destruction of data, each sector address is equipped with a write lockout bit. sector. A/B/C). Setting the bit to a 1 prevents any computer operation from writing in that The smallest segment of storage area that can be locked out is a track (3234The smallest segment of storage area that can be locked out in the 3234-D/E is a sector. Attempting to write in a sector or track that has the write lockout bit set generates an abnormal condition (write lockout error). The write operation is then performed in a normal manner except that no writing or erasing takes place (the write and erase enable are disabled by the lockout). If selected, the abnormal end of operation interrupt is set and operation ends immediately. Reading from a sector/track that has the write lockout bit set is accomplished in a normal manner (this is not considered an abnormal condition) even though the write lockout status bit sets. No abnormal end of operation interrupt is generated even though it is selected (refer to WRITE LOCKOUT switch under Switches and Indicators). CHECKWORD ERROR During write operations, a cyclic encoder in the controller 'generates a 12-bit check character referred to as a checkword. This checkword is written at the end of each sector (in the disk files and disk drive units); six checkwords are written per sector (track) in the strip file. 2-30 60333300 B During a read/ search operation a new checkword is generated from the data read and compared against the checkword previously written. If the two do not agree, an error has occurred in writing, reading, or transferring of data between the controller and selected storage unit, and a checkword error is generated. On detection of a checkword error, the associated status bit sets, and if selected, an abnormal end of operation interrupt is generated. If the abnormal end of operation interrupt is not selected, the checkword error status bit sets; however, operation continues in a normal manner. NOTE If a checkword error is detected in an address header, an address error is generated and operation ceases immediately (refer to address error). ADDRESS ERROR An address operation consists of two main phases. 1. A seek operation wherein, upon completion of an output operation following a load address function, the R/W heads are positioned to the addressed cylinder. 2. A header verification sequence wherein, upon initiation of an operation, an automatic s·eek operation takes place, and the address header is read and verified for the addressed sector. An address error can occur in either the previously initiated seek operation or during the header verification sequence. Therefore, since initiation of an operation (and header veri- fication) may occur anytime and is semi-unrelated to any preceding seek operation, an address error can occur at two distinct and mutually independent times (upon detection of a seek error or upon detection of a header verification error). SEEK ERROR An address error occurs during a seek operation if either of the following conditions is detected. 1. The controller address register contains an illegal address. (The controller is being requested to operate beyond the address limitations of the selected storage unit.) (Applicable to all models, indicated by status 0215) 60333300 B 2-31 2. The seek operation initiated by the preceding load address function and buffer is incomplete as indicated by a seek error signal for the selected unit. (Applicable to all models, indicated by status 0015) HEADER VERIFICATION ERROR An address error occurs during a header verification sequence if any of the following conditions are detected. 1. The content of the controller address register does not compare with the address being read from the selected storage unit (on all 3234 models, indicated by status 0215). 2. A checkword error is detected in the address being read (on .the 3234-A/B/C models, indicated by status 0215; on the 3234-D/E models, indicated by status 0255). 3. An address in which the defective sector bit is being read (on the 3234-A/ B/ C models, indicated by status 4215; on the 3234-D/E models, indicated by status 4205). 4. An address in which the incorrect sync bit has been detected, causing the checkword error, defective sector, and write lockout bits to be detected (on the 3234-A/B/C models, indicated by status 4315; on the 3234-D/E models, indicated by status 4355). On detection of an address error the associated status bit sets, and if selected, the abnormal end of operation interrupt is generated on initiation of an operation. (If the abnormal end of operation interrupt is not selected, the channel hangs up on initiation of the operation.) DEFECTIVE SECTOR In order to prevent reading or writing in a sector which contains a defective (faulty) recording medium, a defective sector bit is normally written in each faulty sector. The sector is defective when this bit is a 2-32 o. 60333300 B On detection of a defective sector, the associated status and address error bits set (except for the 3234-D/E models in which no address error occurs). and if selected, the abnormal end of operation interrupt is generated upon initiation of an operation (refer to DEFECTIVE SECTOR switch under Switches and Indicators). In the 3234-D/E models, individual sectors may be labeled defective by software. ADDRESSING After transmitting the connect code, the computer may transmit a load address function code. If the operation is to start or continue at the address location presently held in the controller address register .. the select and I/O operating mode codes are set up on completion of the connect. The address currently held in the register is the one in which the next I/O operation takes place. This address is automatically incremented after each sector is written or read from the selected storage unit with the following exc eptions. 1. The equipment becomes not ready during an oper ation. 2. The abnormal end of operation interrupt is selected and any abnormal condition occurs. 3. An address error occurs. This process takes place throughout the entire file .. pack .. or cell from the starting address up to and including the last available address. If the I/O operation attempts to cause incrementation to continue beyond the last available address .. an address error occurs .. and if selected .. the abnormal end of operation interrupt is generated. Once the last available address has been utilized.. the address register must be reset. This may be done' by loading a specific address (via the load address function) or by executing a master clear or a clear function instruction. In the latter two cases .. the register is set to zero. However.. if a new address is necessary for the next operation.. the load address code must be sent to the controller. LOAD ADDRESS On initiation of an output operation following receipt of the load address function code, the controller commences loading 12-bit bytes from the data channel. These bytes form a 24-bit address word which is automatically loaded into the address register by the controller. The first byte of the address is loaded into the upper· portion of the regis- ter; the second byte is loaded into the lower portion of the register. 60333300 B The output from 2-33 the computer may consist of several words. The controller continues to load the bytes into the address register as previously described until the end of the operation. Thus, the last two bytes transferred comprise the address remaining in the address register. NOTE The 3000 series computers disassemble words upper byte first. ADDRESS FORMAT The address format structure varies within the various peripheral types (refer to Figures 2-4, 2-5, and 2-6). The byte address specifies (in the case of the disk file and disk drive units) the cylinder, track, and sector within the cylinder at which the next operation takes place. From a programming viewpoint the lower address portion may be considered to specify the addresses of all sectors in a single cylinder/ strip. In the disk file and disk drive, the lower order bits of the lower address specify one of the sectors in a track; the upper order bits of the lower address specify one of the tracks in the cylinder. During multiple-recorded transfers (operations of more than one sector in length), the address is automatically augmented to select the next sequential sector location. This process takes place throughout the entire file, pack, or cell up to and including (but not beyond) the last address (operation is not end-around within a file, pack, or data cell). DISK FILE UP·PER PORTION LOWER PORTION Ir-------------~A~__________\lr--------------~A~----------~, 23 1918 .~ ° 12 II 7 BITS I 12 ·BITS ~-----"V--------JI\~----------~vr-----------~~ CYLINDER SECTOR Figure 2-4. Disk File Address Format sectors in a cylinder (addresses 0 through 7777 ), 8 10 Bits 0 through 4 specify one of 32 sectors per track (addresses 0 through 37 ); bits 8 10 5 through 11 specify one of 128 tracks per cylinder (addresses 0 through 177 ), 8 10 Bits 0 through 11 specify one of 4096 2-34 60333300 B Bits 12 through 18 specify one of 128 177 ), 8 cylinders per access (unit) (addresses 0 through 10 Addresses above 1778 are illegal and cause an address error to be generated. DISK STORAGE DRIVE I~ LOWER PORTION UPPER PORTION __________~A~________~\/~__________~A,~__- - - -__---,\ 23 20 19 12 II ~ 8 7 0 ~ SBITS 8· ·BITS ------~v-------J V SECTOR CYLINDER Figure 2-5. Disk Storage Drive Address Format sectors per cylinder (addresses 0 through 237 ), 10 8 are illegal and cause an address error to be generated. Bits 0 Bits 0 through 7 specify one of 160 Addresses above 237 8 through 3 specify one of 16 through 7 specify one of 10 sectors per track (addresses 0 through 17 ), Bits 4 8 tracks per cylinder (addresses 0 through 11 ). 10 8 10 Bits 12 through 19 specify one of 100 cylinders (addresses 0 through 143 ) in the 10 8 853 Disk Storage Drive units or one of 203 cylinders (addresses 0 through 312 ) in 10 8 the 854 Disk Storage Drive units. Addresses above 143 in the 853 or above 3128 in 8 the 854 are illegal and cause an address error to be generated. STRIP FILEt I~ UPPER PORTION __________ ________ -JA~ 20 19 23 1 4 - BITS L-- ~ LOWER PORTION __________ __________ 12 II I --1\~ ~'\/~ I-BITS V STRI P Figure 2-6. 8 7 ~ ______ r--____-J' CELL JA~ ~\ 0 8-BITS \....._ _ _---,. _ _ _ _--' - v- TRACK Strip File Address Format Bits 0 through 7 specify one of 100 143 ), 8 Addresses above 143 8 tracks (sectors) per strip (addresses 0 through 10 are illegal and cause an address error to be generated. Bits 12 through 19 specify one of 200 strips per cell (addresses 0 through 307 ), 10 8 Addresses above 3078 are illegal and cause an address error to be generated. t Applies to 3234-A/B/C models only 60333300 B 2-35 Bits 20 through 23 specify one of 10 cells per array (unit), (addresses 0 through 11 8 ). Addresses above 118 are illegal and cause an address error to be generated. The strip file addresses for track/ cylinder (lower portion) and subcell/ strip (upper portion) are continuous binary addresses. Table 2-2 indicates decimal to octal conversion of the addresses used by the strip file. TABLE 2-2. STRIP FILE ADDRESS CONVERSION CHARTS TRACK (head selections) Cylinder (head position) LOWER ADDRESSES 6 7 8 9 10 o 1 2 3 4 5 0 0 1 2 3 4 5 6 7 10 11 12 1 24 25 26 27 53 30 31 32 33 34 35 36 11 12 13 14 15 16 17 18 13 37 14 15 16 17 20 21 22 23 40 41 42 43 44 45 46 47 72 73 54 55 56 57 60 61 62 63 64 65 66 67 70 71 77 100 101 102 103 104 105 106 107 110 111 112 113 114 115 4 120 121 122 123 124 125 126 127 130 131 132 133 134 135 136 137 140 141 2 50 51 52 3 74 75 76 SUBCELL Strip 2-36 o 1 2 3 4 5 0 0 12 24 36 50 1 1 13 25 37 2 2 14 26 3 3 15 4 4 5 UPPER ADDRESSES 6 7 8 9 10 11 12 13 14 15 19 116 117 142 143 16 17 18 19 62 74 106 120 132 144 156 170 202 214 226 240 252 264 276 51 63 75 107 121 133 145 157 171 203 215 227 241 253 265 277 40 52 64 76 110 122 134 146 160 172 204 216 230 242 254 266 300 27 41 53 65 77 111 123 135 147 161 173 205 217 231 243 255 267 301 16 30 42 54 66 100 112 124 136 150 162 174 206 220 232 244 256 270 302 5 17 31 43 55 67 101 113 125 137 151 163 175 207 221 233 245 257 271 303 6 6 20 32 44 56 70 102 114 126 140 152 '164 176 210 222 234 246 260 272 304 7 7 21 33 45 57 71 103 115 127 141 153 165 177 211 223 235 247 261 273 305 8 10 22 34 46 60 72 104 116 130 142 154 166 200 212 224 236 250 262 274 306 9 11 23 35 47 61 73 105 117 131 143 155 167 201 213 225 237 251 263 275 307 60333300 B STRIP FILE t The following considerations apply only to the strip file. CHECKWORD LOCATION Six checkwords are written in each sector (track) in the strip file. These checkwords are located in the following byte-positions within each track. 106 874 362 1130 618 1386 (last byte of the track) RESTORE, The unit select must be dropped (either by releasing. clearing, or disconnecting) or a restore function performed in order to prevent excessive strip wear. If a select is dropped for more than O. 8 microsecond. an automatic restore is initiated in the file. TEST CELL A special customer engineering test cell is provided to aid in strip file maintenance procedures. It becomes program responsibility to prevent damage (writing) of the information within this cell when the test cell is in the array. CELL REPLACEMENT! Individual cells can be installed in any of the 10 positions of the array. a cell. care must be taken to ascertain that it is in the proper position. When installing The cell ac- cessed by the previous operation is under the access station and not at the replacement station (Figure 2-7). DISK STORAGE DRIVES The following considerations apply only to the disk drive units. Compatibility exists between the 1738 disk drive subsystem and the 3234 mass storage subsystem. However. the compatibility is limited to those features necessary for data recover ability. To accomplish this. all record gaps are of the same size. the address t Applies to 3234-A/B/C models only 60333300 B 2-37 headers are recognizable by both sUbsystems. and the data areas contain the same number of bits. Both subsystems use a 12-bit cyclic code (checkword) for error de- tection. The following differences exist in the two SUbsystems. 1. Data transfer: 3234 12-bit byte 1738 16-bit byte Computer formatting is necessary for conversion of word size between the two systems. 2. The 3234 subsystem utilizes a write lockout bit in the address header (refer to write lockout); this bit is not used or recognized by the 1738 subsystem. 3. The 3234 subsystem has an EOR mode available; the 1 738 subsystem does not use or recognize the EOR bit. UNSAFE CONDITIONS The presence of an unsafe condition indicates the selected storage unit has one or more of the following fault conditions and causes the controller to become not ready. 2-38 1. More than one R/W head selected 2. Both the read and write controls set 3. Erase and write driver on (strip file only) 4. Both write drives on (strip file only) 5. Read or write on and not ready set 6. Read and erase drivers both on 7. Write driver on and erase driver off (disk file and disk drive only) 60333300 B SECTOR TIMING DISK FILE AND DISK DRIVE UNITS ADDRESSED SECTOR ADDRESSED SECTOR ,~ . SECTOR REVOLUTION I I I I fJ I I 1' II 2 3 4 15 16 SECOND FIRST POSITIONER READY (SELECTED UNIT) ----1r-------Iff -------I MUST~OME POSITIONER READY UP ON SECTOR CAN COME UP ~ BEFORE .-Ims ---, ON SECTOR (SELECTED UNIT) ~ 2131411~ r!I 1--- ~ ~ L---jf , I ___ ~----£~f_______ ON SECTOR DROPS WHEN BUFFER IS INITIATED AND DOES NOT OCCUR AGAIN UNLESS THE UNIT IS READDRESSED. INTERRUPT OCCURS AT THIS TIME SEEK INTERRUPT STRIP FILE hr'--------------------~H~------ t j.--- ---I SO ms INDEX AREA 8.2 :t .2 MS REVOLUTION r-- FIRST - +.- ... SECOND POSITIONER READY. I,....- - - - - -.....11' (SELECTED UN[ T) ~ :1.1 i I ON SECTOR (SELECTED UNIT) , n ~I~ SEEK INTERRUPT t Applies - - - - - -If- - - - - - - - - - - - - - - - - ON SECTOR DROPS WHEN A BUFFER IS INITIATED AND DOES NOT OCCUR AGAIN UNLESS THE UNIT IS READDRESSED. __________ ~rff ___________________________ .", to 3234-A/B/C only 60333300 B 2-39 BUFFER TIMING INPUT I~ BYTE A ~I --.n'--____ ---'n. . ____ By_T_E_B_ _ _ _ ---In. . ____ .:::.BY.:...;T:..=E'-C=--_ _ _ ...:::B..:...YT:..:;E;.....:;:..D_ _ _ _ _ , DATA SIGNAL , ~ IA (iJ' ~ ® ~ I-- Jf-I'- - - - - - , ~ 0 B 0' 1 I ~® ~ 200NSEC ~ REPLY ~ ~ ~ 200 NSEC 1 1 --i ~ 300 NSEC ~ , 1 I I 1 LATEST DATA SIGNA~ MUST COME UP EARL Y ENOUGH TO ALLOW THE REPLY TO DROP 300 NSEC BEFORE THE START OF THE 3RO READ CYCLE ~ , 1 ~ @) ~ ~ 1-200 NSEC --------~.I ALL TIMES ARE MEASURED AT THE CONTROLLER I NTERF ACE CD ® DISK FILE DISK DRIVE STRIP FILE 10.2 IJSEC 9.6IJSEC 27.41JSEC LATEST REPLY, 700 NSEC IF BUFFER REGISTER IS EMPTY @ CABLE DELAY AND TURN AROUND TIME @ EARLIEST REPLY, 340 NSEC IF BUFFER REGISTER IS FULL OUTPUT CD I~ 1 ~ •1.- BYTE A ~ ~ --t ® ~ . 1 340 NSEC REPLY ~ I' J @ ®' ~ C t-- 200 IBl NSEC U 1 , 1 MAXIMUM TIME FOR ~'~-----NEW DATA SIGNALIF'----~~ BUFFER REGISTER IS 1 FULL 1 U I t-- 320 NSEC I 101 200 340 ----lOI ~ !'NsEC NSEC~ 200 NSEC NSEC~ 1 B I 320 NSEC 340 I DATA SIGNAL ~I I ® t-- ~ --t 'w.. ~ DATA SIGNAL 1 ~I 340 --t ~ t-- 200 NSEC r I NSEC REPLY - - - - l L--f CD n ....._ _ _ _B_Y_T_E_B_ _ _----'n'-_ _ _-B-Y-TE-'-C------'n BYTE D t-- 320 NSEC ~ .1-1 CD D 0' --t®~ to&- 200 NSEC rcl , , 340 --t NSEC ® 0' ~ ~ 200 NSEC 101 f ALL TIMES ARE MEASURED AT THE CONTROLLER INTERFACE 2-40 CD DISK FILE DISK DRIVE STRIP FILE ® IF BUFFER REGISTER IS EMPTY DATA SIGNAL MUST COME UP 320 NSEC BEFORE ITS ASSOCIATED WRITE CYCLE 10.2 JJSEC 9.6 ",SEC 27.4 JJSEC @ CABLE DELAY AND TURN AROUNDTLME @ EARLIEST REPLY TO A DATA SIGNAL. 340 NSEC IF BUFFER REGISTER IS EMPT.Y. ® LATEST REPLY TO A DATA SIGNAL. 500 NSEC FROM START OF NEXT WRITE CYCLE IF BUFFER REGISTER IS FULL 60333300 B MANUAL OPERATIONS Operation of the mass storage subsystem is under program control from the computer through the use of the connect ~nd function codes. Initial manual starting procedures include turning power onto the equipment, selecting the various switch settings, and loading a disk pack or data cell into the associated unit (if none are presently loaded). Disk packs, disk files. and data cells must have headers recorded before they can be used for programmed operation. POWER SEQUENCE The following procedure brings the subsystem up from a deadstart (no power applied to the storage units or controller). 1. Turn on the unit circuit breaker and then turn on the main power circuit breaker t in the controller. The disk drive units and $trip file automatically enter a power-on sequence. Each disk drive unit requires approximately 30 seconds to come up to speed and land the R/W heads. In a multiple unit system. each succeeding disk drive motor is automatically sequenced. The sequencing overlaps and a full complement of drive units should be up to speed in approximately 1 minute. The last disk drive (or controller if no disk drives are in the subsystem) initiates the power-on sequence in the strip files. A strip file requires approx- imately 10 to 15 seconds to become ready and sequence the next unit; however, the power-on sequencing does not overlap and each file begins only upon completion of the previous file becoming ready. As soon as the disk drive and strip file units become operable (R/W heads landed) and the READY indicator (located in the unit designation indicator in the disk drives) lights, operation for both the disk drives and strip files may begin; however, a 15-minute warmup period is recommended when commencing from a cold start. 2. Normally the disk file and associated hydraulic unit should have all switches in the proper position and only step 7 should be necessary to activate the units. NOTE Pressing the normal STOP switch requires ·a 20-second delay before shutdown. Pressing EMERGENCY OFF shuts down the system instantly. t On 3234-A/C/D models, CB1 is located on the front of the power panel. On 3234-B/ E models, Sl is located on top of the power distribution box. 60333300 B 2-41 CAUTION Except in emergencies the blowers should remain on at all times to prevent dirt accumulation on the disk surfaces. If the units must be started from a completely shutdown position, the following steps should be performed. 1. Open front doors on the disk file and hydraulic unit plus the right side door on the disk file. 2. Place all front panel circuit breakers of both units to the ON position. 3. Set both positioner switches to the NORMAL position. 4. Set all manual/ automatic switches to the AU TOMA TIC position. 5. Set FILE OPERATION switch to the AUTOMATIC position. 6. Set main circuit breaker (located below power distribution panel) to the ON position. 7. Press START and observe START indicator light. light temporarily: The following indicators LOW FILE TEMP, STACK RPM 0, STACK RPM 1, and dc voltages. The disk file requires approximately 3 minutes to come up to operating speed; the hydraulic unit requires approximately 15 minutes (on water-cooled unit). L.OADING AND UNLOADING PROCEDURES The storage medium of the disk file units cannot be changed. Use the following proce- dures to load or change the storage medium in the disk storage drive and strip file units. DISK STORAGE DRIVE. 1. If the unit is operating, turn power off by pressing the STAR·T switch located on the front of the disk drive unit. 2. Wait for the spindle to stop rotating. Lift the disk drive unit cover upward as far as pOHsible to provide maximum loading clearance. 3. 2-42 Load or unload the pack as follows: 60333300 B For loading" place the pack on the spindle and turn the cover handle clockwise to a full stop position. The pack should now be tight on the spindle and the protecti ve cover lift off easily. For unloading, engage the protective cover over the disk pack and rotate the cover three times in a counterclockwise direction. The pack releases from the spindle and can be lifted from the drive unit. 4. Close the disk drive unit cover and press the START switch on the front of the drive unit. This causes the unit to perform an initial seek operation which positions and loads the R/W heads and brings the unit to the ready state. (Operation cannot commence until this step is performed.) STRIP FILE't To facilitate a cell replacement, five operating aids are provided (Figures 2-7 and 2-8). 1. An entry door which permits access to the cell array. An interlock in the door molding prevents any machine motion while the door is open. 2. The cell location indicator which identifies (by number) the cell located in the replacement position. It also identifies the cell presently positioned at the access station. 3. SPINDLE RELEASE lever which allows the operator to manually rotate the array to any position. 4. The cell locking lever which prevents unit operation if any position is lacking a c ell (normal or ballast). 5. Cell elevating door which raises the cell to facilitate cell replacement. To remove or change a cell" use the following procedure. 1. Open the entry door (DRIVE READY indicator should go out). tt 2. Open (pull out) the spindle release lever and rotate the array to place the desired cell at the replacement position. 3. Press the cell locking lever. 4. Lift the cell elevating door (the cell lifts upward about an inch). t Applies to 3234-A/ B/ Conly tt Door cannot be opened if the unit is selected. 60333300 B 2-43 5. Place a cell cover over the top of the cell and rotate the cover locking lever counterclockwise. The cell is now engaged by the cover and may be lifted from the unit and placed in protective plastic container (Figure 2-9). To load the unit, reverse the events previously listed starting with step 4. When loading is completed and the entry door is closed, the indicator light pattern should be: AC-ON DRIVE OP DC-ON DRIVE READY Figure 2-7. 2-44 Cell Indicator and Elevating Door 60333300 B SPINDLE RELEASE _~~~ LEVER CELL ELEVATING DOOR ~~~ CELL LOCKING LEVERS Figure 2-8. 60333300 B Cell Array 2-45 Figure 2-9. Cell, Cell Cover, and Protective Container READY STATE The mass storage subsystem requires approximately 15 minutes to become operative (ready) depending on the type of storage in the subsystem configuration. The controller is always ready when power is available; however, subsystem readiness (indicated by the ready status bit) depends upon the state of the selected storage unit. 2-46 60333300 B DISK FILE The disk file units are ready only when all of the following conditions are present. 1. Power is available to the disk and hydraulic units. 2. The hydraulic and air systems are opening. 3. Hydraulic temperature (100oF) and pressure (1150 pounds per square inch) are up to normal. 4. The disks are spinning and up to speed (1180 revolutions per minute). 5. The positioner is in the landing area and the R/W heads are landed and in an operating position. The disks require approximately 5 minutes to come up to speed; the hydraulic unit requires approximately 15 minutes to come up to the proper operating temperature. DISK STORAGE DRIVE The disk drive units are ready only when all of the following conditions are present. 1. A pack is loaded. 2. The cover is closed. 3. The disks are spinning and up to speed (2400 rpm). 4. An initial seek has been performed to position the R/W head (refer to Loading and Unloading Procedures). The disk drive requires approximately 30 seconds to bring the disks up to speed and perform the initial seek operation. STRIP FILEt Strip file units are ready only when all the following conditions are present. t 1. A full complement of data cells or ballast (dummy) cells is in the array. 2. The cell locking lever is engaged. 3. The array door and cell elevating door are closed. 4. The spindle release knob is disengaged. 5. The drive is operative (power-on and all proper settings, etc.). 6. A strip ready condition exists (the drum is selected and the array stationary). Applies to 3234-A/B/C only 60333300 B 2-47 The strip file requires approximately 10 to 15 seconds to perform a. power-on sequence and become ready; however, a 15-minute warmup period is recommended when commencing from a cold start. In a multiple unit system, each succeeding strip file is automatically sequenced; however, sequencing does not overlap and each file commences only upon completion of the previous file becoming ready. PROCEDURES FOR WRITING ADDRESS HEADERS Address headers are normally written from the maintenance panel. In order to write the header, defective track, and write lockout bits, the following procedure should be followed. NOTE Do not use the REPEAT mode switch when writing headers. Ensure that the REPEA T switch is OFF. 1. Turn the maintenance panel keyswitch ON. 2. Patch the desired PHYSICAL UNITS jumper wire into LOGICAL UNIT O. 3. Set the DEVICE SELECT switch to the desired device type. 4. Select the desired ending mode by pressing the appropriate switch/indicator (TRACK, CYL, etc.). 5. Enter the starting addresses into the upper and lower portions of the address register. (Select the appropriate portion of the address register via the REGISTER SELECT switch and enter the starting address by pressing the appropriate register indicator/pushbutton switches.) 6. If there is no good or usable data in the sectors where new headers are to be written, the HEADER and DATA switches must be set to the WRITE position. If there is useful data in the sectors and only headers are to be written, the HEADER switch must be set to the WRITE position and the DATA switch to the READ position. 7. If write lockout and/ or defective track bits are to be written, these switches should be set to ON. 8. Press the SEEK ADRS. switch. Step 8 initiates the operation which ends when the address of the last sector of the selected storage area has been written or an abnormal condition occurs. Any abnormal condition is displayed in the status register. 2-48 60333300 B NOTE When in read data mode, the data read is not displayed in the data register. On completion of the operation, select READ HEADERS and READ DATA and read (verify) the area just written. Error conditions are displayed in the status register. When complete, return all switches to their normal positions. PROCEDURES FOR WRITING DATA Data can be written from the maintenance panel. and write data. The procedure is to read headers To accomplish this, the following procedures should be followed. 1. Turn the maintenanc e panel keyswi tch ON. 2. Patch the desired PHYSICAL UNITS jumper wire into LOGICAL UNIT O. 3. Set the DEVICE SELECT switch to the desired device type. 4. Select the desired storage mode by pressing the appropriate switch/ indicator (TRACK, CYL, etc.). 5. Enter the data to be written into the data register. Select the appropriate data register via the REGISTER SELECT switch and enter the data by pressing the appropriate register indicator/pushbutton switches. 6. Set the HEADER switch to READ and the DATA switch to WRITE. 7. The WRITE LOCKOUT and/or DEFECTIVE TRACK switches should be OFF. 8: Press the SEEK ADRS. switch. Step 8 initiates the operation which ends when the last sector of the selected storage area has been written or an abnormal condition occurs. Any abnormal condition is displayed in the status register. On completion of the operation, select READ HEADERS and READ DATA and read (verify) the area just written. Error conditions are displayed in the status register. NOTE When in read data mode, the data read is not displayed in the data register. When complete, return all switches to their normal positions. 60333300 B 2-49 SAMPLE PROGRAM ROUTINES To aid in understanding the 3234 Mass Storage subsystem, two sample programs (A and B) are included. Each program is preceded by a short explanation of the program and an associated flow chart (Figures 2-10 and 2-11). Both routines are written in 3200 COMPASS language. PROGRAM SEQUENCE Basic Programming Sequence COMMENT SEQUENCE 1. Connect Connect code selects controller and peripheral unit (if available). Select response If the connect operation is successful, a reply is returned; if not (controller or peripheral unit unavailable), a reject is returned. Return to connect after determining the cause of the reject. 2. Select interrupts Any desired interrupt conditions should be s elected at this point. 3. Load address If a new address location (other than the one presently held in the controller address register) is desired, it should be sent prior to the execution of an operating mode selection. 4. Select operating modes Function code selects the desired modes of operation. 5. Set up R/W Function code selects a read or write function. 6. Initiate operation Receipt of a read or write signal initiates operation within the selected unit. Completion of the operation may be detected through the • end of operation interrupt or dropping of the busy status bit. 2-50 60333300 B PROGRAM A This routine assumes the controller to be on channel 3 as equipment 5 with an 853 Disk Storage Drive as unit 0 and an 813 Disk File as unit 1. The routine utilizes the overlap seek feature by initiating positioner movement in first the disk drive. and then initiating positioner movement in the disk file. The rou~ine then loops. searching for a successful positioner ready status on either unit. As soon as a positioner ready status is located. that unit is ;eaddressed and one sector of data is written. The unit is readdressed again and a compare is performed to d~termine whether the data was correctly written. (If the data is incorrectly written, the write and compare operations are performed again. ) Upon completion of a successful compare, a similar operation is performed on the other unit. After both units have been written on and successful compares made, the program stops. 60333300 B 2-51 OUTPUT ADDRESS ( INITIATE SEEK OUTPUT ADDRESS ( INITIATE SEEK) WRITE (OUTPUT) ONE SECTOR OF DATA ~ITE (OUTPUT) ONE SECTOR OF DATA Figure 2-10. 2-52 Flow Chart. Progr am A 60333300 B AA A B ENA 0 STA FLAG 1 STA FLAG 2 CON 5010B.3 UJP A EXS 1, 3 UJP *+2 UJP *-2 lOB, 3 Load address function OUTW B 3. ADDRESS 1, ADDRESS 1+1 Output disk drive address UJP *-2 CON 5021B, 3 SEL UJP C D DD UJP D SEL lOB, 3 UJP *-1 OUTW 3, ADDRESS 2, ADDRESS 3+ 1 UJP *-2 CON 5010B, 3 UJP *-1 EXS 200B, 0 UJP E CON 5021B, 3 UJP *-1 EXS 200B, 3 Connect disk drive zero on equipment 5" channel 3 Check for ready status Connect disk file one on equipment 5, channel 3 Load address Output disk file addres s Reconnect disk drive Check for the first unit that will have positioner ready Reconnect disk file Check positioner ready UJP F DDD UJP DD E CON 5010B, 3 UJP ':<:-1 SEL lOB, 3 Load address UJP Reset the address register for the disk drive ERR OUTW *-1 3, ADDRESS I, ADDRESS 1+1 UJP *-2 SEL 41B, 3 UJP ':<:-1 OUTW 3. DATA, DATA +64 UJP ':<:-2 PAUS 40B UJP *-1 60333300 B Select write function Write one s ector of data Wait channel inactive 2-53 F ERR 1 ENA 1 STA FLAG 1 SEL lOB, 3 UJP >:<-1 OUTW UJP 3, ADDRESS 1, ADDRESS 1+1 ~~- 2 SEL 52B, 3 UJP ~~-1 SEL 42B, 3 UJP ~~-1 Select equal mode Select compare function OUTW 3, DATA, DATA +64 UJP ~~-2 EXS 2, 3 UJP >:<-1 EXS 20B, 3 Compare error UJP ERR MIst rewrite data Output data for compare Wait not busy LDA FLAG 2 Has disk file been written on? AZJ,NE H (If yes, flag 2=1) CON UJP 5021B. 3 >:<-1 SEL lOB, 3 UJP ':~-1 OUTW 3, ADDRESS 2, ADDRESS 2+1 UJP >:<-2 SEL 41B, 3 UJP >:<-1 OUTW 3, DATA 1, DATA 1+64 UJP >:<-2 Reposition the disk file Write one sector of data ENA 1 STA FLAG 2 PAUS 40B Wait channel inactive UJP *-1 lOB, 3 Reposition the disk file SEL 2-54 Reposition the disk drive UJP *-1 OUTW 3, ADDRESS 2, ADDRESS 2+1 UJP *-2 SEL 52B, 3 UJP *-1 SEL 42B, 3 Select equal mode Select compare function 60333300 B UJP *-1 OUTW 3, DATA 1, DATA 1+64 Output data for compare UJP EXS *-2 2, 3 Wait not busy UJP *-1 EXS 20B, 3 Compare error UJP ERR 1 Must rewrite data LDA FLAG 1 Has disk pack been written on? AZJ,EQ E (If yes, flag 1=1) End H HLT ADDRESS 1 OCT * 01400220 Disk drive address ADDRESS 2 OCT 01200776 Disk file address PROGRAM B This routine assumes the 3234-A/B/C controller (equipped with Special Option 60076) t to be on channel 3 as equipment 1 with a strip file as unit 2. Two tests are performed which calculate the time (in milliseconds) to perform the following: 1. The time required to pick one strip 2. The time for the drum (which rotates the selected strip past the R/W heads) to make one revolution The r.outine loops 10 times during the first test (operation) and then types out the maximum, average, and minimum time required to perform the operation; a similar loopi,ng and typeout is then performed for the second test (operation). The process time and typeout sub- - routines are not actually included in this program sample. t Special option 60076 is the adapter for the 2321 Data Cell (strip file). Applies to 3234-A/B/C models only. 60333300 B 2-55 OUTPUT ADDRESS (INITIATE SEEK) OUTPUT ADDRESS (INITIATE SEEK) PROCESS TIME STORED IN INITIAL AND f"INAL CLOCKS TYPEOUT MAXIMUM, AVERAGE, AND MINIMUM TIMES TO PICK STRIP. (TIME IS IN MILLISECONDS) TYPEOUT MAXIMUM, AVERAGE, AND MINIMUM TIMES FOR DRUM TO MAKE 1 REVOLUTION (TIME IS INMILLISECONDS) Figure 2-11. CON 1002B, 3 UJP SEL *-1 5, 3 UJP *-1 DINT ENI 2-56 Flow Chart, Program B Connect to unit 2, Equipment 1, Channel 3 Function clear to knockdown on sector status Dis able interrupts 0, 1 Counter to repeat 10 times 60333300 B A SEL 1, 3 UJP ~c-l EXS 1, 3 *-1 22B 200 Wait controller not busy STA LDQ STOR Wait 200 milliseconds to make sure restore is complete TMA 22B ~:c_ 2 UJP TMA INA AQJ, LT Restore strip STOR TMA lOB, 3 *-1 3, ADDRESS~ *-2 22B STA CLOCK I Store initial clock EXS UJP lOB, 3 *+2 *-2 22B Wait on sector status SEL UJP OUTW UJP UJP TMA STA SEL UJP Select load address ADDR~SS+l Output address (initiate seek) Pick up clock as soon as output initiated On sector .. store final clock CLOCK F EXS 5, 3 *-1 1, 3 UJP ~''''''' 1 Function clear to knockdown on sector status Wait controller not busy Process time stored in CLOCK I q.nd CLOCK F. lSI 9, 1 UJP A Repeat 10 times Typeout maximum. average, and minimum times to pick one strip. SEL 0,1 lOB, 3 UJP OUTW *-1 3, ADDRESS~ ADDRESS+l Output address (initiate seek) ENI B Select load address UJP 60333300 B 2-57 EXS 1. 3 UJP *-1 SEL 40B. 3 Wait controller not busy Select read function UJP *-1 INPW Read one word UJP 3. BUFFER. BUFFER+1 ':<-2 PAUS lOB Wait channel inactive UJP ':<-1 TMA 22B STA CLOCK I SEL 40B. 3 UJP *-1 INPW 3. BUFFER, BUFFER+1 UJP ':<-2 PAUS lOB UJP ':<-1 TMA 22B STA CLOCK F EXS 1, 3 UJP ':<-1 Channel inactive - store initial clock Select read function Read one word Wait channel inactive Channel inactive - store final clock Wait controller not busy Process time stored in CLOCK I and CLOCK F lSI UJP 9, 1 Repeat 10 times B End HLT Typeout maximum. average. and minimum times for one drum revolution in milliseconds. ADDRESS OCT 10330044 STaR OCT CLOCK I OCT CLOCK F OCT o o o BUFFR BSS 500B 2-58 . Address cell=2, strip=33, track (sector)=44 Read buffer 60333300 B 3 SWITCHES AND INDICATORS MAINTENANCE PANEL The maintenance panel (Figure 3-1) is controlled by means of a locking keyswitch; however, the following items on the panel are active at all times (regardless of whether the panel keyswitch is in the ON or OFF position). 1. The eight LOGICAL UNITS and PHYSICAL UNITS designation jumpers 2. The two equipment number switches (CONTROL A and CONTROL B) and their associated CONNECTED, RESERVED, and PARITY ERROR indicators 3. The REGISTER SELECT switch 4. The register indicator/ switches NOTE The maintenance panel switch/indicators (TRACK, CYL/STRIP, etc.) light up when pressed even when the keyswitch is in the OFF position. The switch function is active only when the keyswitch is in the ON position. WRITE WRITE DATAe HEADElle READ ON FECTI~ SECTOR'U! o0 211 ZIO 0.0 0 0.0 0 0,0 000 2' 27 Z, 2S Z4 22 21 ZO CLR •• 8' LOCI(OUT~ OFF COMICCTEQ IlESERVED MRITV ERROR \ 4 DISI( DISI( SSTA:~T OF~,""CI( FIL~--: UPP£R LOWER o 0 S DAT@ADDRE I FILE STATUS 4 0 Z@' o DEVICE SELECT 0 J 7 CONTROL. A I STIY o RESERVED PARITY ERROR CONNECTED 5 Figure 3 -1. 60333300 B Z, 011 WRITEt.iii\ OFF RE.ISTER SELECT 2' READ 4 0 , 2§' 7 o STaY CONTROL I o ON OFF •• 7. L081CAL MAINTENANCE MMEL UNITS Maintenance Panel 3-1 The purpose and method of use for the v;;l.rious switches and indicators on the maintenance panel is explained in the following paragraphs. While the panel is designed primarily as a maintenance aid, it is necessary to activate the panel (turn the keyswitch ON) in order to write address headers or to enter defective track or write lockout bits in the address headers. NOTE When the keyswi tch is ON and an ending mode switch (TRACK, CYL etc.) is selected, all programmed and/ or manual master clears from the computer are inhibited. This prevents the possibility of the computer clearing the registers just before a seek is initiated from the panel. MAINTENANCE PANEL For normal programmed operation, the MAINTENANCE PANEL (keyswitch) switch must be OFF. When the switch is in the ON position, all switches and indicators on the panel' are active and override any progralTI settings in the controller. The switch is used to activate other switches of the panel. CONTROL A AND CONTROL B CONTROL A and CONTROL B are two sets of switches and indicators which provide individual control for each data channel physically connected to the controller. Each set consists of a rotary switch which permits varying the equipment (controller) designation from 0 through 7 and STBY, and a CONNECTED, RESERVED, and PARITY ERROR indicator which lights when the associated condition exists in the controller. LOGICAL/PHYSICAL UNITS These jumper wires provide for varying the LOGICAL/ PHYSICAL UNITS (unit designation jumpers) designation relationship. The units of the subsystem are permanently wired in at the controller; however, their logical designation may be varied by plugging the appropriate PHYSICAL UNITS jumper into the desired LOGICAL UNITS jack. files are included, they must be physical units 0 through 3. If strip (If less than four strip files are in the subsystem, the higher order numbered units should be removed first.) NOTE All panel oper~tions are through LOGICAL UNIT 0; therefore, when operating from the panel, the PHYSICAL UNIT to be operated on must be plugged into LOGICAL UNIT O. 3-2 60333300 B DEVICE SELECT For normal programmed operation, this switch must be OFF. This rotary switch selects the device type (disk drive, disk file, or strip file) from the maintenance panel. When the keyswitch is ON, this switch instantly overrides any previous (present) programmed selection. REGISTER SELECT This rotary switch provides the capability for displaying and entering information via the register indicator/switches to or from the register selected. For the 3234-B/C/D/E models only, the UNIT position indicates which physical unit is connected. For the 3234-D/ E models only, the FILE STATUS position causes the file status to be displayed. REGISTER INDICATORS/SWITCHES Thirteen indicator/ switches (CLR and bits 0 thr<;>ugh 11) provide for displaying and/ or entering information to and from the register indicated by the setting of the REGISTER SELECT switch. The CLR switch clears only the selected register. DEFECTIVE SECTOR For normal programmed operations, this switch should be OFF. This toggle switch is used to write a defective track bit in each sector of the address headers (refer to Procedures for Writing Address Headers). DATA (WRITE) For normal programmed operations, this switch should be in the READ position. This toggle switch provides for writing data from the panel via the register indicator/ switches. The same data (byte) is written in the entire storage area specified by the mode selection (for example, TRACK, CYL, etc.). DATA (READ) Provides for reading data from the selected storage area. The data read is not displayed in the data register when operating from the panel; however, a checkword error and other errors are sensed. and are displayed jn the status register. 60333300 B 3-3 HEADER For normal programmed operations, this switch should be in the READ position. This toggle switch provides for writing or reading address headers in the storage area selected during a control panel operation by the setting of the ending mode switches (refer to Procedures for Writing Address Headers). WRITE LOCKOUT' The WLO bit recorded in the address header prevents the computer from writing in a sector except when the WRITE LOCKOUT switch is off and the maintenance panel keyswitch is on. This combination allows the computer to override the WLO without re- writing the address header. under program control. This is the only time the keyswitch should be on when All other switches (DEVICE SELECT, TRACK, CYL/STRIP, and FILE/CELL) must be off. Data is not protected during panel operations. Six indicator/pushbutton switches provide for selecting the ending point of an operating mode (storage area), repeating the selected operation, seeking, and master clearing from the panel. off. For normal programmed operations, these switch/indicators should be The following are the various switch/indicators and their functions. TRACK The TRACK switch/indicator selects the track mode in which the entire track designated by the content of the address register is operated upon. NOTE Operation ceases at the end of the selected mode (that is, the track or cylinder operated oli is read/ written only once). (Refer to REPEAT switch.) CYL/STRIP This switch/ indicator selects cylinder (or strip in the strip file) mode wherein the remainder of the cylinder (strip) designated by the content of the address regis ter is operated upon, starting at the address held in the address register. (The STRIP portion applies to the 3234-A/B/C only.) 3-4 60333300 B NOTE While it is possible to operate on an entire file, pack, or cell, the entire array cannot be operated on without individually selecting each of the ten cells. When using this switch, the REPEAT switch should be off since a file (ceU) is not repeated automatically. FILE/CELL This switch/indicator selects file (or cell in the strip file) mode wherein the remainder of the file (cell) designated by the content of the address register is operated on, starting at the address held in the address register. (The CELL portion applies to the 3234-A/B/C only.) REPEAT The REPEAT switch/ indicator is used for troubleshooting in a particular mode. When on, this switch causes the selected mode of operation to be continuously repeated; that is, when in track mode, the same track is continuously read or written. The repeat operation halts at the end of the selected storage area when the switch is off. When in file/ cell mode, this switch should be off. A file/ cell is not repeated. NOTE The previous switch must be used for maintenance purposes only. When this switch is used, the controller does not stop for any abnormal conditions unless it is unable to continue. SEEK ADRS. This switch/indicator initiates a seek operation on the selected unit to the address indicated by the content of the address register. I/O operation commences. On completion of the seek, the selected (On the 3234-0!E models, when the indicator is lighted, it indicates that the MODE switch on the switch panel is not in the NORMAL position.) MASTER CLEAR This switch/indicator cause~ a master clear in the controller and selected disk drive or disk file. Master clear is not used by the strip file. If a master clear is performed while a write operation is in progress, operation ceases immediately. That sector now contains two portions of useless information and generates a checkword error if a read operation is attempted. 60333300 B (A no-op is generated in the strip file.) 3-5 CARD SWITCHES The card switches (Figure 3-2) t are used in conjunction with the maintenance panel. CARD LOCATION 1005 ~ REPEAT SEEK ~ REPEAT SEEK OPERATE @ @ NOT USED NOT USED Figure 3-2. Card Switches REPEAT SEEK This switch is for maintenance purposes only and causes the disk unit to seek alternately between cylinder zero and the selected cylinder under the following conditions. 1. It is on (up position). 2. Track or cylinder mode is selected. 3. The SEEK ADRS. switch is pressed. REPEAT SEEK OPERATE This switch is for maintenance purposes only and causes the disk unit to seek alternately between cylinder zero and the selected cylinder and perform an operation under the following conditions. t 1. It is on (up position). 2. The repeat seek switch is on. Applies to 3234-A/B/C models only 3-6 60333300 B 3. Track or cylinder mode is selected. 4. The SEEK ADRS. switch is pressed. The controller stops for an error condition and the REPEAT switch does not override an error. SWITCH PANEL The switch panel (Figure 3-3) t is an extension of the maintenance panel and is used in conjunction with it. REPETITIVE ~ SEE~OPERATION ~"ORMAL ~~~~~ ... 1 _2--S~T~RQ~'~E..--...,...+....I 14 2 I O""S£T Figure 3-3. ~AR8IN SELECT MODE Switch Panel MODE The MODE switch is used for maintenance purposes. When in any position other than NORMAL" the SEEK ADRS. indicator on the maintenance panel lights. The positions are described as follows: NORMAL: Inactive position REPETITIVE SEEK: Causes the disk unit to seek alternately between cylinder zero and the selected cylinder when track or cylinder is selected and the SEEK ADRS. switch is pressed REPETITIVE OPE:RATION: Causes the disk unit to seek alternately between cylinder zero and the selected cylinder" and to perform an operation when the track or cylinder mode is selected and the SEEK ADRS. switch is pressed MARGIN SELECT When pressed" the MARGIN SELECT switch transmits to the file positioner the values selected on the OFFSET ano STROBE switches. t Applies to 3234- D/ E models only 60333300 B 3-7 OFFSET AND STROBE The OFFSET and STROBE toggle switches are used to simulate the margin selection function offset and strobe values. The switches are set to the desired value and the MARGIN SELECT switch is pushed to pass the information to the file. FAULT CLEAR The FAULT CLEAR toggle switch.. when uP.. clears fault status conditions in the file when the MARGIN SELEC T switch is pressed. 3-8 60333300 B COMMENT SHE.IT Control Data® 3234 Mass Storage Controllers MANUALTITLE ____~~--____~~__----~____~~~------______- -_________ Hardwar~ Reference Manual REVISION _ _ _ B=--_ __ PUBLICATION NO. _6_0_3;;...3. . .3..: .,.3. : .,.0. : .,.0---_ _- - - FROM~ NAME: _ _ _~-........._ _ _ _ _- _________- - - - - - - - - 8USINESS AOORESS: ________~--__------------------------------------- COMMENTS: This form is not "intended to be used ~s ~n order blank. Your evaluation of this manual will be welcomed by Control Datil Corporation. Anyerrorl!. suggested additions or deletions, or general comments may be mt;lde below. Please includ~ page number references and fill in pubUcation revision level as shown by the last entry on the Record of Revi$ion page at the front of the manual. Customer engineers are urged to use the TAft. l&J Z :::J C!) z 9« ....:::> u 0> <0 '- NO POSTAGE STAMP NECESSARY IF MAILED IN U. S. A •. FO..,O ON DOT"FED LINES AND STAPLE STAPLE STAPLE FOLD FOLD -------------------------------------------~ FIRST CLASS PERMIT NO. 8241 MINNEAPOLIS, MINN. BUSINESS REPLY MAIL w Z NO POSTAGE STAMP NECESSARY IF MAILED IN U.S.A. :::i C) z o 4. POSTAGE Will BE PAID BY I- :::> u CONTROL DATA CORPORATION Technical Publications Depto 4201 North Lexington Avenue Arden Hills, Minnesota 55112 FOLD FOLD
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 Producer : Adobe Acrobat 9.0 Paper Capture Plug-in Modify Date : 2018:05:08 19:15:49-07:00 Create Date : 2018:05:08 19:15:49-07:00 Metadata Date : 2018:05:08 19:15:49-07:00 Format : application/pdf Document ID : uuid:7be845b8-df6f-674b-9a2e-e2f10f0aa3f5 Instance ID : uuid:e6a493b2-73c3-2944-b142-0d921d3e92eb Page Layout : SinglePage Page Mode : UseOutlines Page Count : 91EXIF Metadata provided by EXIF.tools