IBM PS/2 Model 70 Troubleshooting Guide
.jpg)
The IBM PS/2 Model 70 (machine type 8570) is a Micro Channel machine, and IBM diagnoses it with the Micro Channel checkout procedure and symptom-to-FRU index in the PS/2 Hardware Maintenance Manual, which cover the Models 50, 55, 56, 57, 60, 65, 70, 76, 77, 80, 85, 90 and 95.[1] Routine service, part numbers and removal procedures are on the IBM PS/2 Model 70 Maintenance Guide.
Identifying the planar
[edit | edit source]The Model 70 was built on three planar designs, and the 25 MHz planar was sold in 386 and 486 versions. The type decides which memory, processor board and Reference Diskette the machine needs.
| Planar | Processor and models | Memory sockets | Clock chip |
|---|---|---|---|
| Type 1 (full size) | 386DX, 16 MHz or 20 MHz | Three, J30-J32 | MC146818A at U150 |
| Type 2 (reduced size) | 386DX, 20 MHz | Three, J7, J9, J10 | MC146818AFN at U81 |
| Type 3 (25 MHz) | 386DX-25 on a processor board with an 82385 cache controller (A16, A21, A61, A81) | Four, J17, J18, J20, J21 | MC146818AFN at U143 |
| Type 4 (25 MHz) | 486DX-25 on a processor board (B16, B21, B61, B81) | Four, J17, J18, J20, J21 | MC146818AFN at U143 |
[1][2][3][4] IBM's parts list gives system board FRUs by speed and serial-number range: 96F7309 or 41G3984 for 16 MHz boards, 96F7308 or 41G3985 for 20 MHz boards, 92F0580 with processor board 15F7659 for the 386 25 MHz machines, and 41G3979 with processor board 64F0123 for the 486 machines. IBM's HMR shows full-size and reduced-size boards for both the 16 MHz and 20 MHz machines.[1][5] The 386 25 MHz machines (A61 and A21) could be upgraded to the 486 by fitting IBM's 486/25 Power Platform, which came with two EPROMs for the system board.[4] The HMM also lists a 486DX33 processor upgrade (FRU 92F0436) and a 486DX33 and 486DLC2 system board upgrade for the Model 70.[1]
The clock chip on every Model 70 planar is a Motorola MC146818A, backed by the separate 6 V lithium pack in the battery/speaker assembly. There is no Dallas clock module to replace.[2][3][4]
Starting the checkout
[edit | edit source]IBM's general checkout for Micro Channel machines begins with the computer and all external devices switched off, all cables and power cords checked, and no diskettes in the drives. Switch on the external devices, then the computer. A healthy machine gives one or two short beeps and then readable instructions or the Reference Diskette's main menu. If a POST error code appears, press Pause and write it down, then press F1 to continue.[1]
- Start the Reference Diskette from power-on with the diskette in drive A. Ardent Tool warns that starting it with Ctrl-Alt-Del can give false errors and a false report that a power-on password is already set.[6]
- If several codes appear, deal with the first one first; it can cause the others, especially where configuration errors are involved.[6]
- Advanced Diagnostics are reached by pressing Ctrl+A at the Reference Diskette's main menu. At least 896 KB of memory must be working for the diagnostic tests to load. If the tests start but do not finish, replace the device being tested.[1]
Beep codes
[edit | edit source]| Symptom | FRU / action, in order |
|---|---|
| One long and one short beep | Display adapter; system board; video memory (if used); bus adapter; power supply |
| One long and two short beeps | Display adapter; video memory (if used); system board; adapter video memory; bus adapter; ACPA/A audio adapter; power supply |
| One long and three short beeps | Display adapter; video memory (if used); system board |
| One or two beeps with a blank or unreadable screen, or a blinking cursor | Display adapter; video memory (if used); system board; adapter video memory; display; bus adapter; ACPA/A audio adapter |
| Continuous beep | System board; power supply |
| Continuous beep and a blank screen | Display adapter; video memory (if used); system board |
| Repeating short beeps (check the keyboard for a stuck key) | System board; ACPA/A audio adapter |
[1] On a Model 70 with no display adapter fitted, "display adapter" means the VGA on the system board.
No beep
[edit | edit source]| Symptom | FRU / action, in order |
|---|---|
| No beep, power-on light off, fan not running | Power supply; control/speaker assembly; system board; any device or adapter; bus adapter |
| No beep, fan runs, computer hangs during POST with no message | System board; any device or adapter; bus adapter |
| No beep, fan runs, computer hangs during POST with a message | System board; processor board; power supply |
| Memory count hangs during POST | System board memory; memory option (if fitted); system board; processor |
| No beep, fan runs, and 000215XX displayed | System board memory; system board |
| No beep, computer otherwise works | Control/speaker assembly; system board |
| No beep, computer does not work | System board; power supply |
[1] Entries that apply only to the Models 90 and 95 (processor board and cache) and to machines with tamper-evident switches are left out of this table.
Dead system
[edit | edit source]Ardent Tool's summary of IBM's dead-system procedure: the most likely cause of no POST, a blank screen and no beep is a device shorting the power supply, such as a memory module fitted wrongly or a defective adapter or device. The system board must return Power Good to the supply within 150 ms, or the supply shuts down.[6]
- Check that the mains cord is fully home at both ends. Loose or badly seated cords cause intermittent faults, including unexpected resets during operation. Unplug the cord before working on its contacts.
- Remove all adapters, options and extra memory.
- Disconnect the keyboard, mouse, display and other external cables.
- Switch on and listen for a beep.
- If there is no beep, check the speaker for continuity. If it is good, change the system board memory and try again. If there is still no beep, measure the supply voltages: correct voltages mean the system board, wrong voltages mean the supply.
- If there is a beep, refit the adapters one at a time and try again after each. When the beep stops, the last item fitted is the suspect. There is no need to reconfigure between steps.
The Model 70 supply latches all its outputs off after a short circuit or an internal over-voltage fault, and it restarts only after the fault is gone and the power switch has been off for at least ten seconds. A machine that clicks off and will not restart immediately may simply be latched.[7] The green power-good light on the front comes on only when +5 V, +12 V and −12 V have passed their under-voltage sense levels of 4.5 V, 10.8 V and −10.2 V.[5] Voltage limits for the supply test and the fuse details are on the maintenance guide.
Numeric error codes
[edit | edit source]IBM writes error codes as eight characters: a four-digit major code (largely the same as the PC/AT codes) and a minor code, with zeros to fill. A code shown as three, four or five characters is the same code abbreviated: 301 is 00030100 and 1701 is 00170100.[6] The full Micro Channel list, with IBM's FRU order for every code, is on pp. 33-59 of the HMM. The codes most often met on a Model 70 are below.[1]
Configuration and battery (160-170)
[edit | edit source]| Code | Meaning | FRU / action, in order |
|---|---|---|
| 161 | Configuration lost; battery caution printed against this code | Battery; system board; bus adapter |
| 162 | Check that all devices are powered on; if the configuration has changed, run Automatic Configuration | Any device; battery |
| 163 | Date and time error | Set the date and time |
| 160, 164, 165, 167 | Configuration error | Set Configuration and Features; system board |
| 166 | Reseat all adapters | Run Advanced Diagnostics |
| 168 | Real-time clock error, not always a hardware fault | Run Advanced Diagnostics |
| 173 | Possibly a weak battery | Set Configuration and Features |
[1] This is the Micro Channel list. The AT-bus list for the Models 25 and 30, which names a clock module, does not apply to the Model 70.
For 161 and 163 on a Model 70, test the 6 V pack out of its holder: IBM retain tip H024809 says to replace it below 5.5 V and names the battery/speaker assembly as the other cause of intermittent 161, 162 and 163 errors.[8] A weak battery loses all the setup; if only part of the setup is lost, the battery is probably not the cause.[6] A clock that stands still while the configuration survives is an early sign: the RTC oscillator stops before the CMOS contents are lost, and 161 and 163 follow.[8] A clock that only drifts is usually software, since the time-of-day clock is interrupt driven; check the time from the Reference Diskette's Set Features menu before blaming the hardware.[1]
165 after adding, removing or moving a card means the stored configuration no longer matches the slots. Copy the card's option diskette to the working Reference Diskette so its ADF is available, then run Automatic Configuration. Replacement adapters must go back in the slot the old one came from, or the configuration must be set again.[5][1]
Memory (2xx)
[edit | edit source]| Code | Meaning | FRU / action, in order |
|---|---|---|
| 201-205 | Reseat system board memory | System board memory; system board |
| 210, 211 | Check memory | System board memory; system board |
| 214, 215, 216, 221, 226, 235, 240, 241 | Memory may be the wrong type or in the wrong slot | System board memory; system board; bus adapter |
| 225 | Memory may be the wrong type or the wrong speed | System board memory; system board; bus adapter |
[1] On the Models 50, 55, 60, 65, 70 and 80, IBM finds a failing kit by running the advanced memory test and then exchanging kits one at a time with a known-good kit; if the kits are good, suspect the riser card and memory expansion adapter (if used) and then the system board. The socket-location digit in the error code applies only to other models (53, 56, 57, 76, 77, 85, 90 and 95).[1]
If POST finds an error in the first 512 KB, it disables the first 1 MB block and moves its addresses to the second block if there is one; errors elsewhere are reported and can be disabled with the customer diagnostics on the Reference Diskette.[5] Ardent Tool reports that the 25 MHz planar accepts only 2 MB 80 ns parity SIMMs, and that one owner got 215 before the memory count when he started filling sockets at J17 or J18 instead of J21 and J20; 225 indicates unsupported presence-detect coding or SIMMs in the wrong place.[4] IBM identifies a 2 MB kit's speed from resistor positions R1 and R2 on its end: R1 alone is 85 ns, R2 alone 80 ns, both 100 ns.[1]
Keyboard, mouse, diskette and display
[edit | edit source]- 301 and 302: keyboard cable, then system board. 303 and 304: system board, keyboard, keyboard cable.[1] The keyboard and mouse sockets use the same interface logic; the keyboard socket is the one marked with a keyboard symbol.[5]
- 305, keyboard voltage error: the fuse, then the keyboard cable. IBM says to replace the system board if there is no fuse.[1] Every Model 70 system board has a 3 A fuse, FRU 72X8561.[1]
- 162 with the diskette drive dead at POST: Peter Wendt's usual causes are a sticking head carriage, a dirty track-0 sensor or a head damper pad that has come loose inside the drive.[9]
- 601: diskette, drive, system board. The 606-660 group is "generally media errors": try a known-good diskette before replacing the drive. 6xx not otherwise listed: drive, system board, cable.[1]
- 7xx: the math coprocessor, then the 486DX33 upgrade if fitted, then the system board.[1]
- 2401 and 2402: if the screen colours change, the display; if the colours are right, the system board and then the display.[1]
Fixed disk (104xx)
[edit | edit source]| Code | Meaning | FRU / action, in order |
|---|---|---|
| 10450, 10451, 10490, 10491, 10492, 10499 | Read/write or drive controller error | Run Advanced Diagnostics; riser card; fixed disk; system board |
| 10452 | Seek test error | Run Advanced Diagnostics |
| 10453 | Wrong drive type? | Information only |
| 10454 | Sector buffer test error | Run Advanced Diagnostics |
| 10455, 10456 | Controller error | Run Advanced Diagnostics |
| 10461, 10462, 10464 | Drive format error, controller seek error, drive read error | Run Advanced Diagnostics |
| 10467, 10468, 10469 | Non-fatal seek error, fatal seek error, soft error count exceeded | Run Advanced Diagnostics |
| 10473 | Corrupt data; a low-level format may be needed | Information only |
| 10480 | Fixed disk (ESDI) | Fixed disk; drive cable; controller; system board |
A drive that will not spin up after long storage may be suffering stiction. Ardent Tool reports it as common on the Seagate ST-177I 60 MB drive in some 70-061 machines, and gives Peter Wendt's method: unplug the system unit, shake it up and down a few times, turn it sharply clockwise a few times, and switch on. Do not drop the front of the unit on the desk. Copy the data off as soon as the drive runs.[10]
Engineering change announcements
[edit | edit source]Three ECAs are recorded for the 8570, all on 25 MHz machines:[4]
- ECA 021: POST error 110 with five question marks after a memory upgrade on an 8570-A21 system board, P/N 15F7657. The board is replaced.
- ECA 029: bus-master adapters such as the IBM SCSI adapter fail on some 25 MHz boards. Check the barcode label on the system board; a prefix from B1HC"A" to B1HC"H" qualifies the board for replacement with P/N 15F7657.
- ECA 092: a possible data-transfer problem on 25 MHz machines with an 80 MB or 160 MB fixed disk (Models A81, A16, B81 and B16) under DOS or OS/2, fixed by an ABIOS patch diskette.
Undetermined problems
[edit | edit source]When the diagnostics find nothing, the Devices List is wrong, or the machine does not work at all, IBM's procedure is: check the supply voltages and replace the supply if they are wrong; otherwise switch off and remove or disconnect, one at a time, any external devices, a surge suppressor, the modem, printer and mouse, each adapter, the drives, the memory kits, the bus adapter and the math coprocessor, switching on and reconfiguring after each. A Micro Channel machine needs 1 MB of memory to run. If the problem remains with everything out, suspect the system board.[1]
For intermittent faults IBM adds: look for damaged cables and connectors, reseat all adapters, drives and modules, check the system fan, check the supply voltages, and run the tests repeatedly with an error log.[1]
Related pages
[edit | edit source]- IBM PS/2 Model 70
- IBM PS/2 Model 70 Maintenance Guide
- IBM PS/2 Model 70 Capacitor Replacement Guide
- IBM PS/2 Model 50 Troubleshooting Guide
- IBM PS/2 Model 80 Troubleshooting Guide
- Capacitor Failure Symptoms
- Recommended Tools
References
[edit | edit source]- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.18 1.19 1.20 1.21 1.22 1.23 IBM, IBM Personal System/2 Hardware Maintenance Manual, S52G-9971-02, October 1994. Hosted on this wiki as IBM Personal System 2 Hardware Maintenance Manual (S52G-9971-02, October 1994) (PDF). General Checkout (Micro Channel), pp. 19-23; symptom-to-FRU index (Micro Channel): no-beep symptoms, pp. 27-28; beep symptoms, p. 29; miscellaneous symptoms, pp. 30-32; numeric error codes, pp. 33-59 (160-170 on p. 36; 2xx on pp. 38-40; 3xx on p. 40; 6xx and 7xx on p. 41; 24xx on p. 43; 104xx on pp. 46-47); memory problems, pp. 124-126; power-supply voltage check, pp. 138 and 141; "Real-Time Clock Problems", p. 150; "Undetermined Problems", p. 183; Model 70 system board locations, pp. 203-205; Model 70 parts list, pp. 255-256.
- ↑ 2.0 2.1 "8570 Type 1 Planar", Ardent Tool of Capitalism. 16 MHz and 20 MHz planar: FRU 96F7309 or 41G3984; J30-J32 SIMM sockets; U15 VGA controller 15F6864; U150 MC146818A; U145 2 K × 8 configuration SRAM; BIOS 15F8302/15F8303 of 11 April 1988; Reference Diskette image rf7080a.
- ↑ 3.0 3.1 "8570 Type 2 Planar", Ardent Tool of Capitalism. 20 MHz reduced-size planar: FRU 96F7308 or 41G3985; F1 keyboard/mouse fuse; J7, J9, J10 SIMM sockets; U81 MC146818AFN; U79 configuration SRAM; BIOS 33F4906 (May 1988) and 64F3513 (December 1989).
- ↑ 4.0 4.1 4.2 4.3 4.4 "8570 Type 3 & 4 Planar", Ardent Tool of Capitalism. 25 MHz planar: J15 processor-card socket; J17, J18, J20, J21 SIMM sockets; U143 MC146818AFN; BIOS 33F9988/33F9987 (386, 20 February 1989) and 64F3126/64F3125 (486, 1 December 1989); memory notes; 386 processor board FRU 15F7659 with 82385-25 cache controller; 486/25 Power Platform FRU 64F0123 and its use as an upgrade for the A61 and A21; ECAs 021, 029 and 092; Reference Diskette image rf70486a.
- ↑ 5.0 5.1 5.2 5.3 5.4 IBM, IBM PS/2 Model 70 Hardware Maintenance Reference, S64F-3993-00, March 1990. Hosted on this wiki as IBM PS2 Model 70 Hardware Maintenance Reference (PDF). Sections 1.1.2 (power-on password), 1.2.2 (CMOS RAM and extension), 1.2.8 (keyboard and pointing-device connectors), 3.1 (power supply and under-voltage sense levels), 3.3.1 (system-board memory) and 3.3.4 (memory errors).
- ↑ 6.0 6.1 6.2 6.3 6.4 6.5 "PS/2 Error Codes", Ardent Tool of Capitalism. Beep codes, the eight-digit error format, the Reference Diskette cold-boot note, the multiple-error rule, the dead-system procedure (including the 150 ms Power Good requirement and the line-cord note) and the battery note.
- ↑ IBM, IBM Personal System/2 Model 70 386 Technical Reference, October 1990. Hosted on this wiki as IBM PS2 Model 70 Technical Reference (October 1990) (PDF). Power supply outputs, p. 1-15; output protection and voltage sequencing, p. 1-16.
- ↑ 8.0 8.1 "60, 65 SX, and 80 – Common Devices", Ardent Tool of Capitalism. IBM retain tip H024809, "Low Battery Voltage May Cause Config Errors" (battery 72X8498, assembly 33F5950, minimum 5.5 V DC measured out of the assembly); Alfred Arnold's note that the RTC oscillator stops before the CMOS contents are lost.
- ↑ "8580 Diskette Drives", Ardent Tool of Capitalism, content by Fred Spencer and Al Savage. Peter Wendt on diskette drives that cause a 162 at POST.
- ↑ "Known Problems – Experience, RETAIN Tips, ECAs", Ardent Tool of Capitalism. Peter Wendt on hard-disk stiction and the Seagate ST-177I in the 8555-061 and some 8570-061 machines.