IBM PS/2 Model 80 Troubleshooting Guide
.jpg)
The IBM PS/2 Model 80 (machine type 8580) is a Micro Channel machine. IBM's Model 80 Hardware Maintenance Service pamphlet of September 1990 carries a symptom-to-FRU index for this model, and the 1994 PS/2 Hardware Maintenance Manual covers it in its Micro Channel checkout and index.[1][2] Part numbers, removal procedures and cleaning are on the IBM PS/2 Model 80 Maintenance Guide.
Identifying the system board
[edit | edit source]| Type | Speed and models | Memory connectors | Standard memory kit (1990 HMS) | Math coprocessor (FRU) |
|---|---|---|---|---|
| 1 | 16 MHz: -041, -071 | J15, J16 | 1 MB, 72X6670 | 72X6673 |
| 2 | 20 MHz: -081, -111, -121, -161, -311, -321 | J15, J16 | 2 MB, 90X7391 | 90X7393 |
| 3 | 25 MHz with 64 KB cache: -A16, -A21, -A31 | J8, J9 | 4 MB, 34F0023 | 15F7661 |
[1][3][4] Programs can tell the boards apart from the model and submodel bytes: F8/00, F8/01 and F8/80.[5] All three boards use a Motorola MC146818A clock and the separate 6 V lithium pack in the battery/speaker assembly; there is no Dallas clock module.[3][4][1]
Starting the checkout
[edit | edit source]IBM's Micro Channel checkout starts with the computer and external devices off, cables and power cords checked and no diskettes in the drives. Switch on the external devices and then the computer. A working machine gives one or two short beeps and then readable instructions or the Reference Diskette main menu. If a POST error code appears, press Pause, write it down and press F1 to continue.[2]
- 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 of a power-on password.[6]
- Deal with the first of several error codes first; it can cause the others.[6]
- Advanced Diagnostics are started with Ctrl+A at the Reference Diskette main menu, and need at least 896 KB of working memory to load.[2]
- After a 161 and 163, Peter Wendt notes that the first load of the Reference Diskette can take up to 90 seconds, and that older Model 80 BIOS levels may need F1 pressed to force it.[7]
Beep and display symptoms
[edit | edit source]| Symptom | FRU / action, in order |
|---|---|
| No beep, system otherwise works | Speaker; system board |
| No beep, system does not work | System board; power supply |
| Continuous beep | Power supply; system board |
| Repeating short beeps | Keyboard (stuck key); system board; power supply |
| One long and one short beep | Display adapter; system board; power supply |
| One long and two short beeps | Display adapter; system board; power supply; display |
| One or two short beeps and a blank or unreadable screen | Display adapter; system board; display; power supply |
| One short beep, then the diskette prompt, or a program loads from the fixed disk | Diskette drive; system board; power supply |
| Changing colours | Display |
[1] For the "no beep" and "continuous beep" cases IBM says to go through its undetermined-problem procedure before replacing parts, and for the display symptoms to run the display self-test first: switch off, unplug the display's signal cable, switch the display on with contrast at maximum and brightness at the centre detent. The screen should be white with a black margin; if it is not, the display is at fault, and if it is, the display adapter or system board.[1] On a machine with no display adapter, "display adapter" means the VGA on the system board.
Dead system
[edit | edit source]Ardent Tool's summary of IBM's dead-system procedure: the likeliest cause of no POST, a blank screen and no beep is a device shorting the supply, and the system board must return Power Good within 150 ms or the supply shuts down.[6]
- Check that the mains cord is fully home at both ends; loose cords cause intermittent faults and unexpected resets. Unplug it before working on its contacts.
- Remove all adapters, options and extra memory, and disconnect the keyboard, mouse and display.
- Switch on and listen for a beep.
- If there is no beep, check the speaker. If it is good, change the system board memory and try again; if still silent, measure the supply voltages. Correct voltages mean the system board, wrong voltages the supply.
- If there is a beep, refit the adapters one at a time until the beep stops; the last item fitted is the suspect.
The supply latches all its outputs off after a short or an internal over-voltage and will not restart until the fault has gone and the switch has been off for a while (one second in the 1987 technical reference, ten seconds in the 1990 edition).[5][8] Before opening the supply of a dead machine, check the power switch: Peter Wendt reports that about half the failed Model 80 supplies he saw had a broken switch.[9] IBM's test voltages and the Power Good thresholds are on the maintenance guide; IBM replaces the supply if any voltage is out of limits or the fan is not running.[1]
Numeric error codes
[edit | edit source]IBM's codes are eight characters, a four-digit major code and a minor code; a three-, four- or five-character code on screen is the same code abbreviated (301 is 00030100).[6] The table gives the Model 80 entries from the 1990 HMS; the 1994 HMM's Micro Channel list on pp. 33-59 has the full range.[1][2]
| Code | FRU / action, in order |
|---|---|
| 110 | System board memory; system board (see memory, below) |
| 111 | Adapter memory; memory adapter |
| 112, 113 | Any adapter; system board |
| 161 | Battery |
| 162 | Any device; check the enable and disable settings in Set and View SCSI Device Configuration first |
| 163, 164, 165 | Set Configuration and Features; system board |
| 166 | Any adapter |
| 1xx not listed | System board |
| 201 | System board memory; adapter memory; memory adapter (run Advanced Diagnostics first) |
| 215, 216, 221 | System board memory; system board |
| 301, 302 | Keyboard; keyboard cable |
| 303 | System board; keyboard; keyboard cable |
| 304 | System board; keyboard |
| 305 | Keyboard cable |
| 401 | System board |
| 602 | Defective diskette |
| 6xx | Diskette drive; system board; power supply |
| 7xx | Math coprocessor; system board |
| 1780 | Fixed disk C; ST-506 adapter; system board |
| 1790 | Fixed disk C; ST-506 adapter |
| 17xx not listed | Fixed disk; ST-506 adapter; system board; power supply |
| 2401, 2402 with changing colours | Display |
| 2401, 2402 without changing colours | System board; display |
| 10480 | Fixed disk C; ESDI adapter; system board |
| 10490, 10491 | Fixed disk C or D; ESDI adapter |
| 104xx not listed | Fixed disk; ESDI adapter; system board; power supply |
| 96xx | SCSI adapter with cache; any SCSI device; system board |
| 112xx | SCSI adapter without cache; any SCSI device; system board |
| 208xx | Any device; check for duplicate SCSI IDs |
Configuration and battery
[edit | edit source]For 161 and 163, 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.[10] A weak battery loses all the setup; if only part of it is lost, the battery is probably not the cause.[6] Peter Wendt reports that dust under the supply fan, over the clock chip and the 8042, gives lost configuration and a wrong clock that a new battery does not cure; clean the board before replacing it.[10] The 1994 HMM adds 168 (real-time clock error, run Advanced Diagnostics) and 173 (possibly a weak battery; set configuration and features).[2]
A 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 and run Automatic Configuration.[2] Replacement adapters must go in the slot the old one came from, or the configuration must be set again.[11]
Memory
[edit | edit source]IBM's procedures for the Model 80 memory codes:[1]
- 00011000, system-board parity check: with two kits fitted, remove the one in J15 (or J8). If the error goes, replace that kit; otherwise replace the kit in J16 (or J9), then the system board.
- 00011100, memory-adapter parity check: remove memory adapters one at a time until the error goes, then refit the kits on the last adapter one at a time until it returns; replace that kit, then the adapter.
- 00021500: the kit in J15 (or J8), then the system board.
- 00021600 or 00022100: the kit in J16 (or J9), then the system board.
For an intermittent 110 that Advanced Diagnostics do not find, IBM retain tip H006554 says to swap the two cards and run the memory test again, because the diagnostics load into the first 256 KB and cannot test it.[7] On an -A16, -A21 or -A31 with 8 MB on the board and intermittent OS/2 Trap 0002 or NMI errors, check ECA 051 before replacing memory: if the modules at U24 or U34, near J8 and J9, are marked FCT841, the board qualifies for replacement.[4] Fred Spencer found that 1 MB and 2 MB cards give a 225 error on the 25 MHz board.[12]
Keyboard, coprocessor and diskette
[edit | edit source]- For keyboard codes, IBM's check is to disconnect the keyboard cable at the keyboard, switch on, and measure the voltages at the cable connector; if they are wrong, the cable or the system board is at fault, and if they are right, the keyboard.[1] The keyboard and mouse sockets use the same interface logic; the keyboard socket is the one marked with a keyboard symbol.[11]
- 7xx: IBM warns that the wrong coprocessor for the board's speed can cause intermittent errors. The FRUs are 72X6673 (16 MHz), 90X7393 (20 MHz) and 15F7661 (25 MHz).[11][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 loose head damper pad.[13] 6xx errors can come from failed electrolytics on the drive's board or from dirt on the diskette controller under the supply fan.[13] ECA 069 covers a 162 with a 601 on early PS/2 machines with down-level drives.[7]
Fixed disk
[edit | edit source]- 1780 on an 8580-041: IBM retain tip H064300 says to check that the fixed-disk cables are installed correctly.[10]
- A second 314 MB drive in a -311 that will not come ready: ECA 040, a fault in the drive logic board.[7]
- In the 1994 HMM, 010454XX is a sector-buffer test error, 010455XX and 010456XX controller errors, 010462XX a controller seek error and 010464XX a drive read error, all "run Advanced Diagnostics".[2]
- A drive that will not spin up after long storage may have stiction. 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.[14]
- On SCSI machines, a device missing from the Devices List: IBM's procedure is to disconnect the SCSI devices, terminate the adapter, run Automatic Configuration, and check whether the adapter appears and passes its tests.[1]
Engineering changes
[edit | edit source]The full list is on the Model 80 page. For fault-finding, the ones to know are ECA 031 and ECA 048 (bus-master adapters on early 20 MHz boards, usually a 110 parity error), ECA 051 (FCT841 modules on 25 MHz boards with 8 MB), ECA 040 (second 314 MB drive) and ECA 069 (162 with 601).[7][4]
Undetermined problems
[edit | edit source]IBM's procedure in the Model 80 HMS: switch off; remove or disconnect one item at a time (non-IBM devices, external devices, the math coprocessor, each adapter, the fixed-disk or diskette drives, the memory kits), switching on after each, until the symptom goes. If everything is out and the problem remains, check the supply voltages: correct voltages mean the system board, wrong voltages the supply.[1] For intermittent faults the 1994 HMM adds: look for damaged cables and connectors, reseat adapters, drives and modules, check the fan and the supply voltages, and run the tests repeatedly with an error log.[2]
Related pages
[edit | edit source]- IBM PS/2 Model 80
- IBM PS/2 Model 80 Maintenance Guide
- IBM PS/2 Model 80 Capacitor Replacement Guide
- IBM PS/2 Model 70 Troubleshooting Guide
- IBM PS/2 Model 50 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 IBM, IBM PS/2 Model 80 Hardware Maintenance Service, S15F-2199-01, September 1990. Hosted on this wiki as File:IBM PS2 Model 80 HMS.pdf. Non-SCSI and SCSI device checks (2.0), power-on password (3.0), undetermined problem (4.0), power supply voltages and fan (5.0), enable and disable settings (6.0), display self-test (7.0), keyboard voltage check (9.0), memory (11.0), symptom-to-FRU index (12.0), parts (15.0).
- ↑ 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 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; numeric error codes, pp. 33-59 (160-173 on pp. 36-37; 104xx on pp. 46-47); memory problems, p. 125; Models 60, 65 and 80 test voltages, p. 142; "Undetermined Problems", p. 183; parts list, pp. 251-253.
- ↑ 3.0 3.1 "8580 Type 1 Planar", Ardent Tool of Capitalism. 16 MHz planar: slot map, J15 and J16 memory connectors, J18 battery/speaker connector, U187 MC146818A, U188 8042; Model 70/80 Reference Diskette image rf7080a.
- ↑ 4.0 4.1 4.2 4.3 "8580 Type 3 Planar", Ardent Tool of Capitalism. 25 MHz planar: J8 and J9 memory connectors, U54 82385-25 cache controller, U96 386DX-25, U132 MC146818AFN, fuses F2 (1 A), F3 (1.5 A) and F4 (2 A), and the text of ECA 051.
- ↑ 5.0 5.1 IBM, IBM Personal System/2 Model 80 386 Technical Reference, October 1990. Hosted on this wiki as IBM PS2 Model 80 Technical Reference (October 1990) (PDF). Model and submodel bytes and system board features, section 1; output protection, p. 1-13.
- ↑ 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, and the dead-system procedure including the 150 ms Power Good requirement and the line-cord note.
- ↑ 7.0 7.1 7.2 7.3 7.4 "8580 Common Devices", Ardent Tool of Capitalism. Dust on the diskette controller, retain tip H104756, Peter Wendt's 161/163 notes, memory error procedures, Peter Wendt's summary of ECAs 008 to 069, retain tip H006554 (intermittent 110), the 64 MB configuration limit, retain tip H125078.
- ↑ IBM, IBM Personal System/2 Model 80 Technical Reference, April 1987. Hosted on this wiki as IBM PS2 Model 80 Technical Reference (April 1987) (PDF). Section 5, "Power Supply": output protection, auto restart, power good timing and sense levels.
- ↑ "60, 65 SX, and 80 – Power Supply", Ardent Tool of Capitalism. IBM's test voltages, the supply variants, and Peter Wendt on power switch failures.
- ↑ 10.0 10.1 10.2 "60, 65 SX, and 80 – Common Devices", Ardent Tool of Capitalism. Retain tip H024809 (battery minimum 5.5 V, measured out of the assembly); retain tip H064300 (1780 on the -041); Peter Wendt on dust under the supply fan; Alfred Arnold on the clock stopping before CMOS is lost.
- ↑ 11.0 11.1 11.2 IBM, IBM PS/2 Model 80 Hardware Maintenance Reference, S64F-3995-01, 25 September 1990. Hosted on this wiki as File:IBM PS2 Model 80 HMR.pdf. Keyboard and pointing-device connectors (1.2.8), memory errors (4.1.4), adapters (8.4), math coprocessor (8.6).
- ↑ "8580 Planar Memory", Ardent Tool of Capitalism. Fred Spencer's table of which memory card sizes work on each planar.
- ↑ 13.0 13.1 "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; dirt on the diskette controller; failed surface-mount electrolytics on Mitsubishi drives.
- ↑ "Known Problems – Experience, RETAIN Tips, ECAs", Ardent Tool of Capitalism. Peter Wendt on hard-disk stiction.