IBM IntelliStation Troubleshooting Guide

This guide documents fault diagnosis for the IBM IntelliStation family — IBM's professional workstation brand. IntelliStation troubleshooting straddles two distinct error-code traditions: the classic IBM POST numeric error codes inherited from the PS/2 / AT era (e.g. 161 / 162 / 163 CMOS battery cluster), and (on POWER IntelliStations) the 3-digit LED operator panel inherited from the IBM RS/6000 family.

Reference Documents

The authoritative per-machine reference is the IBM Hardware Maintenance Manual (HMM) for that machine type. HMMs are still available from IBM Support — search hardware maintenance manual ibm intellistation type xxxx on https://www.ibm.com/support/pages.

Key per-MT IBM Support pages used in this guide:

• M Pro 6849 — POST error codes.
• M Pro 6233 / 6850 — Beep and no-beep symptoms.
• M Pro 6868 / 6878 — HMM.
• Cross-family — Troubleshooting POST numeric error codes.
• Cross-family — POST beep and no-beep errors.

Initial Diagnosis

Power on the system and observe in order:

1. Fans spin up. PSU is delivering at least the +5 V SB rail and the +12 V rail.
2. Single short beep at the end of POST. Successful POST.
3. Video appears on screen showing BIOS splash.
4. Boot proceeds.

If any of these does not happen, stop and diagnose at that stage.

Stage 1 — No Power, No Fans

• Mains lead seated and switch on rear is ON.
• Internal PSU 5VSB green LED on planar lit? If dim or off, PSU 5VSB regulator dead. Replace PSU.
• Front-panel power button cable seated on planar header? Try shorting the PSON pin pair on the planar with a paperclip (or shorting the green-and-black wires on the 24-pin ATX connector with PSU disconnected from the board); if PSU runs standalone, suspect front-panel switch or planar.

Stage 2 — Fans Spin But No POST / No Video

• Wait 30 seconds. Some IntelliStation generations are slow to POST (M Pro 6868 with RDRAM in particular takes 20 + seconds to count memory).
• Listen for beep codes.
• If beeps present, match the pattern to the table below.
• If silent and no video, suspect CPU, planar, or VRM (the canonical "capacitor plague" symptom).
• Try removing all PCI cards, memory except one stick in slot A, and disconnect drives — bare-system POST should at least beep an error.

Stage 3 — POSTs, Beeps Single Short, No Video

• Verify graphics card is seated. Reseat AGP / PCIe x16 card.
• Try a different graphics card (any PCI VGA card works as a smoke test on x86 IntelliStations).
• On POWER models, video output is the IBM GXT card — verify the GXT is enumerated by the firmware (boot to SMS, "I/O Configuration").

Stage 4 — Video Present, Halts at BIOS Splash with Error Code

Look up the POST error code in the table below.

Stage 5 — BIOS Boots But OS Won't Load

• Boot order incorrect? Enter Setup (F1 on most IntelliStations); verify boot device order.
• SCSI / SATA controller not enumerating drives? Check ServeRAID firmware vs drive firmware.
• AIX 5L on POWER won't boot? Boot from AIX install CD and use the maintenance shell.

POST Beep Codes

IntelliStation BIOS is Phoenix-derived on most generations (AwardBIOS on 9229). IBM documents the following beep patterns as video-related faults on most IntelliStation MTs:

IntelliStation beep codes (cross-family)

Beeps	Meaning
1 short	POST passed. Normal boot.
2 short	Configuration error. Numeric error on screen.
1-1-3	CMOS read / write failure.
1-1-4	BIOS ROM checksum bad.
1-2-1	Timer chip failure.
1-2-2	DMA controller failure.
1-2-3	DMA controller page register failure.
1-3-1	RAM refresh test failure.
1-3-3	First 64K RAM chip failure.
1-3-4	First 64K RAM data line failure.
1-4-1	First 64K RAM address line failure.
1-4-2	First 64K RAM parity failure.
2-1-1 to 2-4-4	Bit-position memory test failure within first 64K.
3-1-1 to 3-1-4	Slave DMA / master DMA / interrupt mask failure.
3-2-4	Keyboard controller failure.
3-3-4	Display memory test failure.
1 long, 1 short	Planar fault.
1 long, 2 short	Display adapter fault (missing / failing video).
1 long, 3 short	Display adapter / DAC fault.
2-2-2	Video / display failure.
2 long, 2 short	Video / display failure.
Continuous	Power supply or planar fault.
None, no display	Planar or PSU fault before video init. Frequent symptom of capacitor plague.

Numeric POST Codes

IntelliStation BIOS displays a 3- or 4-digit numeric error code on screen for non-fatal POST faults. Codes inherit from the classic IBM PS/2 / AT POST series.

1xx — Planar / System Board / CMOS

Planar / CMOS POST codes

Code	Meaning	First action
101	Interrupt failure	Planar IRQ controller fault
102	Timer failure
161	CMOS configuration empty (dead RTC battery)	Replace CR2032 on planar
162	CMOS checksum bad / configuration mismatch	Re-enter Setup; if recurs, replace CR2032
163	Time and date not set	Re-enter Setup
164	Memory size mis-match with CMOS	Re-enter Setup after memory change
165	PCI / option ROM mismatch	Update BIOS or reset CMOS
173	Real-time clock failure	RTC chip itself (rare; usually battery)

2xx — Memory

Memory POST codes

Code	Meaning
201	Memory test failure (address shown is the failing block)
202	Memory address line error
203	Memory address line error
215	Wrong memory speed for board
218	Continuity RIMM (C-RIMM) missing in an RDRAM slot
225	Wrong memory speed
229	ECC memory error

The 218 code is specific to the RDRAM-era IntelliStations (6849, 6868, 6850, 6231, 6233, 6229, 6866) — all RIMM sockets must be populated with either a RIMM or a Continuity RIMM (C-RIMM) for the channel to terminate. A missing C-RIMM in slot 4 (when slots 1–3 are populated) will halt POST with code 218.

3xx — Keyboard / Pointing Device

Keyboard POST codes

Code	Meaning
301	Keyboard not responding
305	+5 V fuse on planar blown (keyboard / mouse fuse)
365	PS/2 mouse fault

8xx — Math Coprocessor

• 801 — Math coprocessor test failed. Modern IntelliStations have integrated FPU; an 801 indicates the integrated FPU is non-functional and the CPU is defective.

17xx / 18xx — SCSI / IDE

Storage POST codes

Code	Meaning
1701 / 1702	SCSI drive 0 / 1 failure
1701-xxx	SRN-formatted SCSI fault (xxx specifies the FRU)
1762	SCSI configuration mismatch
1780 / 1781	Drive 0 / 1 boot failure
1810	IDE / SATA drive not detected

24xx / 26xx / 28xx — Video

• 2401 / 2402 — onboard video failure.
• 2410 / 2420 — graphics adapter failure.
• 2601 / 2602 — alternate video adapter failure.
• 2801 — secondary video adapter failure.

ServeRAID-Specific Faults

ServeRAID controllers report errors during POST in their own splash banner:

• 0xc0000005 / adapter not ready — controller firmware corrupt. Reflash from ServeRAID Support CD.
• Configuration mismatch — drives moved between cards. Boot ServeRAID Support CD; "Copy Configuration from Drives".
• Drive offline — single drive failed. Rebuild from hot-spare or replace drive.
• Array critical — RAID-1 / RAID-5 in degraded state. Replace failed drive immediately.
• ServeRAID firmware vs drive firmware mismatch — install ECA037 / ECA036 firmware updates on 9.1 GB / 18.2 GB SCSI drives in 6889 / 6897 / 6893 to fix "may fail to spin up after power-off".

POWER IntelliStation LED Operator Panel

POWER IntelliStations (185, 265, 275, 285) carry the 3- or 4-character LED operator panel inherited from the IBM RS/6000 family. The codes are the same family as documented in the IBM RS/6000 Troubleshooting Guide — refer to that page for the full table.

Key codes specific to IntelliStation POWER:

• 200 — Mode switch in SECURE position (POWER 275 has a keylock).
• 201 — Checkstop during IPL.
• 202 / 203 — NVRAM fault.
• 260 — Diagnostic mode requested.
• 299 — IPL complete, AIX kernel handed control.
• 888 flashing — kernel halt sequence (see RS/6000 guide for 888 102 xxx yyy decoding).

Capacitor Plague Symptoms

The M Pro / Z Pro Pentium 4 / Xeon generation (2001–2004) is in the timeframe of the capacitor plague. Symptoms:

• System won't POST at all — bulged or shorted VRM cap.
• POSTs intermittently, stable cold but reboots when warm — caps approaching end of life under thermal stress.
• Random reboots under load — VRM ripple exceeds CPU tolerance under heavy load.
• System fails to leave standby — 5VSB cap on the planar failed.
• Memory training fails (POST halts at memory count) — capacitor on the memory subsystem failed.
• USB drops out, then PCI / PCIe drops out, then graphics, then crash — staged failure of caps on different rails as the worst-affected ones go first.

Confirmed-affected boards include the Z Pro 6221 (Nichicon HM or Rubycon MBZ around the VRM).

Rubycon MBZ has a documented "no-bloating" failure mode — a failed MBZ may look perfectly fine. If the system is unstable under load and the date codes on the VRM caps are 2001–2005, plan for recap regardless of visual condition.

Recap procedure: IBM IntelliStation Capacitor Replacement Guide.

Memory Faults

• 201 with address in low memory — first DIMM / RIMM. Reseat or replace.
• 218 — missing C-RIMM in RDRAM slot. Populate every slot.
• 225 — wrong memory speed. Verify the planar matches required PC-rating (e.g. M Pro 6868 needs PC800 RDRAM, not PC600).
• Memory degraded entries in BIOS event log — single-bit ECC errors. Run extended memory test; replace the indicated DIMM.

Graphics Faults

• No video, system POSTs to single beep — graphics card not enumerated. Reseat AGP / PCIe card; try a different card.
• Video at POST but blank when OS loads — driver loaded for card that's not present. Boot Safe Mode (Windows) or runlevel 3 (Linux) and remove the wrong driver.
• Quadro FX 1400 / 3400 / 4500 thermal shutdown — clean fan; re-paste GPU thermal compound.
• Wildcat Realizm 800 hangs during boot — BIOS / driver mismatch. Reflash card; update host BIOS to latest.

PSU Faults

• Dead — no fans, no green LED: PSU 5VSB or primary failed. Replace PSU.
• Fans spin briefly, click-retry: Power Good not asserted. PSU fold-back, or shorted planar tantalum, or capacitor plague.
• Boots cold, fails when warm: aged secondary electrolytics on PSU.
• Audible whine, smell of fish: X2 mains suppression cap venting.
• Rails low / high: PSU feedback path issue.
• +12 V rail sags under load: drive motors + CPU VRM exceeding PSU rating; consider PSU upgrade or replace.

Reset CMOS / BIOS Recovery

If the BIOS is corrupted or the CMOS is unrecoverable:

1. Power off, unplug.
2. Locate the CMOS reset jumper on the planar (per HMM; typically labelled JBAT1 or CLR_CMOS).
3. Move jumper to "Clear" position; wait 30 seconds; return to "Normal".
4. Remove and re-fit CR2032.
5. Power on; enter Setup; re-configure.

Some IntelliStations support BIOS recovery from a boot diskette:

1. Power off.
2. Insert recovery diskette / CD with the BIOS image.
3. Hold the recovery jumper (per HMM).
4. Power on. The recovery code in the boot block reads the image from the floppy / CD and reflashes the main BIOS.

Diagnostic Workflow

1. Power on. Watch for fans, beeps, video.
2. If no beep / no video, check PSU rails first, then suspect planar capacitor plague.
3. If beep code present, match to the table above.
4. If numeric POST code, match to the table above; if 161/162/163, replace CR2032.
5. If POST passes but ServeRAID errors during boot, boot ServeRAID Support CD.
6. If POWER IntelliStation halts at a 3-digit LED code, refer to IBM RS/6000 Troubleshooting Guide for the full code table.
7. Run vendor diagnostics: IBM Enhanced Diagnostics CD for x86 IntelliStation; AIX diag for POWER.

Related Pages

• IBM IntelliStation
• IBM IntelliStation Maintenance Guide
• IBM IntelliStation Capacitor Replacement Guide
• IBM IntelliStation BIOS Mod Guide
• IBM RS/6000 Troubleshooting Guide — LED operator panel reference for POWER IntelliStations

References

• IBM Support — Troubleshooting POST numeric error codes.
• IBM Support — Troubleshooting POST beep and no-beep errors.
• IBM Support — POST error codes M Pro 6849.
• IBM Support — Beep and no-beep symptoms M Pro 6233 / 6850.
• IBM Support — M Pro 6868 / 6878 HMM.
• IBM Support — ECA037 SCSI HDD firmware update for M Pro 6889 / 6897.
• IBM Support — ECA036 SCSI HDD firmware update for E Pro 6893.
• IBM Support — ServeRAID Support CD v7.12.14.
• Capacitor plague — Wikipedia.
• Badcaps — Z Pro 6221 recap. Field report of Nichicon HM / Rubycon MBZ failure.
• Badcaps — Rubycon MBZ no-bloating failure mode.