VOGONS


Unstable 486 - Need assistance in fixing it.

Topic actions

First post, by JAKra85

User metadata
Rank Member
Rank
Member

This will be a long(er) post.
I have this 486-66MHz PC in my Retro PC Gallery.

The attachment 20260902_155612.jpg is no longer available

It is using an undocumented and rare mainboard:
https://theretroweb.com/motherboards/s/matra-486-vip

It's unstable and I want to fix is, not change it/throw out.

Issues:
1) The DOS crash:
(Soft?)Freeze in DOS. All looks fine, I want to start Norton from DOS, type "NC" ->hit Enter -> Crash.
(Soft?)Freeze = Enter a command in DOS and PC freezes command not executed, DOS cursor still blinking. Responding to NumLock key with huge delay. Press NumLock... 30-60 seconds or more and Numlock led is ON. If I press a bunch of letters the PC Speaker gives an alert not to molest further and freezes for good, no more interaction between us.
Ctrl+Alt+Del not responding.

2) The Windows crash:
BSOD in Win95. All sorts of other errors in Windows95 like registry corruption and explorer application crash with 2 buttons (Ignore or Restart)...

3) The Funny crash:
Memtest+ runs for hours and no issues.
"Pass complete, no errors, press ESC to Exit"
I press ESC, it says:
Testing: "132K - 16MB Halting...".
Little red "+" sign next to Memtest86 is blinking for hours; the little shit crashed. 😁

Things I've done to it are unimaginable. Changed and crosschecked:
- L2 cache (on/off in BIOS)
- L2 cache (removed from mainboard)
- Different Cache & TAG chips ranging from 128K till 512K
- CPU
- Memory modules (2x4 60ns and 2x4 70ns)
- PSU
- Video card
- Running bear minimum: One SIMM (4MB), no CF (HDD), no cache, one VGA.
- HDD controller
- HDD type (LBA vs. CHS)
- Power Management disabled in BIOS.
- TB SW does nothing observable.
- There is a "Green" mark silkscreened near the TB button and TB led. Does nothing.
Update:
- Tantalum caps are not shorted or in an exploded state.
- Removed the "modern" PC Speaker.
- Tried 2 AT keyboards.
- Tried basically all BIOS settings.
- Memory Slots 0-1 and 2-3.
- Bus speed 33 and 25. (Currently set to 25)
- Tried 3 KBC chips, all working. System still crashes the same way.
- Solder joints (individually probed) for main chips.
- Program load/exit loop below -> freeze.
I could only stop the loop with Ctrl+Ald+Del or Reset button. Based on the count size system did not crash. Keyboard crashed.
"
@echo off
:top
command /c exit
echo x >> count.txt
goto top
"

- Append-only loop below -> freeze.
I could only stop the loop with Ctrl+Ald+Del or Reset button. Based on the count size system did not crash. Keyboard crashed.
"
@echo off
:top
echo x >> count.txt
goto top
"

Pure CPU loop below -> freeze.
I could only stop the loop with Ctrl+Ald+Del or Reset button. No count.txt I guess system did not crash. Keyboard crashed.
"
@echo off
:top
goto top
"

Observations:
- Most of the crashes are random and instant and sometimes it happens after a keyboard input during an idle period.
- Windows idles for hours and no crash. Open My Computer->Explorer app crash or a BSOD.
- Would not crash during benchmark running. Running for over 12 hours.
- Would not crash during game running. Games run for hours in a loop and no crash.
- Memtest+ runs for hours and no issues.
- Power Management disabled in BIOS. However! I have never seen such a PM on a 486. If Power Management is enabled it turns off the CPU and video even if benchmark or a game running and there is no keyboard input. Never seen Doom with 1-2FPS on a 486 right before video is cut(going to sleep). After a keystroke all comes back... sometimes fine sometimes with a crash.

===================================
Observation update 1:
I think I narrowed down the issue to something keyboard related.
- Boot with keyboard connected.
- NumLock LED lights up.
- Boots and Memtest+ started.
- Press "C" for configuration, works fine. Also, I can navigate in this menu system.
- Pressing "Space" activates the Memtest+ scroll lock.
Now:
- Pressing NumLock button does not turn off the NumLock LED. It has no effect.
- Pressing "Esc" freezes the PC.
- Restart. Boots and Memtest+ starts again.
- Pressing "Ctrl+Alt+Del" ... well, I loose the keyboard but the Memtest+ continues to run fine, no freeze.
- Re-connect the keyboard, acts as nothing happened. I can press "C" I can navigate in the menu system, I can press "Space", so on.
- Pressing "Ctrl+Alt+Del" disconnects the keyboard, again. 😁 Pressing "Esc" freezes the PC.
===================================
===================================
Observation update 2:
One more thing I noticed during memtest+.
I press NumLock button and LED not turning OFF (stays ON).
However there is a delay here that I can "play with."
Maybe someone experienced similar and can figure out what is happening here.

I press NumLock button and then press "c" for configuration menu -> the menu pops up in an instant.
I press NumLock button quickly 5 times and press "c" for configuration menu -> the menu pops up with a small delay. (~7-10 sec)
I press NumLock button quickly 10 times and press "c" for configuration menu -> the menu pops up with a BIG delay. (20+ sec)
I press NumLock button quickly ~20 times and press "c" for configuration menu -> the menu pops up with a delay almost reaching 1 minute! Huge delay.

When pressing NumLock button in a quick succession there could be a small (NumLock) LED dimming once or twice. Barely noticeable, in the millisecond range. - not sure how relevant this could be.
===================================

Since the mainboard has no documentation it's like jumpering in the dark with eyes closed.

Jumper findings:
-Besides the BUS I could not change/jumper anything CPU wise: multiplier, CPU type, so on.
CPU1: 80486 DX-2 66MHz - works fine at 66MHz.
CPU2: AMD Enhanced Am486 DX-2 66 MHz (Am486DX2-66V16BGC) - works at 100MHz (with a 33MHz BUS) and 75MHz (with a 25MHz BUS), for some reason it's using an 3x multiplier. This one has the 5x85 core, I presume that is the reason for this behaviour.
- JP2 = FSB select (25/33/40 MHz) — silkscreened (the only thing).
- JP15 = Tag pin 3 shorted to cache pin 2 (cache enable?) - this must be closed at all times when cache chips are installed.
- JP16 = Single/dual rank L2 cache. 1-2 dual. 2-3 single.
- JP82 = 512 KB cache size select.
- JP24 = This one is strange, not seen before on any other mainboard. Roughly halves bandwidth. L1 is fine, L2 and memory gets -50%. It was the only setting that enabled BANK0 only L2 cache to work, however with huge penalty to both L2 and memory. 2-3 - Normal; 1-2 - Reduced bandwidth (-50%).

This board has major L2 Cache issues:
When only BANK0 cache is used, BIOS/POST shows the right amount of cache.
CACHECHK or SPEEDSYS shows no L2 Cache. After L1 it's memory speed.
Whatever I tried with BIOS (timings and everything) or jumpers it would not work.
Using BANK0 and BANK1 for a total of 256K cache.
Currently all cache chips are removed, crashing still occurs.

I welcome any ideas on what to check where to look: except the trash bin. (for now)

Regards,
Attila

Last edited by JAKra85 on 2026-09-04, 19:04. Edited 12 times in total.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 1 of 60, by wierd_w

User metadata
Rank Oldbie
Rank
Oldbie

The board has tantalum caps.

are any of them internally shorted?

Reply 2 of 60, by Kekkula

User metadata
Rank Member
Rank
Member

I had this random 486 board that had similar instability issues, after swapping pretty much everything, I eventually switched crystal to something I had laying around, I think it was 40mhz, after that board has been rock steady.

Reply 3 of 60, by rasz_pl

User metadata
Rank l33t
Rank
l33t

>no CF (HDD)

I was going to blame CF, but you already tried 🙁 another thing I would do is resolder all the legs of 5 'big' chips, cache sockets, all the rp resistor ladders

https://github.com/raszpl/sigrok-disk FM/MFM/RLL decoder
https://github.com/raszpl/FIC-486-GAC-2-Cache-Module (AT&T Globalyst)
https://github.com/raszpl/386RC-16 ram board
https://github.com/raszpl/Zenith_ZBIOS Zenith Z-386 MFM-300 ZBIOS disassembly

Reply 4 of 60, by Shponglefan

User metadata
Rank l33t
Rank
l33t
rasz_pl wrote on 2026-09-02, 15:22:

another thing I would do is resolder all the legs of 5 'big' chips, cache sockets, all the rp resistor ladders

I would recommend diagnosing the issue first before randomly resoldering everything. There is always a risk of causing damage when soldering so it's better to only do what is necessary.

Same with recapping. I've seen people damage boards blindly recapping everything when the caps weren't even an issue.

Pentium 4 Multi-OS Build
486 DX4-100 with 6 sound cards
486 DX-33 with 5 sound cards

Reply 5 of 60, by Shponglefan

User metadata
Rank l33t
Rank
l33t

For the OP, do you have any electronics diagnostics equipment (multimeter, etc)?

Intermittent issues can be a pain to diagnose. Best thing is to systematically check everything board testing for shorts, continuity, voltages, etc.

You also test components for heat stress including by heating or cooling them down while in use.

Pentium 4 Multi-OS Build
486 DX4-100 with 6 sound cards
486 DX-33 with 5 sound cards

Reply 7 of 60, by JAKra85

User metadata
Rank Member
Rank
Member
Kekkula wrote on 2026-09-02, 14:42:

I had this random 486 board that had similar instability issues, after swapping pretty much everything, I eventually switched crystal to something I had laying around, I think it was 40mhz, after that board has been rock steady.

There are 2 crystals:
-Near the SIMM.
-Near the IDE connector.

The attachment Screenshot 2026-09-02 181432.png is no longer available

Are these measurable or only a change would tell me if it's defective or not?

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 8 of 60, by JAKra85

User metadata
Rank Member
Rank
Member
Shponglefan wrote on 2026-09-02, 15:59:

For the OP, do you have any electronics diagnostics equipment (multimeter, etc)?

Intermittent issues can be a pain to diagnose. Best thing is to systematically check everything board testing for shorts, continuity, voltages, etc.

You also test components for heat stress including by heating or cooling them down while in use.

For the OP, do you have any electronics diagnostics equipment (multimeter, etc)?
- Yes.

You also test components for heat stress including by heating or cooling them down while in use.
- That would be harder to do... I don't think it could be heat related. Sometimes it crashes 2 minutes after boot stuff pretty much still cold. Then other times it runs for hours on a benchmark I exit and start doing something and freezes.

Last edited by JAKra85 on 2026-09-02, 16:29. Edited 1 time in total.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 9 of 60, by JAKra85

User metadata
Rank Member
Rank
Member

One additional observation. Could be important.
I can crash the system on purpose any time with the 3rd "The Funny crash" method.
It runs stable for hours, once I want to exit and restart it crashes.

Last edited by JAKra85 on 2026-09-04, 09:42. Edited 1 time in total.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 10 of 60, by wierd_w

User metadata
Rank Oldbie
Rank
Oldbie

Re:L2 cache

The cache chip sockets are not populated. 'No L2 cache' is correct. (Unless those SMT chips are soldered-on cache?)

Reply 11 of 60, by JAKra85

User metadata
Rank Member
Rank
Member
wierd_w wrote on 2026-09-02, 16:48:

Re:L2 cache

The cache chip sockets are not populated. 'No L2 cache' is correct. (Unless those SMT chips are soldered-on cache?)

Not sure I understand this one.

L2 memory sockets are unpopulated on purpose.
Disabled in BIOS ans jumpers set to no L2 cache.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 12 of 60, by wierd_w

User metadata
Rank Oldbie
Rank
Oldbie

Ahh, I misunderstood. You REMOVED the L2 cache chips as a diagnostic step. Gotcha.

I was under the mistaken presumption the image was "before troubleshooting"

Reply 13 of 60, by Nexxen

User metadata
Rank l33t
Rank
l33t

Have you checked if all the legs of the various chips soldered to the mobo are actually still firmly so?
It happens a lot to find those "barely making contact but good to the eye". 😀

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K

- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 14 of 60, by rasz_pl

User metadata
Rank l33t
Rank
l33t
JAKra85 wrote on 2026-09-02, 16:17:

There are 2 crystals:
-Near the SIMM.
-Near the IDE connector.

24 is for SuperIO chip
14 is for FSB

https://github.com/raszpl/sigrok-disk FM/MFM/RLL decoder
https://github.com/raszpl/FIC-486-GAC-2-Cache-Module (AT&T Globalyst)
https://github.com/raszpl/386RC-16 ram board
https://github.com/raszpl/Zenith_ZBIOS Zenith Z-386 MFM-300 ZBIOS disassembly

Reply 15 of 60, by JAKra85

User metadata
Rank Member
Rank
Member
wierd_w wrote on 2026-09-02, 14:09:

The board has tantalum caps.

are any of them internally shorted?

Hmmm, did I found something?

Checked all tantalum caps.
All of them are consistent(based on similar sizes).
Except one: TC2

The attachment 20260902_203937.jpg is no longer available

What is the opposite of shorted? - Open?
This TC2 is open.
All tantalum caps have a few Ω of resistance. The big ones 0,5KΩ the smaller ones are 0,2KΩ
TC2 is in the MegaΩ range.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 16 of 60, by JAKra85

User metadata
Rank Member
Rank
Member
Nexxen wrote on 2026-09-02, 17:23:

Have you checked if all the legs of the various chips soldered to the mobo are actually still firmly so?
It happens a lot to find those "barely making contact but good to the eye". 😀

Did a thorough visual check, my method is to set my phone camera to 3x optical zoom and with good light it is way better than just with eyes.
Few dirt and dust between chip legs/pins.
Nothing strange, not cracks or legs that make no contact I can see.
The solder quality looks quite damn good.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 17 of 60, by JAKra85

User metadata
Rank Member
Rank
Member
rasz_pl wrote on 2026-09-02, 18:32:
JAKra85 wrote on 2026-09-02, 16:17:

There are 2 crystals:
-Near the SIMM.
-Near the IDE connector.

24 is for SuperIO chip
14 is for FSB

I don't have an oscilloscope to check it.
It's good to know for later. They are easy to change probably.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0

Reply 18 of 60, by rasz_pl

User metadata
Rank l33t
Rank
l33t
JAKra85 wrote on 2026-09-02, 18:50:

Hmmm, did I found something?

no. Tantalum caps have two states of existence
- fully working
- legs sticking out and some charring on PCB from the explosion

JAKra85 wrote on 2026-09-02, 18:54:

Did a thorough visual check, my method is to set my phone camera to 3x optical zoom and with good light it is way better than just with eyes.

not enough. You have to take a tool (toothpick is safe) and actively press on every leg to see if its moving/wiggling, like this for example https://www.youtube.com/watch?v=syAhMC4H5aY

https://github.com/raszpl/sigrok-disk FM/MFM/RLL decoder
https://github.com/raszpl/FIC-486-GAC-2-Cache-Module (AT&T Globalyst)
https://github.com/raszpl/386RC-16 ram board
https://github.com/raszpl/Zenith_ZBIOS Zenith Z-386 MFM-300 ZBIOS disassembly

Reply 19 of 60, by JAKra85

User metadata
Rank Member
Rank
Member
rasz_pl wrote on 2026-09-02, 19:07:

not enough. You have to take a tool (toothpick is safe) and actively press on every leg to see if its moving/wiggling, like this for example https://www.youtube.com/watch?v=syAhMC4H5aY

Done, that was exhausting. All pins have proper contact.

Can a keyboard crash a system?
I start to think that the crash is somehow connected to the keyboard.
The keyboard I currently use with old systems is an original IBM Model M made in 89 with a PS2->AT converter.
(had no issues with any mainboard [30+ maybe more] so far)

If I remove the keyboard... there are additional pings form the PC speaker.
-One long beep during POST. - Same when keyboard is connected.
-One "minor"/short beep when memory count is done. - When keyboard is connected there is no such thing.
-One "minor"/short beep when booting OS - When keyboard is connected there is no such thing.

I don't have any other AT keyboard at the moment.
With USB (USB->PS2 and PS2 -> AT ) no keyboard BIOS message... of course. 😀
For fun I tried a the Model F (XT) keyboard... PC speaker "ringing" when memory count happens, system crashes after memory count with PC speaker beeping.

One additional observation, keyboard related:
When I want to exit/restart from the memtest. The Numlock is on by default (set in BIOS). When I press ESC the system crashes and at the same time the NumLock led turns OFF.
This NumLock LED OFF is intermittent sometimes it does this with the IBM keyboard. Not with the other one.

Last edited by JAKra85 on 2026-09-07, 10:33. Edited 3 times in total.

Retro PC Gallery:
https://docs.google.com/document/d/1ara32t4rV … nU/edit?tab=t.0