Beeping means CPU does run, if even a little.
No post codes is worrying. Seeing its a BGA chipset and problems started when swapping hardware - did you press hard on mobo during the swap by any chance?

Hi, I didn't bend anything. I mean, I hope I didn't. I removed the whole mainboard carriage when I swapped the CPU and changed the jumpers.
Buft after that, the screen stayed black. I'ven even tried different PCI graphics cards.

Edit: I've also switched the PSU, no change.
With the Pentium MMX 166 there's same beep code sequence.
I guess so something broke when configuring the board for dual-voltage, but I'm not sure what it is.
The electrolytic capacitors seem to be fine, too.

Edit: Here's what the POST card shows..

Hmm I didnt even know POST cards could just display half the number 😀 I dont think thats right, as in it shouldnt be possible. Have you tried ISA slot? and without the card? maybe just the post card died?

rasz_pl wrote on 2026-06-15, 09:58:

Hmm I didnt even know POST cards could just display half the number 😀 I dont think thats right, as in it shouldnt be possible.

Correct, it shouldn't be. They are simply showing the last 8-bit write to port 80h, after all.

According to beep code tables, high-frequency beeps = temperature alarm. Could it be that the board requires a fan to be connected?

Regarding the too high regulator voltage, did you measure it with a CPU inside the socket? Without a load (CPU) installed, the regulator is essentially off, and what you're measuring is just a small leakage current from the regulator's input voltage side.

asdf53 wrote on 2026-06-15, 16:13:

According to beep code tables, high-frequency beeps = temperature alarm.

afaik not in this era of hardware. Temp monitoring was introduced somewhere around 1998-99.

jmarsh wrote on 2026-06-15, 15:07:

rasz_pl wrote on 2026-06-15, 09:58:

Hmm I didnt even know POST cards could just display half the number 😀 I dont think thats right, as in it shouldnt be possible.

Correct, it shouldn't be. They are simply showing the last 8-bit write to port 80h, after all.

Thanks, guys! I've just checked the ISA/PCI POST card in another PC and it shows same reaction..
To be honest, I've never used this card before. In the past I've used a vintage POST card meant for PC bus.
It can be seen in this old video, btw: https://www.youtube.com/watch?v=xXE51QGmkg0
That's how this GA-586S motherboard used to POST correctly.
Unfortunately, I can't find this vintage diagnostic card right now. 😅

Edit: I've checked the PCB of thr POST card..
It seems the solder joints were bad and there was at least one short.
After fixing the solder joints, the card still doesn't work. Broken address decoders, maybe?

Edit: Good news! I've searched the attic and found my old debug card from the 1980s.
It shows different error codes and the last one that stays is error code 41.

Hi Jo22,

On the Award BIOS, the 01, 05, 0C, 0D, 41 POST code sequence is part of the Bootblock recovery. So if you see POST code 0D before 41, the BIOS signals a BIOS ROM Checksum error!
POST_41 is where the BIOS searches for a bootable floppy, to attempt a BIOS flash from a recovery disk. The high/low chime tells you a bootable media wasn’t found and the recovery is halted.

When you install an ISA videocard, you should be able to see these error messages on screen.
Apart from a corrupted BIOS, there are other possible sources that can cause this error:
1) The BIOS EPROM chip may have corroded legs or isn’t properly seated in its socket.
2) Strangely enough, the most common cause of this BIOS ROM Checksum error is bad RAM. So check your memory sticks and SIMM/DIMM slots for corrosion or bad contacts.

Because these BIOSes are compressed, the BIOS Bootblock contains a decompression engine that decompresses all modules into RAM. It then does integrity checks on these decompressed BIOS sections and jumps to recovery when a checksum error is found. A compressed module inside the EEPROM may have been corrupted.
But when the RAM is bad, the same will happen.

I hope this helps.
Cheers, Jan

Hi Jan! Thank you a lot for the information! ^^
I think I'll try to check the BIOS with an EPROM programmer, I think it should contain the WIMS patched BIOS.
So I can compare it with the image file available online, maybe.
Not sure what to do about the RAM, though. I'm not sure if I have some spare PS/2 SIMMs here.
I do have some ISA VGA cards, however.

About error 41 (floppy error).. Could it be that the multi i/o chip is the culprit, too?
It's an ite it8663bf.
From what I've read, it also measures voltages and handles A20 gate and other stuff.
So if communications with that chip fails, it perhaps would halt at error 41h, too?

Edit: This *might* have been the BIOS I was using.
https://www.wimsbios.com/forum/viewtopic.php? … 0f7472a149095ac

Edit: Oh, here's the "32GB HDD patch" version. It matches the BIOS message during POST.
https://theretroweb.com/motherboards/s/gigabyte-ga-586s

Chkcpu wrote on 2026-06-16, 14:15:

But when the RAM is bad, the same will happen.

Not directly related to this motherboard but I believe when EDO RAM is put in a 486 motherboard that does not support it and has this style of BIOS, the exact same thing happens.

Jo22 wrote on 2026-06-17, 02:51:

Hi Jan! Thank you a lot for the information! ^^ I think I'll try to check the BIOS with an EPROM programmer, I think it should c […]
Show full quote

Hi Jan! Thank you a lot for the information! ^^
I think I'll try to check the BIOS with an EPROM programmer, I think it should contain the WIMS patched BIOS.
So I can compare it with the image file available online, maybe.
Not sure what to do about the RAM, though. I'm not sure if I have some spare PS/2 SIMMs here.
I do have some ISA VGA cards, however.

About error 41 (floppy error).. Could it be that the multi i/o chip is the culprit, too?
It's an ite it8663bf.
From what I've read, it also measures voltages and handles A20 gate and other stuff.
So if communications with that chip fails, it perhaps would halt at error 41h, too?

Edit: This *might* have been the BIOS I was using.
https://www.wimsbios.com/forum/viewtopic.php? … 0f7472a149095ac

Edit: Oh, here's the "32GB HDD patch" version. It matches the BIOS message during POST.
https://theretroweb.com/motherboards/s/gigabyte-ga-586s

Okay, you have an (E)EPROM programmer, nice. Please upload a dump of your present BIOS here, and I will check it for any corruption.

When you have confirmed that the BIOS is going to recovery mode, either via a message on screen through an ISA videocard or seeing POST code 0D before code 41, then a failure of the multi i/o chip is unlikely. Stopping at code 41 here usually means it can’t find a bootable media.

Checking the FDC inside the multi i/o chip is easy. Just hook up a floppy drive and insert a bootable MSDOS disk. When the POST card cycles from code 41 to FF, the FDC is working normally and the board boots to DOS via the limited BIOS run-time code in the Bootblock.
If this works, your screen should look like this:

Jan

Chkcpu wrote on 2026-06-17, 09:23:

Okay, you have an (E)EPROM programmer, nice. Please upload a dump of your present BIOS here, and I will check it for any corrupt […]
Show full quote

Jo22 wrote on 2026-06-17, 02:51:

Hi Jan! Thank you a lot for the information! ^^ I think I'll try to check the BIOS with an EPROM programmer, I think it should c […]
Show full quote

Hi Jan! Thank you a lot for the information! ^^
I think I'll try to check the BIOS with an EPROM programmer, I think it should contain the WIMS patched BIOS.
So I can compare it with the image file available online, maybe.
Not sure what to do about the RAM, though. I'm not sure if I have some spare PS/2 SIMMs here.
I do have some ISA VGA cards, however.

About error 41 (floppy error).. Could it be that the multi i/o chip is the culprit, too?
It's an ite it8663bf.
From what I've read, it also measures voltages and handles A20 gate and other stuff.
So if communications with that chip fails, it perhaps would halt at error 41h, too?

Edit: This *might* have been the BIOS I was using.
https://www.wimsbios.com/forum/viewtopic.php? … 0f7472a149095ac

Edit: Oh, here's the "32GB HDD patch" version. It matches the BIOS message during POST.
https://theretroweb.com/motherboards/s/gigabyte-ga-586s

Okay, you have an (E)EPROM programmer, nice. Please upload a dump of your present BIOS here, and I will check it for any corruption.

When you have confirmed that the BIOS is going to recovery mode, either via a message on screen through an ISA videocard or seeing POST code 0D before code 41, then a failure of the multi i/o chip is unlikely. Stopping at code 41 here usually means it can’t find a bootable media.

Checking the FDC inside the multi i/o chip is easy. Just hook up a floppy drive and insert a bootable MSDOS disk. When the POST card cycles from code 41 to FF, the FDC is working normally and the board boots to DOS via the limited BIOS run-time code in the Bootblock.
If this works, your screen should look like this:

The attachment Monitor_1_20251031-155441-305.png is no longer available

Jan

Hi, it's exactly as you said. Thanks! I got an output via a Trident 8900C (ISA).
I'm curious, I guess that's because it needs no initialized PCI bridge?
But if so, how comes the ISA VGA card works without the PCI-ISA bridge being initialized yet?

Anyway, I'm glad that it "lives". I've also dumped the BIOS, but the verify failed each time.
But that could also because of my vintage programmer, a TL866.
It uses flash memory to buffer data. Maybe that flash memory has degraded over the years, not sure.
The BIOS itself does also report a checksum error, however. So there's indeed something wrong, obviously.

At the moment, I still have no floppy drive connected. Because drive A: is an 1,2 MB 5,25" drive (B: is 1,44 MB 3,5").
I would need get another cable and a 3,5" drive for testing. I've got some in the attic, I think.

I really wonder why the BIOS got corrupted exactly after I had switched CPUs..
I was very carefully and touched the chassis to avoid ESD etc.
Should I try to later reprogram the EPROM and try installing the Pentium MMX again?
Or should I leave the motherboard configured for single voltage operation and leave the Pentium 133 installed?

Jo22 wrote on 2026-06-17, 20:18:

I'm curious, I guess that's because it needs no initialized PCI bridge?
But if so, how comes the ISA VGA card works without the PCI-ISA bridge being initialized yet?

in this generation ISA was native and didnt need initialization

Jo22 wrote on 2026-06-17, 20:18:

Should I try to later reprogram the EPROM and try installing the Pentium MMX again?
Or should I leave the motherboard configured for single voltage operation and leave the Pentium 133 installed?

coincidence, no reason not to go with MMX
I worry POST card went bonkers and send weird voltage on data/low address bus killing BIOS and maybe something else

rasz_pl wrote on 2026-06-17, 20:43:

in this generation ISA was native and didnt need initialization

Hi rasz_pl! Ah, I see, okay. I wasn't entirely sure about this.
Back in the day™, I got the impression that 386/486 marked the end of classic AT architecture.
It didn't help that our PC upgrade/repair books were from before Windows 95 days, I guess! 😅

rasz_pl wrote on 2026-06-17, 20:43:

coincidence, no reason not to go with MMX

Thanks, glad to hear!
The Pentium 133 as such is fine, I think, but being able to run ZSNES or Impulse Tracker with MMX support would be nice! 🙂

Would also help a bit with SVGA titles, since on this motherboard (256kb L2 cache) the Pentium 133 reaches about 17 fps in PC Player benchmark (640x480 256c, lfb).
That's not bad, but 20 to 25 fps would be a bit smoother.

rasz_pl wrote on 2026-06-17, 20:43:

I worry POST card went bonkers and send weird voltage on data/low address bus killing BIOS and maybe something else

Hi, yes, that sounds reasonable.
Though I haven't used the ISA/PCI card at the beginning when it all started.

But maybe it did some damage, too.
The floppy drive somehow doesn't seem to work for the Award's emergency BIOS.
I've tried three different floppy drives and floppy cables now.

The MS-DOS boot disk I've tried is an 360k or 720k formatted floppy, though..
Maybe the minimal BIOS has "detection" issues, thus.

I think I'll wait a bit and think about what to do next.
Maybe I'll try to reflash the existing EPROM or try to get a new one..

Hi Jo22,

I’ve checked the BIOS you’ve dumped and it is indeed corrupted!
I’ve compared it to the Wims 32GB patch from TRW and found only 27 different bytes in the compressed main BIOS module. The other modules and the bootblock are correct.

The locations of the corrupted bytes, and their bit-patterns, very much look like bit-rot in the EEPROM. So there is a good chance that reprogramming this flash chip with the copy from TRW will fix the BIOS.
Try this first and let your TL866 do several compares after the reprogramming to confirm the EEPROM is still good.

So now we found a corrupted BIOS, the RAM is probably fine.
But I share @rasz_pl’s worry that the faulty POST card may have caused further damage, but hopefully fixing the BIOS will save this board.

Cheers, Jan

Jan and everyone else - thank you very much for your help! ^_^

I've erased the flash chip (intel P28F00 TBX?) multiple times and then flashed it with the WIMS patched binary image.

Everything seems to work fine now! The Pentium MMX is detected, too! I'm happy!
I can now boot from a DOS boot-up disk, too and run SpeedSys!

I'll reassemble the PC tonight!

Thanks again,
Jo22

Great, another retro board saved! 😀
Also thanks to the TRW project for hosting all that retro data and BIOSes!!

Cheers, Jan

Thanks. 😃