eton975 wrote on 2025-02-26, 03:19:

I got this board from an Ukrainian seller last year and just recently tried to put it to use with 2 16MB SIMMs and an AMD 5x86-133ADZ.

So it's a 486 mobo? These are always very picky about RAM type. Make sure you put FPM modules in there, most these mobos will not work with EDO.

Deunan wrote on 2025-02-26, 11:16:

eton975 wrote on 2025-02-26, 03:19:

I got this board from an Ukrainian seller last year and just recently tried to put it to use with 2 16MB SIMMs and an AMD 5x86-133ADZ.

So it's a 486 mobo? These are always very picky about RAM type. Make sure you put FPM modules in there, most these mobos will not work with EDO.

Just tried with a MCM417400J60-based 16MB FPM SIMM, but no dice.

And yes, this is a 486 (with PCI) board.

32-bit or 36-bit FPMs? I've got a very similar mobo (AOpen AP43) and while I can't remember the exact details of running it (including 5x86 CPU) I found some photos and the memory I've used is 3x 4MiB stick of 36-bit FPM. I know because I've put stickers on these SIMMs and I can see that exact amount of RAM in the HWINFO.LOG and other files I've made during benchmarking.

This is not the mobo that I know for sure that requires SIMMs with parity, that would be VLB SIS471, but it's also SiS chipset. And I'm pretty sure that mobo would stop at 0A or 0B code with wrong memory. So do make sure you are running compatible sticks in it. Preferably try 4M FPM stick with parity, I've yet to find a mobo that would reject those. While 16M sticks should work in a PCI mobo (8M ones are hit or miss in general) make sure you've tried SIMMs with parity chips.

Deunan wrote on 2025-02-26, 12:18:

32-bit or 36-bit FPMs? I've got a very similar mobo (AOpen AP43) and while I can't remember the exact details of running it (including 5x86 CPU) I found some photos and the memory I've used is 3x 4MiB stick of 36-bit FPM. I know because I've put stickers on these SIMMs and I can see that exact amount of RAM in the HWINFO.LOG and other files I've made during benchmarking.

This is not the mobo that I know for sure that requires SIMMs with parity, that would be VLB SIS471, but it's also SiS chipset. And I'm pretty sure that mobo would stop at 0A or 0B code with wrong memory. So do make sure you are running compatible sticks in it. Preferably try 4M FPM stick with parity, I've yet to find a mobo that would reject those. While 16M sticks should work in a PCI mobo (8M ones are hit or miss in general) make sure you've tried SIMMs with parity chips.

It's a stick with 8 MCM417400J60 chips, so it should be 32-bit non-parity memory, as shown below in this post's attachment.

Anyways, I have ordered 4x 4MB FPM 70ns 72-pin SIMMs off eBay which I will test with, from 2 different sellers, but the chips are the same between them.

Based on running your BIOS under emulation, this is long after the BIOS has already decompressed itself into RAM. It should not be a RAM issue.

Between 0B and 0C, the BIOS indeed verifies the CMOS checksum. It also sends commands AD (disable keyboard) and AE (enable keyboard) to the keyboard controller.

The CMOS and KBC are already used earlier in the post, before code 0B is reached. So I don't have any answer for it hanging. Make sure you have a fresh coin cell installed and that the holder still has continuity, I guess? Is the KBC the original one for the board?

jakethompson1 wrote on 2025-02-27, 06:12:

Based on running your BIOS under emulation, this is long after the BIOS has already decompressed itself into RAM. It should not be a RAM issue.

Between 0B and 0C, the BIOS indeed verifies the CMOS checksum. It also sends commands AD (disable keyboard) and AE (enable keyboard) to the keyboard controller.

The CMOS and KBC are already used earlier in the post, before code 0B is reached. So I don't have any answer for it hanging. Make sure you have a fresh coin cell installed and that the holder still has continuity, I guess? Is the KBC the original one for the board?

KBC is an AMIKEY-2, and is original as far as I can tell. The coin cell CR2032 battery reads 2.9V according to my multimeter.

Also note that no beeps happen with the Matrox Millennium II PCI installed, but with no video card or an ISA video card it does 1 long 3 short beeps which is code for a video card error.

eton975 wrote on 2025-02-27, 07:20:

Also note that no beeps happen with the Matrox Millennium II PCI installed, but with no video card or an ISA video card it does 1 long 3 short beeps which is code for a video card error.

Your board is probably incompatible with the Matrox card. I have two older boards that just hang at boot without any errors when using newer PCI video cards. I would recommend sticking with the ISA cards for debugging.

asdf53 wrote on 2025-02-27, 08:28:

eton975 wrote on 2025-02-27, 07:20:

Also note that no beeps happen with the Matrox Millennium II PCI installed, but with no video card or an ISA video card it does 1 long 3 short beeps which is code for a video card error.

Your board is probably incompatible with the Matrox card. I have two older boards that just hang at boot without any errors when using newer PCI video cards. I would recommend sticking with the ISA cards for debugging.

It's possible that it is a video card incompatiblity, but it can't be the only thing. I have tried with a 2nd ISA video card and the POST code display stops at '31 30' for a few seconds before finally displaying '4E 4D' now. It's the same regardless of which RAM I use (hopefully this will change once I get the 4x 4MB FPM SIMMs).

The POST code 4E is 'Reboot if Manufacturing Mode; if not, Display Messages and Enter Setup'. Code 4D is 'Reserved'.

Pressing F1 or enter on the keyboard does nothing, but Ctrl-Alt-Del does reset the system for what it's worth. When this is done, I get 1 long and 2 short beeps.

For me 06 or 0b as you want to call error codes are always related with video card initialization

As stated here some video cards are incompatible and may not "beep" error codes because there was some initialization and the problem is a system crash so the CPU doesn't issue any more error codes
I would try different video cards on different slots and pass some deoxid on the contacts before investigating further

Good luck

Just tried with 2 HY514400AJ-70 based 4MB SIMMS in various slots, no dice. 4E 4B is the code that's now produced. Neither the ISA video card nor the PCI Matrox card seem to work; the ISA card producing the 3 beeps and nothing for the Matrox.

I am beginning to wonder if I actually did put the P8/P9 AT power supply connectors round the wrong way, and if that could have fried something. The PSU I'm using is an EVGA 430W1.

For AWARD the 4E is "Show all error messages on screen", somewhat logical if you have issues but not so bad that the POST would fail completly. It's just there is no screen output at all. Did you try pressing F1 when 4E is ecountered? Does it do anything?

Look for another PCI SVGA card perhaps. Could be the ISA is somehow damaged but these early PCI mobos were very picky about the cards. And Matrox is great when it works but if you want compatible you need S3 Trio64 card. Fortunately those were made in such numbers that it should be possible to get one cheap. Keep this RAM, in general the higher the error code on POST card the better.

Deunan wrote on 2025-03-07, 12:01:

For AWARD the 4E is "Show all error messages on screen", somewhat logical if you have issues but not so bad that the POST would fail completly. It's just there is no screen output at all. Did you try pressing F1 when 4E is ecountered? Does it do anything?

Look for another PCI SVGA card perhaps. Could be the ISA is somehow damaged but these early PCI mobos were very picky about the cards. And Matrox is great when it works but if you want compatible you need S3 Trio64 card. Fortunately those were made in such numbers that it should be possible to get one cheap. Keep this RAM, in general the higher the error code on POST card the better.

Finally it works with a spare S3 ViRGE/DX PCI. The BIOS chip and its socket needed some reseating and cleaning though. I can also confirm that EDO RAM seems to work with the board, at least as far as passing POST.

So now the big questions: What's the maximum HDD size the board can detect/work with, and will it be as picky about ISA sound cards as it was video cards?

Nice!

The maximum HDD is likely 8.4GB, though it may have a display bug on the sys config screen if you have a drive over 2GB. Sometimes that can just be cosmetic. The good news is it will not have the 32GB hang bug, since it's too old to even see the fields for drives over 8.4GB. So you could put in a larger drive and just treat it as 8.4GB.

It shouldn't be picky about sound cards.

jakethompson1 wrote on 2025-03-08, 04:53:

Nice!

The maximum HDD is likely 8.4GB, though it may have a display bug on the sys config screen if you have a drive over 2GB. Sometimes that can just be cosmetic. The good news is it will not have the 32GB hang bug, since it's too old to even see the fields for drives over 8.4GB. So you could put in a larger drive and just treat it as 8.4GB.

It shouldn't be picky about sound cards.

Well, I spoke too soon. The thing now gives a Floppy disk(s) fail (C0) error. From digging around this usually means that the FDC is damaged or something is set up wrong. I've tried 2 different floppy cables (making sure to use the segment after the twist), 2 different 3.5 inch 1.44MB floppy drives of which 1 I know for a fact works on another 486 machine, and have used 2 different power connectors (one Molex to berg, the other Berg straight from the PSU).

Interestingly, I hear a soft click from the floppy drives just before this floppy drive error shows up in the BIOS splash screen, so something is clearly happening between the floppy controller and the drive. Any ideas about how to fix this, or is this likely a toast Winbond W83787F/W83768F?

Edit: Oh, and obviously I have tried swapping the floppy drives in BIOS settings, setting drive B to the 1.44M drive setting, disabled floppy seek (no error is shown but system still does not boot), etc.
I have also heard it could be a bad crystal oscillator. I don't really have anything besides a Cat IV 4000 count digital multimeter to test those with though.

Picture of the POST screen attached.

It also seems to hang on detecting my 8GB CF card in a Startech CF-IDE adapter, or after detecting the nonexistent secondary master. Even with floppy drives disabled.

Cursed board?

eton975 wrote on 2025-03-08, 09:30:

It also seems to hang on detecting my 8GB CF card in a Startech CF-IDE adapter, or after detecting the nonexistent secondary master. Even with floppy drives disabled.

Cursed board?

Well, everything is possible, but since doing some tests doesn't cost you anything, I would start with only the IDE-CF and memory card, although it would be better to try with an HD of about half a GB.
And then I would go to the BIOS, I would do the detection of the disks, then if it does not start, it is to be understood if it detects it correctly (the CYL Head numbers etc...), or if they are ok, what codes does the post card show? When it crashes.
Another attempt you can do is to go to the BIOS and disable the L2 cache, often so they start, then because it doesn't work is another problem to face, and it may not be easy to solve.

eton975 wrote on 2025-03-08, 04:47:

Finally it works with a spare S3 ViRGE/DX PCI. The BIOS chip and its socket needed some reseating and cleaning though. I can also confirm that EDO RAM seems to work with the board, at least as far as passing POST.

Yeah, S3 Trio or Virge is a good card to have on hand for these mobos. I wonder if the earlier code issue was due to BIOS chip socket after all? Though it's curious because BIOS does checksum test pretty early and will die if that is wrong, but it's a simple checksum, not CRC or anything like that.

eton975 wrote on 2025-03-08, 04:47:

So now the big questions: What's the maximum HDD size the board can detect/work with, and will it be as picky about ISA sound cards as it was video cards?

I still think your mobo might have some ISA issues. It's very odd for ISA video card to not work, I mean I've seen that on 286 but that is due to broken ALE, that shouldn't really happen on 386 and later systems. As for the HDD, try something below the 504MB limit if you have any smaller cards. If not then 1-2GB, reason being the BIOS might pre-date the LBA and established CHS translation methods and will not be able to deal with bigger HDDs. That being said it might just be dumb enough to be fooled by some safe values entered by hand (rather than using IDE auto-detect). So try that: up to 1023 cylinders, 16 heads, 63 sectors. That won't boot if the card is already partitioned with correct parameters but should at least try. Note, I've seen some mobos (though earlier than PCI) spit out FDC errors when the problem was HDD instead. Try booting with no HDD/CF connected to see if you can get the floppy drive working at all.

Deunan wrote on 2025-03-08, 22:48:

Yeah, S3 Trio or Virge is a good card to have on hand for these mobos. I wonder if the earlier code issue was due to BIOS chip s […]
Show full quote

eton975 wrote on 2025-03-08, 04:47:

Finally it works with a spare S3 ViRGE/DX PCI. The BIOS chip and its socket needed some reseating and cleaning though. I can also confirm that EDO RAM seems to work with the board, at least as far as passing POST.

Yeah, S3 Trio or Virge is a good card to have on hand for these mobos. I wonder if the earlier code issue was due to BIOS chip socket after all? Though it's curious because BIOS does checksum test pretty early and will die if that is wrong, but it's a simple checksum, not CRC or anything like that.

eton975 wrote on 2025-03-08, 04:47:

So now the big questions: What's the maximum HDD size the board can detect/work with, and will it be as picky about ISA sound cards as it was video cards?

I still think your mobo might have some ISA issues. It's very odd for ISA video card to not work, I mean I've seen that on 286 but that is due to broken ALE, that shouldn't really happen on 386 and later systems. As for the HDD, try something below the 504MB limit if you have any smaller cards. If not then 1-2GB, reason being the BIOS might pre-date the LBA and established CHS translation methods and will not be able to deal with bigger HDDs. That being said it might just be dumb enough to be fooled by some safe values entered by hand (rather than using IDE auto-detect). So try that: up to 1023 cylinders, 16 heads, 63 sectors. That won't boot if the card is already partitioned with correct parameters but should at least try. Note, I've seen some mobos (though earlier than PCI) spit out FDC errors when the problem was HDD instead. Try booting with no HDD/CF connected to see if you can get the floppy drive working at all.

Floppy drive by itself does not work. Spits out the 'Floppy disk(s) fail (C0)' error message, with the correct setting of 1.44MB 3.5 inch for drive A.

Cleared CMOS also to no avail.

The CF card by itself does not boot either. Sometimes it gets detected as size 'H620' or similar in the BIOS's IDE HDD autodetection wizard, other times the correct size '8011'. Usually it's detected wrong when the CF-IDE adapter is at the very end of the IDE cable, but correctly on the closer connector.

Re the ISA issues: Your guess is as good as mine. All I can say is the two ISA cards I have tried did not seem to work. I could try again I guess.

That would suggest some mobo damage. Inspect all the big ICs for bent pins and the like. Also the bottom of the mobo for deep scratches that might have cut traces.

Newer mobos with proper south bridge would usually have a separate channel to the NB/CPU for the integrated IDE/FDC and other stuff (I/O ports, keyboard controller, perhaps USB too). On older ones this is very often tied with ISA bus, which means if there are any problems with ISA it will also affect the integrated stuff. Could be the opposite too, the chip that does the IDE/FDC/IO (and interrupts, DMA) might be faulty and in turn messing up ISA slots. It would be the smaller ones behind PCI slots on this mobo (W83787), but inspect the big ones too.

If everything looks OK then there is a possibility the W83787 or the one nearby (not sure what that is, serial port(s) driver?) is faulty. I never had a case like that myself but there are a few topics where people had to replace these because of floppy drive not working.

BTW your photo shows the POST card in PCI slot if I'm not mistaken? Did you check it in ISA slots? Does it show codes, and are all voltages present? I've seen VGA cards that require -12V to output picture, though usually these will not stop the system from booting if that rail is missing but the screen will be very dark or completly black.