Deunan wrote on 2021-04-19, 23:53:

I think the voltage is OK, it's not uncommon to have it a bit higher to compensate for drop with load and/or stability. So long […]
Show full quote

figbash wrote on 2021-04-19, 18:29:

I'm able to get it down to 3.6 (but not 3.3) by messing with the voltage setting jumpers with the CPU removed, though it still doesn't post and the 3.8 is the official jumper layout.

I think the voltage is OK, it's not uncommon to have it a bit higher to compensate for drop with load and/or stability. So long it's not over 4V the CPU will be fine. And it's working at least well enough to show some codes (and reacts to missing memory with C1 error, that also seems correct).

Try lower CPU clock (20/25 MHz) to see if that will make it go any further. Go over all the mobo jumpers (I know, tons of them), especially those that you didn't touch or are not sure what they do - and try to establish what each of them do and if it's set correctly. Perhaps one is missing or something.

And again, these mobos are really quirky about RAM. Most will not tolerate EDO sticks, only FPM ones, and even that is no guarantee. You only need one 72-pin RAM stick to get a 486 mobo to work. See if you have any 4MB sticks, I don't think any of those were EDO, too old for that.

Ok, I tried several different RAM sticks that are all FPM from googling chip numbers and are all different, nothing seems to make a difference there.

I tried 25 and 20 MHz clock but nothing there either.

I went over the jumpers again, as far as I can tell everything is correct. It is very jumper heavy though ha.

Could it be a capacitor issue? Anyone know of a good way to check the capacitors for shorts in circuit? There are 3 tantalum capacitors that seem to read 0 for resistance whereas the others slowly go up.

Your mobo has tantalum capcitors, those rarely go open. The usual mode of failure is cracking due to thermal cycling, mechanical stress or just manufacturer defect to begin with - and that invites moisture and oxygen. Tantalum is pretty reactive, the capacitor will typically self-destruct in short order. I've yet to see one fail slowly. So if there are no explosions, smoke, and glowing remains, I'd say the caps are fine. You can also look for one that is very hot (usually to the point it will change color) - that happens if there is enough resistance in the circuit to prevent an outright short to power lines. Typically the first ones to go are the +/- 12V line caps next to power connector and these look fine (plus most PSUs will either shut down on short or just turn the offending cap into a small sun).

I've just noticed both CPUs you've tried are DX4 - I think it might be worth trying an older CPU. Preferably some old 5V SX, these are cheap enough. In case it's some weird BIOS issue or maybe some VRM problem that doesn't clearly manifest when testing the voltages. BTW you can pretty safely run a 486SX with no heatsink - it will get hot but not above 65C or so.

figbash wrote on 2021-04-20, 04:19:

I went over the jumpers again, as far as I can tell everything is correct. It is very jumper heavy though ha.

Is JP3 jumpered for normal cmos mode?

Does the POST always stop at the "0D" code?
If yes, that's related to the video initialization, so it might be worth to investigate this part, maybe test other video cards in different slots.
If you're using a VLB video card, it might be worth to also try an ISA card.

Deunan wrote on 2021-04-20, 10:21:

Your mobo has tantalum capcitors, those rarely go open. The usual mode of failure is cracking due to thermal cycling, mechanical stress or just manufacturer defect to begin with - and that invites moisture and oxygen. Tantalum is pretty reactive, the capacitor will typically self-destruct in short order. I've yet to see one fail slowly. So if there are no explosions, smoke, and glowing remains, I'd say the caps are fine. You can also look for one that is very hot (usually to the point it will change color) - that happens if there is enough resistance in the circuit to prevent an outright short to power lines. Typically the first ones to go are the +/- 12V line caps next to power connector and these look fine (plus most PSUs will either shut down on short or just turn the offending cap into a small sun).

I've just noticed both CPUs you've tried are DX4 - I think it might be worth trying an older CPU. Preferably some old 5V SX, these are cheap enough. In case it's some weird BIOS issue or maybe some VRM problem that doesn't clearly manifest when testing the voltages. BTW you can pretty safely run a 486SX with no heatsink - it will get hot but not above 65C or so.

Alright I can try to grab a CPU

weedeewee wrote on 2021-04-20, 11:58:

figbash wrote on 2021-04-20, 04:19:

I went over the jumpers again, as far as I can tell everything is correct. It is very jumper heavy though ha.

Is JP3 jumpered for normal cmos mode?

Yeah it is, I also tried resetting the cmos earlier, sadly it’s not that.

bloodem wrote on 2021-04-20, 12:01:

Does the POST always stop at the "0D" code?
If yes, that's related to the video initialization, so it might be worth to investigate this part, maybe test other video cards in different slots.
If you're using a VLB video card, it might be worth to also try an ISA card.

Yeah it always stops there. I tried a couple ISA video cards which both work in a 386 and a VESA card, and I’ve tried every slot with the ISA. It doesn’t react at all to video cards.

Tried DX2 66 and 50 CPUs, and also flashed the BIOS with the newest one with the help of a friend.
It still has the same issue, but one single time when it didn't have the video card in, we saw it get way farther on the POST card and heard it do the memory check beeps. We plugged in a video card excitedly, but it never got past 0D again and now it's back to being broken with the video card in or not. It doesn't seem to be loose chips/components, and the jumper settings are obviously correct enough for it to POST the one time. We looked at it with a thermal camera and nothing looked wrong there.

Anyone have any ideas with that new info?

figbash wrote on 2021-04-23, 15:58:

Anyone have any ideas with that new info?

Verify all the signal lines between the ISA slot & chipset.

You heard the RAM count beeps. Means it's a matter of time to get it working.
Have you tried removing all cache chips? If legs are oxidized it could prevent booting.
Any loose components?
You could try holdnig the motherboard down by pressing in the middle or other parts...

figbash wrote on 2021-04-23, 15:58:

Anyone have any ideas with that new info?

Did you try all 3 VESA slots? Typically you'd use the one closest to CPU but I have a 386/486 mobo that seems to have jumpers for 50MHz operation on slot 1, and that is actually the second one from CPU.
And what other VGA card have you tested? A Trident TVGA 8900D perhaps? If not, did you try different slots (in case there is some damage to mobo between slots).

Nexxen wrote on 2021-04-23, 17:51:

You heard the RAM count beeps. Means it's a matter of time to get it working. Have you tried removing all cache chips? If legs a […]
Show full quote

You heard the RAM count beeps. Means it's a matter of time to get it working.
Have you tried removing all cache chips? If legs are oxidized it could prevent booting.
Any loose components?
You could try holdnig the motherboard down by pressing in the middle or other parts...

The cache chips are new and I didn't see any oxidation on the sockets, but not sure.
There are no loose components I can find, but I'm very suspicious that is the issue. We tried pressing down all the various chips while booting. Starting to think it may be a bad solder joint that isn't obvious to the eye.

Deunan wrote on 2021-04-23, 21:59:

figbash wrote on 2021-04-23, 15:58:

Anyone have any ideas with that new info?

Did you try all 3 VESA slots? Typically you'd use the one closest to CPU but I have a 386/486 mobo that seems to have jumpers for 50MHz operation on slot 1, and that is actually the second one from CPU.
And what other VGA card have you tested? A Trident TVGA 8900D perhaps? If not, did you try different slots (in case there is some damage to mobo between slots).

The issue doesn't seem video card related (it got past that bios code the once without a video card in it), but I have tried an ISA cirrus logic I know is working and a VESA cirrus logic as well, in all slots.

It's because it only passed that code without a video card that I'm asking. Obviously it shouldn't just die like that with VGA installed but it does - so there has to be something related to that too. And I doubt you'd be interested in running this system headless.
Last time I had issues like this was on a 286 mobo that didn't properly generate ALE signal on ISA slots. This prevented Trident 8900D from working, and that card works in every other mobo I've tried (and I tried many, 386 to Pentium). I had to buy 8900C model just for that mobo. Try removing the BIOS extension ROM from the VGA and try it without. The BIOS should not detect the card and again go past that code. This is to test if it tries to init the extension and hangs, or just having the card inserted at all hangs it. Does your friend have any VGA cards that are not Cirrus Logic? In case it is a problem with ALE signal?

figbash wrote on 2021-04-23, 22:09:

The issue doesn't seem video card related (it got past that bios code the once without a video card in it), but I have tried an ISA cirrus logic I know is working and a VESA cirrus logic as well, in all slots.

Out of curiosity, and since I suggested verifying all the traces between isa slot and chipset, do you have an 8 bit video card to test?

Deunan wrote on 2021-04-23, 22:22:

It's because it only passed that code without a video card that I'm asking. Obviously it shouldn't just die like that with VGA installed but it does - so there has to be something related to that too. And I doubt you'd be interested in running this system headless.
Last time I had issues like this was on a 286 mobo that didn't properly generate ALE signal on ISA slots. This prevented Trident 8900D from working, and that card works in every other mobo I've tried (and I tried many, 386 to Pentium). I had to buy 8900C model just for that mobo. Try removing the BIOS extension ROM from the VGA and try it without. The BIOS should not detect the card and again go past that code. This is to test if it tries to init the extension and hangs, or just having the card inserted at all hangs it. Does your friend have any VGA cards that are not Cirrus Logic? In case it is a problem with ALE signal?

Sorry I’m confused, it only got past the 0d code one single time, and doesn’t get past the 0d code with or without a video card installed.

There isn’t currently a video card installed, as the only time I’ve seen it randomly get past the code there was no video card installed so I’m trying to limit variables.

I just tried an ISA trident 8900d I had laying around in every slot and it didn’t change anything unfortunately, the code is always 0d still.

weedeewee wrote on 2021-04-23, 16:53:

figbash wrote on 2021-04-23, 15:58:

Anyone have any ideas with that new info?

Verify all the signal lines between the ISA slot & chipset.

Ok I can check that. I don’t have any 8-bit video cards.

figbash wrote on 2021-04-23, 23:38:

There isn’t currently a video card installed, as the only time I’ve seen it randomly get past the code there was no video card installed so I’m trying to limit variables.

Ah, I misunderstood. I thought it was always passing the 0D code so long there was no video card installed. If it hangs exactly the same without VGA then it's unlikely to be ALE signal problem. BTW 8900D would not work for testing that, you need a card that has ALE not connected. Some CL cards impement it, some don't, but if you are curious it can be verified by looking at the edge connector B28 contact - it should be missing or not connected, this means ALE is not used on that card.

That kinda leaves you with suspected mobo damage I'm afraid. Now this is me just guessing but it could be hanging on 0D when trying to figure CPU speed. For that you need a timer to measure it, and all basic I/O (timers, DMA, interrupts) are integrated in one chip. This is usually 82C206 type but on your mobo it's called 85C407 and it also has RTC built-in. Inspect that chip, and anything nearby for any signs of damage (bent pins, cracked solder, some tool marks that cut a trace - there is a mounting hole nearby, etc). And with that I'm out of fresh ideas.

I bet it is one of the 4 tantalum capacitors near the power connectors. Measure them with a multimeter and if it shows OL, replace the faulty capacitor. I have revived many dead motherboards this way.

OL? You mean close to zero ohms I guess?

I have yet to see a shorted tantalum cap that would not burst into flames when power is applied. You really can't miss that. PSU might shut down instead if the short is on -12V line but the +12V line has more than enough amp rating to just feed the fire and not care. And obviously that should be indicated by the POST card LEDs, that one rail is missing due to short. On the other hand, if the cap went open, it would not prevent the mobo from booting. It's just a 10uF and there's plenty of other local bypass caps to make up for that.