HandOfFate wrote on 2020-01-31, 10:45:

It was always missing. But there are some jumpers for setting the keyboard controller to "internal" and "external".

The newer chipsets can emulate these, true, but I've never seen mobos with an empty socket. Usually the manufacturer would just not solder it at all, and the jumpers would be hardwired to one position at factory. I mean, where is the end user going to find a compatible keyboard controller chip anyway? But I'm basing this on the mobos I've seen, perhaps this is not the case in other parts of the world, so good chances are your mobo is just like that and the socket is meant to be empty.

The BIOS chip is either EPROM of Flash, maybe EEPROM, but in the end it doesn't matter much. It's going to have the same pinout as 27C010 (or 020/040) part: https://www.futurlec.com/Memory/27C010-150.shtml
You want to check pins 22 and 24 (again, make sure it's not a direct connection to GND on one of them) for activity.

Ah, great, that saves me from some peeling. I will report back in 10 hours or so. Thanks 😀

I'm measuring a series of square wave-looking signals on pin 22 in the first few seconds after powering on. After that it's steady for a bit (at low voltage), then a few more pulses and then steady again. And then I turned the power off. So the board is not completely fried, I guess.

Is this how it should be? Mostly 'active low' and a few 'high' pulses?

@SquallStrife I've been so focused on the BIOS that I forgot to try your method for measuring the speaker. I will do that next.

Edit: when turning on the power supply, the speaker immediately gets 5V and stays there. I'm not seeing any movement, whatever hardware configuration I start up with (RAM, no RAM. (No) video card)

Hi! According to some handwritten notes on a manual for a different board, using the AMD Enhanced 486 DX4 should be correct. And, your speaker should be working normally, from the data you gave, by connecting it as usual. But, the manual says that the only recorded beep code is graphics failure, so if something else is failing it might not beep.

Now, on to the useful part: Looking at your pictures I see only Bank 0 of the cache is populated. Looking at your cache settings jumpers (JP11-14) you seem to have some undocumented configuration. It could just be that I can't see the jumpers correctly, but check it. Pins 1 or 1 and 2 of the jumpers are marked on the board, so you can configure it based on the manual or the text next to the cache chips on your board. EDIT: Just in case, before changing any jumpers, document what you had!

Please, keep us posted on updates!

350606 wrote on 2020-02-01, 16:52:

Now, on to the useful part: Looking at your pictures I see only Bank 0 of the cache is populated. Looking at your cache settings jumpers (JP11-14) you seem to have some undocumented configuration. It could just be that I can't see the jumpers correctly, but check it. Pins 1 or 1 and 2 of the jumpers are marked on the board, so you can configure it based on the manual or the text next to the cache chips on your board.

Yep ! those are 32k X 8 cache chips in bank 0 for 128k total, appears the jumpers are not set proper.

350606 wrote on 2020-02-01, 16:52:

Hi! According to some handwritten notes on a manual for a different board, using the AMD Enhanced 486 DX4 should be correct. And, your speaker should be working normally, from the data you gave, by connecting it as usual. But, the manual says that the only recorded beep code is graphics failure, so if something else is failing it might not beep.

Good to know about the settings (edit: and that handwritten note is a nice find!). Weird that the error beeping was so different between boards back then. My other 486 motherboard also has an Award BIOS and does beep if something's up with the memory. But this gives me some more more hope that it isn't dead.

Now, on to the useful part: Looking at your pictures I see only Bank 0 of the cache is populated. Looking at your cache settings jumpers (JP11-14) you seem to have some undocumented configuration. It could just be that I can't see the jumpers correctly, but check it. Pins 1 or 1 and 2 of the jumpers are marked on the board, so you can configure it based on the manual or the text next to the cache chips on your board. EDIT: Just in case, before changing any jumpers, document what you had!

Please, keep us posted on updates!

Interesting observation. I haven't touched them at all since I got the board, assuming that it was fine (but you know what they say about assumptions). I will check the settings and let you know 😀

What I did find surprising is that the cache chips don't take up the entire socket. I found some other random photos online where that was also the case, but does the way they're inserted make sense?

In the overview photo that I posted, the notch on the chips, and marking on the board, are positioned on the left side. The chips are aligned to the right on the socket, leaving a gap on the left. Should they be like this, or be moved to the left side? (I believe there are traces going to all holes so that doesn't give a hint)

Thanks for the help so far everyone!

HandOfFate wrote on 2020-01-31, 22:17:

I'm measuring a series of square wave-looking signals on pin 22 in the first few seconds after powering on. After that it's steady for a bit (at low voltage), then a few more pulses and then steady again.

And that's pretty much all you can do with a scope. It tells you the CPU is running, but as to what code and to what end - you can only guess. You won't get much further without a POST card. Well, unless you just happen to find what is wrong by poking around.

I checked the cache jumpers and I think the confusion was caused because a) my terrible phone camera and b) because the seller was nice enough to include the jumpers but leave them disconnected. So none of the pins on JP11, 13 and 14 are actually closed.

I also tried starting with the disconnected ones removed, just in case. But no improvement.

It would be nice if that POST diagnostics card would arrive soon. But seeing that China is getting closed down more and more makes me think it might take a while.

That's a bummer! Just going on a limb here, and not having ever messed with AMD's 586s, what's your RAM? Do those 586s like having just one slot populated? Are you following the motherboard's DRAM configuration rules?

Deunan wrote on 2020-01-30, 23:49:

U30 socket (left bottom side) is empty - did you remove that chip for some reason before taking the photo?
Typically this is where keyboard controller is. If it's missing the mobo might not boot at all, these chips provide (as a secondary function) control over the CPU reset line. Might very well be why it appears dead.

For some motherboards this is normal.. Because the keyboard bios is intergrated in the bios rom. Also some other 486 boards with UMC chipset have the same non keyboard bios lay out.

350606 wrote on 2020-02-01, 22:28:

That's a bummer! Just going on a limb here, and not having ever messed with AMD's 586s, what's your RAM? Do those 586s like having just one slot populated? Are you following the motherboard's DRAM configuration rules?

I think I'm following the rules.

The manual says that SIMM slot 1 has to be used before you can use SIMM 2, SIMM2 before you can use SIMM3, etc. I've just now tried to use some other combination of slots (only 2, 3, 4 and all of them) but no change in video or PC speaker beeping signal, unfortunately.

The manual also says the speed should be <=80ns. None of the RAM that I have, have their speed marked anywhere. So possibly that is it? Edit: I found the specs of one of my RAM sticks and it is 60ns, so this is probably not the issue.

It's also possible of course that the first SIMM slot is damaged. It looks fine but since the metal holding clip is weak maybe the last RAM was removed by force?

Because I'm anxious to see if the board shows any other signs of life, I've ordered another POST diagnostics card just now from eBay. The seller claims it will arrive between 6 (this Thursday) and 14 February.

The POST card from eBay arrived a day earlier. It didn't include the manual listing the display codes (*sigh*) but I think I know what '0000' means: dead.

So unless I happen to find the BIOS somewhere I guess there really isn't anything else that I can do.

It always makes me sad to see old hardware perish 🙁

Bummer ! I think you should try any 486 BIOS from a similar board with same UM8886AF + UM8881F chips + PCI slots. I do not see a floppy port in the picture and is missing a I/O chip which is a bit odd since the board has IDE ports.

HandOfFate wrote on 2020-02-05, 19:32:

The POST card from eBay arrived a day earlier. It didn't include the manual listing the display codes (*sigh*) but I think I know what '0000' means: dead.

So unless I happen to find the BIOS somewhere I guess there really isn't anything else that I can do.

It always makes me sad to see old hardware perish 🙁

I am a bit late, I guess you did try out everything already.
Did you try the board with a 5V cpu already? If you have one at hand try a intel 486DX-33, it does not need a heatsink and heats up nicely 😁 Does it get warm?

HandOfFate wrote on 2020-02-05, 19:32:

The POST card from eBay arrived a day earlier. It didn't include the manual listing the display codes (*sigh*) but I think I know what '0000' means: dead.

'0000' shouldn't really be displayed, unless your card works in a different way than the ones I've seen. First of all, does it start with zeros or dashes? Remove CPU and power the mobo with the card in a slot. Is it '0000' or '----'?
If you get dashes now then the BIOS does get executed but dies or loops very early. There are 4 digits but a code is only 2, what you see is the current code and the one prior to it. These shouldn't be the same. And it depends on the BIOS but 00 is such an early code that it could only be a very early test that fails - CPU or maybe CMOS.

Also, is that an ISA/PCI combo card? Try both but ISA is preferred (PCI might not even work properly on some mobos).

HandOfFate wrote on 2020-02-05, 19:32:

The POST card from eBay arrived a day earlier. It didn't include the manual listing the display codes (*sigh*) but I think I know what '0000' means: dead.

Or it means you wasted money on a fake card that has the segments hardwired to display that.

Horun wrote on 2020-02-05, 23:36:

Bummer ! I think you should try any 486 BIOS from a similar board with same UM8886AF + UM8881F chips + PCI slots. I do not see a floppy port in the picture and is missing a I/O chip which is a bit odd since the board has IDE ports.

The missing floppy connector surprised me too. I wonder what the idea behind that was, because you couldn't really live without a floppy drive back then.

CoffeeOne wrote on 2020-02-06, 00:03:

HandOfFate wrote on 2020-02-05, 19:32:

The POST card from eBay arrived a day earlier. It didn't include the manual listing the display codes (*sigh*) but I think I know what '0000' means: dead.

So unless I happen to find the BIOS somewhere I guess there really isn't anything else that I can do.

It always makes me sad to see old hardware perish 🙁

I am a bit late, I guess you did try out everything already.
Did you try the board with a 5V cpu already? If you have one at hand try a intel 486DX-33, it does not need a heatsink and heats up nicely 😁 Does it get warm?

I don't have one unfortunately. The 5x86 isn't getting warm at all in any case, even if I leave off the heatsink.

Deunan wrote on 2020-02-06, 00:09:

'0000' shouldn't really be displayed, unless your card works in a different way than the ones I've seen. First of all, does it s […]
Show full quote

HandOfFate wrote on 2020-02-05, 19:32:

The POST card from eBay arrived a day earlier. It didn't include the manual listing the display codes (*sigh*) but I think I know what '0000' means: dead.

'0000' shouldn't really be displayed, unless your card works in a different way than the ones I've seen. First of all, does it start with zeros or dashes? Remove CPU and power the mobo with the card in a slot. Is it '0000' or '----'?
If you get dashes now then the BIOS does get executed but dies or loops very early. There are 4 digits but a code is only 2, what you see is the current code and the one prior to it. These shouldn't be the same. And it depends on the BIOS but 00 is such an early code that it could only be a very early test that fails - CPU or maybe CMOS.

Also, is that an ISA/PCI combo card? Try both but ISA is preferred (PCI might not even work properly on some mobos).

It's showing '0000'. The card is PCI only. I tried it in another machine just now and it did show some numbers (probably showing "successful boot" because the system works). Still nothing on the 486 motherboard.

I must also say that I'm not measuring that nice square wave on pin 22 of the BIOS anymore (CPU still works in other board). Now I did measure a square wave-looking thing again, but it seems the pattern is different each time I turn the PSU on. Maybe my cheap scope is being weird or the board gets stuck at different points each time I power it on?

I will try the other POST board when it will finally arrive, which will be both ISA and PCI, but I'm starting to think it's time to give up.

Have you checked the voltages at the AT connector, at the transistor below the CPU socket and at the socket itself already?

Doornkaat wrote on 2020-02-06, 11:28:

Have you checked the voltages at the AT connector, at the transistor below the CPU socket and at the socket itself already?

I haven't. Where should I measure the last two points exactly?

The idea of measuring the AT connector itself makes me a bit nervous. How can I do that safely without killing the PSU or my multimeter?

Just don't short anything and you'll be fine. If you aren't trusting your hands then connect the probes first, then power the system on.
Also, not every PC PSU will start or output correct voltages with no load - and that goes especially for the older ones.