Ouch...

Good chance the CPU is toast. The board "might" still live, but consider yourself lucky if that is the case 🙁

Here's the damage. The burning smell has nearly gone. Stupid question - do you think it's worth switching it on again? The "black blood" is only on two pins. Worth a go, or not? Does anyone know what those 2 "charred" pins do? You can work out the orientation of the socket, because the "CPU corner" part is in the top right part of the photo.

The mobo might be able to "limp along" without the need for those 2 pin bits to work. Or it might just start a fire.

Make sure that the board is still working, then have a look at these boards. The jumpers are configured for Am5x86 as far as I remember...

I just tried the CPU that was in the mobo when it started to burn, and it still works! (I tested this CPU in a different mobo.) I haven't tried the "burnt" mobo yet. BTW, all of the other components taken from the burnt work too - PSU, RAM, VGA card, etc. I doubt if the burnt mobo will work, but I might try it tomorrow morning. I could power it on with nothing in it first, to see if it starts to burn again...

Just noticed on the photo that your board doesn't have a voltage converter (see the silkscreened rectangles just east of the CPU socket), which means it runs all CPUs with 5V. So be careful and only try "cheap" CPUs.

^agrees

Try it with an expendable CPU.
One odd thing, why are 2 of the most outer pins burnt? Those pins are only used by the POD?
Could it have been an incorrectly placed jumper?

Thanks a lot for the info, people. Guess what? It's alive! RG100 destroyer of old hardware? I think not! I just wounded it, that's all! 😉 I think that I have been incredibly lucky. I went to this webpage. There, I went to page 7. On it, it shows you a "pin out" diagram of a 486 33MHz CPU. I have copied a bit of it, and it appears below.

I tested the board with an enhanced 486 DX2-66, non writeback model, and it seems to work OK. Tetrium, do you think the location of the 2 burnt mobo socket holes would affect a POD83 CPU? If so, I won't try one. Instead, I'll just use this board with less powerful CPUs.

Regarding the jumper settings for a 5x86, I believe that there are two issues with this. Firstly, as 5u3 pointed out, there is an absense of a voltage converter. Also, the other issue with using a 5x86 is that I have version 1 of this mobo, and reading about this mobo on the net, I believe that it has no support for this CPU.

@rg100
I damaged my first computer in the same way as you - motherboard did not survive but CPU is still alive! (:

I think that you shorted Vcc (positive supply voltage) to Vss (ground) by putting CPU in wrong orientation. Then the current flew and the weakest wire burned. However your motherboard can still operate if your CPU doesn't use damaged wire. I'm sure that not all CPUs make use of all Vcc and Vss pins.

Regards! [:

EDIT:
Good investigation! (: I'm glad that your motherboard works correctly.

udam_u wrote:

... I'm glad that your motherboard works correctly.

So am I! 😀 I booted up the mobo using Windows 98 SE. It seems to work fine. Here's a quick test, using WinTune. It's nothing special, but just shows that it works. It's useful to have a couple of "crappy" mobos lying about, for testing purposes. I like this mobo, because it survived after having been "shot in the arm". 😉

Lucky git!! ;D

And with such burn marks (and plenty other boards available 😉 ) I wouldn't even consider "just trying" a POD83 "just for the heck of it" 😉

5u3 wrote on 2011-05-04, 23:40:

Just noticed on the photo that your board doesn't have a voltage converter (see the silkscreened rectangles just east of the CPU socket), which means it runs all CPUs with 5V. So be careful and only try "cheap" CPUs.

I have the same motherboard with no voltage regulator and the place of the voltage jumper is hardwired to closed position. But the mb works with Intel DX4-100 and AMD 5x86-133-P75 CPUs so it must have autodetect feature. I have not tested it long, but it did not kill the CPUs on instant, and Quake and Doom timedemos run well. Sadly these are old motherboards with very little information on the internet and even those documentations does not fit the motherboard...
With the AMD 5x86-133: the WB-WT jumper have to stay open, in one setting the mb did not post, in the other it detected only 640k of memory. One interesting thing is that my CPU was detected with cpuid 4E4 before I started to try the WB/WT jumper, and now it detects as 4F4 (it was only 1 jumper and I removed it, so nothing was changed permanently). It may be some instability, the CPU socket is not perfect, but interesting because it totally works, only with different cpuid.

(I know the thread is old, but these forums are useful until the last of these old motherboards die)

Does anyone have a BIOS dump for this mainboard? v1.0 specifically.
http://www.win3x.org/uh19/motherboard/show/5568 has a few newer ones but those don't seem to work;
I've tried a v5.0 ROM with the chip inserted the weird way as on the picture in the first post but the board doesn't POST past the RAM test ("D3D2" via debug card) no matter if RAM is inserted or not. No beep or video.

Background: Found two of these boards both fitted with an AMD 80MHz 486 DX2 but someone took the cache chips and the BIOS chip. Cache was easy to fix, but the BIOS isn't available anywhere on the net at this point ...

EDIT: Someone has v3.0 (attached here for convenience) but that doesn't work either, same behavior.

EDIT#2 for future visitors and time travellers: I was unable to obtain a QDI 4P895P3 / SMT V1.0 BIOS. But turns out that UH19 has a very similar board, TMC PAT48PG4, with an Award BIOS but for the same chipset (OPTi 82C895). That one works! Attached just in case.

EDIT #3 there is also a OPTi895 variant of MR-BIOS in the big bundle at vogonsdrivers, but that doesn't seem to work on my board (POST gets stuck very early).

EDIT #4: Turns out that with the above BIOS normal ISA VGA cards work fine but my Speedstar Pro VL is extremely unrealiable on this board and it only POSTs when capacitors are discharged after leaving it alone for so long. I tried a few BIOS chips from other mainboards and the one from a MS-4123 (Sis896 aka 82C896 chipset) also works and POSTs everytime now. So this is a good alternative to try.

EDIT #5: Previous attempts either didn't detect RAM properly or were unstable. After trying ALL of the currently available BIOSes for that chipset on UH19 i've settled with the Award BIOS for this board. Mirrored file:

Verdict: This is one bitchy board. I'm not surprised they needed over 5 revisions to fix their shit.
Some AMI BIOSes outright don't POST at all. Others do but only detect 640k RAM. When they boot they work fine with the speedstar card, but won't POST when any sort of IO expansion card is present, VLB or not.
The Award BIOSes seem to have more trouble booting, most don't like the VLB card. With the last linked BIOS i've got it to a state where it is fine with the speedstar VLB card and a non-VLB IO card. It won't POST with two VLB slots used but at least it's in a state now where the hardware is somewhat usable.

looks like we have a duplicate in UH19..
http://www.win3x.org/uh19/motherboard/show/4271

I'm having so much trouble with this board myself actually, the rev 1.0 like in the original post. I can get it to post at 50mhz on a dx2 and that is it. I can not for the life of me get the jumper settings right on the dx4 100.

I thought I had it figured out when I got the 50mhz to post but now I'm at a loss. The 100mhz will post at 5v but obviously that's not a good idea. Is the voltage regulator dead in mine or something?