tikiellner, someone else on Vogons happens to be currently working on the same board you're having trouble with atm, in this thread Another high-end 486 back in action (ASUS GX4 conversion success! Tons of pictures)
He seems to be very knowledgeable of the VL/I 486SVGX4 and its variants. Perhaps you could ask him for help or advice?

Thanks! Very cool, I'll ask him for advice!

@Feipoa - the man I contacted says he has successfully erased this kind of chips with long (~3 hours) exposure from a single UV bulb he owns.

Is a UV bulb advertised as such? Or is it a grow-op bulb, a tanning bed bulb, or what? If it is cheap enough, I'd like to buy one to erase my UV EPROM chips.

@Feipoa

In my hardware store (Castorama) there are indeed bulbs advertized as "UV bulbs". I got myself one lately for around 1 USD equiv. You could also get this cheap UV lamp that is designed for hardening "hybrid nail polish". I found some of them for cheap - like 10 USD equiv.

I went to the shop where my board is being serviced today - the guy happens to have one of these dedicated UV erasers (erasing takes ~30m in one of these). He should be done burning the new version on the chip so in the late afternoon we should know more.

I'm not really optimistic, as it seems to me that the chances that BIOS got somehow corrupted is rather thin and we're slowly running out of options. We'll see.

They have replaced the electrolytic caps, but to no avail.

Where did you get this board from? I wouldn't be surprised if the previous owner wanted to get an Am5x86 working and tried to flash the BIOS to update it to the latest version. Then the BIOS data got corrupted. I've done that before. The board would no longer turn on. The chance that this has happened here are pretty good, perhaps 40%.

http://allegro.pl/plyta-procesor-486-dx-33-i6589647924.html

This was some anonymous offer from one of these sellers who does not even know what he sells as he sells so many items at once.

The board came equipped with 486 DX33 and configured like on the picture under the link. I do really hope you're right with the BIOS 😀

Small update:

There seems to be something wrong with BIOS chip.

It was erased and all the registers have "0" value. I was not able, however, to successfully burn the image on the chip. The chip seems faulty.

I found an old 386 AMI bios chip (working) and will try to burn the BIOS on it tomorrow (it's and AMD am27c12, so the proper size - 64kbit x8).

You must use an EEPROM chip. I've never seen a 386 with an EEPROM chip.

Save yourself a lot of frustration and order Winbond W27C512-45. Order at least 2.
http://www.ebay.ca/itm/1PCS-EEPROM-IC-WINBOND … rMAAOSw3KFWc7VF

and

write BIOS image 401A0-0402-1 to it.

Ok, I will :--)

I just thought that if the particular chip - as the one in question - has the window (and thus is erasable in the UV light) it may then be used again. I might be wrong though, as I'm not particularly strong with elecronics.

Wouldn't hurt to try. If it works, maybe I'll go to the drug store looking for a UV lamp for nail polish.

At any rate, it is always good to have a few different sized EEPROM's on hand.

tikoellner wrote:

Small update: […]
Show full quote

Small update:

There seems to be something wrong with BIOS chip.

It was erased and all the registers have "0" value. I was not able, however, to successfully burn the image on the chip. The chip seems faulty.

I found an old 386 AMI bios chip (working) and will try to burn the BIOS on it tomorrow (it's and AMD am27c12, so the proper size - 64kbit x8).

If I'm not mistaken erasing one of those chips with UV should leave all the bits in 1 (each cell should have value "FF" after erase). When you program the chip some of those 1s are turned into 0s as needed. Generally the programmer software has a utility to properly check if the chip is "blank".

I don't know what programmer you are using, but in order to write to that kind of UV-erasable chip you need a higher voltage than what you use to read. Be sure to check the datasheet for the correct voltages (Vpp/Vcc) and set those on your programmer accordingly.

Do you have a graphics card installed during your tests? Didf you try to exchange it?

Yes, tried about 4 cards, 2 ISA and 1 VLB.

Update:

The BIOS was successfully written on a chip and verified.

Still, the board remains dead. No beeps, nothing. I think I'm done here, as I don't know what more should I do. I think all this may suggest some invisible component damage.

Are you using the jumper settings for CPU voltage which I mentioned? And try alternate FPM modules? I found this board a little particular about which RAM it would take.

Now that you got the BIOS flashed correctly, you really need to start the troubleshooting process all over again.

Yes, I just did. I tried DX, DX2, DX4 settings, and three different FPM modules (1x4mb, 2x4mb, 1x16mb and 1x32mb). I also played with voltage jumpers (JP16). Tried with and without video card, etc. No beep, no video.

If you didn't live in Canada, I would happily send the board to you for revision 😀

I wonder if something went wrong with the BIOS writing. Try those EEPROM chips I recommended. You should also verify what voltage is being received at the CPU, that is, the potential between Vcc and Vss on the CPU socket. Best to check this under load, but you can start with an open circuit test (no CPU).

Some update.

While I have totally given up trying to fix this board, I was searching for some replacement. Finally I found some cheap strudy 486 setup with similar motherboard. It's an older ASUS VL/I-486SVGO that came with 486DX2ODPR and 4mb RAM onboard.

The good thing is that the board works.

I thought I could try to swap BIOS chips and see what happens. Here is what I found out:

- VL/I-486SV2GX4 BIOS that I have burnt to an old chip works with the other board - the board, however, keeps resetting after memory test is completed (where the table should appear) - can you tell me what BIOS can I use to get 5x86 support? (it's 1.2 revision...).

- on the other hand, VL/I-486SV2GX4 still woun't post with BIOS chip from VL/I-486SVGO. IMO this confrims that the problem is not related to BIOS.

I do not have any experience with the ASUS VL/I-486SVGO so I cannot say which BIOS gets 5x86 support. I"m not sure what is going on with your VL/I-486SV2GX4, but perhaps there is some hardware failure somewhere?

Rev 1.5 is squarely parallel to 486SV2G(X4) 2.0 and BIOS compatible with it. I don't know what the 1.2 revision looks like, but I guess you've shown that the BIOS that brings official 5x86 compatibility won't work with older revs.

Also, the non-X4 SVGO/SV2G is missing support for CPU voltages below 5V.