I have ~5 of these cards and only one works. I've never seen a series of so many defective cards while visually looking fine - maybe also Radeon 9800PRO.

I'm curious what will you find out 😉

Do you clean pci contacts with stationary eraser?

Hi, Babasha, thank you for the fast reply,

Yes, I used contact cleaner. I don't see how it could POST with a bad connection to the PCI bus. As far as I know, BIOS checks the first 256k of video memory so the card has some kind of communication with the main board. More ideas are appreciated.

you can measure FB2 FB3 FB4 ferrite beads for connection, if you had oscilloscope you could directly look at the potentially generated signal
you can try poking all s3/ramdac pins with a toothpick under magnification looking for moving ones, or start computer while squeezing s3 /ramdac chips between your fingers

kixs wrote on 2023-09-29, 06:08:

I have ~5 of these cards and only one works. I've never seen a series of so many defective cards while visually looking fine - maybe also Radeon 9800PRO.

I'm curious what will you find out 😉

I wonder why this is so. IMO the DRAM cards of the same era are pretty much bomb proof. For example, I have one 868 which came with my Pentium system in 1995. It was in use several years and at some point I dismantled the system and recycled most parts. This card was one of the parts which I kept for testing purposes, but after that, it spent years in cold storage and was definitely not kept in optimal conditions.

It is currently sitting in one of my 486/PCI system and works 10/10.

Update on my problem: Tested continuity from the RGB output of the DAC, from the pin on the IC to the output of the ferrite I got continuity (good). As I know from playing with Ethernet cards with XTIDE BIOS, as soon as all IDE drives are tested, the system runs all devices ROMs. So my suspicion of a failure to run the VBIOS made me review my BIOS, and it looks ugly. Opening VBIOS files of that card (found online) and similar DIAMOND Stealth cards, it is possible to read copyright text that appears on screen. See the attached images for clarity.

Now I am not so sure about the DAC, something before that is broken. I checked the clock for the DAC and it is working, so here is a list of what I already checked:

• Big capacitors are fine (not in complete short, just ramp up to a stable impedance, but not completely open).
• General output resistors (checked values), ferrite beads, continuity from DAC to output.
• Crystal that serves as DAC clock, fine to
• Tested with a recently programmed EPROM

If someone has an idea of what should be my next step, please let me know!

I just did some news tests, I am more lost now!

With the old BIOS, I can only hear one short beep on POST (the only video card is the S3 968). With another card as my main display adapter, it locks just after all IDE devices are listed. With the new BIOS, the POST code is one long and three short beeps. With another card (one beep on POST), I see Windows loading, but the system locks on Windows 98 loading screen. With safe mode, it freezes with a corrupted screen (on the working card). Maybe I should test on a more modern Windows with better multi-display capabilities?

Ok, more information:

I check again the BIOS chip and I found out that the original one is a 64k so there are two copies of the original binary code. So if a read it as a 27512 ROM it shows a readable text (Stealth 64 Video VRAM and so on). So the BIOS is not the problem. I checked the continuity from BIOS pins to a pair of 744F374 logic chips, and it looks fine. So now I am considering a problem with RAM. With no BIOS on the card, I got a post error, and with the original bios, my computer has ONE beep on POST. So could it be the first ram chip that failed or has some trace broken? I will try swapping the memory chips from the top row to the bottom row, maybe I can get a better idea of what is happening.

Did you check the resistor arrays marked as RP1-RP10, there's a chance some of them are open.

A quick update on my case. Checked all resistor arrays, and nothing is open, all have 33 ohms. Checked all resistors. Nothing is burning on the touch, I don't have a thermal camera, but the PU and DAC are lukewarm. Memories are cold. Removed all ram in despair, as I can't find other issues. I plan to install only one set of four chips, if I can find good-priced sockets I will try them, I think it will make it easy to test.