VOGONS


First post, by dionb

User metadata
Rank l33t++
Rank
l33t++

I have a beautiful old card - but it is flawed

It's an EIZO MD-B12 add-in TIGA card, so a separate 1990 vintage GPU that sits alongside a regular VGA card (think: Voodoo, just 2D and truly programmable,not 3D and fixed function). It connects with the VGA card via VESA feature connector. The idea is to plug the monitor into the MD-B12 and it displays VGA when the GPU isn't active and it's own output if it is

So, what's the problem?

Firstly the passthrough has big quality issues.

Here's the original output on an S3 805 VLB card:

PXL_20240306_214034873.jpg
Filename
PXL_20240306_214034873.jpg
File size
893.55 KiB
Views
88 views
File license
CC-BY-4.0

(The yellow haze is due to it having crap analog circuitry)

And here's what the same screen looks like via the TIGA card:

PXL_20240306_214022054.jpg
Filename
PXL_20240306_214022054.jpg
File size
875.29 KiB
Views
88 views
File license
CC-BY-4.0

Better colour, but the picture is distorted. This is static/stable, so not a moving distortion.

If I set the card to a test pattern, or CGA or MDA emulation, my monitor does not sync:

PXL_20240306_214214850.jpg
Filename
PXL_20240306_214214850.jpg
File size
1.49 MiB
Views
88 views
File license
CC-BY-4.0

I've tried with two different monitors, a CRT and an HD-capable TFT and neither can sync with this. Conversely both can sync with a different (Siemens HiGraph II) TIGA card at high resolutions, so I'm pretty sure it's the card, not the monitor that is to blame. Note that despite lack of sync it's clear the card is doing something and the output responds to input, so the problem seems related to the card output, not it's inner workings.

So, what have I tried? The card was filthy when I received it, but without any visible damage. I cleaned it, externally with water, then the socketed GPU and RAMDAC with contact cleaner. My first guess was that I messed something up with the RAMDAC. So I removed that and carefully cleaned socket and chip again, without any change.
I also tried multiple cards for the passthrough and a few other flat cables. No changes there either.

Now, if I'd had a decent oscilloscope, I'd check the output of the 126.5 and 50MHz crystals. But I don't so I didn't..

Any ideas as to what I could try?