Did you try www.stason.org? 😊

I can't find it there 🙁

Maybe the memory is bad.

Is the problem with above memory chips (to fast) or with the card itself?

Memory chips will run at the speed the card tells them to run. They just need to be 70ns or better, and the correct type. The Tseng card probably wants FPM, and those 45ns chips are almost certainly EDO, so it doesn't surprise me that those don't work. The NEC ones should be fine though, if they're good.

I'd check the card thoroughly for damage, clean all the contacts and re-seat all the socketed chips, then continue messing with the jumpers and see if things straighten out. If you can't get anything to work, then it'd probably be a good time to start suspecting bad memory chips.

Just for an idea what you're dealing with, JP12 is probably for IRQ enable/disable, and the one up near the VGA connector may be for setting a VLB wait state. Given the proximity of JP8-11 to the ICS2494N, those probably control your memory and dot clock frequencies... rather than relying on trial and error you could check what pin of the chip each jumper connects to and cross reference it with the datasheet.

All memory chips are from working cards stated above. So memory chips by themselves are not the problem - unless the wrong type, but then I doubt the card would recognize 2MB and actually work in DOS txt mode. The artifacts are the same with 45ns and 60ns chips. With on-board 1MB the card works perfectly.

I'll look at the datasheet and try to figure out something... thanks

Indeed i doubt that ET4000 would work with EDO.

Cant you source any FPM SOJ 512KB chips on ebay or whatever?
You could use the ISA cards NEC ram, thats FPM.
Faster than 60ns is mostly of the time EDO ram.

As I already said... I used both cards ram to test it and both show the same artifacts.

i have two et4000w32p cards and one with an older bios also had some currupted display modes, this was fixed by changing over the bios eprom to my 1.32 rev.

maybe the firmware revision for your snafu card is causing the same.
Show us a picture of the pixels that are being harmed, we should do something against that, or maybe some electrical path is disconnected from ET4000 to and from the empty ram sockets.
Are there any cuts? seeing your close-up pic's i must say the solder joints are very minimal on that card.
You should resolder most connections possible, if i only had that card in my hands....

If the BIOS went bad, it could be due to bit rot. I have recently been made aware that EPROMs can't hold their contents forever, so it might be smart to dump the ROMs on all your working cards while you still can.

Especially if they are UV erasable and the sticker covering the window is starting to fall off from age - bought an MM-401 that had this 😳

I do not really believe in bit rot, i believe the eproms data would last a life time.
The exposure glass labels are also intact, but still both boot fine, it is only some graphic modes but not witl all games give some corruptec stripes with the older bios eprom.

I think the problem with this topic ET4000 is because of bad solder joints.

Or maybe we should make a VGA bios rom database for all vintage cards, i am up to save some of my data after xmas when building my next intel chipset based 486.

bjt wrote:

Especially if they are UV erasable and the sticker covering the window is starting to fall off from age - bought an MM-401 that had this 😳

It takes pretty intense UV light to erase those things, and at a specific spectrum that you don't get in significant amounts naturally or from standard household lighting.

I left a chip out in the Australian summer sun for a couple of weeks and it was still perfectly readable, before I bit the bullet and bought a UV EPROM eraser and did it properly. Even then it still takes about 40 minutes to fully erase the chip.

It's probably wise to keep the window on EPROM's covered up but it's not really a disaster if the label falls off, either.

From my understanding, bit rot has nothing to do with the sticker coming off the glass window and UV light getting inside. It's caused by the data losing electrical charge due to imperfect insulation. This will happen to every EPROM eventually, and it's only one of several ways a ROM can go bad. If you're not backing up your ROMs (especially those you can't easily download) you could be asking for trouble. We're dealing with hardware that is 20-30 years old or more in some cases.

Just an update on my forgotten thread...

After some searching I was able to find the correct memory chips. The card now works fine with 2MB installed. In DOS there isn't much improvement while in Windows it almost doubles the performance.

Hopefully I'll compare it to a S3 Trio32 2MB VLB in a couple of days. The other S3 Vision 968 4MB VLB I have is unfortunately dead.