i found another card that uses 26pin memory ics.

http://www.amoretro.de/2012/09/trident-tvga90 … -768kb-isa.html

The part numbers exist but seem to be rare too.

DerBaum wrote on 2024-01-21, 07:23:

i found another card that uses 26pin memory ics.

http://www.amoretro.de/2012/09/trident-tvga90 … -768kb-isa.html

The part numbers exist but seem to be rare too.

Nice, I found a couple of similar video cards too, but the ram part numbers and brand were lost to history.

I’ll do some googling.

DerBaum wrote on 2024-01-21, 07:23:

i found another card that uses 26pin memory ics.

http://www.amoretro.de/2012/09/trident-tvga90 … -768kb-isa.html

The part numbers exist but seem to be rare too.

Darn it, I can’t find much about that one either! 😩

It seems all those UMC cards (both 85c405 and 85c408) use SOJ-26 RAM chips (look at the VGA Museum, they have the cards grouped by chipset). Many of them also include footprints for standard DIP-20 chips (extremely likely to be the usual 256K x 4 type). Finding a Trident card using the same SOJ-26 chips is even stronger evidence that these chips are 256k x 4.

Following the traces, the pinout is entirely different from the standard SOJ 26/20 RAM chips. The Trident card you linked has a quite instructive layout: there are 12 lines connecting all 3 banks, so they are most likely /RAS, 9 address lines, /WE and some line I just can't think of. Pin 20 is not common (might be /CAS), as are pins 3, 11, 16 and 24 (which might be the four data pins).You could likely cross-reference that with your card. I suspect that this kind of RAM chip is possibly a common local variant (Taiwanese?).

But, most importantly: I find a couple of pictures of UMC85c408AF based graphics cards that have only two chips, just like your card does. So it likely should work with two chips as well, although possibly at lower performance than originally intended.

Hmm, I wonder if I need to change a jumper on my board to get it working with less ram. I guess it’s possible mine didn’t work to start with and that is why ram was pinched from it.

Well that is interesting, I put the card into my Pentium II and nothing. So I removed the jumper at JP5 and it fired right up!

Hi
We have very similar cards. Mine has soj-26 memory and dip-20 sockets. You can check which legs of the dip-20 are connected to the soj-26 legs. Maybe you will be able to replace the missing soj-26 memory with popular dip-20 memory?

Mine has the footprints for dip sockets too, but I suspect I would need to fill in the bank with the SMD footprints first or the card won’t work.

If you want to try a newer bios, we were recently given this VLB one:

http://vogonsdrivers.com/getfile.php?fileid=2143

Would it work on ISA?

I’d be surprised if it was compatible, VLB is very different from plain ISA.

dJOS wrote on 2024-01-29, 04:05:

I’d be surprised if it was compatible, VLB is very different from plain ISA.

I don't know. BIOS claims to have support for UM85C4x8 chips and some of those early VLB video cards had jumpers to enable or disable VLB.

DerBaum wrote on 2024-01-21, 07:23:

i found another card that uses 26pin memory ics.

http://www.amoretro.de/2012/09/trident-tvga90 … -768kb-isa.html

The part numbers exist but seem to be rare too.

A recent thread got into the topic of those SOJ26 chips again, and in the mean time, I started to get quite confident, that these chips are B-class (or even worse) 256K x 4 chips, purposefully put into an unusual case to prevent these chips entering the "quality chip" market. I noticed that on the UMC 85C405AF and 85C408AF cards, there seems to be a pattern of having 3 413256J chips or 2 414256J chips, and a jumper field routing 8 data lines from the UMC VGA chip to 12 different positions in the RAM array. This looks like picking "working" or "nearly working" bits. In the 414256J, all 4 bits should be fine, in the 413256, only three of them. See Re: [SOLVED] First attempt at "let's make some 30-pin FP memories"... turned out to "let's not make stupid things again" for details.

The Trident card linked here has one 412256J and two 413256J chips per "row", making the extrapolation that in the 41256J only two bits of the four bits are working correctly. So while that card might have a RAM array worth 768K, only 512K of it is actually used.

mkarcher wrote on 2024-01-21, 08:37:

But, most importantly: I find a couple of pictures of UMC85c408AF based graphics cards that have only two chips, just like your card does. So it likely should work with two chips as well, although possibly at lower performance than originally intended.

Hmmmm hope that is the case for the UMC card I turned up the other week also, had two out of four chips removed, when someone figured it may as well only be a 256KB VGA I guess because 512 wouldn't seem to do much for it.

If it's that other problem with only the good half of the chips actually used, that will be "interesting" to figure out. I am not sure how motivated I will be to do so, but it's the very low profile board, and they do have a good CGA mode, so if I find myself trying to make a low profile 286 machine I might get more interested.

BitWrangler wrote on 2024-04-06, 18:21:

Hmmmm hope that is the case for the UMC card I turned up the other week also, had two out of four chips removed, when someone figured it may as well only be a 256KB VGA I guess because 512 wouldn't seem to do much for it.

If it's that other problem with only the good half of the chips actually used, that will be "interesting" to figure out. I am not sure how motivated I will be to do so, but it's the very low profile board, and they do have a good CGA mode, so if I find myself trying to make a low profile 286 machine I might get more interested.

I have no indication yet that there are cards floating around with 2 2-bit chips to substitute a fully working 4-bit chip, so if a card originally had 4 chips, the chips are most likely fully functional. If your card is of the type of 2 soldered chips and 2 sockets, the soldered chips are the first 8 bit, and the socketed chip make up the second 8 bit to fully populate the 16 bit data bus of the UMC graphics chip.

dJOS wrote on 2024-01-21, 04:40:

Anyhoo, the chips in question are IBM 23F9187's - but I cant find any info on them including compatible replacements. They also have an uncommon pin count - I've never seen a ram chip with 26 pins before.

Anyone have any ideas or better search skills that I do?

Hi! I'm new to this forum and I'm still learning English, but I'll try to help if I can.
I have one UMC 1MB VGA card very similar to yours, with eight memory chips. Maybe the memory chips installed on my card can be used to find replacements chips for your card.
Tomorrow I'll post some pictures of my card.

Cheers, some pics would be helpful.

Try these
http://www.o2xygen.com/parts/MM41425608.htm

card

Cheers, I’ll check my footprints against yours.

I have one UMC UM85C408AF 512 kB VRAM videocard:

As you can see, it's with four chips for total of 512 kB VRAM:

I can't find any technical info or data for these "A413256" (or "A414256") chips, but they're paired with 256 K x 4 Samsung KM44C256 chips, so mkarcher confirmed my suggestions that all four chips on my UMC card are 256 K x 4 (128 kB in size).

You can use Samsung KM44C256 (or similar DIP-20 256 K x 4) instead of A413256/A414256. I see that on your card there are four spare places for DIP-20 chips/sockets and maybe some jumpers have to be rearranged, if you use them.

I'm wondering if I can swap these four 256 K x 4 chips with four 512 K x 8 chips with some disabled data lines and used as 512 K x 4 for a total of 1 MB VRAM (that idea came to me from the six chips version of the card for which are used some "low-grade" chips with reduced number for data lines).

Also I have one Trident TVGA9000i-1 512 kB VRAM with four unknown "IBM 93 14" chips:

So these "IMB 93 14" chips are probably 256 K x 4 (128 kB in size).