You need a TL-866 II or similar to test these chips if you're going down this path or it will be a world of frustration.

There are IS61C1024-10N (32-pin, 128kx8) chips on ebay, possibly counterfeit, but for many of us they more or less work. But about 10-15% of whatever you buy will be so defective that the TL-866 II immediately flags them.

Unlike the 32-pin chips, the 28-pins chips are still made or were until very recently, but the biggest you can get in those is 32kx8

Would help to see the cache area of the board, too, to make sure. The above is for a single-bank config (four 32-pin sockets and four 28-pin sockets). If you have eight 32-pin sockets it opens up additional possibilities.

ive added pictures of the cache and tag chips and area

Nice, you can go for 1MB with the chips I mentioned above. Otherwise there are some 64kx8 ones out there labeled as UMC, 15 ns, that might more likely be real than the IS61C1024 ones.

The downside of your board is the CMD640 IDE I think, and no PS/2 mouse, as compared to later revision UM8881 chipsets with the UMC IDE (and slightly later PCI revision support for newer cards). Doesn't matter if you're using SCSI, or an ATA or SATA PCI card.

jakethompson1 wrote on 2024-10-31, 18:52:

Nice, you can go for 1MB with the chips I mentioned above. Otherwise there are some 64kx8 ones out there labeled as UMC, 15 ns, that might more likely be real than the IS61C1024 ones.

The downside of your board is the CMD640 IDE I think, and no PS/2 mouse, as compared to later revision UM8881 chipsets with the UMC IDE (and slightly later PCI revision support for newer cards). Doesn't matter if you're using SCSI, or an ATA or SATA PCI card.

My board is up8810 vio rev 1.0, its an Amstrad PC9486i which ive upgraded to 32mb ram, sound blaster 16, nic, cf to ide and replaced the cyrix dx2/66 to an intel dx4/100