kingcake wrote on 2025-08-12, 22:13:
So if I write 0xFF to a location in the first 64K with debug I'd see it repeated 3 more times at 64K boundaries, right?
Yeah. This is the first 64K of a bank, no the first 64K of the system of course. Furthermore, this is true for any 64K block of a bank, not just the first one. So assuming the card supports and you found the configuration for 256K + 64K + 64K with a start address of 256K (20000 hex), and further assuming that the 256K bank is first:
- Memory access to 0000:0000 to 3000:FFFF is handled by the mainboard, and works as expected
- Memory writes to 4000:xxxx, 5000:xxxx, 6000:xxxx or 7000:xxxx are visible at all of these locations
- Memory access to 8000:0000 to 9000:0000 is handled by the two 64K banks on the card and works normally
rmay635703 wrote on 2025-08-13, 03:29:
January 1987 device, so very likely a cost reduced replacement for an ast six pack
The AST Six Pak series of cards are multi-function cards (although I have trouble counting six functions). The main selling point of the Six Pak series of cards was that these cards conserve slots (remember the 5150 only had 5 slots, and the original mainboard only had 64KB of RAM onboard). This card on the other hand is a memory card only, which there also were plenty of back in the 80's. Nevertheless, you are correct that this card is cost reduced, and the memory expansion function (not expanded memory!) of this card works the same way as the memory expansion portion of the AST Six Pak cards. Due to their age, cards like the Six Pak Plus support 64K x 1 chips only, and thus require 6 banks for the 384K memory size required to bring a 256K mainboard to 640K. Being able to use 256K chips (which is just an educated guess) is a key point to make the card smaller, because with those chips 3 banks suffice.