On 2023-May-11, at 7:41 PM, Cameron Kaiser via cctalk wrote: > it's actually an artifact of the monitor that the upper 6 were clear. > Actually, > the stuck bit is entirely bit 2 (i.e., it goes > > 0 1 2 3 4 5 6 7 8 9 a b c d e f > 0 1 2 3 0 1 2 3 8 9 a b 8 9 a b > > and the high nybble is OK). Now that sounds more like a bad RAM chip, but why > would it be *just* those addresses? Does that sound like a plausible failure > mode?
So I take it the KIM1 uses 6102 1K*1 RAM chips. (I'm seeing some modern redrawings, but is there no original schematic online?) Any RAM chip internally has the bits organized in a matrix. I haven't found a proper 6102 datasheet, but the most likely array size is obviously 32 * 32. Your bad address range of xx80::xx9F is a span of 32 on base-32 boundaries. So a failed internal driver or sense-amp for one row or column of one chip would produce your fault. That's a pretty plausible failure mode. Don't know how much luck you'll have finding a 6102. Cursorily it looks like it's just a CMOS version of the 2102, and pin-compatible, so that might be an alternative. The 2102 comes in low-power flavors (often but not always called 21L02). It also comes in various speeds so you might have to watch that, something like a 21L02-3 might do.
