I think I killed an old option rom from careless handling. We get
spoiled with new parts being pretty robust, but the old parts really are
sensitive to static and careless handling. At least, I think that's what
happened to this chip.
I have an option rom called DISK+ 2.0 from PCSG.
It's you're standard ceramic UV 27C256 on a pinout-adapter pcb.
I got this a couple years ago inside a machine. I dumped it at the time
and I have that image.
Just the other day I was working on a little programming adapter to read
option roms more conveniently in a programmer with an actual socket,
without having to disassemble the module or unbend pins or use a dip28
test clip and 28 wires etc.
I was getting bad reads of this one rom, that I *think* can be explained
if the A4 pin was stuck low. The output looked like every other 0x10
bytes was just a duplicate of the previous 0x10 bytes.
00000000 c3 d1 00 50 43 53 47 00 22 d8 fa e1 d5 c3 4a 00
|...PCSG.".....J.|
00000010 c3 d1 00 50 43 53 47 00 22 d8 fa e1 d5 c3 4a 00
|...PCSG.".....J.|
00000020 cf ab 4b c9 cf 02 f6 c9 85 6f d0 24 c9 30 30 30
|..K......o.$.000|
00000030 cf ab 4b c9 cf 02 f6 c9 85 6f d0 24 c9 30 30 30
|..K......o.$.000|
00000040 ff ff 44 49 53 4b 2b 20 20 00 5e 23 56 23 e3 eb |..DISK+
.^#V#..|
00000050 ff ff 44 49 53 4b 2b 20 20 00 5e 23 56 23 e3 eb |..DISK+
.^#V#..|
00000060 e3 e5 21 ac 6d e3 c3 12 f6 af 3d c9 e5 21 1c f6
|..!.m.....=..!..|
00000070 e3 e5 21 ac 6d e3 c3 12 f6 af 3d c9 e5 21 1c f6
|..!.m.....=..!..|
...
My previously saved dump of that same chip does not look like that.
I decided to desolder the chip from it's little pcb and read the chip
directly instead of going through two pinout adapters, and now it's even
worse, now the first 16K is all FF and the rest has that same stuck A4
appearance like this example, starting right at hex 0x4000
Also this wrecked the pcb. The castellated edge contacts just
practically fell off. No putting a new chip in.
Also along the way I notice that both PCSG roms I have, DISK+ 2.0 and
RAM+ 2.5, the software is actually less than 8K, and the rom is filled
with 4 identical concatenated copies of the same 8K image. I split them
out with dd, and compared them to each other with cmp, and each copy is
identical.
So if my DISK+ chip had failed in a different way, merely killing or
corrupting a range of addresses, instead of a stuck line that affects
the entire address range, I would still have up to 3 other entire copies
of it that could have been extracted from the same bad chip and burned
to a new chip.
I'm glad I dumped that rom right when I first got it because it's well
and truly junk now.
The RAM+ is the same as the one on Club100 romware page, but on club100
there is no info about it.
It's from PCSG and it came with a ram expansion that goes in a M100 bus
socket similar to the PG Design ones.
Both DISK+ and RAM+ have CALL 63012 printed on them.
--
bkw
