On 6/30/05, Jules Richardson <julesrichardsonuk at yahoo.co.uk> wrote:
Most classic SASI controllers seem to be little more
than a handful of
buffer and latch ICs though (only about 6 chips total)
There's a simple SASI implementation buried in the Commodore
D9060/D9090 disk drives - I've done a lot of digging around those
guys, but have yet to break down and completely reverse-engineer the
source, so I am sketching on deep details, but essentially, the D90x0
has a "DOS board" with one 650x CPU talking IEEE-488, the other
talking SASI to a Tandon(?) SASI<->ST506 board, then a Tandon TM603S
or TM602S drive mech.
We've done some playing around, and it's possible to, say, stick an
ST225 in there and get it to format as if it were a TM602S (4 heads,
153 cyls of 32 sectors of 256 bytes). I am now getting hazy on if
this works (I didn't do it myself), but it might be trivial to patch
the firmware to address cylinders up to 255. My biggest concern with
those drives is, of couse, the unobtanium nature of 5MB and 7.5MB
ST506 drive mechs. I've always wanted to enhance the firmware to a)
address disks as large as possible (16MB?) and b) to use a more
available medium than ST506 drives. So if anyone learns more about
how SASI interoperates with ancient SCSI drives, I'd love to hear
about it.
-ethan