EPSON QX-10 20MB Hard Disk
I've been looking for some DEC Rainbow 3rd party hard drives of late. QCS (Quality Computer Services) made an external hard disk for the DEC Rainbow. There's advertisements in Digital Review and other trade magazines at the time. It uses a SASI interface, and likely had a DEC Rainbow specific add-in card that they rewarmed from other designs...
One recently came up for sale on E-Bay. I thought I'd buy it to check it out. There was no interface card with it, alas. But it was a box with a WD1006 SASI to MFM controller in it that could handle two different drives. The drives were LUN0 and LUN1.
SASI, for those that don't know, pre-dates SCSI-1. It's kinda sorta SCSI-1 compatible, if you turn off parity, don't allow the drive to signal attention and restrict yourself to a subset of commands. It also doesn't have INQUIRY so you kinda have to know the size of the drive before hand. Most SASI controller drivers of the day wrote a label to drive with this information since it was always possible to read LBA0 w/o knowing anything else about the drive. Some controllers had ways to at least return a size, though that varied a lot...
Since SASI is kinda hard to interface to modern SCSI controllers, I used a MFM reader board I got from David Gesswein over at https://www.pdp8.net/mfm/mfm.shtml to read the drive. I had hoped to find that it was from an old Rainbow and I'd complete my collection of drivers for third party drives...
Much to my surprise, I was able to read it without any errors until it hit the manufacturing tracks (480-489). I pulled a full image, then downloaded it to my FreeBSD box for analysis.
hexdump -C told me it was a CP/M disk (I recognized the directory format). It was clear right away it wasn't a DEC Rainbow disk, however.
The first thing I noticed was the "Bi-Tech Multi Drive Support V4.02" string which indicated who made the driver for it. I also noticed strings like the following
PT.COM for EPSON QX-10 PeachText 5000 date changed - 02/03/84
and similar references to the QX-10 or EPSON CP/M.
So, this was from a Epson QX-10 CP/M system. Looks to be a soft-water service company from South Bend Indiana. All their books and correspondence from the mid 1980s was on it, along with some interesting disk support software. There's even some bits of Z80 assembler, but they are too disjointed to know what they were for.
I've not been able to get cpmtools to read the disk in a structured way, however, so it's hard to share just the interesting bits. Still working on it.
If you have one of these machines, or are interested in preserving software from it, please let me and we may be able to work something out.
1 comment:
I used to write CP/M BIOS code and do remember that the disk drivers usually encoded the sector interleave for all supported floppy disk and hard disk formats.
Some times tables were used, but very often there was an algorithm that converted from logical sector number to the sector position on the track. Often there was an extra offset for the first sector on a track to compensate for the time to step to the next track.
There have also been schemes that made a drive with two heads first step in using only head 0, then from the innermost track back out again using head 1.
Since CP/M used to allocate sectors in a strictly sequential logical number sequence, the logically last track used might allow to derive a lot of information about the sector layout, by detecting the pattern of sectors that still hold the data pattern left over from the initial formatting of the drive (generally with "E5" bytes for both floppy disks and hard disks).
An alternative method was to format the drives with non-sequential sector numbers that embedded the interleave into the sector layout of the tracks. But if that was the case, you'd probably have been able to read the data files off the drive.
Post a Comment