15 Jan
2005
15 Jan
'05
6:31 a.m.
The early, original Qumetrak 142 drives have my
nomination for the second
to worst drive ever, coming in right after the early versions of the BASF
2/3 height drives.
The Qume drive was SO slow, that PC-DOS 2.10 had to extend the times to
prevent timeouts on normal disk I/O.
Shugart developed a much worse drive that thankfully never made it into
production. It was assembled on a bent-metal frame instead of a casting,
and instead of a stepper motor for head positioning, it used a mechanism
similar to an 8-track tape. Each step pulse advanced the head position
by one track, but only in one direction, and it was *incredibly* slow.
Once you got to the last track (I'm not sure if that was 34 or 39), the
next step pulse would return the carriage to track 0, which took even
longer.
This was intended to be a very low cost drive, but would have required
software changes since disk controller chips and normal software expect
to have bidirectional seek.
We played with a prototype for a few hours and decided that our
customers wouldn't want it.