Unfortunately, the second drive still did not work, and returned “Drive Not Ready”
errors via the BIOS, although it did respond to stepping command issued via a
monitor (the excellent public domain Z8E.COM). By this time there were no more
drives available, but Scorpio very kindly suggested I keep the dud for spares and
accept a partial refund. This I did, but was able to send the refund back, as the
problem was mechanical and (relatively!) easily solved. The main stepper motor
operates a lead screw which moves the head assembly via a “nut” (which is biased
with a set of small springs to keep the head assembly on its guide rails) and an
intermediate cupped washer. This washer had become displaced, and as a result
the track 0 position of the stepper motor left the heads several track widths outside
the genuine track 0.
Replacement of this washer brought the second drive into action, and I was off. Or
so I thought! Both drives could now read and write reliably, and repeated
verifications of the disks were successful. However, one of my database applications
involves an alphabetic printout of a large random access file, and during this the
BIOS would crash out with a “Drive not Ready” error. Numerous fiddles with head
settling time etc. failed to cure the fault, which tended to occur at the same pages
of the printout each time, but occasionally did not happen at all. I also tried all sorts
of extra decoupling of the power supply to no avail.
Eventually the penny dropped! During the printout, there were times when the head
was unloaded after a period without disk activity, and then immediately loaded
again as a new sector was requested. If the head-load/door-lock solenoid had not
returned to its de-energised position before being re-energised, it could not quite
reload the head. The Drivetec derives the index pulse from pre-formatted data on
the disk, not from the index hole; the drive motor time-out period elapsed without
an index pulse being received and the drive appeared to be empty.
The solution was to test the head-load status before any disk read or write was
performed, and if the head was unloaded to wait for a further 100 mS before
reloading it. This is a rare event, and has no significant effect on performance.
I am very pleased with the addition of these drives to my system, which now offers
the disk capacity of a small Winchester system but allows me to remove data files
completely for security and backup purposes. As far as I can tell, they are now
performing up to specification. I had anticipated some problems in obtaining (and
paying for!) disks, but these are readily available and cost £18 for 5 plus VAT, which
is not a lot when their capacity is taken into consideration.