29 Oct
2007
29 Oct
'07
3:30 p.m.
On 29 Oct 2007 at 14:42, Atsushi Takahashi wrote:
When reading, you could sense where the data starts
and ends but to
write to a new disk you'd have to make sure the motor speed was
constant before trying to generate fake sector hole signals. Maybe
some kind of PLL circuit. Any idea what jitter is allowed?
It would depend on the format, but there's always a run-in (sync) gap
on these things.
Here's what I've been thinking.
Time two or three revolutions of a disk after motor on and select
before generating the index pulse. If it's a soft-sector disk,
you'll see a real index pulse every 200 msec. or so and you can
subdivide the rotation accordingly and update your rotational rate
metric every revolution. If you see many index pulses, just pass them
through as you're handling a hard-sector diskette. You'd need to
look at drive select, index and perhaps ready as inputs and ready and
index as outputs.
At first blush, it appears as if a PIC 12F629 8-pin DIP might just be
able to do the job.
Cheers,
Chuck