24 Dec
2007
24 Dec
'07
10:33 a.m.
From: ard at p850ug1.demon.co.uk
The normal way to separate index form secotr pulses
was to have a
non-retriggerable monostable that was triggered from the index sensor and
which had a period a little less than the time between sector holes. It
was triggered at the end of each hole. Normally it timed out after each
sector hole, but was still set when the index hole came after the last
sectore hole. The output of said monostabel sent the output of the index
hole sesnor to either the 'sector' line or the 'index' line.
The period of the monostable is pretty flexible--anything longer than
half the time between sectors and less than the time between sectors
will work. If you're "back of the envelope-ing" it, you might shoot
for 75% of the time between sectors. This would allow a pot-less
design, as the component tolerances would be well within the "slop"
allowed in timing.
Some drives, particularly 3.5" ones, have an
adjustable monostable in the
index circuit which lets you delay the index pulse. It's adjusted when the
drive is aloighed to give the correct timing between the index pulse and
the data ffrom the ehad -- twiddlign the pot is easier than moving the
index sensor. Such a circuit could msess up what you're trying to do, but
this IBM drive doesn't have anything like that.
Most newer 3.5" drives subsume the "ready" circuit into a monolithic
hunk of silicon with the rest of the drive control. Index output
pulses are blocked until the period between indexes satisfies some
internal "ready" standard. Also, 3.5" drives can have *very* wide
index pulses when compared to 5.25" and 8" drives.
Of course, any 5.25" drive with this logic (and it's very common)
will view a hard-sector disk as not being anywhere near the correct
speed, so will not only not come ready, but will block *all* index
pulses going out.
This feature has a curious implication if you decide to replace an
older drive without the "block index until ready" logic with a newer
drive with that logic. Many controllers (such as the WD1770) or
software will count up a few sectors after select or motor on before
commencing an operation. The result is that everything still works,
but the latency after selecting a drive increases significantly.
Note that once the drive been selected and come ready, this is not an
issue.
For what it's worth,
Chuck