Re: FDD speed and capacity

[mjmahon@aol.com (Michael J. Mahon)]

Bryan Parkoff replied:

>   It has nothing to do with the CPU to manipulate disk drive for the
>speed.  It is only hardware that is built in disk controller card to
>manipulate the head, arm and the motor stepper.

Actually, on the Apple it is the software that determines the head
positioning and its timing.

RWTS actually controls the head position stepper directly, and by
doing so, it is able to move the head faster than Shugart's hardware
head controller (not present on Apple drives).  It does this by using
software-implemented acceleration/deceleration timing variations
so that the head can actually move across the disk faster than it
could moving at a constant speed.  (Steppers "slip" if you try to
step them too fast--inertia and all that.)

Interleave is another topic--related to disk throughput under certain
assumptions.  The ideal for minimizing the effects of rotational
latency would be a whole-track buffer.  Interleave attempts to
get some of the speed advantage without additional buffering,
by placing the sector most likely to be read next in a postion
on the track where the head will soon encounter it at the moment
it is needed.  Not too hard for large sequential block reads, but
very difficult in any more general scenario.  Of course, most disk
read time on an Apple _is_ for sequential block reads.

-michael

 Email:  mjmahon@aol.com
 Home page:  http://members.aol.com/MJMahon/