Date : Tue, 08 May 1990 22:50:35 GMT
From : mcsun!unido!balu!tilmann%cosmo.UUCP@uunet.uu.net (Tilmann Reh)
Subject: floppy drive hardware
Hello.
> Can anyone shed light on this?
I hope so. Generally, the 3.5" and 5.25" drives are IDENTICAL, concerning
their electrical and logical details. There is NO WAY to distinguish a 3.5"
drive from a 5.25" drive by software. As far as I know, the pinout is also
identical (however, the 'AT compatible' drives have some special pinout, for
both 3.5" and 5.25"). But, most 3.5" drives use pin connectors instead of
the card-edge 5.25" types.
The different capacities are the result of different data rates. The 720k
formats (both 3.5" and 5.25") are obtained using a data rate of 250 kbit/s
in MFM mode, so one track will contain 9 512-byte sectors per side.
The 1.44M format (only 3.5", as 5.25" drives for that data rate aren't
manufactured) uses a data rate of 500 kbit/s (like 8" drives) with MFM, so
18 such sectors fit onto one track.
The 1.2M format (5.25") is strictly an 8" format: the 5.25" HD drives are
rotating with 360 rpm, so that they behave exactly like an 8" DS DD drive.
Again, there is NO WAY to distinguish by software...
When using different capacities (and/or) different drives, always be sure
to use disks and drives only within their specifications! The HD disks
(5.25" 1.2M and 3.5" 1.44M) use stronger magnetic fields than the normal
MFM disks, so you can't crossover HD disks with MFM formats and vice versa.
However, if you try to format a normal disks for HD, it may be permanently
damaged (with resident magnetism caused by the strong fields).
Besides, always be aware that the capacities 720k and 1.44M (as well as 360k)
do NOT use all legal resources of disk and drive. It's just that IBM (who else)
wasted about 10 / 18 percent of the possible capacity FOR NO REASON!
Formatting MFM disks with 10 x 512 byte or, even better, 5 x 1024 byte, would
give 400k (40 track) or 800k (80 track) net capacity, and all that within
all specifications of the IBM 3740 disk format.
Formatting HD disks with 20 x 512 byte would give 1.6M, with 11 x 1024 byte
1.76M (that is 22% more than IBM!), also legal.
Again, one example how performance is WASTED by a company of which still
some people think they were able to design computers...
Tilmann