Subject: Re: VAXstation 2000 Progress --- MFM controllers?
To: Allison J Parent <allisonp@world.std.com>
From: Tom Ivar Helbekkmo <tih@KPNQwest.NO>
List: port-vax
Date: 02/18/2000 08:30:25
Allison J Parent <allisonp@world.std.com> writes:

> Ah, ST506 is a Shugart 5mb drive(6mb unformatted), ST412 is a
> Shugart 10mb (guess what 12mb unformatted) drive that also has the
> same interface as a ST506.

...and so do an awful lot of other disks and controllers, which is
why, when we say ST506 these days, we tend to mean the interface, not
the original disks.  ST506 is an electrical interface and its logic.
MFM and RLL are two different ways of encoding information into the
magnetic surface of the disks.  Both generate a magnetization pattern
that encodes a sequence of bits, of course, but RLL is more efficient.
The encoding (and decoding) is performed by the controller, so there 
is no electrical or functional difference in the interface or disk.

The difference is, basically, this: magnetic encoding won't work
unless it changes frequently along the track (so that you get an AC
signal when reading it), and while MFM handles this by making sure
that there's a change for every bit (which is slightly wasteful, but
necessary to protect against cases where too many bits are the same),
RLL (run-length limited) has a smarter algorithm that allows a few
bits to be the same without an magnetization change.  This lets you
pack the bits more closely, but it's slightly more demanding of the
quality of the disk.  Thus, you got disks like the Seagate ST-225, a
20MB MFM disk, that had an RLL counterpart (the ST-231, IIRC) with
about 50% more capacity -- but it was the exact same disk assembly,
the only difference being that one was certified by the manufacturer
for RLL use.  Standard PC formatting was 17 sectors per track for MFM,
26 for RLL.

Using nominally MFM disks with RLL controllers was usually no problem.

-tih
-- 
Popularity is the hallmark of mediocrity.  --Niles Crane, "Frasier"