Re: Apple 3.5" drive questions

[Eric Smith <eric-no-spam-for-me@brouhaha.com>]

I wrote:
> No, because the PC disk controller only supports MFM (and FM on some
> but not all PCs).

"GarberStreet Enterprises" <willy46pa@comcast.net> writes:
> Ok, first of all, we're not really clear on the difference between MFM and
> GCR.
> Let's go back to that period when hard drives were written in MFM, then RLL.
> Both MFM and RLL drives used the same controller, but different drivers.
> What everyone here is trying to say or not say as the case stands is, if the
> very
> same controller can write MFM, RLL, CHS, etc...., why not GCR. It's IMHO,
> easy to just say, NO IT CAN'T BE DONE, just because.

It can't be done, because the uPD765/8272 floppy controller chip in a PC
(or a descendent of that design inside a SuperIO chip or South Bridge)
has *exactly* two data encoding/decoding modes, and those are FM (single
density) and MFM (double density).  It doesn't have a GCR mode, and it
doesn't have a "roll-your-own-arbitrary-format" mode.  It can handle
IBM standard formats, and minor variations on those - e.g., different
sector sizes and counts.

The "other" family of FM/MFM floppy controller chips, of which the
Western Digital 1771 and 1793 are perhaps the most famous instances,
has some small chance of being able to read (but not write) Apple GCR
format.  But those aren't used in normal PC floppy controllers because
they aren't register-compatible with the uPD765/8272.

Your reference to RLL has nothing to do with PC floppy controllers; RLL
was used on some PC hard drive controllers, but those were entirely
distinct from the floppy controllers (aside from possibly being both
installed on the same circuit board).

CHS is a method of identifying the location of a sector on a drive,
and has nothing to do with whether the data in that sector is encoded
using FM, MFM, M2FM, RLL (of which there are many variants), GCR (which
is a form of RLL) or something else entirely.

> Peter Watson didn't
> "just say no" when we were told we could only READ ms-dos disks, he dug in
> there and solved that problem, now we can also WRITE to them.

That's because there wasn't any technical impediment to doing it other
than simply that no one had previously bothered to figure out why it
wasn't working.

> With a little ingenuity and persistence, anything can and will be
> accomplished.

False.  With a little ingenuity and persistence, *much* can and will
be accomplished, but reading and writing Apple GCR with a standard
PC floppy controller can't and won't.

Are there other ways to accomplish this task?  Yes.  But they require
use of hardware other than just a normal PC floppy controller.

If someone has some extra ingenuity and peristence that would otherwise
go to waste (unlikely), I'd really recommend using it to solve some
other problem rather than trying to make PC disk controllers speak
GCR.

Eric