Re: MFM drive gone nuts

From: MikeS <dm561_at_torfree.net>
Date: Mon, 21 Jul 2014 11:50:42 -0400
Message-ID: <5CD63B137D8E4258A7CD39A4055C8CF6@310e2>
----- Original Message ----- 
From: "Clockmeister" <clockmeister@internode.on.net>
To: <cbm-hackers@musoftware.de>
Sent: Monday, July 21, 2014 5:35 AM
Subject: Re: MFM drive gone nuts


>
> I always thought of the PC way of drive selection by a simple cable twist
> was a very good solution.
>
> All you had to do was buy a drive you intended as a second drive (which
> were factory set for drive A:) and plug it into spare connector and hey
> presto, B: drive.

----- Reply: -----

Exactly! (But they were actually factory set for drive B:; the twist
effectively converted B: to A:)

As a matter of fact by the time 3.25" HD drives came along, many drives no
longer had any jumpers or switches at all and higher pull-up resistors had
also removed the termination location issue (which had not really been much
of an issue in practice anyway).

But the real advantage and probably the reason that IBM went to the trouble
was that it allowed individual control of the floppy drive motors.

Most of the CP/M and other systems that preceded the IBM PC connected the
floppy drives the way SD prefers, i.e. the Shugart standard using a straight
cable with up to four drives selected with DS0-3 (or DS1-4) jumpers. Since
there was only one Motor On signal connected in parallel to all four drives
and it was *not* gated with the drive select on drives like the popular
Tandon TM100, activating any drive would turn on all four drive motors; some
of the 8" drives of the day didn't even bother with a motor on signal, using
110/240VAC motors that ran constantly.

The reason for this was presumably to avoid the latency of getting the drive
up to speed every time you selected it, especially with the relatively
sluggish motor/belt drives of the day; imagine a disk copy where you had to
wait 10 times as long for the drive to come up to speed as it took to write
the data every time you switched drives.

To minimize needless diskette wear some drives did have head load solenoids
which only lowered the head(s) when the drive was selected but there was
still the friction of diskettes needlessly spinning in their envelopes, not
to mention the extra power which became even more of an issue when systems
went from the arc-weld-capable linear supplies of the CP/M and UNIX systems
to those early inefficient switching supplies starting to appear in Apples,
the IBM PC etc.

IBM's introduction of the twisted cable and splitting the interface into
separate A/B and C/D cables made it possible to turn each of the four
possible drive motors on and off independently, largely avoiding the
start-up latency by turning on early and delaying turn-off.

SD is certainly entitled to prefer fiddling with jumpers and needlessly
spinning unselected diskettes, but at the time most folks thought individual
motor control and jumperless installation were a good idea, not at all
"stupid" or "utterly sick"; since few people made their own cables, whether
it had a twist or not wasn't usually an issue at all.

Maybe cutting his fingers on cheap cases and being unable to find a way to
route cables is the reason for SD's passionate disdain, but I don't think
that's really the fault of the twist... ;-)

And of course this is all largely irrelevant to the OP's hard disk issue,
since they generally spin constantly just like the 8" floppy drives of 
yore...

m


       Message was sent through the cbm-hackers mailing list
Received on 2014-07-21 16:00:02

Archive generated by hypermail 2.2.0.