stepping machanism of Apple Disk ][ drive (was Re: Heatkit 5

From: Richard Erlacher <edick_at_idcomm.com>
Date: Fri Apr 9 01:17:55 1999

I think there may be a semantic problem at hand. The single-density (FM)
and double-density (MFM) modulation schemes operate on precisely the same
density of flux reversals at the media/head boundary. The difference is
that the FM imposes a flux reversal between every bit window, while the MFM
does not. What this means is that data encoded in MFM can be written to the
media at twice the rate as that at which FM is written without taxing the
media or read/write channel ordinarily limited to FM. There were some
improvements between earlier and later head designs, and media selected for
MFM were required to be better because the bits had half as much material in
which to be recorded. With MFM, if a clock was missing, which happened half
the time a transition was lost, it didn't necessarily make a difference
unless the controller happened to be looking for an address mark.

So, if your definition of "double density" means twice as many flux
reversals per linear inch, well, you're right, I guess, but that's not what
the industry meant when the called it double density. I meant the same
thing the rest of the folks in the business meant, i.e twice as much data
capacity on the same size medium.

It's the same with the difference between MFM and RLL hard disk drives.
There were a few drives which, as a matter of course, didn't work with RLL
encoding. I don't know why this was. I do know that if you use any of a
number of translation schemes, of which ANSI GCR is one, you generate a bit
stream which, though it uses more than half the channel bandwith to do so,
can be recorded at twice the transfer rate as the corresponding NRZ data
without allowing excessive accumulation of charge on the heads as would
occur if too many ones or zeroes in succession were recorded in NRZ. These
schemes don't require the complicated time domain filters and other "neat"
circuits commonly used in read/write channels of that time, and provided
sufficient densities of transitions to allow clock recovery. The ANSI GCR
translates 4 input bits into 5 recorded bits, and recovers them, and ensures
that there are enough transitions to allow clock recovery yet no two
adjacent cells have transitions. (?) I suppose this is easily achievable
with a PROM, and some folks use a state machine to accomplish the same task.
I used a prom. It was easy enough to translate 4 bits received at 5 Mb/sec
into 10 bits at 8 Mb/sec. A number of code sets have been developed over
the years for the purpose of exploiting such "compression" over digitized
voice channels, and many other comm channels.

Like I said, the Perscis are hangar queens. Like a BMW . . . in the shop a
week a month. That's a hyperbole, of course, but it seems that way when
you've paid what the things cost back then only to have to pay that much a
year again to keep it running. They were fast, though, and hard disks cost
a lot more than the Persci drives.

Dick


-----Original Message-----
From: CLASSICCMP_at_trailing-edge.com <CLASSICCMP_at_trailing-edge.com>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Thursday, April 08, 1999 9:03 PM
Subject: Re: stepping machanism of Apple Disk ][ drive (was Re: Heatkit 5


>>with positioning. They now use voice-coil actuators rather than steppers,
>>and therefore can make quite subtle adjustments in head-stack position
>>depending on what is read. Back in the early days, that wasn't so.
>
>Actually, Persci floppy drives in 1976 or so were voice-coil (and quite
>a pain to maintain, even then - these days the glue that holds the
>optical graticules in place is often failing, and gluing and realigning
>from scratch is even harder, even with all the special Persci
>realignment jigs and electronic panels.) And a common modification
>to these drives (at least for folks like me who specialize in data
>recovery) is software-controlled offsets from the normal track positioning,
>something that does use the drive's ability to do fine positioning.
>
>>One interesting thing about the Apple GCR modulation format is that it
>>essentially was a "double-density" technique.
>
>Eric said the same thing, and I disagree with you both. To me (and all
>the tech pubs I've read) the density is how many flux transitions you can
>do per second (or revolution). GCR is a way of getting more real data with
>the same number of flux transitions. Apple GCR drives use single-density
>heads and single density data rates, a considerable cost saving factor
>in 1977.
>
>>cost plenty back then. This was at a time when Radio Shack still stayed
>>with single-density, and Apple exceeded their capacity easily.
>
>While using cheap single-density drives!
>
>--
> Tim Shoppa Email: shoppa_at_trailing-edge.com
> Trailing Edge Technology WWW: http://www.trailing-edge.com/
> 7328 Bradley Blvd Voice: 301-767-5917
> Bethesda, MD, USA 20817 Fax: 301-767-5927
Received on Fri Apr 09 1999 - 01:17:55 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:31:40 BST