TIP: Click on subject to list as thread! ANSI
echo: apple
to: comp.sys.apple2
from: Michael J. Mahon
date: 2009-02-05 12:41:56
subject: Re: Iomega Zip drive MFM or GCR?

Rick Balkins wrote:
> Vanessa, I think I got to that in my follow-up.
> 
> We are mainly talking about the Zip Disk that been around for awhile and may 
> have used an older scheme then the hard drives that came out in 1995 (IIRC 
> about when Zip Disks first came out).
> 
> Like I said in a follow up - it can be PRML or even EPRML. Modern HDs uses 
> this scheme and possibly a custom encoding schem on top of that. I agree 
> with Thomas as well, I think it can be a proprietary/custom encoding scheme 
> (or one of many encoding systems out there) running on top of the PRML/EPRML 
> recording system. As for the use of GCR.... It be suprising to see being 
> used today as it is primarily a "Retro" thing, nowadays.

It may help to add some perspective to point out that Apple GCR is
actually a scheme to increase capacity by using RLL coding principles
on top of a byte-oriented drive interface.  (In doing so we must
distinguish between RLL as a coding technology and "RLL" as a specific
case of RLL used in certain disk recording standards.)

The "standard" single-density floppy encoding was essentially
"4+4"
encoding, with 1's interleaved between each data bit.  This is a
simple RZ (Return to Zero) coding, except that it is "return to one".

Woz's initial GCR approach encoded 5 data bits per disk byte (nibble),
for a 20% increase in track capacity over the standard coding.  The
later 6+2 encoding put 6 data bits per disk byte, for a 40% increase
in track density (and, worth noting, a 40% increase in data transfer
rate).

The greater coding efficiencies of Apple GCR were possible because the
floppy read electronics could reliably read codes in which runs of
consecutive zeros were limited to two per disk byte--a byte-oriented
RLL variant.

It's worth noting that since the high bit of each disk byte is one,
the longest self-sync pattern that can be reliably read is 10 bits
(assuming that the low-order bit is one) since the 2 inter-byte bits
are zero, the run-length limit.

-michael

******** Note new website URL ********

NadaNet and AppleCrate II for Apple II parallel computing!
Home page:  http://home.comcast.net/~mjmahon/

"The wastebasket is our most important design
tool--and it's seriously underused."
--- SBBSecho 2.12-Win32
* Origin: Derby City Gateway (1:2320/0)
SEEN-BY: 10/1 3 34/999 120/228 123/500 128/2 140/1 222/2 226/0 236/150 249/303
SEEN-BY: 250/306 261/20 38 100 1404 1406 1410 1418 266/1413 280/1027 320/119
SEEN-BY: 393/11 396/45 633/260 267 712/848 800/432 801/161 189 2222/700
SEEN-BY: 2320/100 105 200 2905/0
@PATH: 2320/0 100 261/38 633/260 267

SOURCE: echomail via fidonet.ozzmosis.com

Email questions or comments to sysop@ipingthereforeiam.com
All parts of this website painstakingly hand-crafted in the U.S.A.!
IPTIA BBS/MUD/Terminal/Game Server List, © 2025 IPTIA Consulting™.