Path: utzoo!attcan!uunet!tank!ux1.cso.uiuc.edu!uwm.edu!zaphod.mps.ohio-state.edu!usc!elroy.jpl.nasa.gov!jato!mars!kaleb
From: kaleb@mars.jpl.nasa.gov (Kaleb Keithley)
Newsgroups: comp.sys.ibm.pc
Subject: Here we go again: was Re: MFM on RLL?
Keywords: read the Monthly Frequently Asked Questions
Message-ID: <3087@jato.Jpl.Nasa.Gov>
Date: 14 Mar 90 16:40:45 GMT
References: <2881@uwm.edu> <SORC.90Mar13211725@carina.unm.edu>
Sender: news@jato.Jpl.Nasa.Gov
Reply-To: kaleb@mars.UUCP (Kaleb Keithley)
Distribution: comp
Organization: Jet Propulsion Laboratory, Pasadena, CA.
Lines: 50

In article sorc@carina.unm.edu (Paul Caskey) writes:
>paravia@csd4.csd.uwm.edu (Mark David Kakatsch) said:
>
>Mark> Can an MFM drive be connected to an RLL controller? I can get a
>Mark> 40Mbyte MFM drive for free, but I've got a '386 w/ an 65Mbyte
>Mark> RLL. (Gateway). I know that MFM is pretty slow for a 386, but
>Mark> since it would be free, I wouldn't mind...Thanks much.
>
>Very simply, the answer is: "No, not reliably."  IOW, go for it. :-)
>RLL is tough on drives.  If you have a drive that's made for MFM, but
>it's really tough, it may handle RLL.  For instance, my roommate
>recently got an old DEC Rainbow with a 10 meg MFM drive.  That
>computer and the drive in it had been sitting in a snow bank for a
>week before he bought it.  I have a feeling it could probably handle
>RLL.
>
>However, for the average PC drive, chances are that if it was built
>tough enough to be an RLL, it *would* be an RLL.  Therefore I suspect
>that forcing an MFM to do RLL is pretty risky.
>

RLL is *NOT* tougher on a drive than MFM.  RLL uses a different encoding
scheme to achieve a higher *data* density, however the physical bit density
on the platters is exactly the same.  Very old drives that don't hold tight 
timing tolerences (like older Seagate ST-251s) are more apt to fail at RLL 
because RLL relies on hardware timing to work, whereas MFM is "self clocking" 
and doesn't need the precision timing to achieve reliability.

The best advice, is try it and see.  Perhaps it's worth paying a little extra
for an RLL controller with the right to bring it back and get an MFM controller
if RLL doesn't work.  I have first hand experience with RLLing Seagate ST-225s,
newer Seagate ST-251s, and Micropolis 1325s.  I once had an older ST-251 that
just would not RLL.  My experience with the ST-251s would lead me to believe
that U.S. made ST-251s won't RLL, while the oriental ST-251s will.  This is
an *extreme* generalization, based on a sample of two, so probabilities are
just not very good that my generalization is a fair one.  Others have reported
success with other makes and models.

To address the original comment about speed, MFM is only slower than RLL
(at 1:1 interleave) because for a given revolution, 17 sectors go under the
head compared to 26 sectors on RLL.  1:1 RLL reaches xfer rates of about
750,000k(bits or bytes, I can't recall) per second.  1:1 MFM reach xfer rates
of about 500,000k.  My personal impression is that unless you are doing some
very disk intensive stuff, or have a 33mhz 386, or a 25mhz 486, you won't be
able to tell the difference between 1:1 MFM and 1:1 RLL.

kaleb@mars.jpl.nasa.gov            Jet Propeller Labs
Kaleb Keithley

spelling and grammar flames > /dev/null