Xref: utzoo comp.sys.ibm.pc.hardware:9243 comp.sys.ibm.pc.misc:9982 comp.os.msdos.apps:2013
Newsgroups: comp.sys.ibm.pc.hardware,comp.sys.ibm.pc.misc,comp.os.msdos.apps
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!uupsi!rodan.acs.syr.edu!amichiel
From: amichiel@rodan.acs.syr.edu (Allen J Michielsen)
Subject: Re: Low Level HD Formatting Software
Message-ID: <1991May31.182153.11743@rodan.acs.syr.edu>
Organization: Syracuse University, Syracuse, NY
References: <1991May28.021833.5303@cunixf.cc.columbia.edu> <1991May28.133950.14789@cbfsb.att.com>
Date: Fri, 31 May 91 18:21:53 GMT

<1991May28.133950.14789@cbfsb.> mbb@cbnewsb.cb.att.com (martin.brilliant)
>From<1991May28.021833.5303@>, by stone@cunixb.cc.columbia.edu (Glenn Stone):
>> periodic low-level formatting to freshen up the sector markers, he 
>> claims that fading sector marks aren't really the cause of mis-reads;
>> misaligned heads are. (In reference to a magazine article he was quoting)
>Exactly.  I thought the reason for low-level formatting an aging drive
>was to realign the tracks with the misaligned heads, ...

   I tend to agree.  When you take a older hard drive and move it on a axis
it tends to 'lose' format and this should only be due to this problem.  The
drive mfg's say that they are now designing drives with more/any/all axial
locationing designed into them. (meaning you can stand it on end or move it
between on end on on side without any problems)  The NEW design drives, I
have used, seem to support this claim, BETTER than older designed drives.
   I have read in several sources that the actual problem requiring LLF is
most commonly due to data creep.  This is where the controller/drive actuall
wipe out the LLF marks due to miscommunication between the 2 of them, and
due to HOW the system determines WHEN to start writing data bits.  Actually 
the case is that writing data is rather almost done on a by guess by golly
basis, and sometimes the assumptions aren't perfect.  This type of problem is
nonexistant on ide, scsi, & esdi drives due to fundamental design differences
in HOW the controller works  (after all all 3 of these 'standards' were made
AFTER the st506 was well written in stone and has millions of hours of field
tests by users).  Actually IDE is only a st506 mfm or rll drive with much more
of the smarts of the controller included in/on the drive, and much more like
a data/address buffer i/o card for a controller.  However, this means that the
real 'controller' part is not just a standard generic interface, but hopefully
one selected/designed to be perfectly mated to the drive.  It also means that 
the cable lenght and corrected capicitance for signals is known with absolute
certainty with much smaller error rates/lengths.
   However, saying that new designs have eliminated future, end of product
alignment problems AND redesign to reduce/eliminate 'bit migration' is no
excuse for NOT providing the end user with the ABILITY to LLF to their hearts
content.  Don't worry, these drives CAN be formatted by the end user with
almost every available (compatible) controller.  The software drivers are
missing, and in time will become available to everybody.  Don't worry, because
you won't need them for several years probably anyway.  The lucky few that
hopped on at the begining will be the ones to pave the way, and make the
drive mfg's sorry for their mistake.  
al
 
-- 
Al. Michielsen, Mechanical & Aerospace Engineering, Syracuse University
 InterNet: amichiel@rodan.acs.syr.edu  amichiel@sunrise.acs.syr.edu
 Bitnet: AMICHIEL@SUNRISE