Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!ut-sally!husc6!bloom-beacon!oberon!poisson.usc.edu!mlinar
From: mlinar@poisson.usc.edu (Mitch Mlinar)
Newsgroups: comp.os.minix
Subject: Re: xt hard disk driver
Message-ID: <4360@oberon.USC.EDU>
Date: Tue, 8-Sep-87 22:08:43 EDT
Article-I.D.: oberon.4360
Posted: Tue Sep  8 22:08:43 1987
Date-Received: Thu, 10-Sep-87 05:28:32 EDT
References: <1288@cognos.UUCP> <5890003@hplsdla.HP.COM>
Sender: nobody@oberon.USC.EDU
Reply-To: mlinar@poisson.usc.edu (Mitch Mlinar)
Organization: University of Southern California, Los Angeles, CA
Lines: 43

In article <5890003@hplsdla.HP.COM> ritchie@hplsdla.HP.COM (Dave Ritchie) writes:
>>/ allbery@ncoast.UUCP (Brandon Allbery) writes: 
>>As quoted from <1288@cognos.UUCP> by brianc@cognos.uucp (Brian Campbell):
>>+---------------
>>|    Has anyone else had any problems with the 1.2 posting of xt_wini.c?
>>| It now works for my hard drive (Seagate 225 with WD controller), but I
>>| had to swap the drive type calculations in init_params().  Is the
>>| posting wrong, or are my jumpers somehow backwards?
>>+---------------
>>
>>I had this happen also.  I don't know what having write precomp turn on too
>>early does, but check for that as well; I discovered that it was using a
>>value of 128, while my ST225 data sheet says 256.  (This is the starting
>>cylinder for write precomp.)  I added a line to change it in the wini.c
>>I compiled into Minix.
>
>   Try again.... write precomp starts at cylinder 300, not at 256, according
>to the ST225 manual I have for the ST225 on my Ampro.


All those numbers regarding write precomp are also fairly meaningless.  As
a site tester for well over 80 different types of hard drives (whew!), I can
state with some certainty that write precomp on MODERN drives (>1983) has
little effect on the drive data read/write capability.  During tests, the
write precomp was moved from cyl 0 all the way to the last cyl.  What
we found was that write precomp within 50% of the quoted nominal value has
no effect on data integrity.  

When in doubt, set your write precomp to 128 (which is true in all but a
few drives).  Regardless of what the manufacturer said, this value always
worked, although if you know the suggested value, use it.

Regarding the ST225, setting at 128 is at the envelope edge, but still ok.
The difference between 300 and 256 is not worth mentioning.  HOWEVER, what IS
worth mentioning is that once you pick a value and FORMAT your drive that way,
STAY with that value.  The soft error with time increases very slightly if
you moved outside the FORMAT setting by 30%.  If memory serves correctly,
the IBM rounds off to the nearest power of 2 (e.g. either 128 or 256) in its
tables.  (Some controllers round off to the nearest 64 or 128.)

Hope this helps somebody.

-Mitch