[comp.sys.ibm.pc] Problems with ST4096 and Original IBM-AT controller

zielke@romeo.cs.duke.edu (David Zielke c/o hsg) (06/28/90)

  I have just purchased a Seagate ST4096 hard disk (9head,1023 cyl, MFM)
and have spent a significant amount of time trying to format the disk.  
The Seagate Diag program performs the seek test without a problem but
I cannot get the drive to low level format with either DM from seagate
or the IBM diagnostics disk.  (I even tried formatting to a smaller drive
with no avail).  The error is this: Write Fault.  I have called Seagate
and they checked the position of all of the jumpers on the board (on the
disk) and they check out.  Including the Write Fault jumper.  

  The controller works fine with the Rodime 20meg drive which came with
the system.  I even tried formatting a track to see if there was some
problem and it worked fine.

  Seagate suggests replacing the controller.  Is there a reason for this, 
a simple answer (please not `42').  

  Is the 'Write Fault' line interpreted in the drive or on the controller?  
is it possible to do a minor hardware mod to get this beast to work or will
I have to put out $$ for a new controller?

Thanks for your time and attention!

David Zielke
zielke@cs.duke.edu
zielke@phy.duke.edu

________________________________________________________________________
|  Name:		David M. Zielke					|
|  Internet:		zielke@physics.phy.duke.edu			|
|			zielke@cs.duke.edu				|

grege@gold.GVG.TEK.COM (Greg Ebert) (06/29/90)

In article <20539@duke.cs.duke.edu> zielke@romeo.cs.duke.edu (David Zielke c/o hsg) writes:
>
>
>  Is the 'Write Fault' line interpreted in the drive or on the controller?  
>is it possible to do a minor hardware mod to get this beast to work or will
>I have to put out $$ for a new controller?
>

It comes from the drive, indicating an attempt to write under unsafe
conditions, usually an illegal head address or incorrect spindle speed.

It's not difficult to write a low-level format routine. The trick is to
find the technical information. I refer to the 'CP342 Product Specification'
from Conner Peripherals, which has everything you need to know about AT hard
disk controllers. Although I would love to provide copies of the spec for
anyone who asked for it, I don't relish the possibility of being thrown in
jail for copyright violation. If there are enough interested parties, I'll
post the gory details of how to talk directly to the controller. I'm
working on a way to pop it into my S-100 system.

---------------------------------------------------------------------------
#include <stdflame.ibm-pc>
#ifdef FLAME

An now for a flame. I am really angry at the way the PC industry is run by
TECHO-MORONS. Gone are the days when you bought a computer and got all the
schematics, programming details, data sheets, etc. I've been a hardware
nut for 12 years, and I wouldn't trade my 2.5 Mhz Z80 for a 486 if my life
depended upon it.

#endif

grege@gold.GVG.TEK.COM (Greg Ebert) (06/30/90)

There's some good info on hard disk controllers from Intel (82064) and
Western Digital (WD2010 and WD1010). These data sheets most of the nitty
gritty details on how to do most of the programming. I have seen WD1010's
in PC's.

The I/O addresses are 1F0-1F7 for the controller chip, 3F6 is a 'non-invasive'
read-only address for 1F7, and 3F7 bits 0-6 is read-only and gives head/drive
select and write-gate info. You can ignore 3F7.