Defective IDE drive?? ESIX aborts on cylinder > 810

[1593]]

First, thanks for usenet and all the great information.

I'm having a problem loading ESIX.  My clone has Phoenix
bios (I forget the version, but it's for a 386sx), and a Toshiba
MK234FC-1 IDE drive.  When my dealer originally installed
the drive, he set it as Type 37, for 830 cylinders, 10
heads, size 68.  Since the drive is supposed to be about
100mb, I changed the type to 9, 15 heads, size 109.  The
drive booted under DOS, and I was able to format it for the
full 900 cylinders.

For curiosity's sake, I set it at type 41, 917 cyls, 15
heads, size 114.  Although I don't remember if I got it to
format under DOS to that size, the system did boot with no
drive type errors.  I think I partitioned it, I just don't
remember doing the format. (BTW, sec/track=17 in both
instances.)

In any case, when I try to partition the drive 25% to DOS
beginning on cylinder 0, and the remaining 75% to unix
(ESIX), ESIX does a surface analysys of its entire
partition.  When the count gets to about 810 cylinders or
so, the numbers fairly race to the end of the partition, but
ESIX reports no bad blocks.

Next, when the file system is being created, at the last
cylinder (anything over 810 or so), ESIX reports an error.
The error is that block xxxxxx, head 14, sec 16. Sector ID
not found.  (ST506 error)

Originally, ESIX would go to cylinder 830 during surface
analysis, and just sit there.  ESIX support told 
me to partition the drive using Disk Manager, which I did.  
This did allow ESIX to complete the surface analysis, but
each time the ST506 error would appear when the file system
was being created.  Obviously, I want to squeeze as much
space out of the drive as I can.

My questions are:

Is there some data recorded in the Low Level partition(s)
that could be causing ESIX to think the end of the partition
is at the 810 cylinder point?

Are my drive and/or controller defective?

Why can DOS recognize the entire number of cylinders (900
for sure, 917 I bet) and unix can't?

I have read that one should be careful about doing a low
level partition on an IDE drive, for fear of destroying bad
block data that might be stored there.  I guess I'm
wondering whether the initial installation at 830 cylinders
is still haunting the drive somehow.

I did a verify on the drive surface, and Disk Manager
reported no problems.  The version of DM I have is 4.2, and
it has as my exact drive manufacturer and model, but for
that model it reports even fewer cylinders than the dealer
did.

Please help shed some light on this, if you can.

Either reply to the address in the header, or post it.

Thanks,
M. E. Evans  
-- 
    M. E. Evans                           
    UUCP:  swuts!ucmpme or sw1e!ucmpme