On 5 Mar 2002, at 11:38, SatPrem Reddy wrote:

> Hi,
>  Sometime ago (I think in january), there was a post in this list
> about installing a hard drive up side down. There was a reply
> (from Mark Barron, I guess) which said that it is ok as long as all
> levels of formatting are done in the same orientation.
>  I could not understand why this should matter. I thought
> it was more of a problem with bearings and stuff. How does the
> orientation of the disk affect the formatting? Also, what does
> he mean by "all levels of formatting"?
>
>  TIA.
> -Satprem

1.  First, as to "levels of formatting".  The creation of a file
system on a partition is called formatting, and used to usually be
done by running the "format" command.
  However, prior to the creation of entries in the drive's partition
table, there is a process which needs to be done, of going around and
writing some meaningless test pattern to the data spaces of all of
the tracks, and inserting appropriate marker values to subdivide the
tracks into sectors.  This process is generally called "low-level
formatting".  It used to be common for code to do this to be included
in the ROMs for hard drive controllers, and this is still common in
SCSI; ATAPI/EIDE manufacturers tend to recommend that users not
perform this themselves (it is done at the factory before the drive
is shipped), but a few manufacturers have released their utilities
for this purpose to users.

2.  The most significant difference in to a drive that orientation
makes is the direction in which gravity tugs at the various pieces.
This can, of course, impart a tiny bias to the alignment of head
movements and so on.
  The particular risk is that if the low-level or partition formats
are done in one orientation and the drive is actually used in
another, the later tracks written to the disk surface may not
precisely align with tracks written earlier.  Small portions of
earlier data may remain, and might make the process of reading and
decoding the current data slightly more error-prone.

  Whether the risk is currently still at a level where you might
encounter it within the useful life of the drive is a debatable point
-- while drive reliability has been going up, the width of tracks and
strengths of signals recorded have been going down.
  It is quite probable that many people could ignore Mark's advice
without ever encountering a problem, but it *does* make sense for
those who'd prefer not to take unnecessary risks with their data.

David Gillett

                         PCBUILD's List Owners:
                      Bob Wright<[log in to unmask]>
                       Drew Dunn<[log in to unmask]>