On 2014-01-26, Lloyd E. Sponenburgh lloydspinsidemindspring.com wrote:
Martin Eastburn fired this volley in
:

Overwriting it won't always erase. We used to run 10 patterns to
quality a disk and afterwards we had to format the disk.

Low-level formatting will. Always. Martin, actually, that was a silly
example.

Actually -- thermal variations may cause the heads to be offset
a fraction of a track width from one time to the other, and with the
right tools (and a disassembled drive) it is possible to recover data
from the fringes like that. The multiple patterns are to both increase
the chances of overwriting any given pattern, and a chance to have
different track offsets be used to be sure. At a minimum, alternating
0xAAAA and 0x5555 patterns result in each bit position being written as
a one and a zero. Other patterns might make it more difficult to
recover from fringe data left from the first two basic patterns.

There's no reason why a drive with the defective media locations already
mapped would ever need formatting after a checkerboard test.

Perhaps the whole reason for your 10-pattern test was to identify the
defects, then to format with a new defect map.???

More likely to be sure that the fringes of the tracks were fully
overwritten. Start off with the drive cool, and let the heads drift as
it heats up.

On some drives, rather than a stored defect map, defective sectors are
'jumped' over via a linkage table built right into the individual sector
headers. Those sorts MUST be formatted in a way to reflect the bad spots
on the medium.

A real pain in some cases.

Enjoy,
DoN.

--
Remove oil spill source from e-mail
Email: | (KV4PH) Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---