I saw the same problem with fsck (BAD SUPERBLOCK ...) on my
IIcx+Conner1Gb external drive. I looked at the source code and saw, if I
remember correctly, that fsck tries to read the last copy of superblock
from a strange place, get zero block (the drive was brand new) and
then did calculation by zero division. At the end I took the geometry
information from NetBSD tools (newfs -n) and typed it into MacOS
utility. I had also to adjust the partition sizes to be the exact
multiple of cylinder groups. So I think this is the disk geometry
problem. Somehow MacOS utility calculates disk geometry from the whole
disk and Unix utils take into account only partition size. If the disk
has variable number of sectors/cylinder you got the problem. I can not
test those assumptions or repeat the problem without new disk.

	Taras Ivanenko.