As a learning exercise, I am trying to read the data on the hidden
sector of a copy-protected disk.  It is one of those schemes where the last
sector of a track has been marked bad and given an odd set of address marks
so that diskcopy and diskcomp, among common applications will not work.

        Is there any way to trick the controller in to reading the address
marks at the beginning of a sector?  One possible idea might be to tap in
to the serial data as it comes down the cable from the drive to the
controller.  This is where a PIC might come in handy for decoding the bit
stream.

        To share what I have learned so far, each track of a quad-density
1.4 meg floppy is composed of 18 sectors.  A cylinder is both sides of the
disk or 36 sectors.  The tracks on one side are called "head 0" and those
on the other side are head 1.  There are four bytes at the beginning
of each new sector that contain the cylinder, head, sector number, and a
byte that tells whether the sector contains 128, 512, or 1024 bytes.

        Dongles usually put an odd-numbered sector in a track to break the
diskcopy mechanism which is looking for 18 neatly-numbered consecutive
sectors.  It is also possible to change the head and cylinder values in the
address marks for the obscured sector such that one has a huge number of
combinations to try if the trial-and-error method is used.  Being able to
get a peek at the address marks of the next sector would allow one to at least
know what needed to be done next.

        So far, I know how to format a disk with an odd sector so I could
make a dongle if I wanted to, but I am trying to figure out a way to hack
the reading process to give me the raw disk data so as to avoid the possibility
of wearing out the disk by thousands of reads to the same portion while
trying all possible combinations.  Any ideas besides just giving up are
appreciated.

        Several weeks ago, the topic of using floppies as extra memory
for PIC's came up.  I think a PIC might do certain dedicated functions like
running the stepper or possibly decoding certain data, but the control
needed to run a disk controller is probably more suited to a full-fledged
CPU due to memory buffers and lookup tables.

        Again thanks for any ideas or hacks such as ports to look at or
buffers that might have the full sector, marks and all.  When this is all
over, I will know what makes a disk drive tick and I will have a clone of
the disk in question for backup purposes.

Martin McCormick