Software-based floppy disc data separator

[Philip Pemberton]

On 19/06/10 20:59, Chuck Guzis wrote:
>> Chuck is very fond of histograms.
>
> If it doesn't pass the MFM test, that's the first place I go.  2
> peaks=probably FM, 3 peaks = probably MFM, 4 peaks = probably MMFM.
> 5 peaks = maybe GCR.

The "MFM test"? As in, "are there two '1' bits directly following each 
other in the encoded data stream?"

But it's true that histograms are good for figuring out which disc 
formats are likely to have been used... and the math behind the method 
is solid.

> On occasion, I have to build something to handle a peculiar setup.
> Catweasels aren't the end-all, nor are any others.  For example, try
> putting a Catweasel on a 2.8" DataDisk drive.

Are those similar to the Mitsumi Quickdisk (nee Famicom Disk System) 
discs? Basically a continuous stream of binary data, a SEEK_TO_START 
input and a HEAD_AT_START status output?

> I'm surprised that no one's showed up with some floppies from a
> Memorex 651 drive yet.

That looks fairly standard, at least in terms of the I/O interface 
(looks like a Shugart variant to me).

32-sector, soft-sectored, 64 tracks. So probably not a standard track 
pitch then... you're probably pretty screwed if you don't have a 651 
sitting on a shelf.

Although if you were really determined you might be able to "creatively 
misalign" an 80-track drive to read the data, but you'd have to fudge 
the alignment for each track...

> It's great fun if you're the sort of person who can look at a problem
> and have the patience to wait for an "ah hah!" moment rather than
> trying to beat it to death.   If you're the "gotta solve it now" type
> of person, you're probably in for a lot of headaches.

I'm the sort of person that likes to know how things work. Once I've got 
an "interesting project" to play with, I'll follow it to the logical 
conclusion.

Kinda like a small-scale version of "solving the {Enigma,Lorenz} 
cipher", I guess... :)

-- 
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/