Tim's own version of the Catweasel/Compaticard/whatever

From: allisonp_at_world.std.com <(allisonp_at_world.std.com)>
Date: Thu Jul 6 07:16:16 2000

> >and less than 50ppm drift. The typical system was usually within
> >50ppm of exact and drifted less than 25ppm over temperature extremes.
> >Often the actual data rate was far lower than that reference(usually 1/4
> >or 1/8th).
>
> IMHO it's very misleading - or even worse, foolish - to worry about errors in
> the ppm range when the instantaneous rotation speed of the floppy can
> vary a few percent.

Exactly the point. ;)
 
Whatever the clock error may be it's trivial and stable compared to the
mechanical system and the magnetics.

> Look up "PRML" or "Partial Response Maximum Likelihood". It's a
> branch of signal decoding for use in situations where adjacent pulses
> most definitely have measureable effects on each other.

Understood, I was trying to stay out of the land of obscure terms
and letters.

> >MFM, M2FM, RLL or GCR), and previous bits history it should be
> >straightforward enough to predict the likely next transistion(s)
> >be they one or zero.
>
> Most uses of PRML look both forward *and* backwards.

Very true, the 765, 1793 and their heirs select precomp on a bit level for
this on the write case to make the read case easier.


Allison
Received on Thu Jul 06 2000 - 07:16:16 BST