On 1/1/14 8:58 PM, Philipp Hachtmann wrote:
I'm
currently finishing the design of the DECtape G888 replacements to get those
sent off for fab.
Oh, I'd be interested in several of them!
They probably won't be what you're looking for. It is a single double-wide board
that plugs into the
A10/A11 connector with all 5 channels. The signal path has a low pass filter and AGC. It
uses MOSFETs
for write drive. There is a second single-high board that goes into A6/A7 that has the
drive select
and control signals. I'm thinking of adding a microcontroller to handle writing the
timing and
mark tracks, since I need a little test circuit for the write drivers.
I woke up in the middle of the night today and realized why my original attempt at reading
tapes
was unreliable. All I did then was take the output of the G888's into a fifo clocked
by the raw clock
track. I didn't pay attention to the 10uS 'ignore me' one shots on the TP0 TP1
clocks in the
schematics of the original controllers.
You get spurious transitions at the zero crossings. If you look at the app notes for the
MC3470 floppy
disk read channel, for example, that's what the first one shot is for; to ignore
incoming transitions
for about a microsecond. This is mentioned in a couple of places when they're talking
about filtering
the 3rd harmonic before the 90% phase shift in the analog signal, it's also sometimes
called the 'snake'.
I haven't looked at how far back the original FM decoder design on DECtape controllers
goes. Doing
Manchester decoding per channel instead of using a common clock track would have helped
with head
skew issues. I also was pondering what happens to the clock timing when you run a tape in
the reverse
direction. I'm assuming it flips polarity the same way the mark and data tracks do.
Figuring out how to write a DECtape is an interesting problem if you try to do any of the
high speed
part in software and want to have the tapes be compatible with real controllers because of
the use of
the on-tape clock and mark tracks.