[Coco] It's a small win, but a win nonetheless

[RETRO Innovations]

OK, the circuit I have now looks to be relatively solid:

1uF cap from input, other side goes to - comparator input
- comparator input held to ground via 10M resistor
100K/10K voltage divider (.45V on positive pin)
Hysteris from output to + pin, of 200K (pretty wide gap)

I got it working with a comparator IC, then switched to use the 
comparator in the Atmel AVR (lower parts count).

Notes:

Coco3 scope on - comparator input.  Hovers at .12V when idle (I would 
think the cap would remove all DC bias, but something is leaving 
something there)
When active, signal swings from .12 to .8V.  Median is .46, which is 
where voltage divider is set.  With hysteresis, the voltage swings from 
.43 to .65V

Coco2 is .18 to .88V.  Medium is .5, within band

Coco1 is -.28 to .88V.  Median is .3, quite a bit lower than my 
comparator's center, but the data seems to come in correctly.

Every once in a while, the detector messes up on bits, and I get AA AA 
AA instead of 55 55 55, but I am hopeful that my crappy prototype is the 
cause, and a solid PCB will address the issue.

Jim