> Any other ideas? :)

Not really / not yet. Just checking the obvious. Working on my
own little social-life killer

>" to show a constant assertion of Strobe, which in theory should
>be impossible. I don't know, perhaps at this point I'm measuring
>wrong"

Like Olin says, this could be your biggest clue. Start from there,
even if it means breaking the programs down into individually
testable routines and then building the programs up again. You
might think "been there done that", which you may have done, but
the PIC obviously doesn't care about your feelings on the subject

Taking a day off might help - it's so often the case that a solution
magically appears after a break

The little SLK mentioned above is a case in point. One routine,
identical to others, didn't quite work when it should have. Spent the
whole day on it (which wasn't entirely fruitless as it forces you to
look at and clean up other routines) without a major improvement.
The big breakthrough was after a day away from it and then getting
slightly angry and determined to analyse it line by line "properly"
using a scope and marker pulses. Bingo, answer was so obvious
in hindsight

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics