From: "Wouter van Ooijen"
> See it this way: what is most time critical? I assume the servo
> handling. Hence that must be interrupt driven. The main loop can just
> poll the UART as fast as it can. For correctness: prove that the worst
> case serial handling, interrupted as often as possible, will still poll
> the UART fast enough.

That's a good idea. Although in my project the jittery aspect won't matter
because I won't be sending servo commands in such small increments and they
will be spaced in time. Therefore I just stated the problem in my conclusion
as inspiration for someone (or even myself later).

> Or maybe: (I did not follow this thread, so I might just be talking
> nonsense) I assume the servo code is interrupt driven for both the start
> and the end of the pulse. The pulse width itself is critical, the
> inter-pulse period much less. Is it enough to process serial data one
> per pulse, either before or (preferrably) after thye pulse?

This issue was addressed in another topic and I concour that it is the most
reliable way to set the pulse with, but in my case that's not what I am
doing. I am using interrupts to mark the start of the pulse, and then
regular delays to create the pulse. I agree that it is a waste of processor
time, but for this example it works. I would have to carry some experiments
to determine if this decision is what makes it jittery, that's a good
possibility that yes.

Cheers

Padu

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist