Thanks to all who replied to my plea for help.

No-one seemed to have a convincing answer as to why the code in AN512
configured the timer 0 as a counter and then polled the TMR0 regsister.

I decided just to poll RA4 as a normal input, and (this bit of) my code is
now working fine.  The schmitt trigger is still used so I can't see any
reason not to use this method, and it means timer 0 is free for other
purposes (as long as you don't want to use external clock)

Most of my headaches came from the Read Modify Write instructions on ports
that were driving a mosfet H-bridge, I guess the gate capacitance was
slowing things down a bit.  Nothing that a coupla NOPs didn't sort.

My last remaining bug is the PWM (CCP1) output latching up.  I can't set the
PWM output to maximum as the output is AC coupled to a diode pump so it
would stop working.  I limit the maximum value I send to the CCPRL1
register, but occaisionally the output gets stuck high.  When you do a RMW
instruction on port C with the PWM active, does the micro read back the
current state of the PWM output?  If it happens to be high I guess a 1 could
be written back to the output register? I can't believe this is the case as
it would make using the port a nightmare.

My employers decided that buying an ICEPIC2 emulator (with real time trace)
would speed things up on the development front, so I should have a new toy
to play with this week!  Anything I should watch out for when using this or
any problems that anyone knows of?

Regards

Mike Rigby-jones
mrjones@nortel.co.uk