> My simple solution was to take a number of readings and average them.
> In this case, I took 20 readings, which seems now to give me a good
> "average" current output, which is what I was after.  Any comments on
> better ways to do this would be appreciated.

I would have thought that an RC filter from the current sense resistor to
the PIC would be necessary to stop motor spikes damaging the chip. I would
be very loathe to have the motor circuit connected in a more direct way to
the PIC. Suitable values for these components would achieve the same effect
as the software averaging.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.