>One thing I have noticed is that the PWM frequency will be about 20
>Hz, with a 4 MHz oscillator. You count 100 timer0 timeouts for each
>pulse cycle. There is a table of values that returns the length of the
>high part of the cycle but instead of using 100 separate pulse high
>times there are actually only about 30 returned by the table
>lookup. This means that you could use 32 separate duty cycles and
>count only 32 timer0 timeouts instead of 100 and it would give exactly
>the same duty-cycle resolution but increase the PWM frequency by a
>factor of 3.
>
>Even still, I was under the impression that it would be desirable to
>have a PWM frequency in the kiloHertz range and that only 16 or
>perhaps 8 distinct duty cycles would be needed (at least for RC
>cars). Can anyone comment on this? Does anyone know what the time
>constant of an RC motor is?

Malcolm's comments ought to be noted.

The PWM should be 2KHz - 3KHz for these motors.

I have 100+ distinct speeds in my ESC (forward-only), and that's only using
partial range on the input to generate full-range on the output.  It is
wise to have this many, otherwise one can hear the car "shift" and this is
a problem when you are close to the boundary between two speeds.

Andy

======================================================================
Andy Kunz - Montana Design - 409 S 6th St - Phillipsburg, NJ 08865 USA
              Electronics for Industry & R/C Hobbyists
         "Go fast, turn right, and keep the wet side down!"
======================================================================