Gregg Kricorissian wrote:

> Another thing you have to watch with high current loads is keeping the
> drive circuitry and load paths separate ... high frequency PWM can
> cause havoc if the two circuits share much  ground path ... best to
> use single point grounding.

  Which "single point" is actually the source terminal of the FET, that
is, the driver circuit ground and its supply bypass should "hang" from
the source terminal, which is already tied to the negative supply.  The
only problem then is the grounding of input devices.

> After the paper design stage, I prefer to develop a drive circuit by
> monitoring the FET switching waveforms with a scope .... both voltage
> and current, and adjust values to optimize saturation and switching
> times.

  Indeed.  Otherwise you never know!  It may be worthwhile to provide a
special diagnostic "advance synch" output from the PIC for the CRO if it
is generating the PWM.

  Cheers,
        Paul B.