Dan Wrote

> "Resistors Protect Microcontrollers From Capacitively Coupled Spikes".
>
> http://www.elecdesign.com/magazine/1999/dec1799/ifd/1217ifd4.shtml
>
> 16F84 was being used to control gates of MOSFETs [IRL530] which
> provided two-phase variable-frequency drive to a motor. Switching
> spikes were being fed back thru the parasitic capacitances of the
> MOSFETs to the PIC and resetting it.
>
> 10K resistors inserted between the PIC and MOSFET gates fixed
> the problem. Note - they don't show a lot of other protection
> [catching diodes/etc] here, but use of the series Rs looks like
> a good design idea in such ckts.

I am really going to regret bringing this up but it is on my
mind....

That is a large resistance (10k) in series with  typ 3nF
gate capacitance. I think it is fine the the app shown
because the frequency is low.
A very rough calibrated eye ball and rusty brain reckoning
would suggest that you are into serious linear (and hot
mosfet) territory much above 3 to 5 kHz.


I suppose one could calculate the highest value that
would suit ones application freq and see if that is enough
to protect the micro, otherwise zener protection etc or a
mosfet driver or another pre-output mosfet stage will be
needed.

Its just I have never seen this problem before using the
good old tried and true industry standards of 100 ohm
series resistors, gate to drain zener/tranzorb clamping
and back-emf diodes and I have been involved in a fair bit
of motor/transformer work up to 100 watts.



_____________________________

Lance Allen
Technical Officer
Uni of Auckland
Psych Dept
New Zealand

www.psych.auckland.ac.nz

_____________________________