I just was reading David Benson's book "Easy Pic'n" where he mentions the tight
timing requirements of the read-modify-write routine.
He suggests using a NOP to let the pin settle.

Would this eliminate the need for the resistor?
Or is it a current overload that is damaging to the PIC?

Regards, Joe

Regulus Berdin wrote:
>
> Sean Breheny wrote:
> > Can someone please explain why a resistor is needed and how it could matter
> > where it was placed?
> >
> > I could see that the resistor will limit the current which is charging up
> > the gate,but how this would be likely to hurt either the PIC or the FET I
> > don't really understand.
>
> It may disturb a read-modify-write instruction in the software.  If the
> code pulls it high then low and is read fast, the pin still stay charged
> due to gate capacitance and the pic may read it as 1 and the FET may
> remain on.  With a resistor in series, this problem is minimized.  The
> use of shadow register can eliminate this problem.
>
> regards,
> Reggie