> > The closest I can get my micro to 9600bps, is 9420bps, and I have no
idea
> > whether this has a chance of working after taking into account the
> > tolerance of the resonator has on the bps rating.

You have to recognize that the receiver may not be running on frequency
either.
If the drift between both of you is less than half a bit-time over 10 bits,
then you'll be ok.
Since it's only fair to let the other guy use half the error margin, this
gives you a max of 1/4 bit time.

10 bits should take 1.0416 mS and yours would take 1.0615
The difference is .02mS  1/4 bit time is 260uS, so I'd say you're near the
edge. (60uS)
If the drivers introduce any distortion at all, or if the other guy ate more
than exactly half the margin, then you may have problems.

Why not change the xtal so you're on a nice baud-rate multiple, with error
approaching zero?