Hi Steven

If you want reliability over the whole voltage and temperature range I
think INTOSC would be a poor choice. You would have to decide
if avoiding the small complexity of a crystal/resonator is worth the risk

Within a narrow V and T range INTOSC would be workable

One problem is that the base 8MHz (a 10^x frequency) is not conducive
with 9600 baud (a 2^x frequency), ie an error is introduced at the division
(Fosc/baud)/64 before even considering INTOSC accuracy

Using Example 20.1 in the datasheet

((8000000/9600)/64) =3D 13.0208, truncated to 13

8MHz / 9600 -> 0.16% error @ 25C, 5V, INTOSC accurate. 0%
error is achieved with 7.987200MHz

(7987200/9600)/64 =3D 13.0000, no mathematical loss

OSCTUNE can be used to change the frequency (Section 2.2.5.2),
eg compensate for drift, but you'd need a stable reference to do that
with any accuracy. Possibly the 9600 bit times could be used

And note that the datasheet details up to +/- 10% drift at operating
limits. I don't think an OSCTUNE adjustment range figure is given

If comms really are important, and I don't see why they wouldn't be,
it would be worth investigating a more stable clock source, like a
7.372800 or 9.8304 crystal or resonator. It might be possible to
pull an 8MHz crystal to 7.987200MHz, but I've not tried

Joe
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.