>> Can timer2 in a pic18 do what you need?
>=20
> I had thought of a PIC as the source,  but wasn't sure how I'd go
> at higher frequencies

The lower end (0-200Hz @ 0.1Hz, PIC =3D 40MHz fosc) would be
fairly trivial with timers but at the upper end, 20kHz (40,000 half
cycles), one instruction cycle will be 80Hz

10,000,000 / 250 tcyc (25.0us half cycle) =3D 20,000Hz
10,000,000 / 251 tcyc (25.1us half cycle) =3D 19,920Hz

Loading a timer with various sequences would get you a frequency
that is an average of the loadings, in theory with infinitely small
resolution, but at the cost of time to make that average. For example
loading with 250 250 250 250 251 would average 250.1666 per
half cycle over the 6 iterations, or 19,986Hz

You could find the frequencies for which 1 tcyc =3D 0.1Hz or 1Hz
and see if that will fit some of your needs

1/f2 - 1/f1 =3D 100ns, where 100ns =3D 0.1Hz and 1/f2 - 1/f1 =3D 100ns
where 100ns =3D 1Hz

Would still be good up to several kHz (50% duty) but free tcyc
for duty cycle variation is going to run out at a fairly low frequency.
I've a project based on a 40MHz 18F that manages 1.5% - 100%
variation (in 63 steps) at up to 300Hz, doubt I could push it much
more than that without reducing the duty cycle steps

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