At 05:44 AM 17/05/2012, you wrote:
>Tag added, PIC, in case it veers that way

The classic way to do this is with a Johnson counter. See
AoE figure 9.93 for an implementation using a 4015
and an inverter that yields a 16-level approximation to a
sine wave.

>Off the top of my head, once the output of the counter passes 111
>or 1111, the next o/p bit should toggle inverting logic
>
>I think 4 x XOR gates should do it. Each bit of low-order data from
>the counter goes to an A input, the logic control bit is commoned to
>the B inputs
>
>FYI -
>
>Attached is a circuit I use to make a variable AC voltage. The ADC
>pot changes the period of a timer. Output is a fairly good sine wave.
>Filter components depend on the frequency range for best output,
>although a constant-amplification amplifier could be added. My range
>for this circuit is quite low, 5 - 50Hz
>
>Joe

Trivial to do with a PIC. Just generate an interrupt, spit
out the new number, increment the pointer to your table
modulo 256 (say), look up the next number, and set your
comparator for the next interrupt. Your interrupts will be
at your sine wave frequency * 256, so with a 200ns timer
resolution you'll get a resolution of about 0.25Hz at 50Hz.
PICs (at least the smaller ones) have constant interrupt
latency so this is feasible without extra parts or jitter.
If there was variable latency and the jitter was excessive,
you could make a pin toggle right at the comparator hit
and clock an external buffer (pipelined, obviously). Another
part. If I had to that, I'd probably use an HC595 and
use a serial output from the PIC (the 595 is a serial-to-
parallel SR with an edge-triggered output buffer).

There are other techniques (phase accumulator) that can be
used to improve resolution with a relatively modest clock
rate (at the expense of more complexity and some jitter),
but maybe you don't need to do that.

Best regards,

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