On Wed, 20 Mar 2002, shawn wrote:

> I'm looking for some opinions and perhaps suggestions. I would like to
> replace a 625Hz analog sine wave oscillator with a low-end PIC
> microcontroller -- say 16F84 or 8-pin model. Is it better to use a
> look-table approach to replicate the sine wave or can excellent
> results be achieved by generating a square wave then filtering? Any
> advice on the second approach would be appreciated. And finally, would
> the answer change if I were generating say a 1MHz wave?

First question: lookup table plus filter is definitely better.

Second question: yes, if you want to use a PIC. It's quite easy to
generate a pulse train devoid of the 2nd, 3rd and 4th harmonics. This
greatly reduces the filter requirements (suppressing the 3'rd harmonic of
a square wave without affecting the fundamental is practically
impossible). You can continue with an optimized pulse stream to minimize
the first N harmonics. Or you can increase the complexity of the pulse
stream and suppress the 2nd,3rd, 5th and all multiples of them.

Check out:
http://www.dattalo.com/technical/theory/sqwave.html

and specifically usage number 011.

This pulse stream, btw is devoid of 2nd, 3rd, and 4th harmonics. I don't
recall how big the 5'th is.

 ^ v(t)
 |
A|  +--+  +--+                          +--+
 |  |  |  |  |                          |  |
-+--+  +--+  +-----------------------T--+  +-----> t
 0  1  2  3  4  5  6  7  8  9 10 11  0  1  2  3


The idea is to run this through a cheesy low pass filter and you have a
sine wave!

Scott

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.