Take a look here.
**broken link removed**
Hi Mike,
Ok, i did, and found that the design equations and even philosophy
isnt quite right. Let me explain...
First, a fixed frequency sine wave generator made from a square
wave generator followed by a filter (as that article is about) to
smooth the square wave into a sine wave is best by making
the filter a bandpass filter rather than a low pass filter. That
design uses a low pass filter so it's not as good as one that uses
a band pass filter, and since both designs use about the same
number of components, why not use a band pass instead?
The low pass does not produce as good a sine wave, with the
tops bent to the left somewhat instead of an almost pure sine.
It's not too bad i guess, but with the same components a
bandpass would be better.
Second, the design formula for the filter provides for a very
high output which allows too much of the original square
wave to reach the output. This leads me to believe that
the second formula:
R5=1/(8.8856*F*C1)
R6=R5
is also incorrect.
A better formula would be:
R5=2/(8.8856*F*C1)
R6=R5
which effectively doubles R5 and R6.
Third, the design formula for the center frequency of the
square wave oscillator is totally incorrect. This would result
in the calculation of components that dont provide the
correct operating frequency.
The formula given is:
R1=(0.5*F)/(0.693*C1)
but it's obvious with that oscillator that as frequency increases R1 has to
decrease, so the corrected formula is:
R1=(0.5)/(0.693*C1*F)
These formula changes make a better sine wave oscillator,
but again if the output filter was a BP type it would be
even better.