On Nov 21, 1:39 pm, John Larkin > >
Phase-shift oscillators are textbook darlings that usually work badly
in real life.
We're doing 2.5 Hz oscillator right?
Component quality, thermodynamic stability,
typical of oscillators.
That depends on the amplitude.
With anything, keep the amplitude low and
within the linear region of the feedback transistor
and you can get into spec.
Not really once you're in the ballpark, a pot
tweaking the top of one of the resistors will
do ya.
If I was told to make it ultra accurate I'd phase
lock loop off a crystal in a thermally controlled
oven. That way getting 1/10^5 is fairly easy and
1/10^6 is manageable. I find 1/10^8 a bit harder,
but that's a drift of 1 hz at 100 Mhz.
While with Telefunken, we had some 100Mhz
oscillators sent from Germany, (fancy units,
based in crystal controlled ovens), and they
were off 2 or 3 hz, and the damn problem landed
on my desk. From relativity to altitude pressure,
I couldn't figure it out, so I had them re-adjusted
(tweaked), following SOP I blamed Germany.
Of course that's nothing compared to the fella's
who have a dozen atomic clocks that need to
be synchronized.
Fun Stuff
Ken
