On 20/07/2011 14:55, V G wrote:
> I thought about it for 5 minutes, and I'm going to take a wild guess here=
..
>
> The small valued capacitor acts as a "short" for very high frequencies an=
d
> therefore allows the passage of those frequencies into the negative input
> and subsequently causes their increased suppression.
>
> Increasing the value of the capacitor (to 10uF, for example), allows lowe=
r
> frequencies to be suppressed more (which is undesirable in this case).
>
> Does that make sense? Or am I totally off?

Sort of, yes.
The capacitor indeed feeds back the higher frequencies. The reason why=20
this stops it oscillating is that the gain will drop below 1 before the=20
phase shift reaches (and passes) 180 as the frequency increases. After=20
the phase shift passes 180, then the negative feedback becomes positive.=20
The main difference between an oscillator and an amplifier circuit (with=20
feedback) is the polarity of the feedback.

This does not cover it very well though - I would read any good book on=20
opamps and feedback theory. Things like Bode plots, poles/zeros,  and=20
the theory behind why/how phase/amplitude change. You can keep the maths=20
pretty simple and still gain a decent understanding of this, so I'd try=20
to find a book with not too much to put you off, but enough to make it=20
clear. Of course if you're a whizz at maths then it doesn't matter much,=20
but I find that sometimes even if you understand the maths, some books=20
can obscure the "practical" side of things if they concentrate too much=20
on the theory, although both are obviously important. I also recommend=20
you check out (maybe build, very easy and quick) a (opamp based) Wien=20
bridge oscillator and the theory behind it as a good example of=20
intentional phase shift for positive feedback using RC stages.
Also, as Herbert mentioned, be cautious of the spice simulations - in=20
this case it does look like the opamp needs compensating (most don't but=20
some do - usually it will state in the datasheet under what conditions=20
it's stable), but sometimes the models are inaccurate or buggy, or maybe=20
some obscure setting makes it work or not work (e.g. initial DC=20
calculation often throws it for transient analysis)
If you have problems with some opamp models, you can use the opamp2 (I=20
think that's the name) and set the GBW, slew rate and things to the=20
values in your opamps datasheet. Different models use different methods=20
of simulating the IC, usually (I think) with approximations of the=20
things as a whole rather than the actual circuit at a transistor level.

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