Re: Confused about how AGC works?

Jerry Gaffke

Digger previously wrote:
Does the delay in part prevent oscillations by introducing a phase shift?

Consider an amp with positive feedback and no delays anywhere.
It would instantly ramp up to infinity.

If you want it to oscillate, you need to have a delay in that feedback loop
that gives a full 360 degrees of phase shift at the frequency of interest.
So the next cycle at that particular frequency gets reinforced by feedback from the previous cycle.
And other frequencies don't get this positive reinforcement at exactly the right time, so they die out.

If you want the oscillation to be stable, the gain around the entire loop
must be exactly 1.0  (so no gain and no loss).
If it is less than 1.0, the oscillation will die out.
If it is greater than 1.0, the amplitude of the oscillation will keep getting bigger till it swallows the universe.
However, any practical amplifier will have reduced gain as the oscillation gets too big,
so we will soon reach a steady state when the oscillation is so big that the amplifier gain is reduced to 1.0.
In this way the amplifier is considered non-linear, since it treats big signals differently than small signals.

The diagram on that wikipedia page shows a pure amplifier with a voltage gain of A (and no delays).
There is also a feedback network with a transfer function of Beta (a Greek B).
The transfer function mathematically describes all the phase shifts (delays) and attenuations at all possible frequencies,
all wrapped up in a nasty equation.  This is the part that keeps EE students drinking coffee the night before an exam.
But that transfer function is linear, and could be described by a bunch of resistors and capacitors and inductors
all wired up into a big rats-nest.  For example, here's a fairly simple model of the 12mhz quartz crystal
used in the feedback loop of the BFO oscillator on the Bitx40:
https://www.maximintegrated.com/en/app-notes/index.mvp/id/2127

Jerry, KE7ER

On Fri, Apr 19, 2019 at 03:45 PM, Jerry Gaffke wrote:
> Does the delay in part prevent oscillations by introducing a phase shift?

Join BITX20@groups.io to automatically receive all group messages.