take something like crossover distortiuon for example...
No, I don't want crossover distortion.
How about thinking about a distortion that only adds, lets say the 2nd
harmonic to a sine wave. Think about what happens when that is
enclosed in a feedback loop. You take some of that second harmonic
from the output and feed it back into the input. The nonlinear
circuit takes the 2nd harmonic of the 2nd harmonic giving the forth
and sends that out the output. That forth comes back around and
around and around. A nonlinear cicrcuit that only made 2nd a harmonic
is now resulting in an infinite chain of frequencies.
in an open loop amp, crossover dist. creates lots of harmonics.
add neg feedback and they are all reduced. The high order ones are
not reduced AS MUCH as the low order ones, but they are certainly not
increased (assumming a proper design not on the verge of instability
and assuming the feedback componets themselves are linear, resistors
are usually linear for our purposes).
This is not correct. You have to have an extraordinarily large phase
margin to not have a boost in the harmonic near the gain crossover.
If G is the forward gain from the point where the distortion is made
to the output and H is the rest feedback the math looks like:
G /(1 + GH)
Here's the very ugly bit:
The distortion is often created in the output section making the G
part unity or nearly so. A stable servo loop can have a phase margin
of 30 degrees.
1/(1 + 1 * 1@(180-30)) = 1/(1 - 0.866 + j0.5)
= 1/(0.134 + j0.5)
Take ABS()
ABS(1/(0.134 + j0.5)) = 1/sqrt(0.134^2 + 0.5^2) = 1.93
Even though this amplifier is very stable, the feedback loop doubles
the amplitude of the harmonic near the gain crossover.
