I like Tim;s explanation, but since it isn't real clear where you are
heading, there are a couple other points I'd like to make.

"Tim Wescott" wrote in message
...
wrote:

[ BIG CLIP ]

In electronics, amplification of _any_ signal is it's scaling up in
amplitude or power.

Understand that theoretically, increasing the amplitude increases the power.
The reality, especially with RF, and especially with solid state amplifiers,
is not quite so simple. In order to transfer power from one RF stage to
another, or from an output stage to an antenna, we need to match the input
impedance of the following stage with the output impedance of the preceding
stage. If the impedances don't match, not all the power will be
transferred, and the portion that is transferred is likely to be distorted,
perhaps badly.

I turns out that the output impedance of a solid state amplifier is
dependent on the output power, with higher power implying a lower
impedance -- often quite low. Generally, our antennas, and most larger
scale modules, have, by convention, a 50 ohm impedance. (Internal stages
often do as well, but there is no reason a designer might not pick something
else 'inside the box').

Normal practice is to follow the amplifier with a matching network
(typically still inside the amplifier box, however), so that the amp you buy
has a 50 ohm impedance output. This matching network actually distorts the
signal, really, deliberately. The result is what we want for a typical RF
amplifier application.

Your question sounds a bit as if you might not be interested in the
"typical" RF application, so the typical answers might not be quite what you
are looking for.

[ANOTHER BIG CLIP]

We wish. Ideally, yes that is the case. Practically the output signal
will be filtered, delayed, and distorted but not too much.


And again, this linearity is right for typical applications, but again, for
TYPICAL applications. Obviously, if an amplifier produces 1000 watts out
with 100 watts drive, and it is linear, it will produce 500 watts out with
50 watts drive. But if you give it one watt drive, it might not produce 10
watts, and what it does produce might be quite distorted and masked with
noise or other artifacts. Again, within reasonable limits, in typical
applications, these things behave as you would expect. But in some odd
application, or at the liimits of their useable range, the results might not
be close enough to theoretical to satisfy you.

The same is true if your input waveform is not what one would expect. For
communications service, the expected waveform is really quite predictable,
and the design of the impedance matching circuits makes assumptions about
these waveforms. If your input waveform is far from what would be expected,
then you might experience a lot of distortion of the input waveform, even
though a non-critical look appears to be within the operating range. For
example, if you were to drive that amp with a 50 MHz square wave, the output
waveform would probably look more like a sine wave than a square wave.

...