"Walter Maxwell" wrote
... I have explained many times that even though the PA source
upstream of the tank circuit is non-linear (and no one's saying it isn't),
the energy storage in the tank makes the output of the tank a linear
source, no matter what the shape of the current wave form may be
at the input. The output of the tank is proved linear because the
voltage/current ratio at the output is non-varying and the shape of
the voltage and current wave forms are essentially sine waves.
Consequently, the output circuit can be represented by a Thevenin
source that supports both a conjugate match and the maximum
power transfer theorem.
______________

If this statement about the tank circuit being ~ a linear source is valid,
does that mean that any load-reflected power that appears across the output
terminals of the tx stops at the tank circuit, and never sees the
non-linear, non-matching Z of the active PA?

And if so, would that also mean that such a tx would not be prone to
producing r-f intermodulation components when external signals are fed back
into the tx from co-sited r-f systems?

Yet experience shows that this is not the case for ~closely spaced
interfering signals. The only mitigation for this for a PA with a tank
circuit is the rejection of that tank circuit to those off-freq, external
signals, and to the resulting IM products generated by mixing with the main
tx signal in the active (and non-linear) PA stage of that tx.

And the tank has VERY low rejection to load reflections of the signal
bandwidth to which it is tuned.

Also to be considered are the modern broadband (88-108MHz) FM broadcast
transmitters, which have no tank circuits, but except for some designs
incorporating balanced 3 dB hybrid combiners are affected by load
reflections about the same as a tx with a tuned tank circuit.

RF