This is what is meant by "Class F." A Class F amplifier has traps
tuned to various harmonics to allow the use of a longer conduction
angle and/or square wave drive with high efficiency and excellent
modulation characteristics.

There is also Class E-with a SERIES tank that is an open circuit
instead of a short to harmonics. These are run somewhat detuned so as
to cause the circuit voltage to fall to zero at switchon and switchoff
with square wave drive. 90% efficiency not uncommon with both E and F.
I

There are still other such tunings such as "inverse F," where EVEN
harmonics are blocked but odd shorted. This can be as simple as a
push-pull circuit with no center tap in the output tank if the
interelectrode capacitance is low enough at the lower harmonics, as it
often is for an AM broadcast or a 160M ham amplifier. For higher
frequencies this design requires parallel resonating the interelectrode
capacitance at one or more even harmonics abd becomes pretty much a
fixed-frequency affair as a result.

Inverse F(odd) is again capable of 90% plus efficiency.

All of these amps work by blocking the high harmonic current that flows
when a square wave pulse of 180 degrees angle is fed to a normal tank
circuit, which is essentially a short to harmonics. this allows the
active device to turn on and off with essentially no extra
resistance(from the device) beyond full-on in series with the load, for
all of the conduction angle.

The resulting tunings effectively square off either voltage or current
waveforms depending on the partucular tuning used. In other words,
current or voltage in the tank is NOT a sine wave, but the portion that
is trasferred to the load must be a sine wave or else filtered. In
fact, Class F may have been invented when someone put a harmonic trap
in a plate circuit to stop a nettlesome harmonic current from flowing
and then being coupled out. That is in some old RAH editions as a
desperate measure for stoppign harmonic TVI, but also allows more amp
efficiency if part of the design.

Straight Class C is in fact also capable of 90% efficiency, but only
with a very narrow conduction angle so the plate current pluse flows
only when tank voltage is already very low. This requires high back
bias, high drive-and a larger than normal tube or transistor. The
"power density" sucks, and in fact Class B is capable of more power for
a given maximum current and voltage, assuming the heat can be
tolerated. Class E and F amps can match Class B power densities with
early(narrow-angle) Class C efficiency.

For a tube low power density means higher filament power, giving
back some of that efficiency. Going to flourescent lights in the
station may save as much electricity cheaper in that case.



A network with an inductive input will allow a square waveform at the
device output but not waste significant energy in harmonics. I've done
that in designs.

In the 1950's RCA had an AM BC transmitter that drove a tube with a
near square wave, and had a near square wave. The RCA transmitter used
a low-mu triode that had parallel tuned circuits in the grid and anode
set at the third harmonic. It had conventional networks feeding the
grid and to the antenna from the plate resonantor. That transmitter
made over 95% anode efficiency.

73 Tom