WSQT wrote:
If the LOAD on teh device is too high, it is possible for the
"bottoming voltage" or anode voltage necessary to pull two times the
current through it to be more than doubled for double the current.

You are talking about devices that do not follow square law rules. I
specifically restricted my comments to low mu triodes that were set to
switch properly.

ANY modulated stage on a DSB AM transmitter with carrier and 100%
undistorted sine wave modulation has to produce four times the power on
envelope peaks and reach zero output on negative peaks.

The original posted has a pair of tetrode, which are multiple gridded
tubes. Multiple grid tubes require some portion of audio be applied to
the screen, control grid, or driver in orrder to be linear. **This is
because the screen has significant control of anode current.**

In a low-mu class C plate modulated triode, the anode power and output
power follows square law rules virtually independent of loading. This
is why low-mu triodes are the tube of choice in distortion critical AM
transmitters. The major point of concern where a low-mu triode deep
into class C is not linear is when cathode emission is all used up, but
that would be a terrible way to run the tube. Life would be a bigger
worry than linearity.

High plate resistance devices operated with too high a load will NOT
have a constant dynamic anode impedance with varying plate voltage.
This is because the instantanious voltage across the tank circuit
during conduction is the supply voltage minus the tube voltage drop.

If the tube drop is more than doubled when the current(and supply)
voltage are doubled, it means the voltage across the load is LESS THAN
doubled. The load being ohmic, this causes less than double the current
to flow, and cuases downward carrier shift.

What you are saying is the tube has a non-linear plate resistance with
anode voltage changes. The reason for this is screen grid tubes, tubes
that are not deep into class C, and devices like collector modulated
transistors do not make good mixers.

What always must happen when devices with less than desirable
characteristics are used is we have to linearize the response by
applying audio to some other stage or element on that device.

It's indeed true you can have a poor system and fiddle with the loading
and operating point to linearize modulation. That was actually what I
was pointing out, that his plate modulated tetrodes were probably not
working correctly through some type of design shortfall.

If you have to fiddle with loading to linearize the modulation, the PA
design isn't a good one for AM.

73 Tom