[Amps] AM On a linear

[Tom W8JI]

> All of you are overlooking the fact that during the 
> negative swing of the modulator when the plate voltage is 
> near zero  just about all the power will actually go to 
> the grid. The same reason why you should not drive the 
> tube with out the B+ on. You will possibly fry the grid if 
> you don't modulate the driver too.

and

> We are overlloking the feed through RF from the grounded 
> grid configuration.
> A portion of the driving RF will always appear in the 
> output, thus you can never "pinch off" the carrier and get 
> 100% modulation.

Frying the grid isn't the issue because heat is a time 
problem. The heavy grid conduction peaks are far too short 
unless you are using .05 CPS modulation or something.
Even "pinching off" the carrier isn't the biggest issue, 
although it is a small problem.

The issue is we can't easily plate modulate a stage with 
negative RF feedback. The tube has to act like a perfect or 
nearly perfect mixer, going rapidly and sharply  into and 
out of conduction with each RF cycle. In simple form this 
requires a triode with low mu well into class C.

If a tetrode is used, it has to have some other element 
significantly modulated to linearize the control of output 
power by modulation voltage. Otherwise the stage has 
terrible waveshape distortion that creates all sorts of IM 
and harmonic distortion up and down the band, as well as not 
properly reaching peaks.

A cathode driven amp has heavy negative feedback that tends 
to linearize the tube over the RF cycle and hold output 
constant with variations in HV, exactly what we don't want 
in a plate modulated stage. Not only that, most cathode 
driven amps use high mu triodes with softer transitions into 
conduction even when class C. That's exactly what we don't 
want.
The most linear way to modulate a cathode driven amp that 
uses a class AB or B tube would be to apply nearly zero 
volts of modulation to the anode and nearly all the 
modulation to the driver or some earlier stage that acts as 
a better mixer.

Think of building a very linear receiver mixer, where the 
audio is the RF input signal and the carrier is the local 
oscillator. The RF output of the transmitter is like the IF 
output of the mixer.  The carrier has to just switch things 
off and on very hard, like a local oscillator in a mixer. 
The audio then becomes the input and the sidebands that are 
generated must perfectly track the applied audio. We have 
upper and lower sidebands, the sum and difference between 
the steady strong LO and the varying input.

Good AM modulated stages are really just very good 
unbalanced mixers, and grounded grid amps stink when used as 
a mixer (just as anode injected tetrodes do).

73 Tom