Stuart: Walt is perfectly correct as long as we are talking about TUBE
circuits. Tubes represent a severe mismatch to reflected RF, with only the
essentially trivial losses in the tank circuit keeping that impedance from
being infinite in all respects. So for TUBE circuits most reflected power is
actually re-reflected.
Class A, AB, and B transistors are another situation entirely. Somewhere I
have an elementary book on transistors that represents them as a variable
resistor with a motor on a slider for the base. The point being that while a
"vacuum valve" is a one way device whose current is controlled by it's grid
to cathode potential, a transistor is essentially a variable resistor with
its resistance controlled by base current. At RF of course, that is a fast
motor!
Within the limits of the text, this explanation is correct. Even when the
transistor is cut off, some current flows. If the circuit is full conduction
the reverse impedance, for lack of a better term, is very low; there is a
severe mismatch at the device output, and most reflected power is
re-reflected. If the device is as completely cut off as transistors ever get,
the same conditions and re-reflection apply.
BUT for most of a cycle a transistor output stage is in some state of
conduction. So the amount of reflected power that is re-reflected varies
according to the state of the conductive state of the output transistors.
The amount of re-reflected power depends on the relative phase of the
reflected wave and the conditions in the output stage, and range from almost
total to none.
This situation sneaked up on many hi-fi audio designers when germanium power
outputs first came out. Some speakers generated enough back EMF, the audio
analog of SWR, to drive the outputs into conduction, leading to transistor
failures. One combination that comes to mind was several of the early H.H
Scott amps driving the AR3. Transistors didn't last long even when volume
levels were kept to a restrained level.
73 Pete Allen AC5E
|