In brief, how can you have a conjugate match with the source impedance
hopping about trying to follow the load? ;o)
----
Reg, G4FGQ

Reg, I have a myriad of measurements which prove that when the pi-network
is
adjusted to deliver max power for any reasonable drive level into any
reasonable
value of load, the source impedance is the conjugate of the load
impedance. When
adjusting the load control of the network for max output the output
resistance
equals the resistance of the load. And when the plate tuning is adjusted
for max
out the reactance injected into the network equals the reactance opposite
to
that in the load. To be practical, be assurred that the tuning and loading
controls are alternately adjusted continuously until they converge and the
max
power is delivered.

Having performed these measurements many, many times, I assure you that
the
resulting data isn't a series of coincidences. So instead of the source
impedance 'hopping about trying to follow the load', when the network is
correctly adjusted, the source impedance 'finds' the load and stays put
until
the load is changed and the network readjusted to accomodate the new load.

Walt
================================

Walt, forgive me for stating the obvious, but when operating under
non-linear conditions, as tubes and transistors are, I think you'll agree
that when the load resistance is changed, the tubes operating conditions
also change, and so does the tube's internal resistance.

I don't know exactly how you have conducted your measurements. But I suggest
that when you change the load you automatically force the internal
resistance to change with it.

But in any case, whichever way the source changes, the basic idea of a
conjugate match to an unstable source is itself not very sound.

The best way of determining source impedance is to mathematically model and
calculate it. It need not be accurate. It is necessary only to show that it
does change.
----
Reg, G4FGQ