"Reg Edwards" wrote in message ...
....
====================================
Tom, To add a bit more -

50-ohm generators as used in laboratories (so that measured reflexion loss,
mismatch loss etc, mean something) are effectively constant voltage
generators in series with a 50-ohms resistor, or constant current generators
in shunt with a 50-ohm resistor. They may be followed by an ampifier whose
output impedance is held constant at 50-ohms by some automatic means. None
of these circuits bear much resemblance to a pair of 807's and a tuned tank.

The best that can be said about Rg of the usual HF radio transmitter is that
Rg is indeterminate. IT EVEN VARIES AS THE LOAD IMPEDANCE IS CHANGED which
most of the Guru's contributing to this newsgroup appear to be unaware of or
at least choose to disregard. So what does "adjusting RL to equal Rg" mean?
To use it in a description of feeder + antenna behaviour further propagates
myths, including those surrounding SWR, forward power, reflected power, SWR
meters, etc.

Does Terman ever bother to mention Rg of a Tx PA? If he doesn't it can't
matter very much to him. The ARRL handbook, when numerically designing a
transistor linear HF PA, makes no mention of Rg.

Amen, brother. I was thinking after making my last posting to this
thread that the one thing I DON'T bother thinking about when designing
a PA is what source impedance it will present. I worry about
currents, voltages, efficiency, distortion, a network to present the
proper load to the active device(s)... but not Rg.

In precision instrumentation systems, the output is commonly levelled
or monitored through a levelling splitter (not to be confused with a
power divider), so that a virtual zero-impedance point can be
established, with a 50-ohm (or other Zo) resistor from that point to
each output. And network analyzers are commonly calibrated with
precision loads so that the imperfections in their outputs and
reflectometers and cabling can be backed out by the calibration
software.

Cheers,
Tom