Ian Jackson wrote in
:

....
In my simplistic world, this is how I understand things:

Provided the TX is followed by a tuner/matcher, which matches whatever
is attached to the tuner output to 50 ohms at the tuner input, the TX
will be happy.

"Happey" eh!

The power loss in the feeder is essentially a function of its inherent
loss (when matched) per unit length, and the SWR on it.

Wrong.


The SWR is a function of the mismatch between the load on the antenna
end of the feeder, and the feeder characteristic impedance. The

Well, the SWR at the load end is a function of Zl and Zo (both complex
quantities)... but the 'notional' SWR varies along the line as accounted
for by the complex propagation coefficient, in most practical cases at
HF, SWR decreases smoothly from load to source.

greater the SWR and the longer the feeder, the higher will be the loss
on the feeder. 'Lossless' feeder have no loss, regardless of length
and SWR. However, with real-world feeders, the losses rise
increasingly rapidly with increasing SWR.

See your earlier misconception regarding loss being a simple function of
SWR.


The impedance looking into the tuner end of the feeder is the
impedance of the load, transformed by length of the feeder, and is
also modified by the loss of the feeder.

You got that right.


The higher the feeder loss, the closer the impedance seen at the tuner
end will approach the characteristic impedance of the feeder. [Long
lengths of lossy feeder - maybe with a nominal termination on the far
end - can make good dummy loads at VHF and UHF.]

Yes, but is it of practical application in a transmit scenario? If the
input impedance due to a severly mismatched load is at all close to Zo,
then you have lost most of the transmitter power in the line.

The "make a good dummy load" recipe doesn't address the power rating,
especially where most of the power is dissipated in a very short length
of cable.

Owen