John Ferrell wrote:
. . .
The transmitter final was a military surplus cavity with a 4CX250 that
had been "stuffed" to get it up to 440. As a side effect the output Z
was determined to be pretty low by trying several home brew quarter
wave coaxial matching sections. When a good match was made, a lot of
problems went away. Not only was I managing a better signal, but the
polyethylene cooling ducting was taking a longer time to melt down.

Hence, my position that SWR IS important.
. . .

All this demonstrates is that impedance match is important to the
transmitter final. The quality of impedance match is often indicated as
SWR on an SWR meter when in fact the meter reading often has little or
nothing to do with the SWR on any transmission line. Even when it does,
the problems with the transmitter are due solely to the poor impedance
match and not at all due to the SWR on connected transmission lines.

Let me give an example. Connect your transmitter through a half
wavelength of 300 ohm transmission line to a 50 ohm (resistive) load.
The transmitter sees 50 ohms, so an SWR meter at the transmitter will
read 1:1, even though the SWR on the line is in fact 6:1. The
transmitter can't tell the difference between this setup, a direct
connection to the 50 ohm load, or connection to it through a half
wavelength of cable with any impedance and therefore having any SWR. In
all cases, the transmitter sees 50 ohms, which is all that matters. The
line's SWR makes no difference at all.

If for some reason you were really interested in finding the SWR on the
300 ohm line, you'd have to insert a 300 ohm SWR meter at the
transmitter-line junction. It would correctly read 6:1.

Roy Lewallen, W7EL