Richard Fry wrote in news:510d2db0-df70-4433-a216-
:

On Jun 11, 12:09*pm, Roy Lewallen wrote:

Can you write an equation giving the supposed dissipation caused
by "reflected power" as a function of transmission line length?

I'll just quote the following from Chapter 24-11 of REFERENCE DATA FOR
RADIO ENGINEERS, 1975 Edition:

QUOTE
....
The considerably greater loss for 124 feet compared with 24 feet is
because the transmission passes through a current maximum, where the
loss per unit length is much higher than at the current minimum.
END QUOTE

The author acknowledges that loss under standing waves is not uniform,
that in this case, it is higher in the region of a current maximum. That
is simply explained the I^R effects on the conductors (which account for
most of the loss in most practical coax cables at that frequency).

If you treated the forward and reflected waves as travelling
independently, and each attenuated by the matched line loss
characteristic of the line, you would get a different (and incorrect)
result. It happens that we sometimes make that approximation, and under
many practical scenarious, it gives a result with acceptable error, it is
nevertheless less accurate because the model is poorer.


As the input Z for each of these lengths is not the same as their 0.4-
j2000 ohm termination, this shows the dependence of the performance of
such systems on the electrical lengths of the transmission line in
use.

Yes, it does, and not simply on the VSWR as often inferred.

But that does not change the transmitter's behaviour in its output power
being sensitive the the impedance at its load terminals, however that
impedance is derived, and your response has not answered Roy's question,
well at least in my mind.


I realize we have strayed into the real world here, because RG-218/U
transmission line is not lossless even when perfectly matched at both
ends.

Indeed, under some circumstances, the line can have less loss when not
perfectly matched. Yes, throw those treasured graphs of ExtraLoss vs VSWR
away, they depend on assumptions not usually stated.

Cases similar to the example you quoted can be solved using TLLC at
http://www.vk1od.net/calc/tl/tllc.php . An interesting case is to explore
the loss in 1m of RG58C/U at 3.6MHz with 50+j0, 500+j0, and 5+j0 loads.
The latter two are VSWR=10, but have quite different loss. The first two
cases demonstrate that VSWR does not necessarily cause extra loss
(compared to the matched line case).

Owen