"Wayne" wrote in
:

Many years ago, I might have been able to somehow deduce cable losses
from a Smith chart, but nowadays I just use the published curves.

An email correspondent suggested the radially scaled "SW LOSS COEFF"
gives a direct reading loss under mismatch from the chart.

There are assumptions that underly that scale, pre-requisites if you
like for it to be accurate. They are explained in Smith's book: "If a
waveguide is one or more wavelengths long, the average increase in
dissipative loss due to standing waves in a region extending plus or
minus one-half wavelength from the point of observation may be expressed
as a coefficient or factor of the one-way transmission loss per unit
length".

So, it does not apply to a feed line of less than a full wave, or any
residual of less than a full a wave after working out the additional
loss in each of the full wave segments of a line. It is tedious and
accurate only for lines of integral full waves length.

The potential error applying it to arbitrary line length ranges from
insignificant to huge.

The graphs in the ARRL suffer the same problem, and IIRC they do not
state the assumptions.

The Smith chart solution I described in my earlier post works for
arbitrary line lengths, and properly reveals the cases where loss under
standing waves is less than matched line loss. TLLC
(http://www.vk1od.net/calc/tl/tllc.php) also calculates the loss
correctly for arbitrary line lengths.

Owen