Jim Kelley wrote:
Roy Lewallen wrote:

Jim Kelley wrote:
. . .
Voltages on a transmission line do not determine reflection
coefficients.
Reflection coefficients are determined by characteristic impedances, not
virtual ones.

73, Jim AC6XG

I disagree with this. When applied to transmission lines, the (voltage)
reflection coefficient is, as far as I can tell, universally defined as
the ratio of reflected to forward voltage to reverse voltage at a point.


That rho is equivalent to that ratio of voltages is not in dispute. I
might dispute that it's 'defined' by that ratio.

I stand corrected on that point. I did a quick scan of ten
electromagnetics references. Seven quite clearly defined it that way.
One (Magnusson) was vague, and one (Jordan & Balmain) gave pretty equal
weight to both V and Z ratios as a definition. Holt, whom I consider one
my best references, clearly defines reflection coefficient in terms of
impedances. Regardless of definition, virtually all give both the V and
Z relationships which are, of course, mathematically equivalent.

We agree the
reflection is caused by an impedance discontinuity. It is the
relationship of those impedances that determines how much of an incident
voltage will be reflected. From my perspective, one builds a network of
impedances in order to achieve the desired voltage relationships. But
one cannot build voltage relationships in order to obtain a network of
impedances.

Maybe it's another chicken and egg argument.

I think so. One sees different impedances, i.e., V/I ratios, along a
transmission line, so it could be argued that impedances have been
created from voltages and currents. But I'll leave the arguing to those
people more interested in philosophy than engineering. As long as the
principles are understood and communicated, the point of view is largely
a matter of taste.


Only an impedance discontinuity causes reflections, but we can calculate
a reflection coefficient at any point we choose, with its value being
well defined and unambiguous.


Wouldn't the most well defined and unabiguous be at a point of
reflection? ;-)


I don't think so. It's completely well defined and unambiguous at any
point along the line. At any point, there's only one value of Vf and Vr,
and only one value of Z and Z0, so there can only be one value of
reflection coefficient. Knowing the reflection coefficient and Z0, for
example, one knows for sure what the value of Z is at that point. I find
that a number of the text authors freely use the concept of reflection
coefficient at any point along a line, not necessarily a point of
reflection. I don't have any problem with it, either. I see requiring a
reflection in order to have a reflection coefficient as sort of like
requiring dissipation in order to have a resistance. It's not really
necessary, and the calculated value can still have meaning.

Please don't let this detract from the ongoing discussion. It's really a
minor point.

Roy Lewallen, W7EL