Please consider this to be written in red ink, or it may turn out to be my
blood.

The interminable arguing about losses in transmitters, tuners, transmission
lines, antennas, baluns, matching sections, ground systems, etc., never
reach any conclusions because half the time the participants, without being
aware of it, are are not talking about the same things. The internal
impedance/resistance of the source appears in the arguments and then
disappears of its own accord or somebody re-introduces it an attempt to
prove a point and then finds it convenient to forget about it. Confusion
reigns. Hackles rise. Blood pressure soars. Threats of Legal Action are
made.

This is because hardly any of the participants have ever been aware there
are two sorts of loss. Education has been neglected. Both can be described
in terms of decibels or even S-units. The performance of given black box in
a transmission system can be analysed in two different ways, each way
usually having a different loss in terms of decibels but when assessing
overall performance it is vital to understand which sort of losses are
involved.

There are "Transmission Losses" and there are "Insertion Losses."

There will now be raised voices of protest - "This bloody Limey is
insulting our intelligence again". "We have read books about Insertion Loss
as recommended by our professors". Invitations to tea parties can be
expected.

Transmission loss, of course, is that defined by the ratio ( Pout ) / (
Pin ) of a network or a single section of a network. Overall loss being the
product of the ratios of individual sections. The transmitter's performance
is usually unknown, it can be considered to be external to the system and
omitted from an analysis.

Insertion loss applies to individual sections of a number of cascaded
sections. It is the *change* in overall loss which results from inserting
the individual section in a cascade. Numerically it is the difference in
dBs between 'before' and 'after' the insertion.

Consider the tuner, an impedance-transforming network. It is clear the
insertion loss of a correctly operating tuner is in fact a *Gain* and
differs considerably in dBs from overall system loss before it was inserted.

There is an excellent small book which I once read "Transmission Line
Transformers" (baluns wound on ferrite cores), which contains many
transformer measurements of loss versus frequency. The author states
explicitly all measurements are in terms of the loss when a transformer is
*inserted* between specified generator and load impedances. I venture to
guess not a single amateur has ever considered the additional (transmission)
loss due such a transformer which may occur when both the generator and
terminating impedances are unknown or at least are considerably different
from the standard resistances between which the transformer has been
designed.

The moral is - don't mix the two sorts of loss when trying to analyse how a
transmission system works. Don't use insertion loss formulae anywhere in a
system when the transmitter's internal impedance is not known. To use
Scattering Matrix and circulator parameters, intended to simplify
calculations at UHF and microwaves, is asking for trouble and disagreement.
As has been amply demonstrated.

On the other hand, mention of transmitter internal impedance may force an
analysis into considering the insertion loss of the feedline which I'm sure
would not be welcome. If anybody knows what it is for rice-cakes don't tell
anybody.
----
Reg.