Reg Edwards wrote:

You've told us about radiation from the connections to the generator
and
the termination.

Now tell us about radiation from the line.

=================================

Ian, you are falling into the same sort of trap as old wives who
imagine most radiation comes from the middle 1/3rd of a dipole because
that's where most of the current is.

It is self-misleading to consider the various parts of a radiating
system to be separate components which are capable of radiating
independently of each other. They can't.

Actually they can, because that isn't the same as saying...

A system's behaviour must
be treated as a whole.

That is true, of course. Every component of an antenna (or in this case,
a parallel-wire transmission line) interacts with every other component.
The totality of those interactions is what determines how the RF voltage
and current will distribute themselves along the wires.

But once you know the magnitude and phase of the current in each small
segment of the antenna (which need not depend on theory or modeling - in
principle you could go around and measure it) then you have taken
complete account of the interactions. The radiated field from the whole
antenna is then the sum of the fields from the individual components
radiating independently.

However, we weren't originally talking about that...


We have already discussed that the power radiated from a generator +
twin-line + load is a constant and is independent of line length.

No, you have only asserted that.

Total power radiated is equal to that radiated from a wire having a
length equal to line spacing with a radiation resistance appropriate
to that length. The location of the radiator, insofar as the
far-field is concerned, can be considered to be at the load. The
current which flows in the radiator is the same as that flowing in a
matched load. And the load current is independent of line length.

Only if there are no radiative losses from the line itself - and you
have only asserted that, not proved it.

Mathematically, the only way for the total power radiated to remain
constant and independent of line length is for zero radiation from the
line.

Well obviously - but that is a circular argument, based entirely on your
assertion that the power delivered to the load is independent of the
line length.


In summary, the system as a whole BEHAVES as if there is NO radiation
from the line itself - only from fictitious very short monopoles (or
dipoles?) at its ends.

Sorry, but the "behaves as if" argument doesn't wash, because those
short monopoles are real. Since the line spacing is non-zero, those
short transverse sections must always exist, both in practice and in
your circuit model. Each section carries RF current, so it radiates - no
question about that, but it is entirely an end effect. It has nothing
whatever to do with radiation from the main line.

Looking edge-on at the line, we have two conductors carrying equal and
opposite currents, but one is slightly farther away than the other so
their transverse radiated fields do not quite cancel out.

The only question is mathematical: how does the small loss of energy
through radiation translate into a dB/m or dB/wavelength loss along the
transmission line?



--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek