Hi Roy,

Roy Lewallen wrote in
treetonline:

Owen Duffy wrote:
Roy Lewallen wrote in
treetonline:

Owen Duffy wrote:
...
Is NEC capable of modelling the configuration shown at
http://www.vk1od.net/lost/King-22.3b.png (which is the same type
of problem as my figure b)?

A point made by King is that if the three half waves are in phase,
radiation resistance will be quite high (one third current required
for same distant field strength), around 316 ohms against 105 ohms
for three half waves not-in-phase. Presumably these figures are for
free space. . . .

I looked up the section in King, Mimno, and Wing and was pretty
disappointed. It's one of my favorite references, and I usually find
the explanations clear. But the description of that antenna is pretty
vague, with considerable hand waving ("[Operation of coaxial stubs] is
much less satisfactory than that with the open-wire stubs. . ."
without explaining why). And in the explanation of the open-wire

This lesser mortal was encouraged that he noted the difference, but there
really was no explanation. My feeling is that to note the difference but
to be unable to explain it, other than nebuluous conditions like the
coaxial tubes must be large diameter ratio, is incomplete... a problem
yet to be solved.

I have come to the conclusion that the coaxial tubes are not simply a
relocation of a TL as popularly explained. Over the years, I have
accumulated a few projects that were works of art, but didn't work
properly... and they all used coaxial phasing sections.

stubs, the authors seem to state that the wires must carry purely
differential currents. And their models (Fig. 22-4) do show purely
differential coupling from the antenna to the stubs.

I speculate that they really didn't understand how these antennas
worked, had discovered that the coaxial sleeve versions didn't work or
at least didn't work as well -- and didn't show the proper impedance
--, but didn't fully understand why. King, in particular, was and is
one of the giants of antenna theory, and leaves us a lifetime of
brilliant insight and rigorous mathematical analysis. But at least at
the time that book was published, they lacked the modeling tools we
have today.

Understood... but, I think after our discussion on this, NEC is not up to
the task, it may take a more advanced EM field modelling tool.

My suspicion is that NEC's shortfall is that a TL element does not
properly represent the coaxial stub and its interaction with the other
elements near resonance, though well away from resonance, it is possible
that it may be quite ok. King raises the issues of diameter ratios, and
the difference with whether the stub is inboard or outboard of the o/c
end... but it is not resolved quantitatively.


This effect is certainly observable in models using my Fig a) (though
half the respective resistances due to the vertical over perfect
ground).

The feedpoint impedance looks like it might provide a hint as to
whether currents are actually in-phase.

It surely does. Given the currents on and locations of the end wires,
the modification to the center wire can be calculated from mutual
coupling considerations. And I think this is a clue that led King,
Mimno, and Wing to conclude that something was amiss with the coaxial
version.

Exploring that thought, an example (to some extent) of King's Fig
22.3b is the W5GI Mystery Antenna (see
http://www.w5gi.com/images/w5gimster...aschematic.gif ) which
claims to be three half waves in phase at 14.2MHz. It is very similar
to the diagram above in King though I note that the phasing sections
are 105° in electrical length.

The W5GI is fed with a half wave (at 14.2MHz) of 300 ohm line, then
34' of RG8X. W5GI reports impedance looking into the RG8X as
42+/-j18. That suggests the load on the RG8X is 31+j2 or 70-j18. The
feedpoint impedance should be about the same value due to the half
wave of 300 ohm low loss line. Neither impedance is within a bull's
roar of 316+j0, and are so low as to question whether the three half
waves are indeed in-phase. (The highest impedance that would yeild
42+/-j18 on a short length of RG8X would be around 80+j0, closer to
the not-in-phase configuration than the in-phase configuration).

W5GI's reported feed impedance seem inconsistent with three half
waves in phase, and questions whether the phasing arrangement works
as suggested.

Thoughts?

I doubt that it does.

Now W5GI does introduce his antenna with the statement "A multi-band wire
antenna that performs exceptionally well even though it confounds antenna
modeling software".

I know that is almost always a harbinger of bunk, the proverbial "Danger
Will Robinson...", but in fairness, it does appear that one modelling
package, NEC, cannot adequately model the coaxial arrangement near
resonance, though in his antenna, the coax section would be resonant
around 12MHz and King suggests it ought to be much shorter (resonant well
above 14MHz).

That is not to say there aren't other BS warnings in the W5GI explanation
of operation, or claims of performance.

Thanks for your comments, I find this an interesting subject.

Owen