Roy,

I have spent a lot of time exploring different modelling options over
recent weeks.

One view that one might take re my fig a) is that at connection of the
stub with the main vertical, the stub offers low impedance to common mode
current and high impedance to differential current. It leads to thinking
of it as a kind of mode trap that guides the system into in-phase
operation.

I have played around with ways of trying to represent that without using
the wire segments of the stub.

One method was to place a transformer with only one centre tapped
winding. The top and bottom of the winding connect to the upper half wave
and the lower quarter wave respectively, and the centre tap connects to a
horizontal quarter wave. My thinking was that this structure provides low
impedance to common mode current on the horizontal section, but a high
impedance to differential input to the top and bottom of the transformer
winding.

The model achieves reasonably good in-phase operation, but works best
with about 0.35 wave horizontal. I have used an NT card to insert the
transformer windings in the two segments. Here is the deck.

CM
CE
GW 1 15 0 0 0 0 0 5 0.005
GW 2 15 0 0 5 0 0 15 0.005
GW 3 15 0 0 5 7.2 0 5 0.005
GE 1
NT 1 15 2 1 0 0.01 0 -0.01 0 0.01
GN 1
EK
EX 0 1 1 1 0
TL 1 15 2 1 100 0
FR 0 0 0 0 15 0
EN

I then tried changing the horizontal section to two opposed radial wires,
and found that worked well with each radial being about 0.2 wave long.

CM
CE
GW 1 15 0 0 0 0 0 5 0.005
GW 2 15 0 0 5 0 0 15 0.005
GW 3 15 0 0 5 4 0 5 0.005
GW 4 15 0 0 5 -4 0 5 0.005
GE 1
NT 1 15 2 1 0 0.01 0 -0.01 0 0.01
GN 1
EK
EX 6 1 1 1 0
FR 0 0 0 0 15 0
EN

One can achieve similar outcome by wiring an appropriately phased zero
length TL between the segments each side of the horizontal wire.

If these models indicate that the common mode path on the horizontal wire
is important, one loses control of the length of that in the case of the
coaxial configuration because there isn't an o/c end indpendent of the
vertical conductor.

The coaxial construction gives the opportunity to create a high impedance
to differential current between the adjacent segments, but lacks the
ability to create a low impedance common mode path independently of the
vertical structure.

Thoughts?

Owen