[FoxHunt] gamma-matched quad antennas

[Dale Hunt, WB6BYU]

I've used a gamma-matched 2-element quad for transmitter hunting
for many years, and we've built a bunch of them for other folks
as club projects in that time.  This started as the first two
elements of the stiff-wire quad from Joe's book, later adjusted
for better F/B ratio after following the rear lobe a couple
of times.  Now I use the W4RNL optimized quad dimensions
(http://www.cebik.com/radio.html - including a spreadsheet
where you can enter frequency and wire diameter and it will
spit out 2-, 3- and 4-element designs.  Ideal for when you
need a quick antenna to DF on an odd frequency.)  I use insulated
wire for the gamma and wrap the end around the driven element
wire to form an inexpensive, weatherproof variable capacitor
that can be adjusted by unwinding turns, and the wraps can
be slid up and down the wire to find the right tap point.

In the last few months I've been modelling antennas with EZNEC,
including the quads.  It is great for optimizing designs for
maximum F/B ratio or minimum half-power beamwidth.  The quads
all looked good until I tried actually modelling the gamma
match to see if EZNEC would give me the same length of the
gamma rod that I had found experimentally.  Yes, within a
quarter inch or so.  But the pattern looked a bit odd.

Looking more closely at the currents in the elements showed
that the polarization had shifted - it was now about 45
degrees from vertical, as if it were fed in one corner of
the element rather than on the side.  Modelling programs
aren't perfect, of course, so I hooked up a Microhunt
transmitter directly to the BNC jack on one of my quads
(to minimize radiation from the feedline) and checked
the polarization with a yagi and my VK3YNG sniffer.  Sure
enough, two different quads showed skew polarization,
with the horizontal component slightly stonger than the
vertical.

Did the original design have this same problem?  I modelled
Joe's original 4-element design and, sure enough, there is
a slight skew, but much less noticable than for my 2-element.
There are two reasons for the difference:  first, there is
less difference between the horizontal and vertical half power
beamwidths of a beam as it gets longer, so any rotation is
less noticable with the longer quad.  Secondly, the gamma
wire is about half as long on Joe's quad than on the single
quad element that I've been using for investigation.   The
maximum current point appears to be close to the top of
the gamma rod, so a shorter rod (to match a lower impedance)
ends up with less skew.  But it looks like it may be 30
degrees or so on the original, as nearly as I can judge
from the plotted pattern.

I'm not finished with my explorations yet - clearly there
is something odd about the gamma match.  A balanced feed
such as a T or a Delta seems to give a clean pattern and
may be a better approach.  I've often used the delta with
a 4 : 1 coaxial balun, both on quads and yagis, and that
seems to give the expected pattern.  

Clearly such quads still work for direction-finding, but
may have some unexpected quirks in the pattern.  For
example, two such quads pointed at each other may be
cross-polarized, but properly alligned when one is
pointed the other way.  This does funny things to the
effective F/B ratio!

Has anyone else noticed this effect?  Anyone have a quad
that they can test and share the results?  I like having
a continuous driven element without a break at the feed
point, as it makes it mechanically stronger.  So the next
step is to come up with some sort of matching method that
is easy and cheap like the gamma match but doesn't rotate
the current in the driven element.

Any thoughts?

    - Dale   WB6BYU