Yes, I was surprised when I happened onto the same discovery. I think
it goes against intuition, but it's not really surprising once you look into
it. A good way to look at is to first consider a phased array of two close
spaced parallel dipoles where each dipole is fed separately through a
phasing harness. If the phasing harness produces a good current source at
each feedpoint, then the feedpoint sources will actually force the rf on
each dipole to a particular amplitude and phase relationship. If you are
careful to have equal current going to each antenna, and you get the phase
difference between them right, you will have a very nice sharp null off the
back and gain off the front.
It's important to remember that this directive pattern is really just
the sum of two the separate dipole patterns of equal amplitude, originating
from slightly different points in space and having a certain phase
Now suppose you change the length of one of the elements by 5%. You
will find that there is almost no detectable change in the overall array's
pattern. This insensitivity happens for two reasons. One reason is that
the pattern of a dipole is not very sensitive to small changes of length of
the dipole. So changing the length by 5% does not change that elements
pattern contribution to the whole antenna. The other reason is that the
phase and amplitude difference between the two elements of our example is
being forced by the current source feedpoint of each element.
Given the phased array example, it's easy to see why small changes in
the lengths of the driven elements does not change the overall pattern very
much. So now we are faced with the opposite question from what you stared
with? Why then is the pattern of a yagi so sensitive to small changes in
the length of a parasitic element? So, switch the example antenna over to
yagi mode, feeding only one element. The resulting antenna pattern is once
again the superposition of two dipole patterns with a particular amplitude
and phase relationship.
The difference here is that the rf in the parasitic element is no longer
being force-fed. Instead, the parasitic element is sucking off energy from
the field of the driven element and re-radiating it. The amplitude and
phase of that re-radiated signal is now very much subject to the geometry of
the parasitic element. Move it close and it gets more energy, move it
farther away and it gets less. Since it is a "near-resonant" element the
phase of the rf that it reradiates is highly dependent on any reactance in
the element (which in turn is highly dependent on element length). So small
changes in length have a large effect on the phase of the reradiated energy
as you go through resonance. And since the reactance is a function of the
length of the element, therein lies the source of the very high sensitivity
that a yagi pattern has to changes in the length of a parasitic element.
The reason why this shows up especially off the back end of the pattern
is because the null off the back is the result of the subtraction of two
strong dipole patterns. It's much like the difference of two very large
numbers that are very close in value. Since the difference is very small,
small changes in one number produces very large changes in the difference.
In summary, a directive antenna is formed by superimposing the patterns
of individual elements. In a yagi, the shape of the pattern contributed by
any one of the elements is not very sensitive to changes in length.
However, if the element is a parasitic one, the phase of its pattern
compared to the other elements is very sensitive to length changes.
Al, since you are giving a talk on yagi's tonight, I am assuming that you
already know 99% of what I just described. But I got caught up in
explaining it anyway, since its an interesting question. I hope your talk
went well. And I hope at least one person was edified by my use of this
Dudley - WA1X
Date: Wed, 12 Jan 2005 09:24:31 -0800
From: "Al Williams" <email@example.com>
Subject: [TowerTalk] back to antenna subjects
To: "towertalk" <firstname.lastname@example.org>
Content-Type: text/plain; charset="Windows-1252"
Tonight I will be the presenter of the program at the Tacoma Radio Club.
My topic will be antenna modeling (using EZNEC). I intend to include the
results of a discovery that I made a couple days ago and will appreciate
comments from this group if I may have made an "errant"
I modeled an 80m 2 element wire yagi. I chose wire lengths and spacing to
emphasize how the gain pattern can undergo a 180 degree reversal when the
frequency is changed from one end of the band to the other end.
My discovery was that the length of the driven element has little to do
with the gain pattern. Even shortening the driven down to 20 feet or
lengthening it to 100 feet made little difference in the gain pattern.
After thinking about it, it seems quite reasonable to me that the driven
element is primarily just a source for the frequency and phase relationship
to the parasitic element to determine the pattern (except that the driven
element length provides a multiple location source but that has much less
influence on the pattern).
This "discovery" is probably common knowledge to many Towertalkians, but I
would appreciate comments, especially If I misunderstand what EZNEC has
shown, before tonight's presentation.
See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless Weather
Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any questions
and ask for Sherman, W2FLA.
TowerTalk mailing list