W4RNL performed this analysis of the T-match feed of my 5 element 20M
Yagis that I thought might be of interest to some Towertalkians. Although
LB's work may represent the state-of-the-art in modelling the T-match. As
readers will discover, LB put a great deal of effort into this analysis.
73
Frank
W3LPL
donovanf@sgate.com
---------- Forwarded message ----------
To: <towertalk@contesting.com>
Date: Mon, 3 Mar 1997 07:44:18 -0500 (EST)
From: "L. B. Cebik" <cebik@utkux.utcc.utk.edu>
To: Frank Donovan <donovanf@sgate.com>
Cc: towertalk@contesting.com
Subject: Re: W3LPL 5 element 48' boom Yagi: Tee Match
Frank,
Thought I would update you on "progress" in modeling the Tee match, using
your 20 meter beam at 195' as a test model. Since others might be
interested in some part of these notes, I am distributing them to the list
and will do a quick review.
I modeled your 5-element 48' boom Yagi that you derived from YO with a
direct feed and found the following characteristics of the model. (These
will differ very slightly from earlier figures, since I reran the model
with 400+ segments used to test the Tee match so that it makes a more valid
comparison with other results reported here.) All reports are for your
design frequency of 14.2 MHz using average ground and aluminum elements.
Gain: 16.1 dBi @ 5 degrees Beamwidth: 52 degrees (does not change)
F-B: 23.0 dB
Feedpoint Z: 36.3 + 16.1 ohms
This model does not take into account any adjustments required by
attachment to a metal boom or mounting fixtures.
Next, I added your Tee match, using a 48" 0.5" diameter rod spaced 6"
center-to-center (ctc) to the driven element, with no series capacitors.
This is a surface-to-surface (sts) spacing of 5.125". The results were as
follows (given here for a 0.125 connecting rod on each end):
Gain: 15.31 dBi
F-B: 23.13
Feedpoint Z: 262 + 87.9 ohms
Initially, I thought that the gain anomaly was an artifact of modeling,
possibly caused by approaching the NEC-4 limit of diameter-to-segment
length where the rod joined the connector and the connector joined the main
element (1.25" diameter). However, using connecting pieces ranging from
1/8" up to 1" produced negligible changes in gain or in feedpoint
impedance. The trend was the fatter the connecting rod, the lower the gain
(total under 0.1 dB) and the lower the feedpoint Z (under 5 ohms resistive
and under 10 ohms reactive).
Therefore, I recounted the variables involved in the Tee match to see
which, if any, had a significant effect on antenna gain in the model.
1. Driven element length: Changing the driven element length by 1" on
either end could change the feedpoint Z by over 20 ohms resistive, with
only a small impact on the reactance. Lengthening the element lowers the
resistive component, but raises (slightly) the reactive component.
However, gain change for a 6" change at each end was only 0.01 dB.
2. Rod length: Small changes in the rod length altered the feedpoint
impedance. Decreasing the length by 1" lowered the Feedpoint Z by 6 ohms
and increased the reactance by the same amount. However, gain was
unaffected (<0.01 dB).
3. Rod-driven element spacing: Increasing the rod spacing had a
significant affects upon the feedpoint impedance. Going from 6" ctc to
7.5" ctc with the 0.5" rod at the original 48" length raised the feedpoint
resistance 26 ohms while reducing reactance by 8 ohms. A juggling of
driven element length and rod spacing yielded the following performance:
Gain: 15.5 dBi
F-B: 22.8 dB
Feedpoint Z: 215.7 + 87.55 ohms
Note that varying the spacing provides almost 0.2 dB added gain in the
model, the most significant trend so far. Examining the current patterns
for the various models called to my attention the fact that the current
distribution along the combined Tee rod and parallel section of driven
element is not identical to that of the driven element alone. The
consequence is a difference in the current magnitudes and phases on other
elements such that the differentials of the two configurations do not
match. Therefore, I believe the reduction in gain with the original rod
length and diameter is real, although I would not claim that it is
necessarily of the magnitude reported by the comparative models.
4. Rod diameter: one way to alter the current distribution on an element
in close proximity to another is to alter the relative diameters of the two
radiating pieces. Through a series of stepped models, I gradually
increased the diameter of the Tee rod to 2" while retaining the 48" length
and the 7.5" ctc spacing from the driven element. With these elements, the
sts spacing was 5.875". I took the feedpoint impedance down to about 177
ohms and then adjusted the driven element (+/- 214") to raise it closer to
200 ohms, with these results:
Gain: 16.2 dBi
F-B: 23.0 dB
Feedpoint Z: 196.2 + 29.9 ohms
Note that the use of a "fat" Tee rod restores antenna characteristics to
their level without the Tee match modeled, while more closely approaching
the "perfect" 200 feedpoint impedance (SWR = 1.16:1). The Tee rod is not
just a matching element, as might be a lumped capacitance or inductance:
it is also a part of the antenna radiating system.
Question: Is this a correct analysis or a by-product of modeling
limitations?
The model has all the limitations earlier noted of not accounting for
certain factors going into the construction of all-metal beams. Hence,
absolute values cannot be claimed. This is why I have not given precise
measurements except as check points in the progressions.
However, the progressions are sufficiently clear to warrant physical
experimentation by one or more parties with equipment capable of refining
gain measurements to less than 1 dB over some reliable test terrain.
Normal operating situations are unlikely to yield the degree of reliability
needed to validate or invalidate the modeling progressions.
Likewise, no one should alter any antenna (except in the spirit of
experimentation) as a consequence of these modeling progressions. The
progressions have not been validated by empirical results. At the very
most, they are suggestive of tests, but certainly not of practice.
Interestingly, they are consistent with the results of some earlier
modeling I did with vertically arranged linear loading elements of various
diameters and spacings, where the linear load elements were non-cancelling
and hence contributed a small amount to the driven element radiation.
If anyone has access to experimental data related to these notes on the Tee
match, I would be interested in references, as this set of models was done
out of curiosity as to whether the Tee match could be physically modeled
and not out of normal engineering reviews of literature. Even references
to studies invalidating this progression would be educational.
Hope this is of interest to you, Frank, and at least does not bore the
members of the list. I am placing it on the list because a number of
members, like you, use the Tee match with home brew beams, both large and
modest.
-73-
LB, W4RNL
L. B. Cebik, W4RNL /\ /\ * / / / (Off)(423) 974-7215
1434 High Mesa Drive / \/ \/\ ----/\--- (Hm) (423) 938-6335
Knoxville, Tennessee /\ \ \ \ / / || / (FAX)(423) 974-3509
37938-4443 USA / \ \ \ \ || cebik@utk.edu
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Sponsored by: Akorn Access, Inc. & N4VJ / K4AAA
|