[TowerTalk] modeling the 402CD

[L. B. Cebik]

The discussion of results from modeling the 402CD deserve some comment,
because the antenna is not an easy one to model on NEC2.  First, using
EZNEC, with its stepped diameter correction feature, can be misleading
because the feature is not invoked by the program.  The structure of the
antenna, with the spikes just past the loading coils, does not meet the
criteria for automatic employment.  Consequently, stepped diameter errors
are possible.  Second, the values of the loading coils have been handed
down--some reasonable assurance that they represent factory spec values or
correctly measured values is needed.  Something similar is needed to
verify the Q (or series resistance) values also handed down.

The model received here was not satisfactory in terms of segmentation and
hence would not converge within a reasonably low number of segments.
Failure of convergence generally indicates that the results are not wholly
trustworthy, although it does not say that they are necessarily
inaccurate.  Equalizing segment lengths, with a consequent large increase
in the total number of segments in the model, did bring about convergence
in NEC-4.  However, enough questions remain about input values and final
accuracy of the programs with respect to structural complexities to reduce
my confidence in the results of the model.

Note that this lack of confidence is in a model; NOT in the antenna.

The proper range comparison for the determination of antenna gain at a
site with a height is not the ideal free space dipole gain of 2.15 dBi.
Rather, it is the gain of a comparison dipole at the same height.  That
antenna may run 7.5 dBi or more, depending on height and constitution at
its elevation angle of maximum radiation, and the elevation angle of
maximum radiation for the 2-element Yagi may be 1-2 degrees less (where
the dipole may show slightly less gain).  Whatever the precise test
figures in this overall ballpark, the roughly half S-unit of gain is
significant in many operating conditions, as is a couple of S-units of
F-B, especially on a band where having any such gain and F-B is the
exception rather than the rule. Moreover, your terrain features may be
helping or hindering you more than you know.

Incidentally, heights of 5/8 wl and 1 1/8 wl will generally show gain
maxima while 7/8 wl up will show a gain minimum if you raise most
dipole and Yagi antenna models systemmatically.  This phenomenon is
confirmed by standard grund reflection equations.  For 40 meters, gain
maxima occur between 85-90' and 155-160' with a minimum around 120-125'.
F-B maxima and minima do not correspond to those for gain.  And feedpoint
impedance wanders around in its own third curve.  Significant
maxima/minima wash out around 1 3/4 to 2 wl up.  The lower the height, the
steeper the curves.  So if you want to advertise your home brew antenna as
a true winner by virtue of modeling, be sure to model it at a gain
maximizing height, while modeling your neighbor's home brew antenna at a
gain minimizing height.

Modeling is as much art as it is calculational science.  Complex
structures require uncommon care in modeling to ensure reliable results.
The 402CD models give suggestive results--even some guidance in
modifications and adjustments--but even with NEC-4 they require further
refinement to achieve reliability, especially in comparing such models to
models of equally complex but different structures.

-73-

LB, W4RNL



--
FAQ on WWW:               http://www.contesting.com/towertalkfaq.html
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com
Search:                   http://www.contesting.com/km9p/search