January 1967 QST
Table
of Contents
Wax nostalgic about and learn from the history of early electronics. See articles
from
QST, published December 1915  present (visit ARRL
for info). All copyrights hereby acknowledged.

Computer modeling of antenna radiation
patterns has evolved from a relatively simple electric field equation that diminishes
as the inverse of the distance from the source, to exotic, highly sophisticated
numerical methods that account for conducting and dielectric surfaces and volumes.
A spreadsheet can be built rather quickly to calculate and graph the freespace
azimuth and elevation efield patterns for a 1/4wave whip or a dipole antenna using
textbook formulas, but building a model for displaying the 3D radiation patterns
of a cellphone placed next to a human head, or a UHF radio antenna on top of an
aircraft takes some pretty serious computing power. In large part we owe a debt
of gratitude to the Ph.D. types who have labored hard to make such tools available
to us commoners. As with PCB layout software and circuit simulators, chances of
success with a first pass prototype has increased significantly as software has
improved. Even with the advent of nearmiraculous software, there is still a need
to verify empirically that the design matches the predicted results. That's where
taking physical measurements comes in. Fortunately, there is a lot of great software
for automating testing as well, but occasionally, especially for the less well funded
amongst us, measuring points manually is required. Even with automated systems at
your disposal a few quick "sanity check" points are measured prior to submitting
the design to a full barrage of tests. This article describes how the author first
calculated the expected pattern for a 1/4wave vertical whip antenna mounted on
a car, and then went about validating his predictions.
Modeling Radiation Patterns of Whip Antennas
Fig. 1  Coordinate system for the model. Groundplane radials
and vertical element are 1/4 wavelength long.
By Dale W. Covington, K4GSX
The rapid development of efficient transceivers and power supplies has solved
many of the problems of going mobile. Furthermore, old Sol is playing a strong supporting
role by improving propagation conditions on the very bands for which the mobile
antenna is most efficient. While the bandwidth and efficiency of the whip antenna
have been extensively studied and improved, any description of the resulting radiation
patterns has received only light treatment. Such patterns would be useful guides,
for example, in calling DX or in beginning to conclude a contact before making a
major change of route direction. Therefore the intent of the following note is to
call on stage yet another actor portraying a simplified picture of mobile whip radiation.
Actually it is a complicated matter to describe this radiation precisely as a
function of the total elliptical polarization of the radiated Efield, the distorted
currents flowing on sculptured car bodies and loaded whips, the frequency dependence
of the ground conductivity, and so on. On the other hand, the principal features
can be exhibited by using a model of a vertical element over an incomplete ground
plane.
Scalemodel car used for obtaining data plotted in Fig. 3B.
The Model
Employing a model for a complicated analysis usually implies a certain degree
of approximation. The case in point is no exception. Cars are about 1/4 wavelength
long at 20 meters and almost 1/4 wavelength wide at 10 meters. As the whip itself
is a 1/4 wavelength at 10 meters, it seemed appropriate to restrict the analysis
primarily to this 10 to 20meter range. Fig. 1 shows the general shape and
the coordinate position of the model, which had ten 1/4wave ground radials from
0 to 90 degrees beneath a vertical 1/4wave element fixed in the normal180degree
plane. The ground plane was spaced 1/10 wave above ground. Crudely speaking, the
model thus represented a car with a whip mounted on the left rear deck. The driver's
side is along the 0degree direction, and the rear bumper is along the 90degree
direction.
Patterns
The actual calculation of the patterns consisted of computing the far Efield
from cosinusoidal currents flowing on 1/4 wave elements^{1} as arranged
in Fig. 1. All of the resulting vector fields were then added to yield a polar
plot of the radiation patterns as a function of the angle of elevation. Since an
actual whip does not remain truly vertical once the car starts moving, the equations
for the model were solved for the vertical element normal to and tilted away from
the ground radials.
The close spacing between the model and ground requires that ground effects be
included in the analysis. A review of the interrelations between frequency, antenna
height above ground, angle of elevation, and ground constants has been given by
G3HRH^{(note 2)}. Using standard techniques^{3} the Efield expressions
were corrected by the ground factors for 28 Mc. and angles of elevation, Δ,
of 15 and 30 degrees. Higher wave angles are less useful for contacts from 14 to
28 Mc.^{4}. These ground factors revealed that, at their maximum point,
the horizontally polarized Efields from the model over dry soil were 11.7 and 5.9
db. below the corresponding vertically polarized fields for Δ of 15 and 30
degrees respectively. As the conductivity approached seawater values, the horizontal
terms were even smaller: namely, 14.2 and 8.3 db. For simplicity only vertical terms
were retained in the patterns.
Fig. 2A  Calculated patterns of relative Efield strength
for a radiation angle of 15 degrees above the horizon; dry soil. Solid curve, whip
vertical; dashed line, whip tilted 45 degrees.
Fig. 2B  Some for a radiation angle of 30 degrees.
Fig. 3A  Solid points, experimental data taken on model
antenna system shown in Fig. 1, at a frequency of 430 Mc. Open points, 14Mc.
data taken on actual automobile installation.
Fig. 3B  Solid points, experimental data on scalemodel
car shown in photograph, at 430 Mc., whip vertical. Open points, same with whip
tilted 45 degrees.
The patterns of the calculated Efields are presented in Fig. 2 for the
15 and 30degree wave angles. Solid lines show the fields from the vertical element
normal to ground while the dotted lines denote a rather extreme element tilt of
45 degrees. The relative field strengths can be directly compared from one wave
angle to the other; however, directly comparing field values of the normal and tilted
configurations automatically implies a constant input current. It is immediately
noted in Fig. 2 that the quadrant containing the ground plane also contains
the strongest fields. Moreover, these fields generally change only slightly from
0 to 90 degrees.
When the vertical element is perpendicular to the radials, the field pattern
is symmetric about the 45225 degree directions. Here orientation is more important
at the high elevation angle where the pattern undergoes a maximum/ minimum variation
of 6.4 db. compared to a 3.3db. variation at the lower angle. As mentioned before,
the attractive increase in field strength at the higher angle usually cannot be
advantageously employed on the higherfrequency bands.
Pattern symmetry becomes lost as the vertical element tilts back from the normal.
Not only does the direction of maximum field shift from 45 degrees toward 20 degrees,
but also the field strength from the rear of the model is particularly reduced.
Numerically the fields in front of the model are 8.3 and 13 db. stronger at the
15 and 30 degree elevations.
LowFrequency Considerations
Mobile operation on 40 and 80 meters is more difficult to analyze. Even in Texas
cars and whips don't come equipped with 1/4wavelength dimensions. Instead, the
sizes of both the car body and the whip approach small fractions of a wavelength.
Also, in this range the loading coil becomes increasingly important in relation
to the current distribution on the whip. Finally, contacts can be made on these
bands by radiation at fairly high angles of elevation, which complicates the previous
polarization argument by filling in certain parts of the pattern with a significant
combination of vertically and horizontallypolarized fields.
The relative directivity pattern for a very small dipole has only a slightly
greater beam width than the similar figure8 pattern for a halfwave dipole having
1/4wave elements^{4}. Thus it would be reasonable to expect that the character
of the patterns of Fig. 2 would be more nearly omnidirectional because of
the short length of the radiating elements. Consequently this factor along with
the increased usefulness of the higher angles of elevation would reduce any directivity
effects for 80 and 40meter mobile contacts.
It is interesting to speculate about the patterns predicted by the model for
an incomplete ground plane installation at a fixed station. On the lower bands,
particularly, it is not always practical to extend the long ground radials in a
symmetric shape about the base of a vertical antenna. The model should be useful
in understanding such cases if the obstruction limiting the ground plane to less
than a circle does not likewise prevent the vertical element from being installed
in the clear. For example, the patterns for a vertical installed at one corner of
a garden would probably differ from those for a vertical next to the corner of a
house, even though both conditions might have a 90degree area that was unavailable
for ground radials. Basically, the model suggests that a hole or depression exists
in the radiation pattern centered in the area having no radials. Directly opposite
the hole is centered a broad field maximum over the ground radials. The hole is
a function of the angle of elevation, and its maximum depth is of the order of 6
db. or so below the field in the opposite direction at elevations near 40 degrees.
Naturally the hole width could be greatly reduced as the area about the base is
more evenly covered with radials.
Experimental Results
The computed patterns were subjected to several checks. One check utilized an
experimental model of Fig. 1 at 430 Mc. The wire model was located about three
wavelengths from a twoelement beam fed by a 6J6 rig from an old Handbook design.
The detector was a 1N23 crystal operating in the squarelaw region. The measured
Efield pattern is given in Fig. 3A for the vertical element perpendicular
to ground and an angle of elevation of 15 degrees. There is good general agreement
with the calculated pattern of Fig. 2A. Tilting by 45 degrees produced a maximum/minimum
gain of 5 db. An increase of radiation in the forward directions was noted at higher
elevations.
Of course the primary reason for examining the incomplete ground plane model
lay in the degree that it approximated 1020 meter mobile radiation. Included in
Fig. 3A is a mirror image (whip mounted on right rear fender) of some 20meter
Efield data taken on the mobile installation of WA4KQO. While the receiving antenna
was higher than the whipHillman combination, the angle of elevation unfortunately
was not measured. It was less than 5 degrees. The experimental points are characteristic
of the lowangle radiation from the model.
To further confirm the effects of tilting the whip away from the normal, a 1/15.4
scale model of a Toronado was constructed. At this scale, the 430Mc. whip was equivalent
to a 1/4 wave whip on 10 meters. An aluminum foil skin 0.00125 inches thick covered
the balsa stringer shell. Fig. 3B presents the measured field strengths at
a 15degree wave angle. Input power remained constant as the whip was tilted. Again
comparing the experimental data with the curves of Fig. 2A, it is apparent
that the groundplane model does agree fairly well with the scale model. Indeed,
the standard deviation of the measured points was 1.2 db. for both the normal and
tilted conditions.
Conclusions
The radiation patterns of a 1/4wave ground plane model have been employed to
approximate the patterns from a mobile whip in the 1020 meter range. For these
bands the operating experiences of several mobile hams indicate that the field strength
over the car body is on the average 3 to 6 db. stronger than the field in the opposite
direction. This magnitude and direction are confirmed by the model. The model also
predicts that the patterns are more directive when the soil conductivity increases,
when the contact is by means of a short skip, or when the whip curves back from
the normal at high speeds. Low angle DX work is less sensitive to groundplane orientation.
Large variations from the patterns could arise from field distortions produced by
nearby objects, poor electrical contact over various parts of the car body, and
a bumper mount instead of a deck mount. A tilted ground plane instead of a horizontal
ground plane would be a more accurate model in this latter case. The net effect
would reduce the fields over the ground plane and increase the fields in the opposing
quadrant.
In addition to the references listed, helpful ideas are gratefully acknowledged
from two other sources: first, from conversations with K8MBV and a number of other
mobile hams, and second, from the pleasant and informative hours spent in assembling,
testing, and operating mobile equipment with W A4KQO.
1 King, Theory of Linear Antennas, University Press, Cambridge,
Massachusetts, 1956, p. 395, 421, 687.
2 Hills, "The Ground Beneath Us,"
R.S.G.B. Bulletin, June 1966, p. 375.
3 Schelkunoff and Friis, Chapter
Seven, Antennas/Theory and Practice, John Wiley and Sons, Inc., New York. 1952.
4 Chapter Two, The A.R.R.L. Antenna Book.
Posted November 25, 2020 (updated from original post on 12/6/2013)
