Wire 2

Antennas
Made of Wires
Volume 2
Simple & Effective
A Collection from the Works

of L.B. Cebik, W4RNL
Antennas
Made of Wires
Volume 2
Published by
antenneX Online Magazine
http://www.antennex.com/
POB 271229
Corpus Christi, Texas 78427-1229 USA
Copyright © 2010 by Publisher antenneX Online Magazine. All rights reserved. No

part of this book may be reproduced or transmitted in any form, by any means
(electronic, photocopying, recording, or otherwise) without the prior written permis-
sion of the publisher.
ISBN: 1-877992-86-0
Antennas Made of Wire – Volume 2 4
About the Author
L. B. Cebik, W4RNL, passed away in April 2008. He had written extensively

about antennas and antenna modeling (as well as
other electronics subjects) in most of the U.S. ham
journals, including QST, CQ, Communications
Quarterly, QEX, Ham Radio, 73, QRP Quarterly,
Radio-Electronics, and QRPp. Besides the continuing
series of antenna modeling columns he does for
antenneX (continues through 2010), he also wrote a
column for 10-10 News (An-Ten-Ten-nas) and another
for Low Down (Antennas from the Ground Up). A life
member of ARRL, he served as both Technical and
Educational Advisor. Several years ago, LB joined the
position as Technical Editor for antenneX.
L. B. has published over two dozen books, with works on antennas for both the
beginner and the advanced student. Among his books are a basic and
intermediate tutorial in the use of NEC antenna modeling software and
compilations of his many shorter pieces. Some 30 of these books have been
published by antenneX and listed in the BookShelf at our website.
He was a ham since 1954 and also a life member of QCWA and of 10-10
International. He also maintained a web site ( http://www.cebik.com ) on which he
has placed a large collection of entries from his notebooks and publications
sponsored by antenneX. A PhD and a teacher for over 30 years, he retired as
professor emeritus of philosophy at the University of Tennessee, Knoxville.
antenneX is/was very fortunate, indeed, to have had LB as a member of its
writing team and Tech Editor for some 12 years.
I for one, lament daily at the tragic loss of one of my closest friends.
— Jack L. Stone, Publisher
About the Author

PREFACE
“it’s not just wires anymore, it’s an antenna!”
W
hile numerous articles and books have described various wire antenna
designs, but here is a series of new books from the works of antenna
master, L.B. Cebik, W4RNL (SK). He is known the world over for his
unique ideas about new ways to "bend wires" to get the most out of them. With
LB’s guidance, your success is practically guaranteed. It would be a rare
occasion indeed that any design recommended by this author will not work as
described. One can proceed with that confidence in mind.
This book is dedicated to the design, construction and use of antennas of various
types of wire. The reader can save a lot of time and effort by reading these
books. Then, experiment to your heart's content with an aim toward the goal of
achieving the best signal for your unique environment.
With wire, antennas are very simple and easy to build at a very lowest of cost to
achieve one’s goal. This book will demonstrate a number of designs from
conventional antenna wisdom. How satisfying is it to twist and bend wires
together and make connections only to suddenly discover, it’s not just wires
anymore, it’s an antenna!
One book is not enough to describe all of the best-known and LB’s unique
designs, but we shall continue with this second Volume picking up where Volume
1 left off and progress toward the more complex designs. Volume 3 to follow.
Along with some recommended wires, a pair of gloves and simple hand tools,
wonders will sprout from your efforts quickly. And, with wires, such designs can
be made to fit within the closest of environments. Many tips are suggested about
how to make cramped spaces an asset rather than a liability—and keep your
neighbors friendly as well.
We know the reader, newbie or advanced, will enjoy this book by one of the
masters and have fun in the process!
Preface
Table of Contents
Chapter 27: About the Folded Dipole--------------------------------------------------------7

Chapter 28: The Terminated Wide-Band Folded Dipole-------------------------------39
Chapter 29: Wide-Band Multi-Wire Folded Dipoles-------------------------------------76
Chapter 30: More on Wide-Band Multi-Wire Folded Dipoles------------------------105
Chapter 31: About the Folded Monopole-------------------------------------------------142
Chapter 32: Short Folded Monopoles - Some Basics---------------------------------169
Chapter 33: Short Folded Monopoles – Extended Applications--------------------212
Chapter 34: A Doublet as a Good Single Choice Antenna--------------------------256
Chapter 35: Inverted-Vee Doublet Patterns---------------------------------------------274
Chapter 36: About the All-Band Doublet-------------------------------------------------285
Chapter 37: The Zig-Zag Dipole Doublet-------------------------------------------------316
Chapter 38: The Y-Doublet-------------------------------------------------------------------331
Chapter 39: The Doublet - Trap or Not?--------------------------------------------------347
Chapter 40: Some Notes on Triangles of Doublets------------------------------------353
Chapter 41: 40-Meter Vertical Arrays------------------------------------------------------384
Chapter 42: Wire Elements in Lower HF Arrays----------------------------------------428
Chapter 43: The Collinear-Broadside Endfire Array-----------------------------------454
Chapter 44: The Extended Lazy-H---------------------------------------------------------474
Chapter 45: The Expanded Lazy-H--------------------------------------------------------486
Chapter 46: Explaining the Modern Dipole Curtain Array----------------------------500
Chapter 47: The 2-Element Vertical Phased Array------------------------------------530
Chapter 48: The Terminated Vee-Beam and Rhombic-------------------------------557
Chapter 49: The Controlled Current Distribution (CCD) Antenna------------------569
Chapter 50: A Universal HF Back-Up Antenna-----------------------------------------595
Other Publications-----------------------------------------------------------------------------616
Chapter 27: About the Folded Dipole
T
he folded dipole is a simple antenna to build. However, it has
acquired something of a complex web of correct and incorrect
information surrounding it. The points of this note is to sort
out some of the information, with an emphasis upon what it is
correct to say about the folded dipole.
Fig. 1 shows the essential elements of a folded dipole. It consists of

two parallel wires having a constant spacing, S. Each wire has a
certain diameter, d1 and d2. The ends of the parallel wires are
connected to form a continuous loop. The feedpoint is at the center
of the wire having the diameter d1.
Chapter 27
We can construct a folded from common materials ranging from

house wire to parallel transmission line. With such materials, we
can obtain the pattern shown as a free-space azimuth pattern in
Fig. 2--the same pattern as a single-wire dipole. The folded dipole
is a reliable antenna, meaning that we can get it to work without lots
Chapter 27
of finicky adjustments. Something about the ease of building an

antenna seems to go hand-in-hand with not getting a firm grasp on
why it works.
In calling the antenna a folded dipole, we should note that the term
"dipole" is important to our discussion. "Dipole" is a term that we
use as shorthand for a longer characterization of a single wire
antenna. The 1-wire antenna described is a 1/2 wavelength long,
resonant, 2-pole antenna. The reference to length is obvious. Being
resonant means that the feedpoint impedance will have negligible
reactance and hence be purely or close to purely resistive. Having
2 poles means having two transitions from maximum to minimum
current--in this case stating at the current maximum located at the
center of the antenna.
Now all that we need to deal with is the folded aspect of the
antenna. Folding refers not only to the visual appearance of the
antenna, but as well to what folding does. Folding a single wire
antenna (and thereby doubling the amount of wire needed) creates
a combination of an antenna element and an impedance
transformer. The same principle has been used with other antenna
types. For example, the side-fed rectangle--a good vertically
polarized performer for the lower HF bands--has a low feedpoint
impedance. Doubling the loop with a crossover at the far end from
the feedpoint raises the impedance of the antenna.
Using antenna transformer techniques to raise the impedance of an

antenna does not reduce any losses inherent in the antenna
operation. Loss resistances will also be transformed. These losses
Chapter 27
are not significant with the standard horizontal folded dipole, but
have been a major misunderstanding of its cousin, the folded
monopole.
The Folded Dipole as an Impedance Transformer
Our understanding of the folded dipole has been stunted in part by

our use of only a special case within the range of possible
transformations. By using the same diameter wire for both d1 and
d2, we always end up with a 4:1 impedance transformation relative
to a single-wire dipole. However, we usually have no idea why this
is so. Let's start with the general transformation properties and work
our way back to the special case with which we are familiar.
Relative to a single-wire dipole, the feedpoint impedance will be

transformed upward by the ratio R according to the following
equation:
where the terms S, d1, and d2 have the meanings shown in Fig. 1.
Chapter 27
The log of 2S divided by a wire diameter is a complex quantity that

hides some of the consequences of the equation. However,
consider that if d1 and d2 are equal diameters, then the division of
one log by the other log results in a value of 1. Since 1 plus this
value is 2 and the square of 2 is 4, then for wires of equal diameter,
the impedance transformation ratio is always 4:1 relative to the
impedance of a single wire resonant half-wavelength dipole.
In free space the impedance, the resonant impedance of a single

wire resonant half-wavelength antenna that is center-fed is between
71 and 72 Ohms for highly conductive materials like copper. Hence,
a folded dipole using equal wire diameters for both wires will be
about 284 to 288 Ohms.
Now let's note some other aspects of the equation. There are no
rules against using wires of different diameter for d1 and d2. The
wire diameter values always occur as divisors (below the division
line). Hence, the larger the diameter, the smaller will be the
resulting log term. Therefore, we get the following guidelines
(remembering that d1 is the diameter of the fed wire):
• 1. If d1<d2, then R is always >4.

• 2. If d1>d2, then R is always <4.
However, R must always be >1. That is, 1 is the limit of R as the

ratio of the two log values goes to zero, which would imply an
immensely large value for d2 or an infinitesimally small value for d1.
The result is that a folded dipole cannot be used to reduce the
feedpoint impedance relative to a single-wire dipole.
Chapter 27
So far, we have ignored S. Before taking a log for the numerator

and for the denominator of the fraction in the equation, we must
divide twice the wire spacing by the wire diameter(s). This results in
a different value in the numerator and denominator for each
different wire spacing we choose. Hence, the impedance
transformation ratio will also change with every change of spacing.
There is one exception to this consequence of spacing. If the two

log values result in a value of 1 when we divide one by the other,
then the result will always be 4, regardless of the spacing. Hence,
for the case where both wires have the same diameter, the
feedpoint impedance transformation relative to a single-wire dipole
will be 4:1 for any reasonable wire spacing.
There is a limit to how far apart we can place the wires and still
have a folded dipole. That limit, however, is considerably farther
apart than the limit for having an effective transmission line with
confined fields. It also can be a tiny spacing--just enough to prevent
a short circuit between the wires.
Modeling the Folded Dipole
With exceptions that I shall later note, we can use modern antenna
modeling software to calculate the properties of folded dipoles of
many sorts. For folded dipoles using wires of equal diameter, both
MININEC and NEC will yield very accurate results. Remember that
the term "modeling" is used in the mathematical sense of
calculating antenna properties using equations derived from
Poynting Vectors. Hence, the results are very much more accurate
Chapter 27
than the small rules of thumb formulae we find in many antenna

books.
The trick to modeling folded dipoles is to use many segments. The

end wires connecting the parallel wires are a limiting factor. In NEC,
we want the segment lengths in the parallel wires to be less than a
2:1 ratio in length to the segments in the end wires. In MININEC,
we want to use many segments so that the end corners are not
mathematically "cut off" in the calculation. So for the models in this
exercise, I shall use a frequency of 28.5 MHz with 110 segments
along the length of each parallel wire in MININEC to allow a
perfectly centered feedpoint. 111 segments is required in NEC.
These models will fall well within the calculation constraints of each
program type. However, in all cases, the results will apply to bare
wire.
As a test case, let's look once more at the question of spacing. We

shall use 0.1" diameter copper wires throughout for our initial tests.
This diameter is between #12 and #10 AWG wire. I shall present
both NEC-2 and MININEC results for comparison. (For reference,
the NEC-2 results are from NEC-Win Plus and the MININEC results
are from AO 6.5.)
Let's compare the performance of folded dipoles having 3 different

spacing values. A 1" spacing corresponds to the use of ladder line
of common commercial sorts. A 4.14" spacing corresponds to one
recommendation that we use a spacing of 1/100 wavelength.
Finally, a spacing of 13.8" corresponds to another recommendation
that we use a spacing of 1/30 wavelength. In the table below,
Chapter 27
length refers to the resonant length of the folded dipole, while gain
is the free-space gain in dBi. The feedpoint impedance is given in
standard series R +/- jX Ohms terms.
Spacing Length Gain Feedpoint Z

inches dBi R +/- jX Ohms
1"
Mininec 196.56" 2.10 288.0 + j0.0
NEC-2 196.93" 2.12 286.6 - j0.0
4.14"
Mininec 193.70" 2.10 288.0 + j0.0
NEC-2 194.20" 2.12 287.0 - j0.1
13.8"
Mininec 186.94" 2.12 287.0 + j0.0
NEC-2 187.40" 2.13 285.8 + j0.0
Note that the two calculating systems yield resonant lengths within
about a half inch of each other. As well, the predicted gain is never
more than 0.02 dB apart--a truly insignificant amount. Even the
resonant resistance values diverge by less than 1.5 Ohms. The
systems are certainly consistent with each other.
Nothing in the spacing of the wires of folded dipoles could produce

a difference that would be discernable to the most accurate field
measuring equipment available today. There is no aspect of
antenna theory that can justify a claim that one spacing value will
perform better than another.
Private experience might result in other claims. However, private

experience is fraught with many variables of construction and
maintenance, as well as antenna location variables. However,
equivalently well-constructed folded dipoles of different spacing
Chapter 27
values will perform equally well when placed in identical antenna

settings.
Of course, the actual feedpoint impedance encountered by the

builder will vary with the height above ground, just as the feedpoint
impedance of a single-wire dipole varies with height. The two
curves will show a 4:1 ratio in value, but otherwise be congruent.
Before we leave these folded dipoles, let's note the differences in

antenna length. Each antenna was brought to resonance by
adjusting its overall length. The wider the spacing, the shorter will
be the resonant length. The shortening has two major sources.
First, the classic impedance transformation equation does not take
into account the end wires. With wide spacing, these wires begin to
take up a small part of the antenna length. Second, a 2-wire folded
dipole simulates a fat single wire. Just as single-wire dipoles
become shorter at resonance with increasing diameter values, so
too do folded dipoles with increases in wire spacing.
The handy "468/f" rule of thumb that we use for dipoles is actually
only a crude and often inaccurate guide for wire cutting. The
resonant length of single-wire and folded dipoles will vary with wire
size, spacing (for folded dipoles), and height above ground. If we
turn the matter around and cut the antenna according to the old
guide, then we can expect different impedance values--including
differences in both the resistive and reactive components--as we
change wire diameter, spacing, and/or height above ground.
Chapter 27
Wire Size
Let's sample what happens with different wire sizes. We shall keep
d1 and d2 the same, but change both wire diameters together. For
this set of tests, let's use a spacing of 3" between wires. Again, we
shall look at both NEC-2 and MININEC results.
Wire Size Length Gain Feedpoint Z
inches dBi R +/- jX Ohms
0.5"
Mininec 192.10" 2.10 287.0 + j0.0
NEC-2 193.04" 2.13 285.3 - j0.1
#10 AWG
Mininec 194.54" 2.10 288.0 + j0.0
NEC-2 195.09" 2.12 286.9 - j0.0
#12 AWG
Mininec 194.90" 2.10 288.0 - j0.0
NEC-2 195.31" 2.12 287.2 - j0.0
#14 AWG
Mininec 195.16" 2.10 289.0 + j0.0
NEC-2 195.51" 2.12 287.6 + j0.0
#18 AWG
Mininec 195.63" 2.09 290.0 + j0.0
NEC-2 195.88" 2.10 288.5 + j0.0
All values remain within 0.5% of each other for each set of wire
sizes between the two calculating systems. More significantly, there
is no perceptible difference in performance predicted for the range
of wire sizes.
Although there is a change of resonant length as we change wire

size, it is considerably less than the length changes required by
differences of wire spacing in the folded dipole. The range of
spacing in our tests was 13.8:1, while the range of wire sizes was
Chapter 27
12.4:1, comparable ranges. However, the length range was only

about 1.5" for the wire size differences, but 4.5" for the spacing
differences.
With respect to length and performance, a folded dipole acts very

much like a single fat wire. In fact, a single wire dipole will have a
length of about 199.8" using 0.1" diameter copper wire to be
resonant and have a free space gain of about 2.10 dBi. All of our
folded dipoles are shorter, since all are effectively much larger in
diameter. To have a resonant length equal to that of the #18 AWG
wire folded dipole above (about 195.8"), a single copper wire would
need to be just about 1" in diameter. (Reminder: all tests are at 28.5
MHz for consistency throughout this exercise.)
Antenna Currents and What They Tell Us
One very unhelpful conception of a folded dipole is to think of it

solely as some kind of transmission line. As we shall see, there are
currents within a folded dipole that we can call "transmission line"
currents, but this notion has a very limited application and has
misled any number of antenna builders. We have already seen that
folded dipoles will perform normally with wire spacing values
considerably larger than is optimal for a transmission line (13.8" at
10 meters). That condition will neither void nor abet the interesting
currents that we shall later call transmission line currents.
At first glance, a folded dipole operates in ways distinctly unlike a

transmission line. For example, in a properly functioning
transmission line, at any point along the line, the current
Chapter 27
magnitudes will be equal, but the current phases will be opposite,

that is, 180 degrees apart. If a resonant folded dipole acted as a
transmission line, we should expect to see the same pattern of
current values between the two wires.
Chapter 27
To demonstrate this, we can use the model on the lower half of Fig.
3 to derive the currents along the transmission line. A sample every
20% of the way of a line nearly, but not quite, 1 wavelength long is
instructive. We shall present two sets of current phase figures for
wire #2: onset derived from the modeling convention of
continuously developing the model from left to right, the other from
using the dipole junction as the starting point for both wires.
Currents Wire 1 Wire 2

Distance Magnitude Phase Magnitude Phase
0% 1.001 - 0.1 1.001 - 0.0/ 180.0
20% 0.419 18.1 0.419 18.1/-162.0
40% 0.758 173.5 0.758 173.5/- 6.5
60% 0.861 -174.6 0.861 -174.6/ 5.4
80% 0.270 - 31.1 0.270 - 31.0/ 148.8
100% 0.999 - 0.2 0.999 - 0.2/ 179.8
The modeling convention that runs the transmission line wires in

opposite directions shows essentially the same values for each
point on the line. The convention that starts and ends them in the
same direction shows the 180-degree out-of-phase condition.
To illustrate what we actually encounter with a folded dipole, let us

turn to the upper portion of Fig. 3. The markers represent
percentages of distance from the outer end of each wire inward
toward the center. If we plot the current magnitudes and phases for
a typical folded dipole, we end up with an interesting chart. Let's
use our #18 bare copper wire folded dipole with 3" spacing as a test
case. Current magnitudes are relative to a maximum value of 1.0,
while current phases are relative to a feedpoint value of 0.0
degrees. The first current phase figure for Wire 2 is for continuous
Chapter 27
modeling so the end 2 of one wire becomes end 1 of the next. The
second value presumes a model with both parallel wires starting at
the same end of the assembly.
Currents Wire 1 Wire 2
Distance Magnitude Phase Magnitude Phase
0% (end) 0.256 - 74.5 0.244 -106.4/ 73.6
10% 0.463 - 33.8 0.432 -153.0/ 27.0
20% 0.682 - 20.2 0.659 -166.5/ 13.5
30% 0.855 - 12.4 0.842 -172.6/ 7.4
40% 0.964 - 6.2 0.960 -176.1/ 3.9
50% 1.000 0.0 1.000 -177.9/ 2.1
The chart ends at the antenna center point because the opposite
side of the antenna shows virtually identical current values at the
prescribed points. Although the current magnitudes are comparable
(and would be closer had the wire been without any loss at all), the
current phase values show a curious pattern. Corresponding points
along the wires show similar absolute current phase values, but
they are opposite in sign when both wires are modeled from the
same point (e.g., left to right). The pattern is distinctly unlike a
transmission line that is acting like a transmission line, even with
the far end a short circuit.
A phase pattern similar to the one shown is necessary if the folded

dipole is to radiate. Radiation is simply the ability of the fields that
result from the current levels at each point along the antenna to
expand without limit. This condition is unlike that in a transmission
line, where the fields are confined such that radiation is negligible.
For that condition to exist, the current magnitudes would have to be
equal, and the phase values must be exactly opposite. So we have
Chapter 27
a mystery: how can we make sense out of the current magnitudes

and phases along a folded dipole?
The Transmission Line Inside the Folded Dipole
The pattern of current magnitudes and phase angle hides a small

tale, one about the fact that a folded dipole has two sets of
currents. It is the combination of these two current sets that results
in the readings. One set is the radiation currents (Ir), which should
be (if the tale is correct) quite similar to those on a standard dipole.
The other set is comprised of what some call "transmission line"
currents (It). From the modeled current readings, we can separate
the two sets. All we need to do is take half the sum of the currents
at corresponding points along the folded dipole and we get the
value of Ir. If we take half the difference of the currents on each
wire, we arrive at It. Kuecken pointed this out in his book Antennas
and Transmission lines.
The following table provides the modeled values for a bare-wire

folded dipole resonant at 28.0 MHz. The values given are for the
fed wire (Wire 1), the "other wire" (Wire 2), It (transmission line
current), Ir (radiation current), and the corresponding current value
for a single wire resonant dipole at the same relative distance from
the end. The sampled positions are 10, 30, 50, and 80 percent of
the distance from one end of the antenna toward the feedpoint in
the center. For each entry, the format is current magnitude/phase
angle, where the magnitude is relative to a feedpoint current of 1.0,
and the phase angle is in degrees relative to a feedpoint phase
angle of 0.0 degrees.
Chapter 27
Folded Dipole Currents (with Dipole Currents for Comparison)
Folded Dipole Standard Dipole

Position Wire 1 Wire 2 It Ir I
10 0.3740/-59.38 0.3449/+57.79 0.3069/-89.38 0.1878/-4.59 0.1748/-4.44
30 0.5809/-32.42 0.5437/+27.04 0.2973/-89.38 0.4884/-3.77 0.4771/-3.86
50 0.7769/-19.77 0.7506/+14.68 0.4530/-89.93 0.7295/-2.85 0.7228/-3.06
80 0.9649/- 7.61 0.9584/+ 5.63 0.1109/-89.30 0.9552/-1.01 0.9541/-1.60
There is a very good correlation between the folded-dipole radiation

currents as derived by the simple summing method and the single-
wire dipole currents at the corresponding points along the antenna
length. The correlation cannot be perfect, because the simple
summing method does not take into account the currents on the
end wires of the folded dipole. They are short, but significant. The
current magnitude and phase angle both undergo part of their
continuing change in those wires. Nevertheless, the differences
between the folded-dipole radiation currents and the corresponding
currents on a single-wire dipole are small enough that we should
expect to discover any difference in the radiation strength or pattern
between the two antennas. And, of course, we do not.
Ideally, the transmission line currents should all show a phase

angle of -90 degrees. The very slight offset is due both to the end
wire phase angle changes and to the resistance of the copper wire
used in the test model (composed of AWG #18 wire with a diameter
of 0.0403" along with a wire spacing of 1"). Also ideally, the
magnitudes should be the same at each point, but are not for
similar reasons.
The key element in the transmission line currents is their relative

phase angle--almost perfectly -90 degrees out of phase with the
source current. Hence, the transmission line currents represent
Chapter 27
stored energy rather than expended energy, except for the minute
offset from a perfect -90 degrees. As a result, the radiation currents
consume all of the RF energy supplied to the antenna in the form of
its transformation into indefinitely large expanding electromagnetic
fields. Despite energy storage, there is none left over at the end of
a transmission.
The existence of transmission line currents within a folded dipole

has resulted in a number of erroneous practices based on the use
of transmission lines as transmission lines. For example, some
folks have proposed that we short the folded dipole at a position
equal to a quarter wavelength from the feedpoint outward times the
velocity factor of the parallel line used to form the folded dipole.
This practice remains to be modeled in NEC-4, which permits the
modeler to provide each wire with an insulating sheath with a
specified conductivity and dielectric constant.
The first step in considering antennas made from insulated wire is

to consider the normal range of velocity factors that apply to
antennas (in contrast to those that apply to transmission lines). I
began with a 28-MHz dipole that was 204" long and fed in the
center of the bare AWG #14 (0.0641" diameter) wire. Then I
modeled an identical antenna, but added an insulated sheath with a
dielectric constant of 2.5 (about in the middle of the plastics
materials range used for wire) and a conductivity of 1e-10
Ohms/meter (a very good insulator). I made the insulation about
.047" thick--a goodly insulation. Then I re-resonated the dipole at a
length of 195.66". This yielded a velocity factor for the insulated
wire of 0.959, a typical value for heavily insulated wire. Thinner
Chapter 27
insulation would have yielded higher values--or longer resonant

dipoles. This little exercise gives us something against which to
compare a folded dipole composed of insulated wire.
I went through the same exercise with the folded dipole which we
examined in its bareness: 2 AWG #18 wires separated by an inch
center-to-center. The original folded dipole was 198" long at
resonance. Then I covered the wires with insulation that was also
0.47" thick and re-resonated the assembly. The new folded dipole
was 191.5" long, for a velocity factor of 0.968, slightly higher than
our single wire dipole. In both cases--the single-wire and the folded
dipoles, the feedpoint impedance decreased due to the shortening
of the wires. The single wire dipole went from 72.8 Ohms bare to
67.7 Ohm thickly covered. The folded dipole dropped from 289.2
Ohms bare to 274.1 Ohms thickly covered.
I next took down the current readings on both wires so that I could
calculate the radiation currents (Ir) and the transmission line
currents (It) to see if they corresponded to those in the bare wire
folded dipole. The calculations yielded the following table.
Insulated Folded Dipole Currents (with Dipole Currents for Comparison)
Folded Dipole Standard Dipole

Position Wire 1 Wire 2 It Ir I
10 0.3914/-60.83 0.3531/+57.53 0.3198/-89.89 0.1914/-6.57 0.1748/-4.44
30 0.5946/-34.45 0.5435/+25.88 0.2868/-89.87 0.4922/-5.77 0.4771/-3.86
50 0.7845/-21.40 0.7468/+12.52 0.2241/-89.82 0.7324/-4.87 0.7228/-3.06
80 0.9656/- 8.32 0.9546/+ 2.28 0.0889/-89.49 0.9560/-3.05 0.9541/-1.60
Nothing in the new table distinguishes the currents in the insulated

folded dipole from those of the bare wire version, with the possible
exception of a nearly uniform 2-degree displacement of the
Chapter 27
radiation currents. Allowing for the fact that the technique of

calculation is approximate--due to reason noted earlier--nothing in
the table suggests that we should treat an insulated folded dipole
any differently from a bare-wire folded dipole once each is brought
to resonance. In fact, both the bare and insulated versions of the
antenna show the same gain.
Indeed, the current progression in the insulated wires shows only a

single set of curves each side of the feedpoint position, just like the
progression in the bare wire version. The upshot is that we need
not treat a resonant folded dipole made of insulated parallel
transmission line like a transmission line. We can ignore the
"transmission line" velocity factor and simply adjust the overall
antenna length according to the antenna velocity factor created by
the line insulation. The practice of shorting out a folded dipole at the
point indicated by the transmission line velocity factor has never
shown any evidence of doing anything but shorting out the wires at
that point. Finding the resonant length of the folded dipole will be
challenge enough.
Other Impedance Values
The folded dipole is ultimately simply a dipole with an impedance

transformation mechanism built into its structure. As a dipole, on its
fundamental frequency, it provides all of the performance we
expect from a dipole--no more and no less. It tends to have a
slightly wider SWR operating bandwidth (when transformed to our
feedline value) than a single wire dipole because it acts like a fat
wire. But it remains in performance simply a dipole.
Chapter 27
The impedance transformation possibilities, however, should not be

overlooked. The rules of thumb for transformation ratios that are
more than or less than 4:1 can be useful in some contexts. Before
looking at potential applications, let's first look briefly at the levels of
departure from 4:1 as we systematically vary the element
diameters. We shall use the 3" spacing from earlier samples, but
this time, we shall run each wire through a range of 0.1 to 0.5
inches in diameter--with one wire increasing as the other
decreases.
To perform the modeling for this task, we shall set aside NEC-2.
NEC has a known difficulty in dealing with closely spaced wires of
different diameters. Fortunately, MININEC has no such limitation
and handles the calculation task with ease. We shall list the
impedance ratio calculated by the equation, the resultant feedpoint
impedance, and then the modeled values. This should give us a
quick view as to whether the calculations and models reliably
coincide.
--------- Calculated --------------- --------- Modeled ----------------
Diameter Diameter Z Ratio Feed Z Length Gain Feed Impedance
d1 d2 R=Z Ohms inches dBi R +/- jX Ohms
0.1 0.5 7.01 498 193.34 2.09 493.0 + j0.0
0.2 0.4 5.09 361 193.20 2.10 363.0 + j0.0
0.3 0.3 4.00 284 193.10 2.10 288.0 - j0.0
0.4 0.2 3.23 229 193.48 2.11 234.0 - j0.0
0.5 0.1 2.58 183 193.96 2.11 189.0 - j0.0
Given that the calculations do not account for the end wires, the
coincidence of models and calculations is excellent. Incidentally, in
all models, the end wires were sized to match the smaller of the two
Chapter 27
diameters involved. Moreover, the absence of any perceptible

change of gain in the series of models is notable. However we size
the wires in our folded dipole, it gives us dipole performance. To at
least some degree, this convention accounts for the very small
differences in resonant lengths of the models.
In many beams using the dipole as a driven element, the feedpoint

impedance will be far less than 70-72 Ohms. Values from 10 to 50
Ohms are common, although values above 20-25 Ohms are
preferred in order to reduce power losses from the accumulation of
small resistances at connections. Using a folded dipole with
"designer" values for element diameters and spacing, it is possible
to raise the impedance to match almost any value higher than the
initial feedpoint impedance. One option is to use a low
transformation ratio to arrive directly at 50 Ohms. A second option
is to use a higher value to arrive at 200 Ohms and then to use a 4:1
balun at the feedpoint to return to 50 Ohms with an accompanying
reduction on possible common mode currents on the coax.
Although HF use of folded dipole drivers is rare, at VHF they are
still very popular.
The Folded Dipole as a Doublet
Before departing the land of folded dipoles, we should at least

glance at the potential of the folded dipole as a multi-band doublet.
As a sample, we can look at the performance potential of an 80-
meter folded dipole. I resonated a 132.85' version at 3.5 MHz. Wire
spacing is about 2' and the wire are about 0.3" in diameter. I then
checked the patterns and performance on other amateur bands.
Chapter 27
Numbers are rounded, since we are looking only for suggestive

results.
Frequency Max. Gain Feed Impedance Pattern
MHz dBi R +/- jX Ohms No. of lobes
3.5 2.12 287 - j 1 2
7.0 2.27 5 - j 160 2
10.1 3.31 540 - j 750 6
14.0 3.34 25 - j 330 4
18.1 4.71 480 + j 15 10
21.0 4.26 85 - j 530 6
24.9 4.75 550 - j 310 14
28.0 5.03 225 - j 750 8
The gain figures and the number of pattern lobes coincide with
numbers we would obtain from a single-wire dipole pressed into
multi-band doublet service. What differs is the impedance value set.
The difficulty of using a folded dipole on an even harmonic of the
band for which it is initially resonated lies in the very low resistive
component of the feedpoint impedance. By the sixth harmonic, we
have a value that, while low, is well within the capabilities of most
ATUs. In contrast, the second harmonic impedance of 5 Ohms is
likely beyond the reach--or at least the efficient range--of most
ATUs. The 4th harmonic (20 meters in this sample) might well be
matchable, depending upon ATU design.
Chapter 27
Chapter 27
Chapter 27
Fig. 4 shows the free-space azimuth pattern for 14 MHz, with its
typical 2-wavelength 4-lobe pattern. Fig. 5 presents the 6-lobe, 3
wavelength pattern for 21 MHz. The point of these figures shows up
in Fig. 6, the pattern for 18.1 MHz. At about 2.5 wavelengths long,
the antenna shows both the growing lobes for the 3-wavelength
Chapter 27
pattern and the diminishing lobes for the 2-wavelength pattern--for

a total of 10 lobes.
Using a folded dipole as a multi-band doublet--with parallel feedline

to an antenna tuner--thus becomes a matter of matching rather
than of pattern development. Very low impedances may also be
lossy, thus reducing performance even if a match can be obtained
from a given ATU and feedline length.
For multi-band use, a folded dipole offers no advantage over a

single-wire doublet of the same approximate length. Indeed, in the
final analysis, perhaps the only reason for using a folded dipole is
where the impedance transformation is of special interest, that is,
where it may resolve an antenna design challenge. A secondary
use would be to offer a path to discharge static charge build-up and
thus to reduce one (of the many) noise sources. However, there are
other means to this same goal.
Otherwise, the folded dipole performs just like a fat single-wire

dipole.
Two Variations on the Folded Dipole
In its standard form, we normally feed a folded dipole at the center

of one of its two long wires. Which wire we select for the feedpoint
matters only if the antenna performs a "non-standard" impedance
transformation, that is, has two wires with unequal diameters. There
are variations on this feedpoint position, if the impedance
transformation is unimportant. For example, if we are creating a
Chapter 27
VHF or UHF J-pole using parallel transmission line as our material,

we need not do away with the seemingly unused wire that we
separate from the matching section. Instead, we may connect one
or both ends to the radiating element that is continuous with the
matching section.
There are also two interesting variations of the folded dipole, as

suggested in Fig. 7. We may call them the end-gapped version and
the center-short version.
The end-gapped version of the folded dipole simply omits the end
wire, but only one end-wire. A folded dipole is actually two linear
dipoles in close proximity--close enough that the wires show
Chapter 27
transmission-line as well as radiation currents. The dipoles meet at

high-voltage, low-current points as each end. We may open one of
the high-voltage region contacts with very little effect on the basic
antenna properties, that is, on the radiation pattern and the
feedpoint impedance. We might have to readjust the total length of
the antenna if a resonant feedpoint impedance is important to a
given antenna installation. But we would not change the
performance as a radiating element.
The version of the folded dipole with the center short from one long
wire to the other has a special application. Suppose that we feed
the antenna (very slightly off-center, of course) with a transmission
line for which one conductor forms a common or ground lead. We
might connect the common lead to the short and the other lead to
the long wire. In the process, we do not change the essential
performance properties of the folded dipole. However, we do obtain
a means of connecting the structure to the support mast in the VHF
and UHF ranges. That configuration reduces the likelihood that
surges from electrical storms will be conveyed to the equipment.
Although I have run both types of folded-dipole variants through

numerous models at VHF, let's set them up using our 10-meter
model. The standard version uses two AWG #18 wires spaced 1"
apart. With a total length of 198", the model uses 199 segments per
long wire. The model of the end-gapped folded dipole simple omits
one of the 1-segment end wires, but is otherwise identical to the
standard model. The center-short version requires a small set of
changes. The long wires each become two wires that meet at the
center. Each of these wires has 99 segments. I added a new wire
Chapter 27
from one center junction to the other. I then placed the feedpoint or
source on the first segment of the wire extending from the short to
the end. If my descriptions have been correct, we should expect
virtually identical performance from the three folded-dipole
variations.
Modeled performance of 3 folded-dipole variations in a free-space
environment
Version Gain Feedpoint Impedance
dBi R +/- jX Ohms
Standard 2.10 289.1 + j 2.6
End-Gapped 2.09 287.8 + j 1.4
Center-Shorted 2.10 288.0 + j 9.3
The only way to tell the antennas apart--besides the obvious visible
differences--is to perform a current-sorting exercise on the 3
versions. I did this for some VHF folded dipoles. Fig. 8 shows the
radiation currents along the standard and the end-gapped version
of 2-meter folded dipoles. Because the center-short version has
connections at both ends, the currents do not drop as close to zero
in the end segments as they do with one end of the end-gap
version. (Of course, NEC current reports never go quite to zero in a
linear wire end segment because the effective position for the
current is at the center of the last segment, not its outer end.)
Therefore, the center-short and the standard versions of folded
dipole have the same radiation-current curves (within limits that are
too small to show up in these kinds of graphs).
Chapter 27
Where the currents for the 3 antennas show significant variation is

in the pattern of transmission line currents. Essentially, for all three
versions of the folded dipole, the current phase is 90 degrees from
the phase of the current at the source or feedpoint. However, the
transmission-line current magnitude for the standard folded dipole
shows a symmetrical pattern with its minimum at the long-wire
center and maximum values at the long-wire ends. Fig. 9 shows
this pattern for the standard folded dipole, using a VHF model.
Interestingly, both the end-gapped version and the center-short
versions show virtually identical patterns that vary from the
Chapter 27
standard version. From the feedpoint toward the open end of the
end-gap version or from the shorting bar away from the feedpoint,
the transmission-line currents decrease from their feedpoint region
value toward zero. At the same time, the current magnitude
distribution on the feedpoint side toward the closed end (to the right
on the graph) result in higher current magnitudes--in fact about as
much higher than the standard version as the low end is lower than
the standard version. However, at the center of the antenna, the
transmission-line currents have very comparable values.
Chapter 27
The differences in transmission-line current distribution between

standard and variant versions would make a difference only if we
end-feed the folded dipole, as in the many variations on the J-pole,
and then only in the impedance at the end feedpoint. However, the
center-short version of the antenna is normally fed at the center or
as close to it as may be feasible. Hence, we would not be able on
range tests to tell the difference between the center-short version
and a standard version. Equally, the construction difference used
for the end-gap version would hide itself in range test, which would
show only the radiation patterns and field strengths for any tested
version.
You might note that these notes on the folded dipole do not have a
section entitled "conclusion." There have been multiple updates to
these notes. Each time that I think these notes should end, I learn
something new and interesting (at least to me) about folded dipole
behavior. I have no good reason to think that my latest additions
and revisions should be any different. More on this topic in the next
Chapters.
Chapter 27
Chapter 28: The Terminated Wide-Band Folded Dipole
A
s space for antennas continues to shrink in the present era
of smaller urban and suburban yard, hams have begun to
turn to 1-antenna solutions to their operating needs. Among
the choices for a horizontal antenna that operates on all of the HF
amateur bands, the wide-band "folded dipole" (WBFD) has been
gaining popularity. I thought that it might be useful to do some
comparative studies using this antenna as a base-line.
The basic WBFD looks something like Fig. 1.
The antenna design appears to be a folded dipole. However, a

folded dipole is a resonant antenna, while the WBFD is designed to
operate with a low feedpoint impedance across a wide range of
frequencies. Moreover, the WBFD contains a non-inductive
terminating resistor usually located at the point in the loop directly
opposite the feedpoint. Normally, the resistor is in the 800-900 Ohm
range. This impedance is roughly replicated at the feedpoint.
Therefore, builders install a 16:1 RF transformer (either of
Chapter 28
transmission-line transformer or normal transformer design) at the

feedpoint. The result is a low SWR value for 50-Ohm coaxial cable
across the entire frequency range.
For receiving use, such as in SWL service, the terminating resistor

can be a low wattage carbon type. For transmitting service, the
resistor must have a power value capable of dissipating a fair share
of the applied power. The exact amount will vary with frequency,
but commercial versions of the antenna are often rated for reduced
power at the low end of the operating range, where power
dissipation is highest.
Commonly, WBFD antennas are offered in a 90 to 100 foot length

(27-28 meters) for service between 2 and 30 MHz. However, one
can build WBFD antennas in almost any length. Only the effective
operating range of frequencies will change.
Since we may also construct doublets of the same length and feed
them with parallel transmission line to an antenna tuner, it seemed
fair to compare the gain of such a doublet with that of a WBFD of
the same length across the 2-30 MHz range. The model I chose for
the WBFD is 27.2 m (89.24') long, with the wires separated 0.2 m
(7.8"). The terminating resistor is 820 Ohms, a standard value used
in some commercial models. (Other commercial units use 900
Ohms, often composed of 3 2700-Ohm resistors in parallel.) The
wire is #14 AWG. The doublet is a simple length of #14 copper wire
exactly as long as the WBFD.
Chapter 28
Fig. 2 compares the free-space gain of the two antennas at 1 MHz

intervals from 2 to 30 MHz. Since the elevation angle of maximum
radiation will be the same for both antennas for any height above
ground and for any ground conditions, any differences that show up
in the free-space model will also show up in actual antennas at any
height above ground.
Chapter 28
Several instructive notes emerge from the comparison of gain in

Fig. 2. First, the overall average difference in gain between the two
antennas is nearly 6.3 dB, with the advantage going to the doublet.
If we neglect frequencies below 7 MHz, the average difference
diminishes to 5.0 dB. For most of the range of use of the WBFD,
then, there is about a 1 S-unit deficit in gain relative to a standard
doublet of the same length in the same position.
Second, the WBFD gain curve displays a significant knee--a

frequency below which its gain deteriorates rapidly. In the case of
the current model, that frequency is about 6 MHz. At or below the
knee-frequency, the terminating resistor dissipates more and more
of the power. The result is not only a large decrease in gain and
higher temperature stresses on the resistor, but as well, very low
SWR values at the feedpoint. The knee-effect as the WBFD
becomes significantly short relative to the length of a resonant
dipole easily accounts for the need to de-rate the antenna relative
to transmitting power below a certain frequency.
The deficit in gain is not necessarily a disadvantage for receiving

purposes. Modern receivers tend to be equipped with receiving pre-
amplifiers that the user can switch in as desired. The gain may
range from 10 to 20 dB, depending upon design, and in some
receivers may be stepped or variable. Therefore the gain deficit can
be largely made up in the lower HF range. Moreover, the basic
receiver, apart from pre-amplification, already has excess gain that
is rarely used in the lower HF region.
Chapter 28
In addition, one of the major problems in reception in the lower HF

range, especially with respect to SW broadcast stations, is front-
end overload from excessive signal strength. The overload also
tends to produce spurious products within the receiver. Hence,
reduced gain of the antenna can be in some circumstances an
advantage rather than a disadvantage. Combined with the RF
attenuator built into many receivers--which may be a single
reduction value or stepped--the WBFD offers a potential for
excellent lower HF reception, free of some of the problems that
occur with higher gain antennas.
Because the WBFD is also a closed loop with a terminating

resistor, many users claim quieter reception relative to doublets for
a given receiver input signal strength. The degree to which this is
both true and separable from the freedom from front-end overload
is difficult to determine. Nonetheless, SWLs have found the WBFD
a very useful tool for their efforts.
In order to establish that the WBFD has the same pattern as a

doublet of the same length for any given frequency and height
above ground, let's look at a couple of sample free-space patterns.
For example, see Fig. 3.
Chapter 28
The 27.2 meter WBFD and its comparison doublet exhibit a bi-
directional pattern at 10 MHz. The shape of the pattern is identical,
with only the 6 dB gain differential separating the two antennas.
The -3 dB beamwidth points are also virtually identical. Since the
take-off angle (elevation angle of maximum radiation), the reflection
from a given set of ground conditions, and other such factors are
not dependent upon signal strength, the two antennas would also
Chapter 28
show elevation patterns for any equal antenna height that are
likewise congruent.
Fig. 4 shows comparative free-space azimuth patterns for the two

antennas at 25 MHz. The WBFD pattern is simply a "mini" version
of the doublet pattern, with about a 6 dB difference in strength.
Chapter 28
There is an additional point in displaying these patterns. The exact

pattern of lobes and nulls in the azimuth readings for a WBFD is
identical to that of a doublet. As the length of the antenna exceeds
1.25 wavelength and approaches 1.5 wavelength, the bi-directional
pattern at lower frequencies will break up into a collection of lobes
and a collection of nulls. Therefore, the antenna is variably
selective in its favored directions of good signal strength as one
changes frequency. Those who contemplate installing either a
doublet or a WBFD antenna need to consider well the patterns at
key frequencies of interest in order to orient the antenna for
maximum effectiveness.
The antenna type has also been used vertically to provide omni-
directional coverage. However, in this orientation, when the
antenna exceeds 1.25 wavelength in over length, the pattern
begins to show primarily high angle radiation--exactly the opposite
of what one normally desires from the upper HF band. As a result,
some installations may use a pair of vertical WBFDs for full low-
angle HF coverage.
A Note on Knees and Length
The knee of our 27.2-meter wide-band folded dipole occurs at

about the frequency at which a center-fed doublet of the same
length would be self-resonant, that is about 1/2-wavelength. We
know that a doublet that is shorter than 1/2 wavelength exhibits a
feedpoint resistance that declines with length and a capacitive
reactance that increases with length.
Chapter 28
The knee we observed in the gain of the 27.2 meter WBFD is

interesting, since it suggests that we may vary the low frequency
gain by changing the length of the antenna. Altering the length, of
course, will also change the frequency at which the antenna
transitions from a bi-directional pattern into a multi-lobed pattern.
To examine this question, I recreated the 27.2-m antenna model to

perform frequency sweeps on both longer and shorter versions. As
a sample, I ran a 50-m version and a 15-m version. All of the
models used 820-Ohm terminating resistors, #14 AWG copper wire
and a spacing of 0.2 meters.
Chapter 28
Fig. 5 compares the gain of the three antennas from 2 through 30

MHz in 1 MHz steps. As suspected, the 50-m antenna reduces the
knee frequency to about 3 MHz. In contrast, the 15-m version
increases the knee frequency to about 10 MHz. In general, a home
builder may interpolate values for the knee frequency for other
lengths in the overall range.
Chapter 28
The longest of the antenna models shows a mere -10 dBi gain at 2
MHz, a value easily made up by the receiver and only about 1.5 to
2 S-units below the average gain of the antenna. Hence, it is likely
to be more satisfactory as a transmitting antenna in the lower HF
region. In contrast, the 2 MHz performance of the 15-m version is
more than 30 dB lower than the average antenna performance,
making it more suitable for higher HF transmitting.
The variations in gain among the curves in the relatively flat region
of performance are a function of lobe formation. Maximum gain
tends to attach to the major lobes of patterns taken at just higher
than integral multiple of a wavelength, relative to antenna length.
Minimum gain levels tend to be associated with antenna lengths
near the "x+.5" wavelength (where x is an integer) points. When an
antenna is 1.5, 2.5, 3.5, etc. wavelengths long, its pattern consists
of a combination of emerging and disappearing lobes, all of
relatively equal strength. For example, a 1-wavelength wire has 2
strong lobes that are 180 degrees apart and a 2 wavelength wire
has 4 strong lobes that are roughly 90 degrees apart. A 1.5
wavelength wire has 6 lobes, as the 1 wavelength lobes diminish
and the 2 wavelength lobes grow. Hence, coverage is wide, but at a
reduction in maximum strength.
The number of peaks and valleys in the three gain curves is a

function of length. The 50-m antenna passes through many more
transitions from x wavelength to x.5 wavelength (where x is an
integer) across the frequency span than do either of the shorter
antennas. Hence, we should expect more highs and lows in the
gain pattern.
Chapter 28
One question posed by various recommended wire spacing in past

literature is whether wire spacing makes a difference to
performance. Fig. 5a provides something of an answer as it
compares the gain values for models of a 15-meter long version in
0.2-m and 0.45-m spacing. The gain values are insignificantly
different, ranging from 0.2 to 0.4 dB.
Chapter 28
We find that the curve of SWR relative to the value of the

terminating resistor will also show similar transitions according to
WBFD antenna length. Fig. 6 shows the SWR pattern for the three
antenna models. If we look at the most dramatic fluctuations--in the
case of the 50-m antenna, we discover SWR peaks at x+2/3
wavelength points (where x is an integer). In contrast, we find SWR
minimum values at the x+1/6 wavelength points (where x is an
integer). The frequency span between points relative to the antenna
Chapter 28
length is 1/2 wavelength. The shorter antennas show the same

pattern. However, the pattern is less evident because there are
fewer maximum and minimum values to sample.
We may also note that the longer the WBFD, the higher the SWR
excursions for a given value of terminating resistor. However, if we
examine the lowest values of minimum SWR and exclude the
region below the gain knee of the curve, the corresponding low
points in the curve show the longest antenna also to exhibit the
lowest minimum value of SWR. In other words, for a given wire
size, spacing, and terminating resistor, longer WBFDs will exhibit a
larger range within any given SWR cycle. As we approach the
upper HF range, the values may exceed the desired 2:1 SWR limit.
The amount by which a long WBFD exceeds a 2:1 SWR is not

great, but it is noticeable. For receiving applications, mild
excursions beyond the 2:1 limit have virtually no affect on the
received signal strength for any length of 50-Ohm coax. Some
transmitters use automatic power reduction circuitry as the SWR
approaches 2:1 (using an internal reverse voltage sensor), and
some linear amplifiers begin reducing power at lower levels of SWR
in order to protect expensive transmitting tubes.
There are two means of overcoming the potential problems of

"high" SWR. Some manufacturers recommend the use of very long
coaxial cables. Since the losses in the line increase with frequency,
the SWR observed at the station end of the line will be lower at
higher HF frequencies than at lower HF frequencies for any given
value of SWR at the antenna end of the cable. The result of using
Chapter 28
longer coaxial cable runs will then be an SWR curve at the

transmitter output that never exceeds 2:1. Compared to the
reduced gain already inherent in WBFD design, the added losses of
a long cable run are not considered excessive when totaling the
overall system gain.
Alternatively, modern amateur transceivers (and those in other

services) are routinely (but not universally) equipped with automatic
antenna tuner circuitry. Although limited in range compared to a
wide-range external antenna tuner, these tuners are certainly
adequate to handle the modest SWR values presented by even the
longest WBFDs. Hence, the transmitter output circuitry prior to the
tuner will show a very low SWR.
Construction
The decision to use a WBFD involves an evaluation of one's goals

in operating or listening. Only with a set of specifications of this
order can one decide whether the WBFD will meet the needs. The
description of the antenna's advantages and limitations must be set
against the operating specifications and along side other potential
antennas that are candidates. Then selection becomes a matter of
choosing the antenna that does most of the jobs well enough.
If you do decide to use a WBFD, you can purchase one of the

commercially made types. B&W (USA), Giovannini (Italy), and
others produce these antennas in a variety of lengths. Alternatively,
you can build your own.
Chapter 28
The antenna proper uses standard techniques of wide-spaced

folded dipole construction. You will need twice the length of wire as
you determine the antenna length to be. There is nothing critical
about the exact length, although the general length will be a
function of where you decide to place the frequency that forms the
knee separating relative even performance at higher frequencies
from diminishing gain at lower frequencies.
Fig. 7 shows just 2 of many ways to space the wires along their
length. In the 1930s, we might have used wood dowels boiled in
paraffin. Today, we have access to a variety of better materials.
Part A of the sketches shows fiberglass rods, with holes drilled to
Chapter 28
pass the size wire we decide to use. #12 to #14 AWG copper wire
(0.06-0.08" or 1.5-2.0 mm) is likely to be the most common choice.
The end post can be longer to hold tie-off ropes for the assembly.
Fiberglass rods can be purchased from mail order sources.
However, local home improvement centers often carry adaptable
materials. For example, I recently spotted some 1/2" diameter
fiberglass rod under the guise of chimney flue brush extension
handles.
Alternatively, I have also had good luck using 1/2" diameter CPVC,
a thin-wall form of PVC tubing that replicates copper tubing sizes,
shown in Part B of Fig. 7. A hacksaw cut in each end leads to a
hole drilled to pass the chosen wire size. The wire press fits down
the slot and into the hole. If the holes are not deburred, the wires
stay put, although the spacers can be repositioned with fair ease.
These are simply two of many ways to make the required spacers.
Narrow strips of polycarbonate, acrylic, or Plexiglas would also
work. Polycarbonate likely has the best UV resistance of this group.
When adapting materials to a new use and environment, it is wise
to check the structure every so often to ensure that it is wearing
well under the influence of sunlight, precipitation, and temperature
excursions. Of course, cut any spacers that you use to the desired
length--about 8" (0.2 m) between wire holes for the models
examined here. However, this spacing is not very critical.
Locating a non-inductive resistor of sufficient power dissipation is

likely to be the chief problem for WBFD builders who intend to
transmit with the antenna. Unless you can find a suitable resistor at
Chapter 28
one of the surplus outlets, purchasing an antenna may prove

economical in the long run, if we add both cost and parts-searching
time together. Any value in the 800-900 Ohm range--or even
"thereabouts," if a bargain appears--will serve.
Manufacturers use different methods of packaging the resistor into

the antenna assembly. Some prefer a total enclosure to
weatherproof and bug-proof the resistor. However, one might have
to de-rate the resistor's power handling capability under these
circumstances. To maximize power dissipation, the resistor can be
placed within a tube that is about twice the diameter and about 1.5
times the resistor's length. Air passing through the tube provides
cooling, while the tube itself protects against immediate weather
impacts. Since the antenna wire and resistor terminals will attach to
strips of metal bonded to the tube, the resistor itself is relieved of
strain. The down side of this technique is the need to clean out
bugs and others debris on a regular basis. However, semi-annual
inspection and antenna maintenance is always a good policy.
For receiving-only applications, the resistor problem is much

simplified. A series-parallel combination of carbon resistors with a
net value of about 820 Ohms is easy to arrange. 1 to 5 watt non-
inductive resistors provide the sturdiest construction. The assembly
should be mounted in a UV-resistant plastic housing with strong
terminals for connecting the antenna wires.
The other challenging component is the 16:1 RF transformer. The

builder has two general types of transformers to use: a
transmission-line transformer or a standard wide-bandwidth
Chapter 28
transformer using a toroidal core. Transmission-line transformers

are slightly more efficient for transmitting purposes, although they
prefer purely resistive loads. Jerry Sevick, W2FMI, has written
extensively on these units, with instructions on how to build them
for many impedance transformation ratios. In a pinch, one might
place two 4:1 baluns in series.
There are proponents of standard RF transformers using toroidal

cores. Doug DeMaw, W1FB, has written on their use, including
calculating the power-handling capability of various cores. For
receiving-only applications, small cores can be used, and the basic
requirements and calculations are described in recent editions of
the ARRL Handbook, Chapter 6.
Whatever form of RF transformer you use, package it to withstand

weather. A sealed UV-resistant plastic box with a correctly placed
"weep" hole for moisture drainage is a good choice. Obviously, you
will need connections for the antenna wires as well as a coax
connector.
A 3-Wire Version of the WBFD
The WBFD has many variations, and from time to time--as I

encounter and model them--I shall add a few notes on them. The
first addition to the list is a 3-wire WHFD, outlined in Fig. 8, with the
2-wire version shown for contrast.
Chapter 28
For comparative purposes, I have made each antenna 27.2-m long,

with a 0.2-m separation between wires. The 3-wire version places
the terminating resistor in the "center" wire and parallel feeds the
two "outer" wires. Although the arrangement is shown as a flat
configuration, one can, as a variation, create a triangle of wires.
With two parallel-fed wires, the terminating resistor needs to be

about 1.5 times the anticipated center feedpoint impedance. Hence,
the model uses a 900-Ohm resistor, with 600-Ohms as the
expected feedpoint impedance. Of course, the feedpoint
Chapter 28
impedance is a nominal value, since the actual resistive component

of the impedance will fluctuate continuously above and below that
value as we move across the operating span.
In general, the 3-wire version of the WBFD is capable of about 1-3

dB (depending upon frequency) gain advantage over the 2-wire
version. The following brief table provides a glimpse at the
fluctuations for the 27.2-m antennas. All gain values are for free-
space.
Frequency 2-Wire 3-Wire 3-Wire
MHz Gain dBi Gain dBi Advantage dB
5 -4.39 -1.62 2.77
10 -2.51 -1.11 1.40
15 -2.42 -0.39 2.02
20 -0.56 +0.74 1.30
25 -1.20 +0.53 1.73
30 +1.28 +2.58 1.30
Despite these fluctuations, the gain curves for the two versions of
the WBFD are remarkable congruent, as shown in Fig. 9. The dual
or parallel feed system of the 3-wire WBFD increases gain by
feeding 2 wires, but it does not change the main characteristics of
the antenna. Besides the congruence of gain curves, the patterns
yielded by the 3-wire antenna differ from those of the 2-wire version
only in peak gain, but not in strength. Since the gain increase is
marginal, both antennas have patterns that replicate those of a
simple doublet (with its widely varying impedance with changes in
frequency), but remain smaller, that is, have much lower peak gain
values.
Chapter 28
The cost for the small increase in gain is a wider fluctuation in the
impedance about a central impedance value. The reactance tends
to be somewhat higher than for the 2-wire WBFD. Fig. 10 shows
the comparative SWR curves for the two antennas, with each one
using its own reference value: 820 Ohms for the 2-wire version and
600 Ohms for the 3-wire version.
Chapter 28
In addition to having higher peak values in the 2-30-MHz span, the

3-wire version curve slopes differ from the corresponding 2-wire
slopes. The shallower parts of the curves are on opposite sides of
the peaks.
The construction of a 3-wire WBFD can generally follow the same

set of techniques used for the 2-wire antenna, with the separators
enlarged to handle the broader plane. The additional wire will
increase the weight of the antenna proper by nearly 50%, and
center support of the terminating resistor and the feedpoint area is
advisable. Short sections of 600-Ohm (or thereabouts) open-wire
feedline can create the feedpoint junction. Like the 2-wire WBFD,
the 3-wire version requires a balun system or a wide-band
transformer to match a 50-Ohm coax feedline.
Chapter 28
WBFDs can be used in inverted-Vee configurations. However,

expect a decrease in the broadside gain and some vertically
polarized radiation off the ends of the array--just as you would find
in a simple doublet. The steeper the angle of the two side of the
antenna, the greater will be the radiation off its ends. As the
antenna length exceeds 1 wavelength, the patterns may
increasingly differ from those of the antenna used horizontally.
A Common Mistaken View of the Wide-Band Folded Dipole
We may look at the wide-band folded dipole from another

perspective, one encouraged by a 1983 B&W patent (#4,423,423).
The patent sketch shows shorting wires close to the center of the
array. The wires connect the feedpoint to the terminating resistor. In
addition, they also connect the inner or feedpoint ends of each wire
pair.
The result of this variation is an antenna that is NOT a WBFD.

Instead, as revealed by the sketch portion of Fig. 11, the new
configuration places the resistor in a simple parallel connection with
the feedpoint. The wires extending from the junctions of this parallel
connection form virtual "fat" single wires to complete what amounts
to a center-fed doublet.
Chapter 28
The sketch shows a 900-Ohm resistor in parallel with the feedpoint

of the 27.2-m antenna. The SWR curve that accompanies the
sketch reveals a property that is unlike the true WBFD antenna: a
resonant feedpoint at about 5.2 MHz. The actual reported resonant
impedance is about 63 Ohms. The parallel combination of 900
Ohms (the resistor) and 70 Ohms (the typical resonant dipole
resistance) is about 65 Ohms.
Chapter 28
Fig. 12 provides a set of SWR curves across the 2-30-MHz span

that we have used in the examination of the true WBFD. Since the
increment between readings is 1 MHz, the 50-Ohm curves do not
necessarily show the best SWR values possible. However, the 3
SWR minimum points show that the antenna acts completely
normally when treated as a center-fed doublet. The 3 low 50-Ohm
SWR values occur at the fundamental, the 3rd harmonic, and the
5th harmonic frequencies for the wire.
The 900-Ohm SWR curve, of course, shows the opposite trend

relative to the 50-Ohm curve. Minimum SWR values occur at the
2nd and 4th harmonic frequencies of the wire (between 10 and 11
MHz and again between 20 and 21 MHz). At these frequencies, the
Chapter 28
antenna feedpoint impedance would show very high values of

resistance and reactance (except for a tiny frequency region where
the reactance passes through zero as it changes type). Under
these conditions, the 900 Ohm resistor dominates the parallel
combination.
We may compare the gain performance of this "faux" WBFD with a

true WBFD of the same length. Construction would be identical
except for the use or non-use of the shorting wires. At the
fundamental and the odd harmonic frequencies, as shown in Fig.
13, the new antenna shows higher gain. However, at the 2nd and
4th harmonics, the gain values are virtually the same for both
antennas.
Chapter 28
The dominant problem with the faux WBFD is the changing

feedpoint requirement as we alter the frequency of operation. A
direct coaxial feed is applicable in only 3 narrow frequency regions
of the spectrum. Likewise, the 16:1 balun treatment shows equally
limited application. Given the feeding limitations, one might well use
a simple center-fed doublet with a parallel feedline and a wide-
range antenna tuner. Fig. 14 overlays free-space E-plane (azimuth)
patterns for 5.2, 10.4, 15.6, and 20.8 MHz to provide a better
performance comparison among the doublet, the WBFD, and the
faux WBFD.
Chapter 28
In all cases, the center-fed doublet exhibits the highest maximum

gain. On the fundamental and the odd harmonics, the faux WBFD
Chapter 28
is nearly as strong, and the true WBFD is considerably weaker. On

even harmonics, the doublet's gain is very significantly higher than
both versions of the folded antenna with a resistor. The consistent
aspect of the patterns is their shape. They are perfectly congruent
throughout, differing only in the gain value.
The faux WBFD turns out to be a simple center-fed doublet with a

resistor that parallels the feedpoint terminals. Unless one needs an
antenna that operates only on even harmonics of the antenna's
fundamental length, it offers no useful advantages over the single-
wire doublet without the resistor. At many frequencies, its gain
deficit can be a distinct disadvantage.
What Kind of Antenna is the True WBFD?
We have been exploring the behavior of the WBFD so intently, that

we have overlooked an interesting and significant question: What
kind of antenna is the WBFD? We have already shown some
reasons why it is not a true folded dipole. But, we have not placed
the WBFD into an appropriate category of antennas.
The answer to our question is both easy and difficult: the

terminated folded is a variety of traveling-wave antenna. Entire
books have been devoted to traveling wave antennas. See, for
example, C. H. Walter, Traveling Wave Antennas (1965): a classic
and very thorough text on traveling-wave fundamentals for all
relevant types of antennas. More commonly recognized traveling
wave antennas are terminated long-wires, V-beams, and rhombics.
Chapter 28
However, many types of traveling wave antennas require no

resistive or other termination.
For our very limited purposes, we may contrast standing-wave and

traveling-wave antennas in an over-simplified manner. Consider a
transmission line that is lossless or perfect. If we leave the load end
of the line without a load in an open condition, the entirety of the
energy from the source returns to the source. In a transmission line,
we usually view this condition as a system fault, since we obtain no
useful work from the source energy. However, if we separate the
transmission-line wires to create a doublet, something else
happens. We have the energy reaching the line ends and returning.
The result is a set of standing waves along the wire. Since the
separate-wire situation creates a transducer, we obtain useful work
in the conversion of the energy into into a field that expands without
limit. An ideal antenna would show standing waves that reach the
same peak value (however measured) at every peak. As well, all
wave minimums would go to zero.
There is no such thing as a perfect standing-wave antenna.

Consider a 5-wavelength center-fed doublet composed of lossless
wire. It will show a feedpoint current minimum value that is limited
by the impedance at the feedpoint. As well, other current minimum
points will not reach zero except at the very ends of the antenna.
Likewise, the peak values will not be identical along with wire, with
the highest peaks occurring closest to and farthest from the
feedpoint. The top portion of Fig. 15 shows a typical (imperfect)
standing wave current pattern.
Chapter 28
The lower portion of Fig. 15 shows the current distribution along a

terminated end-fed long-wire, one of the simplest traveling-wave
antennas. Since we have a terminating impedance, the ideal
situation would show a constant current magnitude all along the
wire. The termination impedance prevents the return of energy
necessary to create standing waves. For a variety of reasons,
terminated wires fail to achieve a perfect traveling-wave status. The
termination is normally and for highly practical reasons a non-
inductive resistor. However, the required impedance turns out to be
both complex and finicky. So we only approximate a traveling wave,
a basic current level with superimposed standing waves that are
small but detectable. (The example shown uses a technique for
setting up a traveling-wave long wire described by E. A. Laport in
Radio Antenna Engineering, pp. 55-58, 301-339 (1950).)
Chapter 28
The relevance of these notes to the WBFD and related antennas is

straightforward: we may use the current distribution along an
antenna to determine in a general way whether an antenna is a
standing-wave or a traveling-wave antenna. Fig. 16 shows the
current distribution at 10.4 MHz of several antennas, all 27.2 m
long. 10.4 MHz is the second harmonic of the 5.2-MHz fundamental
frequency. The arbitrary feedpoint current for all of the antennas is
1.0 (A) with a 0.0-degree phase angle. All current values are
relative to this convenient value.
Chapter 28
The center-fed single-wire doublet, a prime example of a standing-

wave antenna, shows peak current values about half-way between
the feedpoint and the wire ends. The peak currents are over 5
Chapter 28
times the feedpoint value, a typical situation for a 1-wavelength

doublet. The second antenna is a folded dipole operated at its
second harmonic. It gain is low due to the very low value of the
feedpoint resistance (about 7 Ohms). Hence, even the copper wire
construction removes almost 3 dB of the antenna's theoretical gain
if made from lossless wire. Nevertheless, within these restrictions,
the antenna shows the folded-dipole version of a standing wave.
Relative to the single-wire doublet, the standing wave is displaced
by a quarter-wavelength along the wire due to the fact that folded
dipoles show a combination of radiation and transmission-line
currents. (For more on this subject, see the preceding Chapter 27.)
Note that the current minimums are very close to zero (actually
about 0.02 to 0.03 relative magnitude).
The next two antennas shown in Fig. 16 are alternative narrow-

spaced and wide-space versions of the true WBFD. The narrow
version spaces the wires by 0.2 m; the wide version uses a spacing
of 1.0 m. In the wide version, each current curve is relative in height
to the distance from its associated wire. In both cases, the
antennas exhibit a standing-wave property, but overlaid on a
traveling-wave current value of about .75 relative magnitude. The
standing-wave component is about +/-0.30 relative magnitude.
Hence, the minimum current level is a bit over 0.45 relative to the
feedpoint value. (Values for the 2 versions of the antenna may vary
by about 0.03 in relative magnitude from the listed values or about
5%.) The fact that we see the overlaid standing-wave component in
current graphs may obscure the basic traveling-wave component of
the current distribution until we realize that the current magnitude
Chapter 28
never approaches zero and remains well above zero throughout the
antenna structure.
All traveling-wave antennas that use terminations dissipate energy

in the resistive portion of the termination. Although "sound-bite"
wisdom about terminations uses a dissipation value of 50%, we
should generally ignore that figure. The actual dissipation in the
termination depends on the antenna configuration and frequency of
operation. The actual value may be over or under that value by a
considerable amount. However, any energy dissipated is
unavailable for the radiated field, and hence, we obtain the WBFD
gain levels that fall considerably short of comparable fields from the
center-fed doublet.
The bottom current distribution pattern in Fig. 16 completes the

survey by including the curves for the faux WBFD. The curves have
the general shape of those for the center-fed single-wire doublet
and the minimum values approach zero. The parallel resistor does
not change the antenna's classification as essentially a standing-
wave antenna.
Although the last antenna is indeed a faux WBFD, since is has no

relationship except physical appearance to a folded dipole of any
sort, I have avoided calling the WBFD a faux folded dipole. The
WBFD is not a true folded dipole because it is a traveling-wave
antenna. Nevertheless, it retains more than vestigial traces of its
folded-dipole origins in the overlaid standing wave pattern on the
baseline traveling-wave current level.
Chapter 28
Conclusion
A WBFD antenna is not for everyone. However, gaining some

understanding of its operation, its nature, its advantages, and its
limitations may be useful in the process of choosing an antenna--or
even simply learning more about what various antenna types can
do. The WBFD has its niche among amateur,
governmental/military, and SWL antennas, but that niche is
certainly not universal. The government and the military find the
antenna useful for ALE (very rapid frequency excursions), and
some amateurs are experimenting with these techniques. Within
more normal time periods allowed for frequency changes, antenna
tuner automation is generally fast enough for most contest
environments, and manual tuning suffices for other
communications--using a single-wire doublet.
Receiving versions of the antenna can be home built for not much
more than the cost of the wire, since the materials necessary for
low-power terminating resistors and wide-band RF transformers are
low. However, building a transmitting version of the antenna at
home may be much more problematical, since parts may be hard to
find or hard to fabricate. The alternative, of course, is one of the
commercial versions, in an exchange of bucks for bother.
Chapter 28
Chapter 29: Wide-Band Multi-Wire Folded Dipoles
I
N Chapter 28, I briefly touched on the 3-wire version of the
antenna. In this Chapter, I want to expand a bit on that antenna
as part of a larger look at multi-wire "folded dipole" antennas
using a terminating resistor to extend the SWR bandwidth. In fact,
we shall review and expand coverage of 4 antennas: the standard
single-wire doublet, the most familiar 2-wire terminated version, the
sometimes mis-drawn 3-wire version, and a 5-wire version of the
antenna. The goal is to enlarge our understanding of how these
antennas work and what features count as advantages and
disadvantages of them.
In the process, we shall examine some interesting properties of

models of wide-band multi-wire terminated antennas based on
idealized models. There are some techniques of model formation
that are very useful under certain circumstances. However, if
inappropriately relied upon, they can mislead us. As well, we may
sometimes collect only partial data from an antenna model and be
equally led astray. In this first part of the exercise, let's allow
ourselves to be led and see where the path may wind.
The basic antenna will be 27.2 m (89.14') long. Fig. 1 outlines the
single-wire doublet and the common 2-wire version of the
terminated antenna having the same length. The sketch does not
specify any particular spacing between wires of the terminated
antenna. Any reasonable spacing will work from very close to some
larger spacing that is still only a small percentage of the total
length. The models for this antenna use a spacing of 0.2 m (6.5"),
Chapter 29
which is well under 1% of the antenna's length. However, before we

close, we shall explore the effects of using the relatively narrow
spacing of our initial models and using wider element spacing.
The TR (terminating resistor) element in the 2-wire wide-band

antenna is 820 Ohms in this model. However, values between 800-
900 Ohms are most common, and versions exist with resistors
ranging from 400 Ohms to 1200 Ohms. The resistor must be non-
inductive, ruling out wire-wound power resistors. The resistor
inductance would count as a loading coil, adding very significant
reactance. In fact, the reactance would climb with frequency and
eventually form an effective RF choke. To overcome the problem of
finding a desired resistance power value that will not over-weight
the center of the antenna, many builders use higher value resistors
in parallel. For receiving only, the resistors can be of any power
value. However, for transmitting, the terminating resistor must be
capable of dissipating about one half of the anticipated power
Chapter 29
applied to the antenna. Paralleled resistors must allow for air flow to
carry away heat, while at the same time providing protection from
the weather in which the antenna operates. Weight, heat, and
weather are the three primary enemies of a terminated "folded
dipole."
The terminated "folded dipole" only looks like a folded dipole. The
wire configuration does not perform an impedance transformation
like an ordinary folded dipole without the terminating resistor. As
well, the standard folded dipole is a narrow-band antenna, like the
simple dipole. In a terminated 2-wire wide-band antenna, the
impedance at the feedpoint is a function of the terminating resistor.
In fact, the actual impedance at any frequency is a joint function of
the termination and the antenna length in wavelengths.
The chief reason for using a 2-wire wide-band terminated antenna

is to obtain a satisfactory feedline SWR level across a wide range
of frequencies. For the 27.2-m antenna, I shall sample 2 through 30
MHz as a preliminary operating range. Fig. 2 outlines on the same
graph the SWR of a single-wire doublet and of a 2-wire terminated
antenna. The most common feedline used with a single wire
doublet is parallel 450-Ohm transmission line. Therefore, the
doublet SWR curve uses that reference impedance. For all
frequencies, the SWR ranges from moderately high to very high.
The doublet works by having low line losses, even with fairly high
SWR levels, and an antenna tuner in the shack provides the
impedance transformation to the 50-Ohm input/output of the
transceiver or other equipment. The impedance at the antenna
tuner terminals is unlikely to be the same as at the antenna
Chapter 29
feedpoint terminals. When the feedpoint impedance and the

characteristic impedance of the line are different, the transmission
line acts as an impedance transformer for each electrical half-
wavelength. Hence, the tuner terminals show an impedance that is
a joint function of the original feedpoint impedance, the line
characteristic impedance, and the electrical length of the
transmission line.
The wide-band 2-wire terminated antenna SWR curve uses 820

Ohms as its reference impedance. The precise value is non-critical
so long as it is close to the value of the terminating resistor. The
curve in Fig. 2 shows only minor peaks above the 2:1 level, and
those peaks tend to grow with increasing frequency. However, most
Chapter 29
versions of the antenna employ 50-Ohm coaxial cable as the

feedline. The reference impedance and the cable impedance show
a 16:1 ratio. Although a 16:1 balun is possible, many builders
employ 2 4:1 baluns (although one of them can be a “unun”). Many
balun designs become lossy with rising reactance. Although those
losses may be low compared to the power dissipated in the
terminating resistor, some antenna makers use a standard
transformer to effect the wide-band match. The key factor in any
such transformer is to avoid core saturation. See the end of
Chapter 6 of the ARRL Handbook for a brief characterization of the
2 types of impedance transformers and the basic needs of each
kind. A final alternative involving impedance transformation
concerns the terminating resistor itself. I have heard of using a 50-
Ohm resistor and placing a 16:1 transformer between it and the
wire to create the effect of an 800-Ohm terminating resistor.
To take care of any remnant SWR peaks that exceed 2:1 at the
junction with the coaxial cable and the antenna and its impedance
transforming devices, some wide-band antenna makers
recommend very long lengths of coaxial cable. The rationale is
simple. Coaxial cable losses are real, but will be relatively small--
even at 30 MHz--compared to the losses due to the dissipation of
applied energy in the terminating resistor. Hence, the use of a long
cable is operationally insignificant relative to antenna performance.
Moreover, the long cable will usually prevent the triggering of fold-
back circuitry that protects the final amplifier in the presence of
excessive SWR values.
Chapter 29
The 3-wire terminated wide-band antenna is an extension of the 2-

wire version. The general claim associated with the 3-wire version
is higher gain with equal or better SWR curves. We shall eventually
examine the gain claim in some detail. For the moment, we may
simply see the schematic outlines of the 3-wire wide-band antenna
in Fig. 3. For the sample model, the terminating resistor is 900
Ohms.
The right side of the sketch shows the parallel connection of the
antenna feedpoint terminals on the 2 non-terminated lines. This
configuration yields proper connections for wide-band service. In
some sketches (that do not pretend to be electrical schematic
diagrams of the antenna), I have seen simplified connections that
can mislead the home builder. The sketches seem to show the
center point of each non-terminated wire as comprising each side
of a proper series feedpoint. For normal installations with no special
components, this system will not work. Fig. 4 shows why.
Chapter 29
The left side of the sketch shows the circuit path of the antenna if
we join the center points and connect the source between them.
The path leads at any instant away from one connection and
toward the other through 2 parallel paths. At the ends of the
antenna (obviously not drawn to scale in Fig. 4), the junctions with
the line carrying the terminating resistor at its center have equal
voltage. The net voltage drop across the resistor is therefore zero,
and that component has no function in the antenna, when set up in
this way. In fact, models created using this system show no
difference of pattern, impedance, or termination-line current for any
value of resistor, from zero to very high.
Chapter 29
The system on the right provides the correct path for the
terminating resistor to do its work. When I modeled the antenna, I
used the terminating resistor line as a center point. Then I set up
the fed lines 0.2 m away from the center line. The need to connect
the feedpoints in parallel presents an interesting geometry
challenge if we model the antenna using only wire entries.
However, there is a simple technique of connecting feedpoints in
parallel that may be useful to newer antenna modelers. The
following model description from EZNEC may illustrate the general
technique.
EZNEC/4 ver. 4.0
3-Wire Wide-Band FD 6/24/05 9:28:16 AM
--------------- ANTENNA DESCRIPTION ---------------
Frequency = 12 MHz
Wire Loss: Copper -- Resistivity = 1.74E-08 ohm-m, Rel. Perm. = 1
--------------- WIRES ---------------
No. End 1 Coord. (m) End 2 Coord. (m) Dia (mm) Segs
Conn. X Y Z Conn. X Y Z
1 W4E2 -13.6, 0, 15 W2E1 13.6, 0, 15 #12 69
2 W5E1 13.6, 0, 15 W3E1 13.6, 0, 15.2 #12 1
3 W2E2 13.6, 0, 15.2 W4E1 -13.6, 0, 15.2 #12 69
4 W3E2 -13.6, 0, 15.2 W7E2 -13.6, 0, 15 #12 1
5 W1E2 13.6, 0, 15 W6E1 13.6, 0, 14.8 #12 1
6 W5E2 13.6, 0, 14.8 W7E1 -13.6, 0, 14.8 #12 69
7 W6E2 -13.6, 0, 14.8 W1E1 -13.6, 0, 15 #12 1
8 -0.1, 1, 15 0.1, 1, 15 #20 1
Total Segments: 212
-------------- SOURCES --------------
No. Specified Pos. Actual Pos. Amplitude Phase Type

Wire # % From E1 % From E1 Seg (V/A) (deg.)
1 8 50.00 50.00 1 1 0 V
-------------- LOADS (R + jX Type) --------------
Load Specified Pos. Actual Pos. R X

Wire # % From E1 % From E1 Seg (ohms) (ohms)
1 1 50.00 50.00 35 900 0
Chapter 29
-------- TRANSMISSION LINES ---------
No. End 1 Specified Pos End 1 Act End 2 Specified Pos End 2 Act Length Z0 VF
Rev/Norm
Wire # % From E1 % From E1 Wire # % From E1 % From E1 (m) (ohms)
1 3 50.00 50.00 8 50.00 50.00 0.01 500 1 N
2 6 50.00 50.00 8 50.00 50.00 0.01 500 1 N
Ground type is Free Space
Wire 1 is the terminated line, as indicated by the Load entry. Wires

3 and 6 are the lines with the feedpoints shown in the original
antenna sketch. Wires 2, 4, 5, and 7 are the end-connecting wires.
Wire 8 is the feedpoint wire segments, as indicated by the Source
entry in the description. The wire is only a meter away from the
main wires, but it is so short and thin (0.2 m by AWG #20) that it
does not materially affect the performance figures of the main
wires.
Now note the 2 transmission-line entries. Each runs from the center
of one of the fed wires to the new distant source wire. A
transmission line is not a physical or radiating wire within a model.
It is only a mathematical construct factored into the model after
completion of basic matrix calculations. In fact, the physical or
geometrical distance between the source wire and the wire
segment at the other end of the transmission line does not define
the line length. Instead, we specify the line's electrical length (as
well as the characteristic impedance) when we enter the
transmission line data, and calculations use this length. Note that
the length for each of the 2 lines is 10 cm (about 4"). We could
have made it even shorter (down to something like 1e-10). But 4"
effects virtually no impedance transformation down the line. In fact,
the specified characteristic impedance (500 Ohms) is also non-
Chapter 29
critical, and changing it by a few hundred Ohms creates virtually no

impact on the reported final feedpoint impedance.
The technique of using near-zero lengths of transmission line to

connect wires in parallel is very useful to many types of models that
connect feedpoints in parallel. NEC-2 and -4 have some limitations
with very shallow angles that are often involved in these kinds of
connections. The wire surface can penetrate the center third of the
other wire segment at the junction, creating errors in the current
calculations. The transmission-line technique avoids those
problems. As well--and especially apt to this case--the technique
avoids involving us in extra wires, odd changes of wire direction,
and the inevitable expansion of the model size as measured by the
number of segments. Nevertheless, we must recognize that our
models are idealizations. Any implementation of a 3-wire wide-band
terminated antenna will necessarily have leads to the common
feedpoint at which we find the impedance transformation device.
Whether the ease of modeling with the TL-based parallel
connections serves us well or ill we shall not determine until we
explore Part 2 of the exercise.
I used a similar system to parallel-connect the feedpoints of 4 wires

surrounding a terminating resistor wire. The 5-wire wide-band
terminated antenna appears in schematic form in Fig. 5. The
schematic does not attempt to replicate the modeled physical
structure, but is handy to represent the 4 paralleled feedpoints. The
sketch on the right gives an end-on view of the antenna. The two
new wires are at right angles to the existing wires, forming a cage
Chapter 29
of sorts around the center wire. The square represents the

terminating resistor at the center of the center wire.
As an alternative version, I modeled the 5-wire wide-band

terminated antenna as a planar construct. The far right sketch
shows the general idea, but not at all to scale. The fed wires lie in a
single plane with the termination wire, 2 above and 2 below the
terminating resistor. The model uses the same spacing--0.2 m--
between each wire. As well, the model uses the same short
transmission-line technique to connect the feedpoints in parallel. In
both versions of the 5-wire antenna, the terminating resistor is 800
Ohms. The planar version seems to show slightly better
performance than the square version, so I shall use it for the initial
data-gathering exercise.
We can now examine the SWR bandwidths of the 2 expanded

versions of the wide-band terminated antenna. Fig. 6 shows the
500-Ohm SWR curve for the 3-wire version and the 300-Ohm curve
for the 5-wire version. The 3-wire antenna shows a very well-
behaved SWR curve, with an excursion above 2:1 only between 3
Chapter 29
and 4 MHz, below the frequency that I shall eventually recommend

as the minimum usable frequency for all of these 27.2-m antennas.
However, note a change relative to the 2-wire antenna. For the
smaller antenna, the center of the feedpoint impedance excursions
occurs at about the same impedance as the terminating resistor.
However, the SWR reference for the 3-wire antenna is only about
55% of the value of the terminating resistor (500 vs. 900 Ohms).
Chapter 29
The curve for the planar 5-wire version of the wide-band array is
less well behaved, with 3 humps above the 2:1 SWR level.
Although the terminating resistor is 800 Ohms, the reference level
for the SWR is down to 300 Ohms, or under 40%. As we drop the
necessary reference impedance, due to having more wire
feedpoints in parallel, the reactance at any given frequency plays a
larger role in determining the SWR. Hence, the lower the feedpoint
reference level, the more likely we are to find higher SWR levels.
The 3- and 5-wire antennas shows interesting parallels between the

curves, with peaks near 3-4, 7-8, 14-15, 20, and 25-26 MHz. In this
respect, the curves are only vaguely similar to the 2-wire version of
the antenna, despite the similarities in the terminating resistor
values (800-900 Ohms). For SWR bandwidth, either the 2-wire or
the 3-wire versions of the antenna show the more promise. While
the 2-wire antenna requires a 16:1 impedance transformation for
use with 50-Ohm coaxial cable, the 3-wire antenna requires a 10:1
transformation. The 5-wire version sets up a conflict: the reference
impedance requires a 6:1 transformation, a value within
construction limits for a single balun. However, the SWR excursions
suggest a considerable reactive component that may induce higher
losses in at least some balun designs.
If SWR bandwidth were the only consideration, then the simplicity

of the 2-wire antenna would dictate its use. The addition of more
wires complicates the antennas structure and adds weight,
especially in the separators needed to keep the antenna from
corkscrewing in the wind. So our final question is whether we obtain
Chapter 29
any advantage in turning to a more complex wide-band terminated

"folded dipole."
The brief answer is a qualified "yes:" we obtain more gain from

adding fed wires to the system. The following table illustrates the
gain advantage of adding wires.
Comparison of Maximum Free-Space Gain Values of 27.2-m Antennas at Selected Freq
Frequency 6 MHz 12 MHz 18 MHz 24 MHz
Antenna Free-Space Gain dBi
1-Wire Doublet 2.18 4.37 3.73 4.59
5-Wire Wide-Band 1.32 0.19 2.78 1.56
3-Wire Wide-Band -0.36 -0.80 1.64 0.50
2-Wire Wide-Band -2.59 -2.36 0.04 -1.07
For each of the selected frequencies in 6-MHz increments, the

larger the number of wires in the wide-band antenna, the higher the
gain. In some cases, the 5-wire antenna comes within 0.9 dB of
matching the non-terminated doublet, while in other instances, the
differential is almost 4.2 dB. However, by the time we get down to 2
wires in the terminated antenna, the gain differential can be as
much as 6.7 dB. Fig. 7 provides maximum free-space gain curves
for all 4 antennas from 2 through 30 MHz in 1 MHz intervals.
Chapter 29
The curves show several items of interest. First, the more wires in
the array, the greater the variability of maximum gain across the
entire scanned range. Despite the higher variability of gain in the 5-
wire curve, the gain curves for the 3 terminated antennas show a
tight parallelism. On a few of the sampled frequencies (1-MHz
increments), the 5-wire antenna comes close to equaling the gain
of the doublet used as a standard. However, a 3-dB differential is
more common. The 2-wire antenna averages about 5-6 dB
differential from the doublet values. Note that all of the models set
up their wires parallel to the Z-axis, so the gain reports are taken in
the X-Y plane or broadside to the plane of the wires.
Chapter 29
More significant is the fact that all 3 terminated antennas show a

knee in their curves. Below the knee frequency, the gain drops very
rapidly. The knee frequency occurs when the antenna passes
below an electrical 1/2-wavelength at the operating frequency.
Broadly speaking, the knee for the 27.2-m antennas occurs around
5.5 MHz. The larger the antenna, the lower the knee frequency as
the multitude of wires act like a fatter single wire. However, below
the knee frequency, the 5-wire antenna loses gain faster than the
simpler wide-band versions. Although the doublet exhibits quite
reasonable gain below the knee frequency, at a certain point, the
gain may be unobtainable in practical terms. As an antenna falls to
3/8 wavelength or shorter, the reactance climbs rapidly while the
resistance sinks to a very low value. The combination will show
considerable line loss, even using low-loss parallel transmission
line, and the antenna tuner may have difficulties in effecting a
match to the values that appear at its terminals.
To obtain a better view of comparative gain with the recommended

operating range, we may omit the region at and below the knee
from our graph. Fig. 8 provides the same data as Fig. 7, but the
frequency range restriction expands the Y-axis. With a knee at
about 5.5 MHz, the recommended operating range for the 27.2-m
antennas is about 6-30 MHz. For operation down to 2 MHz, I would
suggest an antenna (of any of the 4 types) that is about 71-72 m
(about 235'). Since we are not seeking precise resonance, the
exact length is not important so long as the antenna is at least 1/2
wavelength at the lowest operating frequency.
Chapter 29
The expanded curves in Fig. 8 show more clearly the parallel

structure of the 2-wire, 3-wire, and 5-wire gain values across the
recommended spectrum. As well, Y-axis expansion shows the high
variability of the 5-wire gain curve. It sometimes almost reaches the
level of the doublet, but on other frequencies, it falls closer to the
level attained by the 3-wire antenna. Of course, the exact structure
of these curves is subject to variation with small changes in
construction--either overall length or spacing--or in the exact value
of the terminating resistor. Nevertheless, the most notable trends
remain intact.
Gain variations are NOT the result of any changes in antenna

pattern as we move from one 27.2-m antenna to another. In fact,
Chapter 29
the antenna patterns are functions of the overall wire length, and
the presence of multiple wires and the terminating resistor does not
affect any other property than lobe strength. (This statement
requires a bit of modification: the wire length is the electrical length
of the antenna rather than its simple physical length. Multiple wires
tend to act like a single fat wire, making the antenna longer by a
wider margin than a single wire. Hence, if all the subject antennas
from 1 to 5 wires are the same physical length, the larger the
number of wires, the electrically longer the antenna. The planar 5-
wire may further complicate the calculation of electrical length due
to the end connecting wires. Fig. 9 overlays the patterns of all 4
antennas at 12 MHz. Perhaps the only difference detectable is at
the junction of the main lobe with each of the minor lobes. The
more wires the softer the curve at the junction. It is likely that the
need for end wires to connect the horizontal wires is the main
reason for the softened junction.
Chapter 29
The pattern test is repeatable on any frequency. Fig. 10 provides a

second sample of overlaid patterns at 24 MHz. Where lobes join (at
pattern nulls), we once more find that if we have more wires and
end wires in the antenna structure, the null points soften into
curves. Other than that single phenomenon, the patterns are wholly
congruent and vary only in lobe strength.
Chapter 29
Before we close this exploration, it may be wise to examine the

effect of spacing between the wires in a terminated array. In a
standard folded dipole, there is no significant difference in
performance between narrow and wide spacing values, at least up
to the point where the distance exceeds a value that will support 2-
wire transmission-line operation. However, as we have noted, the
terminated arrays only look like folded dipoles.
Chapter 29
All of the terminated antennas used so far are 27.2 m long with a
spacing between wires of 0.2 m (about 7.9"). I revised the spacing
for a series of test models to a value of 1.5 m (59"). Because initial
tests using diamond and fan configurations proved unpromising, I
maintained the parallel runs of AWG #12 wires.
Fig. 11 compares the maximum free-space gain values for the 2

versions of the 2-wire terminated array. The wider array shows an
average gain increase of about 1 dB, although that increase does
not appear at every frequency in the test range (using 6 MHz as a
starting frequency). The original narrow version of the antenna
used an 820-Ohm terminating resistor and a similar value for the
SWR reference impedance. Widening the antenna required an
Chapter 29
increase in both values to 900 Ohms. Fig. 12 compares the SWR

curves for both 2-wire antennas on the assumption that each will
use an optimized impedance transformation device for any
adjustment to match a coaxial cable. The SWR curves show no
features that would dictate the use of one antenna version over the
other.
The 3-wire version of the terminated array used the same 1.5-m
spacing in the wide version. The result is an array that is 3-m wide
overall, with the terminated line between the fed wires. Fig. 13
shows the comparative gain curves for the two antennas. The wide-
version curve shows similar characteristics to the corresponding
Chapter 29
curve for the 2-wire antenna. However, the gain differential

between wide and narrow antennas averages close to 1.5 dB.
The terminating resistor for both the narrow and wide antennas is
900 Ohms. However, the best reference impedance for the wide
version is about 450 Ohms, about 50 Ohms less than for the
narrow version. Fig. 14 traces both SWR curves. The curve for the
wider version of the antenna shows larger excursions, but remains
below 2:1 across the entire test range.
Chapter 29
Although the 5-wire array remains the same length, its area grows
by a factor of 7.5. The inner fed wires are 1.5 m from the terminated
center wire, while the outer fed wires are 3.0 m from center. Fig. 15
provides one measure of whether the increased width is worthwhile
in terms of the overall gain advantage of the wider antenna. The
wide array provides almost 2-dB average gain over the narrower
(0.8-m) version. Unlike the 2-wire and 2-wire arrays, where the gain
lines cross at certain frequencies, the wide 5-wire antenna shows a
gain advantage throughout the operating spectrum.
(Remember that the gain measurements are for free space and are
taken broadside to the plane of the wires. Part 2 in this exercise will
Chapter 29
give us good reason to remember these limiting factors of this initial

study.)
Both versions of the 5-wire antenna use 800-Ohm terminating

resistors. As well, the reference impedance for both SWR curves is
300 Ohms. The average 300-Ohm SWR is lower for the wide
antenna, although the specific value may vary at some selected
frequencies. Unlike the narrow array, from 6 MHz upward, the wide
array SWR makes only one excursion above the 2:1 level and
remains below 2.2:1 at 13 and 14 MHz.
Chapter 29
The following short table samples some of these results in tabular

form. For a few frequencies between 6 and 12 MHz, the table lists
the maximum free-space gain of each of the 6 terminated arrays.
For comparison, the table adds a 7th gain column, listing the
maximum gain of a single wire 27.2-m doublet.
Maximum Gain Comparison Among 27.2-m Antennas at Selected Frequencies.
Maximum Free-Space Gain dBi
Frequency Doublet 2-Wire 2-Wire 3-Wire 3-Wire 5-Wire 5-Wire
Narrow Wide Narrow Wide Narrow Wide
6 2.18 -2.59 -2.47 -0.36 -0.05 1.32 1.62
8 2.66 -2.22 -1.27 -0.70 0.58 0.50 1.99
10 3.38 -2.51 -1.15 -1.11 0.55 -0.25 1.91
12 4.37 -2.36 -1.20 -0.80 0.63 0.19 2.35
The table makes clear that the narrow 3-wire array is superior in
gain to the wide 2-wire antenna and that the narrow 5-wire array is
Chapter 29
superior in gain to the wide 3-wire antenna. However, each wide

version is superior in gain to its corresponding narrow version.
However, even the highest-gain terminated array (the wide 5-wire
planar version) is deficient in gain by 0.5 to 2.0 dB relative to the
single-wire unterminated doublet of the same length.
These results emerge from exploring only the plane that is

broadside to the axis of the multi-wire antennas. We have not
looked at the plane that would be edgewise to the wires. Nor have
compared patterns taken broadside and edgewise to the plane of
the wires. Moreover, we have not explored the relative gain of the
various antennas at some common height above a specified
ground quality.
Some Tentative Conclusions
This initial attempt to take a fresh look at multi-wire terminated

wide-band antennas seems to justify a few general conclusions.
These conclusions rest on 2 features of the investigation. One is
the use of idealized models for 3- and 5-wire arrays that use virtual
zero-spacing between the wires forming the parallel connections at
the feedpoint. The second feature is the use of only partial data
from the collection available to us from the models. Nevertheless,
the project appears to show high promise for potentially improving
the multi-wire wide-band terminated antennas that we might build.
1. All terminated wide-band folded dipoles have knee frequencies,

below which the gain drops very rapidly. The recommended
Chapter 29
operating range for any of the antennas is from an electrical length

of about 1/2 wavelength upward in frequency.
2. As we add more fed wires to a terminated antenna, we increase

its average gain over the operating spectrum. The gain increase
never quite reaches the level of a single-wire doublet.
3. As we add more wires to a terminated wide-band antenna, the

center or reference SWR impedance decreases both intrinsically
and with respect to the value of the terminating resistor.
4. 2- and 3-wire terminated wide-band arrays show stable SWR

curves through their operating ranges. However, adding further
wires tends to produces curves with greater SWR excursions
relative to the reference impedance.
5. Terminated wide-band antennas show increased gain by

widening the distance between wires. Spacing adjustments may
require revision of the optimal terminating resistor value and the
reference SWR impedance.
Some of these conclusions are in fact generally true. Others may

be only partially true or simply illusions based on the limiting factors
of model formation and use in gathering data. Next, we shall
explore the available data in greater detail and develop some all-
wire models of the various arrays. That effort will allow us to sort
out which of the conclusions are useful as they stand and which
require modification, revision, or deletion.
Chapter 29
For the moment, we may simply allow ourselves to be enthusiastic

about the potential improvements that we have seemingly
uncovered.
Chapter 29
Chapter 30: More on Wide-Band Multi-Wire Folded Dipoles
I
n Chapter 29, I developed idealized models of multi-wire
terminated wide-band antennas as a pathway to understanding
better their performance. I replicated 2-, 3-, and 5-wire
terminated arrays using idealized techniques of feeding the
antennas and examining the free-space patterns taken broadside to
the plane of the wires. The patterns were all tidy, and the data
seemed to show improved performance as a. we increased the
number of wires and b. we increased the spacing between the
wires.
In this follow-up episode, we shall try to rectify some of the

shortcomings of the initial models. For the 3- and 5-wire antennas,
we shall reform the models into all-wire versions for comparison
with the idealized models. We shall also explore both the broadside
and edge-wise patterns to see if we can find any differences of
note. One of our goals is to improve upon the partial and
sometimes faulty understanding offered by Chapter 29. Another
goal is to stress the need for a full exploration of alternative ways of
forming models and of the complete data set offered by modeling
programs before declaring the work complete.
One major departure from the models in the previous Chapter rests
upon some of the data that we acquired. Although terminated wide-
band antennas offer a good match when we use them at
frequencies below the knee frequency, rapidly increasing losses to
the terminating resistor render such operations marginal to useless.
Therefore, I increased the antenna length to 250', the length
Chapter 30
needed to ensure that the antenna was electrically at least 1/2-

wavelength at the lowest test frequency: 2 MHz. A simple doublet
of this length shows considerable inductive reactance at a center
feedpoint. Hence, we can be assured that the new models operate
completely above the knee frequency. I also increased the wire size
to AWG #10 on the premise that this size is a relatively wise
selection for a 250' span of wire.
2-Wire Terminated Wide-Band Antennas
The traditional 2-wire terminated wide-band antenna is a good

starting point for our work. It does not require special treatment for
the feedpoint, since we may place the modeling source at the
center of the unterminated wire. Hence, we shall have only 2 major
concerns, both of which apply to the models in Chapter 29. First,
we shall look at the effects of taking patterns edgewise to the wire
plane instead of taking them broadside to the 2 wires. See Fig. 1
for an outline sketch of the two orientations.
Chapter 30
Second, we shall review the effects of spacing the wires. We must

increase the basic spacing in proportion to the scaling of the
antenna that moved the knee from about 5.5 MHz down below 2
MHz. I shall employ a simple set of spacing values that will apply to
all models in this part.
Spacing Standards for This Exercise Set
Category Spacing Between Parallel Wires (Feet)
Narrow 1
Medium 5
Wide 15
Most wide-band terminated antennas use a spacing between wires

that is 1' or less. I chose 1' because it allows a reasonable narrow
space for versions of the models scaled for higher frequencies and
shorter lengths. At the other end of the line, 15' is a little under 5
meters, the scaled value that emerges from the wide versions of
the antennas in Chapter 29. (Compare 250:15 and 27.2:1.5) Of
course, I rounded the new numbers for bookkeeping simplicity.
For the new antennas, I also selected a single value for the
terminating resistors in all versions: 900 Ohms. This value is
especially useful, since one might create it from a parallel
combination of 3 2700-Ohm non-inductive resistors. As noted in
Chapter 29, there are other techniques for creating the terminating
resistor.
Regardless of spacing, the 250' 2-wire antenna shows a very

usable 900-Ohm SWR curves from 2 to 30 MHz. Fig. 2 overlays
the curves for the 3 spacing values. The SWR spikes occur at
Chapter 30
intervals a little under 4 MHz, since the antenna passes integral full-
wavelengths (electrically) at those points. The narrow version
shows a few peaks just above 2:1 in the upper part of the overall
passband. However, both wider versions manage to remain below
the standard limit. In fact, the widest version shows a declining
curve with increasing frequency.
Note that the SWR reference impedance is the same as the value
of the terminating resistor. This fact will become more interesting as
we later look at the 3-wire and 5-wire antennas using the same
900-Ohm terminating resistor. For the moment, we may classify the
SWR behavior of the 2-wire antenna--in any width--as quite well
behaved.
Chapter 30
The remaining question is whether the patterns are as well

behaved as the SWR curve. In Chapter 29, we could not fairly
evaluate this facet of performance because we took the patterns
broadside to the plane of the two wires. In practice, at least the
narrow versions of the antenna twist and turn in the wind and
weather and hence may wash out any difference that we might find
between edgewise and broadside behavior. However, the wider
versions of the antenna would likely have a fixed non-twisting
installation set-up. Most likely, that set-up would place all of the
wires parallel to the ground. Hence, it will be useful to look at the
edgewise performance.
The following brief table samples both the edgewise and broadside
free-space gain values of the narrow, medium, and wide versions of
the 250' 2-wire terminated antenna.
Free-Space Maximum Gain Values for 3 Versions of the 250' 2-Wire Wide-Band Antenna
Frequency 2 3 4 5 15
Maximum Gain (dBi) E/W B/S E/W B/S E/W B/S E/W B/S E/W B/S
Antenna
Narrow (1') -3.37 -3.45 -2.67 -2.74 -2.87 -2.88 -2.20 -2.32 1.61 1.47
F/B Ratio dB 0.16 0.13 0.01 0.23 0.29
Medium (5') -3.00 -3.30 -1.71 -1.94 -2.02 -2.08 -1.47 -1.96 2.12 1.98
F/B Ratio dB 0.61 0.49 0.16 1.08 0.48
Wide (15') -2.38 -3.03 -1.07 -1.47 -1.17 -1.53 -0.56 -1.97 3.28 2.20
F/B Ratio dB 1.45 1.01 1.07 3.65 4.50
Reference Doublet 2.08 2.77 3.93 4.99 5.39
Note: E/W = Edgewise; B/S = Broadside; F/B = Front-to-Back
Despite its brevity and incompleteness, the chart generally confirms

the trend noted in Chapter 29: as we increase the spacing between
the conductors of a 2-wire terminated antenna, the gain generally
rises. The rise does not occur for all frequencies at every widening,
as shown by the move from 5' to 15' at 5 MHz, if we look at the
Chapter 30
broadside gain value. Nevertheless, the general gain trend is

upward with wider element spacing.
At the same time, sampling the edgewise gain values introduces a

new dimension to 2-wire performance: a gain differential between
the heading of maximum gain and a heading 180 degrees opposite.
The chart calls this difference a front-to-back ratio. For narrow
spacing, the differential is minuscule and operationally insignificant.
Even at medium spacing (5' for a 250' antenna or 2%), only one of
the listed values is potentially problematical. However, when the
spacing reaches 15' (6%), the differential grows to troublesome
proportions. At 15 MHz, the maximum gain in the favored direction
is 3.28 dBi, but in the opposite direction the gain drops to -1.22 dBi.
For modest front-to-back values, the front-to-back ratio is roughly
twice the differential between the edgewise and the broadside gain.
However, as we increase frequency and encounter higher
differentials, the front-to-back ratio climbs at a faster rate. This
phenomenon suggests that the pattern may undergo some serious
distortion relative to the nearly perfect bi-directional patterns we
expect from the narrow version of the antenna.
Chapter 30
Chapter 30
Fig. 3 provides a small demonstration by providing edgewise

patterns for all three spacing values at 2, 5, and 15 MHz. In all
cases, view the antenna as running up and down the page or
graph. The right side of the pattern is the direction toward the
feedpoint and away from the terminating resistor. The left side of
the pattern is the direction toward the resistor and away from the
source. The 2-MHz patterns show a classical figure-8 pattern with a
growing lobe toward the feedpoint side as we increase spacing. In
contrast, the 5-MHz patterns show their growth toward the
terminating-resistor side of the antenna. In addition, note that the
widest spacing yields a difference in pattern shape. What we would
call the main lobe on the higher-gain side becomes weaker on the
opposite side, and the side lobes grow in strength to equal it.
The 15-MHz plots show major distortions of the pattern established

by the narrow version of the antenna. If you closely examine the
medium-version pattern, you can see a very slight displacement of
the lobes toward the higher-gain side of the antenna. When we
reach the limits of our spacing exercise, the pattern is very
seriously distorted relative to the narrow-spacing version.
Moreover, we find extra lobes. Of course, the overall loop
circumference at the widest spacing is about 5.5% longer than at
the narrowest spacing. As well, the 15' end wires are 0.2
wavelength. At that spacing, the mutual coupling between wires
does not form a single element, but acts somewhat like a distended
loop antenna that is about 530' in total circumference--about 8
wavelengths overall.
Chapter 30
The 2-wire terminated wide-band antenna is notorious for its low

gain relative to a simple doublet. The table lists maximum gain
values for the listed frequencies. The doublet's patterns as virtually
identical in shape to those of the narrow version of the antenna, but
the gain differentials for just the listed values range from about 4 dB
to over 7 dB. Hence, the temptation to widen the spacing to obtain
higher gain is strong. As shown in Fig. 2, the wider versions of the
antenna sustain the SWR curves. However, before embarking upon
the widening process, one must closely examine all patterns to
determine if they will satisfy the needs of the application for the
revised antenna. For general skip communications, the medium (5')
version might fulfill the need, although the gain increment is
marginal relative to the increased complexity of construction.
The 3-wire terminated wide-band antenna is especially interesting

by virtue of its symmetry. Two outer wires, equally spaced from the
center wire that contains the terminating resistor, are fed in parallel.
The balanced layout results in a symmetrical edgewise pattern.
From this perspective, the 3-wire array eliminates the front-to-back
problem that appears in wider versions of the 2-wire antenna.
However, effectively modeling the 3-wire version of the wide-band
antenna presents challenges. Fig. 4 outlines 2 ways to proceed
with the modeling.
Chapter 30
The modeling scheme on the left uses the NEC transmission-line

(TL) facility to create near-zero-length lossless leads to a remote
source wire. I used this method of modeling in Chapter 29 as an
initial assessment of the potentials for the antenna. For this
exercise, I replicated the system with the 250' long 3-wire antenna,
creating 3 variations. The narrow version uses a spacing of 1'
between wires for an overall antenna width of 2'. The medium
version uses 5' spacing for a total width of 10'. The wide version
uses 15' spacing and results in a 30' maximum antenna width.
Nevertheless, the leads from the wires to the combined parallel
source remain very short electrically. Fig. 5 overlays the SWR
curves for the 3 versions of the 3-wire antenna.
Chapter 30
The idealized model provides very well-behaved SWR curves for all
3 versions of the antenna. Note that the reference impedance is
450 Ohms, half the value used for the 2-wire antenna and half the
value of the 900-Ohm terminating resistor. We obtained similar
SWR curves in Chapter 29 with the shorter 27.2-m (89') 3-wire
antenna.
Although useful as a preliminary modeling venture, the idealized

model does not represent structural reality for any of the 3 versions
of the 3-wire antenna. For parallel feeding of the antenna, we must
use wires that reach from the outer element center to a common
feedpoint. Therefore, I remodeled the antenna according to the
right-hand sketch. The limitation of this modeling method is the
Chapter 30
need for the center or source wire to equal in length a segment on

the outer wire and for the segments on the connecting wires to be
as equal in length as feasible to the other segment lengths. These
requirements are not always well met, but the resulting models are
adequate enough to detect general departures from the idealized
model.
Fig. 6 shows the overlaid 450-Ohm SWR curves for the revised
model. The all-wire model suggests that the narrow version of the
array is usable above 28 MHz before the SWR seriously exceeds
2:1. (We shall not consider impedance transformation losses and
cable losses that might show a lower SWR at the operating
position.) The medium version begins to show serious SWR
Chapter 30
excursions from about 22 MHz upward. The widest version starts to

exceed the 2:1 SWR standard at about 15 MHz. These curves are
based on a feedpoint 1' below the terminating resistor and may vary
in detail with different positions. As well, the exact structure of the
feed segment and the connecting wires may further alter the
curves. Nevertheless, we can see that the idealized model gives us
too optimistic a portrait of the SWR behavior of the 3-wire wide-
band array.
The picture is not necessarily bleak, however. Many applications for

a antenna of this sort do not require full spectrum coverage. As
well, numerous receiving applications may use relaxed SWR
standards, perhaps up to 3:1 relative to the reference impedance.
So the 3-wire antenna remains a viable alternative to the 2-wire
wide-band antenna, while offering freedom from the front-to-back
differential that besets wider versions of the 2-wire array. The
question is whether the promise of higher gain will justify the more
complex 3-wire array. As a sampling, I have set up a table similar to
the one used for the 2-wire array.
Frequency 2 3 4 5 15 30
Maximum Gain (dBi) E/W B/S E/W B/S E/W B/S E/W B/S E/W B/S E/W B/S
Antenna
Narrow (1') -0.26 -0.26 -0.96 -0.96 -1.35 -1.34 0.00 0.01 2.85 2.92 6.46 6.53
Medium (5') -0.92 -0.91 -0.65 -0.65 -0.81 -0.76 -0.07 0.03 2.66 2.98 4.93 5.50
Wide (15') -0.80 -0.67 -0.24 0.02 -0.23 0.24 -0.68 0.15 1.49 4.14 2.69 6.76
Reference Doublet 2.08 2.77 3.93 4.99 5.39 8.11
Note: E/W = Edgewise; B/S = Broadside
In virtually every sampled case, the 3-wire gain exceeds the 2-wire
gain, and often by a significant margin. The rough average of the
gain differential between the 3-wire narrow antenna and the doublet
is about 3 dB, just over half the deficit shown by the 2-wire array.
Chapter 30
From a raw gain perspective, the 3-wire array is attractive for

applications committed to a wide-band terminated antenna.
However, the 3-wire array is not immune to pattern distortion. One

form is evident from the tabulated data. As we widen the spacing
between wires and increase frequency, the broadside gain shows
ever-larger values relative to the edgewise gain. The differential
likely makes no great difference up through medium spacing.
However, the wide-space version shows well over a 1-dB
differential from the frequency mid-range upward. Note that the
differential shows itself most vividly in the frequency region in which
the wide-space version shows the largest SWR excursions. As well,
the tabulated data does not show a clear gain advantage over
medium and narrow spacing.
A second form of pattern distortion appears in the wide version of

the array within the upper frequency edgewise patterns themselves.
Fig. 7 shows the 15-MHz patterns for both the broadside and
edgewise planes of the wide version of the antenna. For the
edgewise pattern, visualize the antenna as extending vertically
within the plot. For the broadside pattern, orient the antenna
horizontally with respect to the graph. On the right side of the figure
are the patterns for the narrow version of the antenna. These plots
follow the form of a single-wire doublet, but at lesser strength. The
broadside pattern for the wide version of the antenna almost
replicates the pattern for the narrow 3-wire antenna. However, the
edgewise pattern for the wide version is significantly different. As
well, the tabulated data shows this pattern to be not only weaker
Chapter 30
than the broadside pattern, but also weaker than the edgewise
pattern for the medium and narrow versions of the array.
Chapter 30
To establish that the 15-MHz pattern is not an isolated instance of

more severe pattern distortion, Fig. 8 shows the patterns for 30
MHz, using the same format. Once more, the patterns for the
narrow antenna show little, if any, difference between broadside
and edgewise views. However, the wide antenna shows changes to
both patterns. The broadside patterns show a widening and
shrinking of the peak values of the minor lobes. The edgewise
pattern shows the opposite development, although some careful
observation is necessary to see it. In the narrow edgewise pattern,
careful scrutiny will show some very tiny minor lobes between the
larger minor lobes--almost invisible without either a table of
radiation pattern values or a gross enlargement of the pattern. In
the edgewise pattern for the wide antenna, those lobes have grown
to equal size with the other minor lobes to form a large and
complex set of lobes.
Chapter 30
Unlike the 2-wire wide-band terminated antenna, which showed a

significant improvement of gain as we widened the space between
wires, the 3-wire array does not show the same gain development
when we model it using an all-wire configuration. Rather, we find
the relative gain values of the 3 widths simply to vary across the
Chapter 30
spectrum. In some cases, the narrow version yields the highest

gain; in others, it does not. When we add to this gain variability the
fact that only the narrow version promises a stable SWR curve
across the entire operating spectrum, we begin to approach a
conclusion. Add in the absence of significant pattern distortion and
a relative simplicity of construction and the conclusion becomes
more solid. In a 3-wire wide-band array, the narrow version has
perhaps the most potential of the 3 widths for actual use.
The narrow 3-wire array holds the promise of higher gain by a

significant margin over the 2-wire array, although the actual gain
margin will change from one frequency to the next. However, even
the 3-wire array falls significantly short of the gain offered by a
single wire doublet that uses no termination. In amateur radio
service, where the use of parallel transmission line and a wide-
range antenna tuner in the shack may serve very well to handle
frequency changes, the single-wire doublet is still the antenna of
choice. For short-wave reception--especially in Europe, where
overloading signals are common--the 2-wire terminated system
may be the antenna of choice, since the overall signal reduction
may prevent or at least ease receiver overload and resultant
spurious products. Only where a system needs both to transmit as
well as receive and to be able to change frequencies without any
equipment retuning does the 3-wire system come into its own--so
long as there is excess receiving gain to compensate for the loss of
sensitivity and there is excess power available to make up for the
losses within the terminating resistor.
Chapter 30
The 5-wire terminated wide-band antenna showed great promise of

better approaching the level of gain performance achieved by the
simple single-wire doublet while providing a possibly usable SWR
curve. Of course, like all of the antennas in our survey, the initial
models checked only the broadside free-space patterns and used
the idealized model for initial checks. The leftmost part of Fig. 9
shows the end view of that modeling scheme.
Converted to the scale used in this exercise, the antenna is now

250' long and has a total width of 4' for the narrow version, 20' for
the medium type, and 60' for the widest version, using parallel wire
spacing of 1', 5', and 15', respectively. In addition, we shall orient
the antenna so that the edgewise view is parallel to the ground,
although we shall take interest in the differential between the
edgewise and broadside patterns in free space.
The antenna uses a 900-Ohm terminating resistor, the same value

as used in all of the other wide-band antennas in this exercise. The
required SWR reference impedance turns out to be 300 Ohms for
all variations on the 5-wire antenna. We may note in passing that
Chapter 30
the decrease in the required SWR reference impedance undergoes

a regular progression in its descent as we add wires to the array.
Fig. 10 reviews the 2-30-MHz SWR curves for the narrow, medium,
and wide versions of the 5-wire array. Even under the idealized
feed conditions with near-zero-length leads for the parallel-
connected wires, the SWR curve is somewhat limited. The narrow
antenna provides the best curve, although the SWR is somewhat
high at the low end of the operating spectrum. As we increase
spacing, the curve improves at the lowest frequencies, but the wide
version appears to be usable only up to the middle of the spectrum
(about 16 MHz).
Chapter 30
We may proceed in two general ways to create all-wire models with

more realistic feed systems. The center sketch in Fig. 9 shows a 2-
lead version (A). Wires extend from the inner fed element to the
central feedpoint 1' below the terminating resistor. The outer
elements simply connect to the inner elements to complete the
overall feed system. In effect, the outer elements acquire extra
length compared to the inner wires, with another increment of
length added at the far end of the wires. Fig. 11 shows the resulting
SWR curves for the 3 antenna widths.
Note that the SWR curves begin to gyrate widely and somewhat
wildly for the widest version. In common, the curves show
increasingly high peak SWR values as we raise the operating
Chapter 30
frequency. The narrow version of the array is the only one usable
for most of the operating spectrum, but only if we relax the 2:1
SWR limit standard.
As an alternative, I modeled the arrays with separate leads from

each fed wire to the common feedpoint. The right-most sketch in
Fig. 9 shows the general outline of the 4-lead model (B). The
question was whether this feed system would alter the SWR curves
relative to the 2-lead model. Fig. 12 shows the results of the trials.
Clearly, we do not gain anything by using the alternative feed

method. Although the details differ in terms of the exact frequencies
and values for SWR peaks, only the narrow version of the 5-wire
Chapter 30
antenna shows potential for extended frequency use. The medium-

width version of the antenna is again usable up to about the middle
of the spectrum, and the wide version shows a rapid decay of good
SWR performance above about 7 MHz. Varying the reference
impedance does not alter the performance significantly. The key
problem in all versions of the 5-wire array is the presence of very
significant reactance levels at most frequencies.
Despite the initial optimism offered by the idealized model in terms

of developing a 5-wire wide-band antenna, remodeling the array
more realistically with wire leads to the center feedpoint presents
serious obstacles to further development. Only the narrow version
of the antenna has a bandwidth potential resembling the curves
that we obtained for the 2-wire and 3-wire antennas--and then only
if we relax the SWR standard. However, one feature that gave the
basic idea of a 5-wire array its allure was the potential for
significantly higher gain than either of the smaller antennas.
Despite the SWR problems, we should explore this facet of the
antenna. The following table of sample values parallels the one for
the 3-wire antenna to provide for direct comparisons. As we did
when checking maximum gain for the 3-wire antenna, we shall
provide gain values for both the edgewise and broadside planes
relative to the antenna. In addition, we shall examine values for
both the 2-lead and the 4-lead versions of the array.
Chapter 30
Frequency 2 3 4 5 15 30
Maximum Gain (dBi) E/W B/S E/W B/S E/W B/S E/W B/S E/W B/S E/W B/S
2-Lead Antenna
Narrow (1') 1.59 1.59 0.28 0.28 -0.16 -0.14 1.72 1.74 3.74 3.83 7.14 7.29
Medium (5') 0.93 0.97 0.73 0.82 0.73 0.90 1.56 1.85 3.33 4.19 5.46 6.56
Wide (15') 0.84 1.19 0.92 1.71 1.07 2.60 -0.42 2.50 0.79 6.92 4.93 -0.22
4-Lead Antenna
Narrow (1') 1.44 1.44 0.36 0.37 0.03 0.04 1.74 1.76 3.81 3.91 7.02 7.16
Medium (5') 0.81 0.85 0.53 0.62 0.56 0.73 1.46 1.75 2.54 3.36 4.17 0.78
Wide (15') 0.76 1.11 0.71 1.50 0.94 2.47 -0.42 2.55 1.54 8.82 5.11 -0.71
Reference Doublet 2.08 2.77 3.93 4.99 5.39 8.11
Note: E/W = Edgewise; B/S = Broadside
As expected, the wider the array and the higher the operating
frequency, the greater differential that we find between edgewise
and broadside maximum free-space gain values. In general, there
is no significant difference between the gain behavior of the 2-lead
and 4-lead versions of the 5-wire antenna. Both narrow versions
show a good coincidence between the edgewise and broadside
gain values. However, as we widen the antenna, the upper
spectrum values actually show a decline relative to the narrow
version of the antenna. As well, the differentials grow to
considerable proportions, with one value set showing a 7-dB
differential. Initially, then, the first order conclusion might be that
only the narrow version of the array holds potential for wide
frequency use. Interestingly, this conclusion from free-space gain
data coincides with the conclusion suggested by the SWR data.
The sampled gain values reveal another interesting pattern to

antenna performance. Let's overlay a few patterns for the single-
wire reference doublet and the 5-wire narrow antenna, using the 2-
lead version. However, let's examine both the free-space values
and the values 75' above average ground (conductivity 0.005 S/M,
permittivity 13). Additionally, we shall restrict the frequency range to
extend from 2 to 6 MHz. This last measure operates on the premise
Chapter 30
that not all applications of wide-band antennas require coverage of

the entire HF spectrum. Instead, some operational needs require
no-tune operation of some part of the spectrum. Fig. 13 provides
the SWR curves for the 5-wire array using the 300-Ohm standard
and for the single-wire doublet using a 75-Ohm standard.
The antenna height (75') is only a small fraction of a wavelength at

the lower end of the spectrum. Hence, we find deterioration of the
SWR curve for both antennas relative to the free-space model. The
doublet curves clearly show that the antenna is longer than 1/2
wavelength at the lowest operating frequency. As well, the doublet
curves show the need for extensive impedance matching efforts as
we change frequency. Although the 5-wire SWR curves are not
Chapter 30
perfect, they present far less of a matching challenge to 300 Ohms

(and from that value down to a coaxial cable value via a wide-band
impedance transformer).
Now let's compare the maximum gain values, letting ground-

reflection phenomena settle any remnant differential between
edgewise and broadside gain values that we encountered in free
space. The following table tracks doublet and 5-wire narrow values
every half-MHz for our reduced operating spectrum.
Maximum Gain values for a Single-Wire Doublet and a 5-Wire 2-Lead Narrow Array 75'
above Average Ground
Single Wire Doublet
Frequency 2 2.5 3 3.5 4 4.5 5 5.5 6
Maximum Gain (dBi) 6.84 7.07 7.09 7.30 7.87 8.73 9.42 8.09 7.64
Elevation Angle 90 90 71 55 48 41 36 33 28
5-Wire 2-Lead Wide-Band Antenna
Frequency 2 2.5 3 3.5 4 4.5 5 5.5 6
Maximum Gain (dBi) 6.22 5.77 4.53 3.95 4.19 5.20 6.51 6.11 7.53
Elevation Angle 90 90 72 56 47 41 36 33 29
The table shows that at the low and high ends of the restricted
operating range, the gain values for the 5-wire array virtually match
those of the single-wire doublet. At the center of the range, from 3.5
to 4.5 MHz, we find the greatest difference in values, with a 3.5-dB
deficit. Fig. 14 provides a few sample patterns for both antennas,
overlaid to show the degree of coincidence or difference.
Chapter 30
Once the antenna passes the 1.25-wavelength mark in electrical

length, it of course results in a pattern where the main lobes are no
longer broadside to the wire, as shown by the 6-MHz pattern.
However, we may also view a more fundamental correlation by
jointly examining Fig. 13 and Fig. 14 (or the table). The doublet 75-
Ohm SWR values--and hence, the general level of the feedpoint
impedance--are lowest at or near those frequencies for which the 5-
wire array shows a gain level that most closely matches the
doublet's gain. When the parallel combination of fed wires in the
wide-band antenna would present a very low impedance, the
terminating resistor absorbs (and dissipates) the least energy,
resulting in the highest gain. When the resistor value is low relative
to the feedpoint impedance without it, the resistor handles a
proportionately higher percentage of the power, leaving less for
radiation. This theory of operation applies to all multi-wire wide-
band antennas using a terminating resistor. The 5-wire version of
the antenna simply makes the process more graphically apparent.
Chapter 30
With a restricted operating range, the antenna models make

apparent the danger of simply taking an average value of gain
deficit of a wide-band antenna relative to a doublet. The more
pressing question facing a design engineer is whether the worst-
case deficit falls above or below the level of acceptability relative to
the intended application.
Since we have explored the difference between the SWR curves for
the 5-wire, 2-lead antenna over a limited range, we might also
explore whether the differences apply to other parts of the overall
operating spectrum. Fig. 15 shows the comparative plots for the
narrow 2-wire wide-band antenna. The antenna is 75' above
average ground for all comparisons. As the plots show, the SWR
curves diverge until about 9 MHz. At that frequency, the antenna is
about 0.7 wavelength above ground. Below that frequency, we find
small differences in the curves that likely would not create any
operational problems.
Chapter 30
In Fig. 16, the graph compares SWR curves for free-space and
over-ground versions of the narrow 3-wire wide-band antenna. The
curves overlay each other very well down to just below 10 MHz.
Again, the 0.7-wavelength (or perhaps the 0.75-wavelength) height
marks the beginning of SWR curve divergence. Note that in this
case, the modeled over-ground value for the all-wire antenna just
exceeds 2:1 at 2 MHz. The actual test measurement for the
antenna under these conditions would depend upon impedance
transformation and line losses, as well as construction variables
relative to the model.
Chapter 30
Fig. 17 expands the narrow 5-wire, 2-lead SWR curves shown in

Fig. 13 to encompass the entire potential operating spectrum. Once
more, the free-space and over-ground curves track each other very
well from about 9.5- to 10-MHz and upward. Below the transition
frequency area, the curves diverge, with the over-ground curve
showing a higher SWR value at the lowest frequency in the
spectrum covered. Lengthening the antenna might well move the
SWR from 2 to 6 MHz below 2:1, but at the same time, it would
reduce the frequency span over which we can obtain bi-directional
patterns. The relative importance of each factor is an application-
specific determination.
Chapter 30
Construction Variables
The sensitivity of the multi-wire terminated antenna arrays to

changes in width relative to various performance characteristics
suggests that these notes inevitably fall far short of covering all
possible design variations. However, length and width are not the
only variables available for alteration. One significant variable may
escape attention. The all-wire models place the common feedpoint
1' below the terminating resistor and its element wire. We may vary
that placement and see what might happen. To test the matter, I
created a very narrow version of the 3-wire (all-wire) antenna using
a spacing of 0.5' between wires. Then I set the feedpoint 1, 2, and 3
feet below the antenna, using a free-space model. The key
Chapter 30
question in this exercise focused on the effects of the 3 placements

on the SWR curve. Fig. 18 shows the results.
The very narrow version of the 3-wire array shows the most
acceptable SWR curve with the common feedpoint 3' below the
terminating resistor wire, the limit of this particular exercise. The
450-Ohm SWR does not exceed 2:1 until the operating frequency
reaches 29 MHz, and then not by much. To see whether the
phenomenon is unique to the very narrow spacing or more general,
I repeated the test using the standard narrow 3-wire spacing (1').
The results appear in Fig. 19.
Chapter 30
The results for the standard narrow spaced 3-wire array are less
dramatic. The closest feedpoint spacing shows the widest
excursions of SWR--both high and low--for most of the operating
range. The widest feedpoint spacing appears to improve SWR
performance in the upper range, but not to the degree possible with
very narrow spacing.
Wide-band terminating-resistor arrays are subject to many

construction variables, and these simple exercises provide only a
sample. However, they do show that we cannot assume that any
particular variable is either significant or insignificant until we
examine it in detail.
Chapter 30
Conclusion
At the end of Chapter 29, based on a limited exploration of

terminated antenna properties, I reached a set of conclusions. We
may now assess how well each holds up and what qualifiers we
may have to place on some of them.
1. All terminated wide-band "folded dipoles" have knee frequencies,

below which the gain drops very rapidly. The recommended
operating range for any of the antennas is from an electrical length
of about 1/2 wavelength upward in frequency.
The first conclusion remains correct, although we by-passed testing

it in this exercise by using an antenna that was longer than the
critical minimum length at the lowest test frequency.
2. As we add more fed wires to a terminated antenna, we increase

its average gain over the operating spectrum. The gain increase
never quite reaches the level of a single-wire doublet.
We must heavily qualify this statement. Although the average gain

of the 3-wire array exceeds the average gain of the 2-wire version,
and the 5-wire average gain is higher still, average gain may not be
the key factor in making design decisions. The gain will be highest
wherever the equivalent doublet length shows the lowest
impedance. However, we must keep a sharp eye out for the lowest
gain levels within a proposed operating span to determine if the
gain at critical frequencies is high enough for the proposed
communications application. We may examine that gain as an
Chapter 30
intrinsic value or as a deficit relative to the single-wire doublet,

depending on the frame of reference.
3. As we add more wires to a terminated wide-band antenna, the

center or reference SWR impedance decreases both intrinsically
and with respect to the value of the terminating resistor.
Our extended exercises actually provided a bit more precision to

this statement by standardizing the terminating resistor at 900
Ohms and watching the required SWR reference impedance. The
criteria for setting the SWR reference impedance are not precise,
since the setting requires a judgment call as to what counts as the
smoothest obtainable SWR curve over a given operating region.
Nevertheless, the 2-wire array showed its best curves when the
SWR reference impedance matched the value of the terminating
resistor. The 3-wire array gave the best results when the SWR
reference impedance was 1/2 the value of the terminating resistor.
With the 5-wire arrays, the best reference impedance was 1/3 the
value of the terminating resistor.
4. 2- and 3-wire terminated wide-band arrays show stable SWR

curves through their operating ranges. However, adding further
wires tends to produces curves with greater SWR excursions
relative to the reference impedance.
Once we modified the 3-wire and 5-wire antennas to provide all-

wire model construction, the stability of even the 3-wire curves
began to slip badly as we widened the array. The SWR
performance for the 3-wire array showed wide SWR swings in the
Chapter 30
wide version. The 5-wire all-wire model strongly suggested that it

was useful only over restricted frequency ranges, and then only in
the narrow version.
5. Terminated wide-band antennas show increased gain by

widening the distance between wires. Spacing adjustments may
require revision of the optimal terminating resistor value and the
reference SWR impedance.
The final conclusion in the series requires the greatest modification.

The initial models registered gain as a function of free-space
patterns broadside to the plane of the wires. Hence, they could not
show the growing differential of edgewise and broadside gain as we
increased the spacing between wires. Our exploration of the 5-wire
models in this extended exercise shows that the net gain of a wide
model may not always exceed that of a narrow model. Moreover,
the existence of any differential at all makes a strong
recommendation for modeling a proposed design over ground at
the anticipated height of actual use. At the low end of the operating
range, we may fairly gauge the effects of the mounting height and
soil type on the SWR performance. As we move up the spectrum,
the gain differentials increase, and modeling over ground allows us
to arrive at a single gain value, whether we handle it independently
or in comparison to the single-wire doublet that the wide-band
antenna ostensibly replaces.
To the list of conclusion derived from Chapter 29 and modified

here, we may add a new one.
Chapter 30
6. Due to the many construction variations possible with a multi-

wire wide-band terminated antenna, range testing at the anticipated
use height and over the anticipated ground quality is an essential
ingredient in the development of a successful antenna.
Chapter 30
Chapter 31: About the Folded Monopole
L
iterature about multi-wire monopoles is fraught with odd
labels for the structures. We can find terms like folded
monopole, folded isopole, caged monopole, and skirted
monopole. If we can find legible diagrams for what the labels label,
we are in for something of a surprise: they all refer to the same
class of antenna. However, some engineers prefer to reserve the
title of folded monopole for an antenna with only 2 wires. Others
apply the term more generally to all or most multi-wire monopole
systems. To the best of my knowledge, the skirted monopole
terminology arose when the outer wires served as a means for
detuning the monopole--usually a tower--from its sensitivity to
interact with nearby (near-field) AM BC transmitting antennas.
Hence, detuning skirts are common on urban cell towers. However,
we can also feed the skirt on a transmitting tower and obtain a
measure of impedance transformation and control that turns out to
be useful. We might speculate that the expression caged monopole
arose as a somewhat more politically correct than the term skirted
monopole. Regardless of the humor we may make out of the
terminological morass, we are left with a basic question.
Is the multi-wire skirted or caged monopole a folded monopole or

isn't it?
Let's start at some sort of beginning by looking at some forms of

multi-wire monopoles. Fig. 1 shows a few of the many possible
configurations. On the far left is a standard 2-wire folded monopole.
It forms a touchstone for all that follows. At this stage, I shall note
Chapter 31
only one interesting property of the standard folded monopole.

Edgewise to the wires, we find a very tiny (and operationally
insignificant) asymmetry to the antenna's gain. For ordinary wire
spacing, the differential might be up to 0.04 dB in models, favoring
the feedpoint side of the antenna. The broadside gain is the
average of the edgewise gain values. Next to the 2-wire folded
monopole is a 2-wire folded monopole with an extension on the
return-wire side. (In practice, it makes no difference whether the
extension connects to the fed wire or to the return wire if the
connecting wire is short enough as a function of a wavelength.) I
drew the antenna in the manner shown because it shows the
relationship of a folded monopole with an extension to a gamma or
omega matched tower used by some amateurs on 40 meters. If we
assume that the antenna would be self-resonant without the
extension, then with the extension, we find an increase of both the
resistive and the inductive components of the feedpoint impedance.
If the fed wire length does not result in a self-resonant antenna
without the extension, then we usually call the added fed wire a
matching line. The preferred term here is a function of what we are
trying to achieve more than it is a difference in the electrical
performance of the antenna. Like all folded elements, we shall find
both a transmission line function and a radiating function within the
folded section.
Chapter 31
The sketches jump to 4- and 5-wire structures. However, a 3-wire

monopole system is both possible and interesting. If we place wires
on opposite sides of a return wire/tower and if we feed both new
wires in parallel, we obtain a rudimentary caged or skirted system.
Like the more complex 4- and 5-wire systems in Fig. 1, it produces
a symmetrically circular azimuth pattern. In fact, within very narrow
limits--largely a function of the fact that the more complex cage
Chapter 31
systems tend to result in slightly shorter antennas when they are

self-resonant--the gain of all types of caged, folded, or skirted
monopoles is the same. (Some models of these antennas have
shown more deviant values, but they usually correct to the basic
value when adjusted for the average gain test (AGT) score of the
model.) Over perfect ground using lossless wire, the entire set of
self-resonant monopoles show a median gain of 5.15 dBi, with less
than about +/-0.04 dB variation.
The two sketches on the right of Fig. 1 differ in only one small way.
The 5-wire monopole uses one or more wires that circle or girdle
the outer wires only. In large installations, the connecting wires
serve an important mechanical goal to help rigidify the cage of very
long out wires. Also note that the sketches show a set of
connecting wires at the base with a single feedpoint between the
wire and ground. (We shall presume that all return wires in the
figure return to ground.) In theory and in practice, the single
feedpoint can result in slightly different current magnitudes and
phase angles along the outer wires. The connecting wires create
short circuits along the structure that tend to equalize currents
along the outer wires.
We may replace the energy source with a network and change the
system function from radiation to tower detuning. There are
numerous techniques that allow engineers to use essentially the
same caging wires to detune a tower from most frequencies within
the upper MF range. Some techniques may involve modifications to
the cage of wires as well as to the base network. Tower detuning
via skirts has become a fairly routine and commonplace
Chapter 31
engineering service. Its necessity and profitability has increased

with the proliferation of cell-phone and other UHF/microwave
towers that now pervade urban, suburban, and rural landscapes.
For our purposes, we may bypass the detuning role of skirts and
cages in order to address more directly our initial question: are they
forms of folded monopoles? To approach an answer, let's begin by
seeing what makes the 2-wire folded monopole so special.
The Folded Dipole and the Folded Monopole
Hardly a soul among those interested in antennas does not know

about the folded dipole. Unfortunately, what many folks know is
only a tiny piece of the story. If we parallel two identical wires at a
reasonably close and constant spacing, if we connect the ends and
feed one of the wires, and if we bring the antenna to resonance,
then the feedpoint impedance on the selected fed wire is 4 times
the impedance of a linear resonant dipole. There is much more to
the folded dipole story than this, and I have tried to tell some of it in
Chapter 27.
Part of the story involves the fact that a folded dipole and a linear
dipole have the same gain and pattern. Another part of the story
involves the fact that a folded dipole is two devices in one. It is a
dipole and has radiating currents that almost exactly parallel the
radiating currents of a linear dipole in both magnitude and phase
angle. The folded dipole is also a transmission line (or, counting
from the feedpoint, two transmission lines with a common starting
point) with a relatively constant current magnitude and phase angle
Chapter 31
(that is 90 degrees out of phase with the radiating current). John

Kuecken showed how to separate the two currents in Antennas and
Transmission Lines (pp. 224 ff).
Perhaps the most significant part of Kuecken's account is that he

describes the technique in connection with the hairpin monopole
(another name for the folded-monopole list of aliases). The
technique applies equally both to folded monopoles and to folded
dipoles, since the former is simply half the latter if terminated in a
perfect ground or in a ground plane the approximates a perfect
ground. If we make the folded structure self-resonant using a
perfect ground and lossless wire, a dipole will show about 72 Ohms
for a feedpoint or source impedance (resistive, of course), while a
linear monopole will show 36 Ohms. Folded versions of the two will
shows 288 and 144 Ohms, respectively under suitable conditions.
Fig. 2 shows the correlation of the two antennas and their linear
roots.
Chapter 31
The "suitable conditions" clause of the folded antenna impedance

report presumes that the two wires in the folded structure have the
same diameter. As well it presumes that the two wires are close
enough to form a transmission line rather than simply an open loop
or open half-loop. Unfortunately, too many amateurs are unaware
that we may effect other impedance transformations by varying the
diameters of the two conductors, or the spacing between them, or
both. Fig. 2 hints at that wider range of potentials by designating
the wire spacing as s, the diameter of the fed wire as d1, and the
diameter of the return wire as d2. How these dimensions (all in the
Chapter 31
same unit of measure) go together to effect an impedance

transformation appears in the following equation.
If d1 and d2 are equal, then the right side of the expression in () is

1, and to that number we add 1, to get 2, which squares to 4 as the
value of R, the ratio. Hence, the most common case of a folded
dipole multiplies the linear dipole impedance by 4. A folded
monopole meeting the same conditions does likewise. Next, let's
make the return wire diameter go to an infinitesimal value. We
cannot let it go to zero or we cannot have a return wire, but an
infinitesimal diameter will suffice to leave us with a wire, but one
that reduces the d2-diameter and right side of the expression in ()
to a value insignificantly different from zero. Within the () we now
have a value of simply 1, which squares to 1. This condition sets
the minimum transformation in a folded dipole or monopole. In
other words, a folded mono-/di-pole cannot transform an
impedance downward from the linear value--only upward. On the
other hand, making the return wire very small drives the
denominator on the right side of the expression in () toward zero,
increasing the value of the fraction to an indefinitely high value. For
most antenna work, resulting ratios (R) greater than 10:1 are
seldom encountered. However, we often find conversion ratios
Chapter 31
above 4:1, especially in gamma-match structures, where the

gamma rod is considerably thinner than the main element to which
it attaches.
Fig. 2 also labels the end wires that must be part of any real folded
antenna. I note these wires because they do have an effect on the
physical version of what we calculate from the equation. The wires
must be short enough that their effect is relatively insignificant.
However, their effect is real and shows up in computer models of
folded dipoles and monopoles. One easy way to see the effect is to
create a resonant model of a folded dipole using wires having
different diameters. Now alternately use the fat wire or the thin wire
diameter for the end wires and recheck the required length for
resonance. You may also wish to look at the current tables
available in NEC and MININEC for additional confirmation of the
effect of end wires. To make the effect more vivid, use a lower
frequency with a fairly wide physical spacing (such as 3' at 3.5
MHz) and use enough total segments so that the end wires have
multiple segments.
If we create a multi-wire cage around the center wire, we can

achieve two different orders of phenomena. We place the source
on the center wire, and then the set of cage wires (any number
from 2 upward for a total folded antenna count of 3 upward)
increases the impedance transformation ratio. We rarely encounter
this situation. However, users of cages around a central mast or
tower in the upper MF and lower HF regions often feed the cage
wires in parallel, allowing the fat center wire to serve as the return
wire. This practice serves a number of ends. First, it allows the
Chapter 31
central tower a direct connection to ground. Not only does this

move simplify the tower's mechanical structure, it also is an
important safety measure. Second, using the central mast or tower
as the return wire results in an impedance transformation ratio that
is lower than 4:1. By choosing the correct number of wires for the
cage or skirt and feeding them in parallel, we can obtain an
impedance that is higher than a monopole's 36-Ohm value over
perfect ground but much lower than the 4:1 value of 144 Ohms.
Indeed, a 4-wire cage provides a value that is very close to 50
Ohms, virtually ideal for coaxial cable.
The Root of the Issue
The problem that we encounter with both physical antennas and

their NEC models is that the resultant impedance does not coincide
with the basic folded dipole/monopole equation. The resultant
impedances that we encounter when connecting parallel sources
together for a caged or skirted monopole do not answer to any
simple relationship to the 2-wire folded monopole (or dipole). As a
consequence, some engineers hesitate to identify the caged or
skirted monopole with the 2-wire monopole.
We might initially treat caged monopoles in a variety of ways. For

example, we might consider the structure to be a version of a
coaxial monopole in which the return wire forms a center conductor,
with the outer wires forming the outer conductor corresponding to
the braid on an ordinary coaxial cable. However attractive this
picture may be, we also must recognize that the outer wires leave
mostly empty space. In addition, if we apply the 2-wire equation to
Chapter 31
this situation, then the value of d1, the fed wire, becomes identical
to the value of 2s, that is, twice the spacing between conductors.
The common log of 1 is zero, resulting in an impedance
transformation of 1. Hence, a resonant monopole under this
treatment would show a 36-Ohm impedance. However, cage
monopoles show a higher impedance, with the actual value being
partially a function of the number of wires.
Alternatively, we may model various samples of cage monopoles

and, with due attention to AGT scores, arrive at reasonable
estimates of the resulting impedance. We quickly discover a factor
that the coaxial treatment cannot take into account directly. For any
given return wire (d2) diameter, as we change the diameter of the
outer wires, the resultant antenna impedance will change. The
coaxial treatment can use only one value for d1, taken just now to
mean the overall outside diameter of the cage system. It is possible
to obtain from the basic equation an impedance that is equal to the
modeled (or field tested) value by using a selected value of
diameter for d1. However, this treatment is initially ad hoc. It
involves varying the diameter of d1 until the impedance value
matches the modeled or field value. If we survey enough values,
then we might use some form of regression analysis to arrive at
correlations that would be usable for almost any (practical)
combination of center diameters, spacing values, and cage-wire
thicknesses. Nevertheless, regression analysis is an excellent tool
for establishing a mathematical correlation to a set of curves
derived from observation, but it does not provide an explanation of
the correlation. The equations would not necessarily fit any set of
Chapter 31
known electrical foundation equations to provide a seamless

continuum between 2-wire and multi-wire folded dipoles.
The absence of a means of direct derivability that might ensure that

cage monopoles are a variety of folded monopole--or that might
provide sufficient reason to withhold the connection--does not mean
that we must give up on the problem. There may be other
alternatives yet to be explored. Any such alternative must recognize
that the presence of multiple fed wires that do not surround the
center conductor or return wire will present disturbances to a strict
correlation. Indeed, perhaps part of the past thinking that wishes to
claim that a cage monopole is not a folded monopole has looked
too strictly at what a cage does not do and too little at what it does
do. Granted, a cage lacks the solidity to form a true coaxial surface.
However, let us suppose that each fed wire forms with the central
return wire a 2-wire folded dipole. Under this supposition, we can
expect the fed wires to interact or mutually couple. The degree of
interaction would vary with the diameter of the center return wire,
the spacing, the diameter of each out fed wire, and, of course, the
relative diameters of the fed and return wires. Despite the variables,
we should still be able to detect a pattern of values that we can
trace to the results of a 2-wire folded monopole.
A Test and Its Limitations
Within certain limits, we may set up a fairly simple modeling test.

For the test we shall use a series of monopoles, as shown in Fig. 3.
Each monopole will be resonant at a given test frequency. Since we
shall use lossless wire and perfect ground to minimize the number
Chapter 31
of extraneous variables, virtually any frequency will do. My test

frequency will be 3.5 MHz. When modeling each test structure, the
monopole will be resonated to within +/-j0.1 Ohm.
The outer wires will have a constant diameter, 0.1". I shall vary the
diameter of the center or return wire in steps. Since a zero diameter
is not possible, I shall use 0.1-Ohm as the minimum return wire
diameter. The remaining steps will be linear: 0.25", 0.5", 0.75", and
1". I shall end the progression at this point for modeling reasons. As
the diameter of the center conductor increases, the AGT score
Chapter 31
departs ever further away from the ideal value. Using NEC-4, the
value becomes unreliable more quickly than with MININEC (using a
suitably corrected implementation of the public domain version
3.13), but both programs eventually fail to yield very usable results.
However, we shall be able to track the requisite results over the
range of selected values and arrive at some preliminary
conclusions. The procedure is no less and perhaps no more
approximate than the industry-standard practice of using single wire
substitutes for complex geometries that more adequately reflect
open tower structures. The AM BC industry regularly uses a wire
radius of 0.37 times the face dimension of a triangular tower and
0.56 times the face of a square tower.
Besides varying the diameter of the return wire, the test models will
also vary the center-to-center spacing between the return wire and
the fed wire(s). I shall use spacing values of 12", 24", and 36". The
top wires will use a diameter of 0.1", the same used for the fed wire.
Each top wire will use a segment length of 1' (12"). The vertical
wires will use 60 segments each, a value that produces segment
lengths between about 0.9 and 0.95 of a foot. The very small
disparity between the top-wire segment lengths and the vertical
wire segment lengths does not jeopardize current calculations in
NEC 4.1, since the feedpoint is in the lowest segment of each outer
vertical wire. Tests using segment lengths that are as a close to 12"
as possible result in variations of the feedpoint impedance by no
more than 0.02 Ohm relative to either the resistive or reactive
component.
Chapter 31
One aspect may be unique to these test models, although it is a

common modeling practice. The uniqueness is the attention that we
shall give to it. Most NEC modelers attempt to supply the model
with a single feedpoint to end up with an overall impedance for the
structure. There are a number of ways in which we can perform the
feat. We might elevate the ends of the fed wires and provide
shorting wires between legs ends. Then we may run a single wire
to ground and place a source on this wire. Alternatively, we may
run transmission lines to a remote wire and place a source on it. If
we select a near-zero length for the transmission line lengths
(which are independent of the actual distance to the remote wire),
we effectively connect the four base segments in parallel.
In MININEC we must--and in NEC, we may--simply place sources

at the ground end of each fed wire. Of course, the number of
sources will be one less than the number of vertical wires in the
model. A 5-wire cage monopole will have 4 sources. Since the
impedance values for all of the sources will be identical, the net
impedance will be the resistance and the reactance at each source
divided by the number of sources. This derived value will be the
same as the one we would produce by using the transmission-line
technique of paralleling the sources. What differs is that we shall
have--and pay close attention to--the source impedances of the
individual fed wires.
The test will include 2-wire folded monopoles, even though we may
calculate the impedance transformation ratio from the standard
equation. We shall be as interested in how close the fit may be
between the model and the equation result as for any other form of
Chapter 31
folded monopole. A series of monopole models of lossless wire

over perfect ground yields a resonant impedance of 36 Ohms.
Hence, the reference impedance for each new combination of
return wire diameter (d2) and spacing (s) will be 36 Ohms times the
equation-based ratio. We shall be comparing the modeled
impedance values for each fed wire with the reference impedance
value.
The test will initially yield reported values of source impedance. As

well, each test will have an AGT score. Over perfect ground, the
ideal AGT is 2.000, although the value is 1.000 in free space. The
free-space equivalent of the ideal perfect-ground AGT is simple 1/2
the perfect-ground value. I shall use this free-space equivalent
value in test reports, since it plays an important role in arriving at a
usable impedance value. One reason for resonating the folded
monopoles to such close tolerances is to allow us to use the AGT
score to arrive at reasonably reliable values of resistive impedance.
However, with folded structures, we must alter NEC manual

procedures somewhat. For a linear element--whether a dipole or a
monopole--we normally multiply the AGT times the reported
impedance to arrive at a corrected source impedance. The
correction is most reliable when the impedance is virtually resistive.
With folded dipoles and monopoles, we must reverse the correction
procedure. To arrive at the usable impedance value, we must
divide the reported impedance by the AGT score. As a matter of
course, we shall also report the net impedance of the structure.
More significant for our interests will be the ratio of the adjusted or
corrected impedance value for each leg to the calculated 2-wire
Chapter 31
impedance for a folded monopole having the wires and spacing

applicable to a single fed wire and the center return wire.
Some Test Results
Table 1 provides the results for the series of modeling tests that I
just described. The table lists 5 test sequences, each of which uses
a different diameter for the inner, center, or return wire. The first
column lists the 3 steps of spacing. The second column tells us the
total number of vertical wires in the assembly. The number of fed
wires is 1 less than the total. The third column lists the height in feet
of the resonated skirt-fed monopole. The "Raw R" column lists the
NEC 4.1 source resistance value before correction. The "Raw X"
column provides a record of how close the model came to perfect
resonance. The final column of raw data lists the AGT score using
hemispherical increments of 5 degrees.
The remaining columns list the operations performed on the raw

NEC data. The adjusted source resistance per leg results from
dividing the raw resistance by the AGT score. The net source
resistance appears in 2 columns, the first dividing the raw leg
resistance by the number of source, the second dividing the
adjusted leg resistance by the number of sources. The final column
lists the ratio of the adjusted leg resistance to the calculated 2-wire
impedance.
Chapter 31
Chapter 31
Chapter 31
The resonant antenna height has only suggestive utility beyond

showing the trends that confirm the propriety of model construction.
As we increase the number of wires or the diameter of the center
return wires, or the spacing between any fed wire and the return
wire, the resonant height decreases over perfect ground. Although
the trends are true, the exact resonant height of an actual folded
monopole will vary somewhat with the ground quality and the
number of radials forming the ground system. The AM BC standard
of 120 quarter-wavelength radials with shorter intervening radials
remains applicable to all monopoles, whether linear or folded. With
respect to the source impedance of an assembly, such a system
best replicates perfect ground. (Of course, the overall ground
quality plays a significant role in determining the far field patterns
for the antenna.
Chapter 31
The AGT values provide an indication of the reason for halting the
systematic modeling venture with a 10:1 ratio between the return
wire and each of the fed wires. As we increase the diameter of the
return wire, the AGT score decreases. By a 10:1 ratio, the score
reaches 0.915. Beyond this value, I would not fully trust the
reliability of the reported data, even when applying the corrective to
the per-leg source resistance value. The impedance transformation
ratio for a 2-wire system is largely a function of the physical
properties of the folded monopole. Hence, the use of a large
diameter tower (or its equivalent 3- or 4-face open tower) with
standard wire sizes for the fed cage structure becomes
problematical from a modeling perspective. 0.1" diameter wire
approximates AWG #10, a value that falls about halfway between
AWG #12 and #14, popularly used by radio amateurs, and AWG
#6, sometimes used by commercial installations.
The net adjusted source resistance columns indicate one reason

why cage-fed or skirt-fed monopoles are finding proponents. As we
add fed wires (symmetrically, of course) to a center return wire, the
net source impedance of the parallel-fed outer wires brings down
the overall source impedance. With equal-diameter wires for the fed
and return legs, we have the familiar 4:1 up conversion of the
impedance for each leg. As we increase the return wire diameter,
the source impedance in each leg increases in step with standard
expectations from the 2-wire equation, although we cannot obtain a
precise number for more than a 2-wire system. Nevertheless, when
we parallel the leg sources, the net impedance decreases as we
increase the number of fed outer wires. For any given combination
of fed and return wire sizes, increasing the spacing decreases the
Chapter 31
net impedance. However, for any number of fed sources,

increasing the return wire diameter increases the impedance. A 5-
wire (4-source) cage indicates impedances that are close to optimal
for direct coax feeding with either no matching or minimal matching
requirements. Construction variables relative to both the cage and
the return mast or tower, along with ground and radial system
conditions will modify the values shown. Nevertheless, the trends
are useful in system planning, even if NEC-4 models may fail to be
reliable at the return structure diameter that might be used. For
example, a triangular tower with a 12" face might show an
equivalent diameter of 8.88", which--even with 36" spacing--would
yield models of very dubious reliability.
For any given spacing and return-wire diameter, the per-leg source
resistance shows a relatively tight grouping of values as we move
from a 2-wire to a 5-wire folded monopole. About 50 Ohms
separates the lowest from the highest value in each group. The
variations within each group tend to suggest the level of mutual
coupling among the fed wires and other interactions. Data patterns
alone do not provide the details of the complex interactions.
However, the close proximity of the values within each group
relative to the calculated impedance value for a 2-wire folded
monopole strongly indicates that each fed wire and the return wire
form a 2-wire folded monopole, modified in impedance performance
by the interactions. If each pair of wires in a complex multi-wire
monopole forms a folded monopole, then we may legitimately call
the entire structure, regardless of the number of cage or skirt wires,
a folded monopole.
Chapter 31
We may regroup the final columns of values that show the ratio of
calculated to modeled values of per-leg source resistance and
graph them by reference to the spacing between the fed wire and
the return wire. Fig. 4 shows the results for a spacing of 12". Each
line represents one of the folded-monopole systems. The X-axis
shows the increase in the return wire diameter. Note that the lowest
value (0.1") is not a true linear increment relative to the other
increments on the graph.
Chapter 31
The 2-wire graphed line establishes that using corrected per-leg

impedance values yields models that are always well within 1% of
the calculated value of impedance. The per-leg values for 3-wire
systems are always below those for 4-wire and 5-wire monopoles.
For a spacing of 12", the lowest per-leg value is about 0.83 of the
calculated 2-wire value. In contrast, the highest value is about 1.4
times the calculated value. Although these limits are fairly wide,
each curve shows a much narrower range of departure from the
calculated value, which the 2-wire curve indicates.
Chapter 31
As we increase the spacing to 24", as shown in Fig. 5, the ratio of

calculated to modeled adjusted per-leg impedance increases,
except for the 2-wire model. It continues to show its tight correlation
to the calculated impedance value. Perhaps only the effects of
requiring a top wire prevent the ratio from reaching 1 to 1. The ratio
values for the larger folded monopoles show slight increases for all
of the more complex assemblies. Interestingly, about half of the
data points fall above the 1:1 ratio lines, and half below it.
Chapter 31
The graph for a spacing of 36" between the return wire and each
fed wire appears in Fig. 6. The overall grouping has all of the
properties that we saw in the preceding graphs. However, the
majority of data points now fall above the 1:1 ratio line. The
differences among the 3 graphs are not great, but they are
noticeable as a function of the increased spacing.
Perhaps more interesting than the differences are the similarities

among the graphs. For each size of folded-monopole assembly,
regardless of spacing, the curve shapes are regular and similar.
(Even the 2-wire curve shows a very slight drop in value as we
increase the diameter of the return wire.) As well, the values within
each graph show a relatively tight clustering around the calculated
impedance value. 5-wire systems show the highest level of
departure from the calculated value, a fact that appears to coincide
with the closer proximity of the fed wires to each other. Determining
the effects of mutual coupling would require a different sort of study
from the present effort. In each case, I have resonated the total
folded monopole system in order to arrive at as pure a resistive
impedance as feasible. Interactive analysis of the wires might
require a study using a set of fixed-length wires.
Nevertheless, the data shows that caged and skirted monopoles

have their roots in 2-wire folded monopoles and are an extension of
the basic structure. The variations are insufficient to deny the use of
the term folded monopole as a generic label for all such structures.
In many ways, however, settling the terminological preferences of

engineers has only been a pretext for the true point of these notes.
Chapter 31
Too many amateurs are unfamiliar with the general properties of

folded monopoles, both simple and complex. These notes
represent one small attempt to fill the void and to acquaint the
amateur with the range of labels that he or she might encounter in
basic reading about these antennas. A secondary goal has been to
show the nature and the limits of modeling these antennas using
NEC-4. AGT values for NEC-2 versions of the models would be
even worse. We might extend the range of the sampling of return-
wire diameters by using a well-corrected version of MININEC 3.13,
but even that program will eventually show less reliable results
before we reach the tower dimensions that we might encounter in
reality.
Nevertheless, the trends shown in these notes may be useful in

establishing realistic expectations from multi-wire folded monopole
assemblies. The data shown here cannot eliminate the need for
extensive field adjustment, but they may go some distance toward
reducing the time involved.
Chapter 31
Chapter 32: Short Folded Monopoles - Some Basics
I
n Chapter 31 defining Folded Monopoles, we examined some
basic properties of resonant folded monopoles using 2, 3, 4, and
5 wire construction. When resonant, modeled folded monopoles
show a clear relationship between the reported source impedance
and the calculated impedance using the classical equation.
R is the ratio between the impedance of the new folded antenna

and the impedance of a resonant linear antenna that otherwise has
the same design. The equation applies equally to folded dipoles
and to folded monopoles. Where the diameters of both the fed wire
(d1) and the return wire (d2) are the same, the ratio is 4:1. The new
impedance follows a very simple equation:
Since the feedpoint (or source) impedance (Zlinear) of a linear

monopole over perfect ground is 36 Ohms, the resonant feedpoint
impedance of a folded monopole (Zfolded) is 144 Ohms. If the fed
wire is fatter than the return wire, then the impedance ratio is less
than 4 but always greater than 1. If the return wire is fatter than the
fed wire, then the impedance ratio is always greater than 4.
Chapter 32
The basic principles in our initial foray into folded monopoles

neglected a very important aspect of folded monopole use. Many
antenna builders use lengths that are shorter than the resonant
length. For very short to moderate folded monopole lengths, (where
the resonant length might be considered long), the feedpoint
impedance will show an inductive reactance. In that property, a
short folded monopole bears a resemblance to a transmission-line
shorted stub. However, the stub and the antenna have some
important differences, crudely marked in Fig. 1.
The two structures are similar in that the properties are dependent
on the diameter of the wires and on the spacing between wires.
The letters s, d, d1, and d2 designate these fundamentals that
Chapter 32
make performance dependent to a significant degree on physical

properties of the structure. The folded monopole requires a ground
(or a suitable ground plane) in order to operate. The energy source
is normally positioned in series with the lower end of one wire (d1)
and the ground. In contrast, the shorted stub uses no ground.
Rather, the energy source is placed across the two wires of the
transmission line opposite the shorted end. As a consequence, the
shorted transmission-line stub does not radiate (if properly
constructed and isolated from influences that would create
imbalance between the lines). It ideally shows only transmission-
line currents, which are equal in magnitude and opposite in phase
at points along the line that are equidistant from the energy source.
The folded monopole radiates and therefore exhibits both
transmission-line currents and radiation currents.
At this point, we have to choose a direction for analyzing the

behavior of short folded monopoles. We might legitimately turn to a
mathematical treatment of the structures. However, my goal is not
to replicate texts on the subject, but rather to familiarize you with
the patterns of short folded monopole behavior. Therefore, my
chosen route of analysis is modeling some selected folded
monopoles to develop some patterns in the behavior. The method
will be effective in showing some of the variables that influence the
behavior, while also developing some rational expectations of them.
We shall choose our modeling software to fit the structure that we
are modeling. For our first two case studies, NEC-4 is adequate to
the task, although we shall pay close attention to the Average Gain
Test (AGT) score (where 1.000 is ideal) in order to adjust numbers
as needed.
Chapter 32
Since we shall not be dealing in resonant antenna lengths, we need

an increment of folded monopole length to use for our samples.
One convention used in the AM BC industry is to list antenna
lengths in electrical degrees, where 1 wavelength equals 360
degrees. We may adopt this convention for physical lengths even
though we know in advance that 90 degrees physical is longer than
the resonant physical length that we would call 90 degrees
electrical length. We shall survey folded monopoles every 10
degrees at a standard test frequency of 3.5 MHz for the entire
exercise. To reduce the number of variables, we shall use lossless
or perfect wire along with a perfect ground. As well, in this initial
investigation, we shall work only with 2-wire folded monopoles.
Case 1: d1 = d2 = 0.1", s = 12"
We may begin with a folded monopole structure already explored in

the earlier item. We shall form a set of folded monopoles where
both wires have a diameter of 0.1" and the spacing is 12" between
them (center-to-center). The use of 0.1" diameter wire is not
accidental. It roughly corresponds to AWG #10 wire, which falls
between two practical extremes. Amateurs often used AWG #14 or
#12 wire for such structures due to its availability and relatively low
cost. Commercial installations may use wire that approximates
AWG #6 (0.16").
At a resonant length of 67.25', the modeled impedance is 143.5

Ohms, compared to the calculated value of 144 Ohms. When we
model the antennas in the collection in 10-degree increments, we
Chapter 32
end up with a set of performance values such as those shown in

Table 1.
Besides showing the length of each model in degrees and feet, the
table lists the source resistance and reactance of each model. It
also lists the AGT in terms of the score and the gain adjustment in
dB (where the adjustment value is subtracted from the reported
value). Because all but the 10-degree scores are very close to
1.000, the table makes no adjustments in this case.
There is much to note in the tabular data. We might begin with the
gain values, which we show in terms of the gain broadside to the
plane of the 2 wires and in terms of the maximum gain in line or
edgewise to the 2 wires. Maximum gain occurs in the direction of
the feedpoint. There is always at least a slight difference in the 2
values, but as we make the folded monopole shorter, the
differential becomes very noticeable. Fig. 2 compares the elevation
patterns for 10-degree and 30-degree versions of the antenna. In
each case, the plots overlay the patterns broadside and edgewise
Chapter 32
to the wires. The shorter antenna shows a large difference that

almost completely disappears by the time the antenna is 30
degrees long.
One way to make sense of the remaining data in the table is to

contrast it to corresponding data for an equivalent linear monopole.
Therefore, I created a model of a monopole that showed a resonant
length of 67.25', the resonant length of the folded monopole. The
model required a wire diameter (d) of 2.75" to achieve this goal.
Fig. 3 shows a sketch of the 2 antennas.
Chapter 32
I then sampled the antenna in 10-degree intervals to produce a

table comparable to the one for the folded monopole. The results of
this exercise appear in Table 2. The columns in this table exactly
parallel those of Table 1.
Chapter 32
Both tables list the gain of the antenna over perfect ground using 3
different wire compositions: perfect or lossless, copper, and
aluminum. Copper has a bulk resistivity of about 1.7E-8
Ohms/meter (corresponding to a conductivity of about 5.8E7 S/m).
Aluminum's resistivity is about 4E-8 Ohms/m (conductivity about
2.5E7 S/m). Antenna modeling programs adjust the material losses
for frequency and skin affect in actual calculations. Hence, the gain
values for perfect or lossless wire would reappear at any frequency,
but the gain values will vary a bit as we change frequency if we use
copper, aluminum, or any other real wire material.
Chapter 32
Fig. 4 compares the gain values from the tables for all sampled
lengths. The values for the 2.75" linear monopole are for aluminum
only, since the differences between perfect wire and the worst case
used in the sample are so small. However, with wires that are only
0.1" in diameter, the material losses of copper and aluminum are
exceptionally significant as we reduce the overall length of a folded
monopole. Below a length of about 50 degrees, the thin-wire folded
monopole shows a rate of gain decrease that may question the
practicality of using such a thin, short structure as an antenna
Chapter 32
without specific needs that make the highly reduced gain

acceptable as a trade-off.
At resonant length, the folded monopole may rival the linear

monopole, but at very short lengths, the very low source radiation
resistance becomes only s small fraction of the total source
resistance. The wires are too thin to overcome the resistive losses
of the material and of skin effect. With real wire, the 10-degree
folded monopole shows less than 1% power efficiency. For the
shortest thin-wire folded monopoles in the sample, the use of
phosphor bronze or stainless steel would increase losses even
further.
The monopoles that we have examined are over perfect ground.

Placing them over real (lossy) ground will further reduce the gain
available (as well as raising the elevation angle of maximum
radiation). Table 4 provides a rough guide to the amount of further
gain reduction we are likely to experience over three levels of
ground quality. Very good soil has a conductivity of 0.0303 S/m and
a relative permittivity of 20. Average soil uses a conductivity of
0.005 S/m and a relative permittivity of 13. Very poor soil uses a
conductivity of 0.001 S/m with a relative permittivity of 5. These
widely diverse soil types may let you approximate the additional
losses of your local soil by rough interpolation. The radial systems
use 0.1" wire buried 1' below the ground surface in the NEC-4
models.
Chapter 32
The table is only a first-order estimation device, not a precise

gauge. For increased accuracy, you would need to model a
proposed short folded monopole using both the actual material ad
the actual ground conditions at the proposed site--along with a
model of whatever buried radial system the antenna might use.
Linear monopoles have gained some renown for their low feedpoint
resistance values when they are very short. However, if you
compare Table 1 with Table 2, you will discover that the feedpoint
resistance of the folded monopole does not catch up to the
feedpoint resistance of the linear monopole until we reach a total
length of about 40 degrees for both antennas. Fig. 5 compares the
feedpoint resistance values for both antennas across the span of
surveyed heights.
Chapter 32
The linear monopole resistance values show a smooth progression

from 10- through 90-degree lengths, reaching a final value of about
41 Ohms. However, the folded monopole shows a very large spike
in values between 50 and 60 degrees. The actual peak value is
much higher than the largest graphed value, since the peak occurs
at a height of about 57 degrees. Then the resistance value
decreases rapidly so that between 80 and 90 degrees, it shows an
expected slight rise with increasing length in this region. Even at 90
degrees, somewhat beyond resonance, the resistance is 164
Ohms, about 4 times the linear value for the same length.
Chapter 32
The peaking of the resistance value accompanies a peaking of the

reactance value in folded monopoles. Fig. 6 graphs the reactance
for both the folded and the linear monopole as we increase the
length from 10 to 90 degrees.
The linear monopole shows a smooth curve that traces the

decreasing capacitive reactance as the antenna lengths increases
toward resonance. Since the 90-degree length is long relative to
resonance, the curve smoothly proceeds into the region of inductive
reactance. In contrast, the folded monopole shows very high values
Chapter 32
of reactance at 50 and 60 degrees. One may interpolate a sudden

shift in reactance type at about the length at which the resistance
reaches its maximum value. Of course, at this point, we would find
a very small region of height at which the reactance would be
nearly zero. However, that specific height is unlikely to be achieved
in any practical installation. Even if achieved, a slight temperature
change would alter the antenna height enough to throw the
reactance into a high value region--either inductive or capacitive,
depending on the direction of the temperature shift.
The very high value of resistance and the sudden shift of reactance
from inductive to capacitive are typical behaviors of horizontal
antennas as they pass the 1-wavelength mark or of ideal vertical
monopoles as they pass through the 1/2-wavelength mark.
However, the folded monopole is less than 0.16-wavelength long.
Fig. 6 contains reactance values for one extra case: a shorted

transmission-line stub constructed according to the same wire
diameters and spacing values that we used for the folded
monopole. I arbitrarily cut off the table at +j5000 Ohms since the
reactance value of a shorted stub increases without limit at exactly
90 degrees. Table 3 shows the calculated values of the stub's
reactance at the line lengths that correspond to the folded
monopole's lengths in the survey.
Chapter 32
The table rests on calculating the characteristic impedance (Zo)

from the physical dimensions of parallel wires, as indicated in Fig.
1. One common equation for the calculation uses common logs.
However, there is a more precise equation that uses natural logs.
For the case in point, the impedance is high enough (657.22 Ohms)
that the two equations yield essentially the same results. The more
Chapter 32
precise equation becomes essential where the wire spacing is very

close. Our 12" spacing is not close for 0.1" diameter wires. To
calculate the inductive reactance of the stub, we may use another
common equation. The l term represents the line length in electrical
degrees (or radians).
The shorted transmission line that uses 0.1" diameter wires and a
12" spacing has a calculated characteristic impedance (Zo) of
657.22 Ohms. For any given length, the inductive reactance is a
direct function of the Zo value, and that value always apears as the
inductive reactance at a length of 45 degrees for lossless lines.
I note these equations only to supply the basis for the tabular and
graphical results shown. They suffice to show that the source
impedance behavior of the short folded monopole--at least for the
sample used here--is quite unlike the behavior of a linear monopole
and the behavior of a shorted transmission line stub. To test these
behaviors and to check for any variability, we need at least one
more sample.
Case 2: d1 = d2 = 0.5", s = 12"
As a check on our work, let's sample a second folded monopole

through the same total height steps. In this case, we shall increase
the wire diameter to 0.5", a 5-fold increase over the initial sample
folded monopole. The increased wire diameter applies to all parts
Chapter 32
of the antenna and so will have no effect on the impedance

transformation ratio. The resonant impedance of a modeled
antenna was 143.1 - j0.1 Ohms at a height above perfect ground of
66.81' (or 99.3% of the height of the resonant version of the first
model).
Although the use of fatter wires with a 1' spacing does not change
the folded monopole impedance transformation ratio, it does
change the characteristic impedance of the line if used in a shorted
stub configuration. The new Zo is 464.17 Ohms (compared to
657.22 Ohms for the version using 0.1" diameter wires). For
reference, Table 5 presents the calculated inductive reactance
values for the sampled lengths of the stub in 10 degree increments.
The 45-degree entry allows a quick reference to the line Zo.
Chapter 32
Except for the lowest 2 heights (10 and 20 degrees), the NEC-4
models of the antenna produce excellent AGT values. Hence, they
require no adjustment in the tabular data. Table 6 provides the
information gathered from the test runs using the same format and
column entries that we used for the thinner model.
To facilitate comparisons, I also created a linear monopole over

perfect ground. I selected an element diameter that would achieve
resonance at the same height (66.81') as the folded monopole in
question in order to develop a relatively fair comparator. The
required diameter was 5". The diameter is 1.8 times the diameter
(2.75") of the comparison linear monopole used for the folded
monopole with 0.1" elements. Both folded monopoles use the same
center-to-center wire spacing. The exercise establishes that finding
an equivalent linear monopole diameter to a given folded monopole
structure requires attention to the wire diameter as well as to the
wire spacing of the original folded structure.
Chapter 32
Table 7 shows the linear monopole data, again using the format
established earlier. Of course, the linear monopole requires only
one gain figure, since the pattern is uniform in all azimuth
directions. Our initial comparisons will be internal to the new
sample. We shall make cross-sample comparisons a bit later.
Chapter 32
The relative gain values appear in Fig. 7. Since the gain of the
linear monopole varies so inconsequentially over the range of real
materials, the aluminum gain curve suffices as a substitute for 3
overlapping lines. The pattern of gain deficiencies with real
materials for the folded monopole below a length of about 50
degrees reappears in this graph and in the tabular data. However,
the 5-fold increase in folded monopole wire diameters shows up as
a significant reduction in the deficiency level.
Chapter 32
In Fig. 8 we find the comparison of linear and folded monopole

source resistance values. The linear monopole resistance
increases in a regular (but not linear) fashion. In contrast, the folded
monopole shows a very sharp peaking of resistance at about 60
degrees. The position of the adjacent resistance values suggests
that the true peak may be at a length very close to 60 degrees.
Chapter 32
Fig. 9 tracks the progression of reactance values. One line shows

the inductive reactance of a shorted transmission-line stub. As I did
earlier, I cut off the Y-axis arbitrarily, since the 90-degree reactance
value increases without limit. (The tables show an exceptionally
high but not limitless value for 90 degrees due to computer
conventions for avoiding errors.) In contrast, the linear monopole
shows a continuous, regular decrease in the capacitive reactance
as the antenna grows toward its resonant height. Of course, at a
physical height of 90 degrees, the antenna is slightly long relative to
resonance, and so we find an inductive reactance.
Chapter 32
The reactance curve for the folded monopole shows peak values of
reactance at 50 and 60 degrees, with a transition region between
them. The actual transition region is very small, as the reactance on
either side climbs to values much higher than those recorded at the
sampling points. In the length region that is about +/-10 degrees
either side of resonance, the reactance curve shows a seemingly
normal curve that moves from capacitive to inductive as we pass
through the resonant folded monopole length.
A Tentative Comparison of Two Folded Monopoles
We have looked individually at two folded monopoles that use the

same 12" center-to-center wire spacing. The only difference
between them is the diameter of the wires: 0.1" vs. 0.5". Hence, the
impedance transformation ratio for both monopoles at a resonant
height is the same: 4. Since a linear monopole using either wire
diameter--or using the equivalent diameters in the comparators--
has a resonant impedance of 36 Ohms (+/- 0.03 Ohm), we expect a
resonant folded monopole impedance of 144 Ohms. The thin and
think folded monopoles show 143.5 and 143.1 Ohms, respectively.
We may consider these values to be very much on target, given the
fact that our basic impedance transformation equation does not
take into account the diameter or the length of the end wires. The
models cannot exist without taking these end-wire factors into
account.
When we look at short folded monopoles using the same basic

structure, we have to recognize that we cannot expect a precise
equivalence. The length increments use the physical length of the
Chapter 32
structure, not the electrical length relative to the resonant length.

The 0.1" antenna was resonant at a length of 67.25', while the 0.5"
version resonated at 66.81'. However, the lengths are less than 1%
apart, which minimizes any differentials in this area of concern.
For a sample of the gain differential, Fig. 10 compares the modeled

maximum gain edgewise to the wires for both folded monopoles
using copper wire. The thinner-wire folded monopole shows a 7-dB
gain deficit compared to the fatter-wire model at the shortest length.
As we increase the length of the folded monopole, the gain
Chapter 32
difference decreases in a smooth curve so that by the time we

reach a length of 50 or 60 degrees, the differential disappears
(depending on our standard of when a differential is too small to be
notable). The root source of the differential lies in skin effect. In
small loop antennas, builders commonly use the largest practical
conductor (sometimes round, sometimes flat) to reduce to the
lowest possible level any losses due to the resistivity of real
materials. The losses of the linear monopoles suggest that a wire
diameter of well over 2" may be needed by folded monopoles
shorter than about 50 degrees in order to reduce these losses
effectively. Unfortunately, such diameters are not practical for NEC-
4 models.
Chapter 32
Fig. 11 superimposes the source resistance and the reactance

curves for the two folded monopoles. Due to the paucity of
sampling points, the peak values appear to coincide. However,
such curves can be somewhat misleading if we seek more than
general guidance. Let's look at each folded monopole.
0.1" Model: A series of models using finer length graduations set

the height at which the reactance passes through zero at about
42.321' or 54.215 degrees. The resistance in the immediate length
region rose to 19540 Ohms. The length of the folded monopole was
Chapter 32
about 62.9% of a resonant version or about 0.157-wavelength

electrically.
0.5" Model: A similar series of models produced a height of 44.270'

or 56.711 degrees at which the reactance passed through zero. At
this height, the source resistance report was 10480 Ohms.
However, the source resistance peaked (10510 Ohms) at about
44.235' or 56.667 degrees. The zero-reactance model was about
66.26% of a resonant version of the antenna or about 0.166-
wavelength electrically.
We may note several interesting items about these numbers after

observing a significant caution. The numbers derive from models
that show an ideal AGT score, but which remain subject to all of the
limitations to which the antenna modeling software (NEC-4) is
subject. Hence, the numbers are useful for comparisons, but not
necessarily for trying to build a short, resonant, very-high
impedance folded monopole.
The thinner-wire folded monopole shows a much higher peak

source resistance than the fatter-wire model. In fact, the ratio of
peak source resistance values is 1.86:1. Although the tests did not
specifically seek out the peak values of reactance that occur on the
limits of the transition region, the thinner folded monopole appeared
to reach values at least 1.6 times higher than the thicker model in
the sequence of test model height. In both cases, the transition
region was about 2 degrees of height, 56-58 degrees for the 0.5"
model and 53 to 55 degrees for the 0.1" model. The transition
region includes heights at which the initial peak inductive reactance
Chapter 32
value begins to decrease and--at the opposite end--the height at

which the capacitive reactance increases toward but does not
reach its peak value.
The height at which the folded monopoles pass through zero

reactance seems initially counter-intuitive, since we might expect
the folded monopole using fatter element wires to pass that point at
a shorter height in keeping with the slightly shorter height of a 1/4-
wavelength resonant version. As well, the 0.5" model's slight
difference between the zero-reactance height and the maximum
source resistance height may also seem somewhat counter-
intuitive to our understanding of high-impedance resonant points. In
both cases, the most likely candidate to serve as the source of
these interesting results is the end wire. The potential corner
coupling of the 0.5" model might account for both phenomena.
However, the models are not self-explanatory in this regard. As
well, we cannot say from the model alone whether the phenomenon
is an artifact of modeling or a real phenomenon. Models that mix
wire diameters at angular junctions quickly become unreliable in
NEC (both -2 and -4).
The differences in behavior between the two folded monopoles

using equal legs, but of a different size for each model, are small.
More significant is the general behavior trends that show the
progression of antenna behavior with increasing length. First, the
very low source resistance of shorter lengths (up to and perhaps
beyond 30 degrees) results in a very lossy structure when using
real materials of even the finest quality. Second, the very high
impedance resonance in the 54- to 56-degree region is notable for
Chapter 32
both its potentials and its limitations relative to using a short folded
monopole. The short folded monopole acts neither like a shorted
transmission line nor like a short linear monopole.
Short Folded Monopoles with Dissimilar-Diameter Legs
Many folded monopoles make use of an existing structure for one

leg and add a wire for the second leg. At resonant 1/4-wavelength
sizes, we expect the monopoles to closely approximate calculated
values of source impedance. We can do some initial modeling to
see what happens as we sample shorter lengths, but we cannot do
so reliably within NEC. We must turn to a version of MININEC. For
the following examples, I used Antenna Model, a highly corrected
version of MININEC 3.13. Fortuitously, the program provides AGT
scores as a matter of course.
To limit the stresses upon the limitations of the core, we may

restrict our initial investigation of folded monopoles with unequal leg
diameters to only 2 leg sizes: 0.1" and 0.5", the sizes that we used
in the first two cases. As shown in Fig. 12, the top wires for the new
cases will use the thinner wire size. The black dots represent the
location of the model source. In NEC, we generally construe the
source location to be along or at the center of the lowest segment
in the relevant leg. In MININEC, the source is at the junction of the
leg with the ground.
Chapter 32
The ratio of leg diameters is 5:1. However, the equation that

calculates the impedance transformation does not use that ratio
directly, but incorporates the leg diameter within a ratio with twice
the space between legs and then takes the common log of both
space-diameter ratios. Hence, the resonant linear monopole source
resistance becomes (by calculation) about 210 Ohms when we
feed the thinner leg and about 105 Ohms when we feed the fatter
leg. The proximity of modeled impedances to the calculated ones
becomes a second test (in addition to the model's AGT) of the
reliability of the reported data on the short folded monopoles.
Chapter 32
Table 8 provides the data for the version of the folded monopole
with the source located on the thinner wire, Case 3. The table omits
information on gain values for real wire materials, since that
information would largely parallel the data for Cases 1 and 2. The
perfect-ground gain data generally parallels the corresponding
information for folded monopoles with equal-diameter legs.
However, the broadside-to-edgewise gain differential is slightly
greater. Although the impedance information for the shortest length
appears quite reasonable and the AGT is only slightly off ideal, the
gain data appears to need further study before we accept it at face
value. In the following notes, gain will not be our main focus.
Chapter 32
Table 9 presents the comparable data for Case 4, which feeds the
0.5" leg of the folded monopole. Although the AGT scores are
virtually ideal (at least through 3 decimal places), the gain data is
subject to further scrutiny. However, the broadside-to-edgewise
gain values more closely parallel those we obtained from the first
two cases.
Whatever the reservations we may apply to the gain data, the most
significant data resides in the resistance and reactance columns of
the tables. Like the equal-diameter cases, both of the new cases
shows exceptionally low source resistance values at the shortest
lengths. The tables provide an entry for the resonant length
information. In both new cases, the modeled resonant impedances
are within 1/2 of 1% of the calculated values.
Chapter 32
The curiosity of the folded monopoles with unequal leg diameters

appears clearly in Fig. 13. When we feed the thinner wire, the
source resistance peaks close to 60 degrees, or higher than either
of the length values that we found in the first two models with
equal-diameter legs. In contrast, when we feed the fatter leg, the
length that shows peak source resistance is closer to 50 degrees.
This length is shorter than we found for the first two cases.
Due to the very high peak resistance values, the lower source
resistance values form almost straight lines, forcing us back to the
Chapter 32
tables for the interesting properties. When we feed the thinner leg
of the folded monopole, we discover a very rapid increase in
impedance, beginning at a thousandth of an Ohm and climbing in a
span of about 0.16 wavelength to an exceptionally high value.
Predicting the source resistance of a physical version of the short
Case-3 folded monopole would be a daunting task at best. Even
temperature changes might yield significant resistance excursions
at the feedpoint unless we use a wideband or a lossy matching
network. Case-4 monopoles fair no better for lengths from 10 to 50
degrees. However, the region from 60 through 90 degrees shows a
much slower rate of resistance change from one step to the next.
Unfortunately, this region also shows consistent capacitive
reactance.
Chapter 32
Fig. 14 overlays the reactance curves for the two new cases. In
general outline, the curves follow the pattern established by te first
two cases that use equal-diameter legs. However, the transition
regions call for some special attention. Cases 3 and 4 both show
that the transition region occurs between lengths of 50 and 60
degrees. Within that 10-degree span, the two new structures differ
considerably. By tracking the level of the 50- and the 60-degree
peak values, we can obtain a fairly close approximation of the
difference. For example, if we feed the thinner wire (Case 3), then
the capacitive reactance at 60 degrees is close to -j3000 Ohms, but
Chapter 32
the inductive peak at 50 degrees is only a little over j1000 Ohms.

The zero-crossing point must therefore occur much closer to the
60-degree mark.
In contrast, in we feed the fatter wire (Case 4), the inductive

reactance peaks at about j8000 Ohms at the 50-degree level. By 60
degrees, the capacitive reactance is about -j1000 Ohms. By the
same reasoning, we must conclude that the zero-crossing point for
this configuration is not much above 50-degrees physical length.
More generally, we see a widening of the possible range for

reactance transitions and resistance peaks with these cases than
when we simply change the diameter of equal-diameter versions of
the folded monopole. A thin-fed wire to fat-return wire situation
tends to push the transition point to a higher length level. In
contrast, a fat-fed wire to a thin-return wire pushes the zero-
crossing point to a lower total fold monopole height.
Some Tentative Conclusions
Modeling demonstrations do not yield proofs of performance.

However, from the general trends that we have observed with our
four case studies, we may draw some conclusions that we may
think of as reasonable expectations of folded monopole
performance.
1. At very short lengths (up to 30 to 40 degrees), 2-wire folded

monopoles of any description show very low source resistance
values. The values are lower than we obtain with linear monopoles
Chapter 32
of the same height. In fact, it is in this region that we find the closest
coincidence between folded monopole and shorted transmission-
line stub behavior with respect to the inductive reactance. When we
translate the models to real types of wire, material losses alone are
sufficient to create very high gain deficits. Below a total height of
about 30 degrees, the losses may be high enough to jeopardize the
utility of the structures for communications.
2. Between heights of 50 and 60 degrees, we find a transition

region in which the resistance rises to a very high peak value. In
the same region, the reactance peaks both inductively and
capacitively, with a very small height region in which it crosses the
zero point. The behavior closely resembles the behavior of center-
fed linear horizontal antennas as they approach 1 wavelength or of
linear monopoles as they approach 1/2 wavelength. In folded
dipoles, the behavior occurs with lengths between 0.14 and 0.17
wavelength.
3. Although the region of peak values between 50 and 60 degrees

dominates graphs of the folded monopole's feedpoint performance,
the entire span from 40 to 70 degrees shows rapid changes in both
resistance and reactance with only small changes in folded
monopole height. These behaviors can make the matching of a
folded monopole in this height region a very finicky task. In all
cases, one must experimentally determine if the settings used will
be stable through the entire set of environmental conditions that the
antenna may face.
Chapter 32
4. The most stable region of folded monopole performance occurs

at physical lengths between about 70 and somewhere above 90
degrees. In virtually all cases, resonance will occur at physical
heights between 80 and 85 degrees. Within the overall region, the
resistance changes per unit of height change are relatively small.
Reactance changes are likewise small and follow the normal
progression from capacitive reactance below resonant length to
inductive reactance above resonant length. Therefore, matching
networks and wide-band impedance transformers will tend to show
the same performance characteristics that they display when used
with linear antennas.
5. For every resonant folded monopole, there is a linear monopole

of some diameter that will resonate at the same length. The
required diameter for the linear dipole is a mutual function of both
the wire diameters and the wire spacing of the folded monopole.
Since the equivalent linear monopole will be very large relative to
the diameters of the wires within the folded monopole, it will exhibit
far lower losses with real materials than the folded monopole using
the same real materials, especially at very short lengths. However,
the folded monopole may offer a considerable total weight
reduction relative to the linear monopole, especially at longer, more
stable overall heights.
In the tables, the numbers recorded are overly precise relative to

the general reliability of the models with respect to reality. My
reason for recording the reported modeling data in these terms was
to ensure accurate graphing. At best, these notes serve as a
general guide to reasonable expectations from folded monopoles.
Chapter 32
By omitting real ground types and real materials from the

calculations, the notes do not qualify as guides to building a
physical antenna. Nonetheless, the consistency of the general
trends may provide some insight into short folded monopole
behavior.
In the cases that we have explored in this set of notes, both legs of
the short folded monopoles used the same length, and the top wire
shorted the upper end of the structure. However, short folded
monopoles often find application where the return wire is longer
than the fed wire. We have learned to name such applications, but
it is less clear that we have developed any reasonable expectations
of behavior. Therefore, we have another trail to explore through the
forest of folded monopoles.
Apprendix: A supplementary Exercise in Current Analysis
The exercises that we have explored in developing some basic

properties of folded monopole antennas focused upon antenna gain
and the feedpoint impedance as guides. There is an alternative
approach that may provide additional insights into the behavior of
short and long folded monopoles. We may analyze the currents in
the legs of the folded monopole into two component currents, often
called radiation currents and transmission-line currents. At any
point along the length of the folded monopole, the sum of the two
current magnitudes and phase angles (one on each leg) result in
the radiation current, while the difference yields the transmission-
line current, assuming that we set up the model wires for the legs in
parallel fashion, that is, counting from the ground up (or the top
Chapter 32
down) for both legs. See the Antenna Modeling series of articles,
#123, for details of how to set up the calculations. Although NEC
output files list the currents in terms of both real and imaginary
components and of magnitude and phase angle, EZNEC current
tables list only the magnitude and phase angle. Thus, the first step
is to convert the given values to real and imaginary components,
then to perform the additions and subtractions, and finally to
reconvert the values back into magnitudes and phase angles. A
repetitive spreadsheet is, of course, the most convenient method
for automating the required machinations.
If we perform the operation on a resonant folded monopole at 3.5

MHz, we obtain some interesting results. For the subject antenna
using 0.1"-diameter elements and a separation of 1', the resonant
length is 67.25 or 86.15 degrees. For comparison, let's also set up
a resonant linear monopole of the same length. The linear
monopole must have a diameter to 2.75" to be resonant at the
prescribed physical length. For simplicity, we shall set both
antennas against a perfect ground and use perfect (or zero-loss)
conductors. NEC, of course, will provide a direct reading of currents
on the linear monopole because it has no currents that we can
analyze as transmission-line currents.
The results of the exercise appear in Table 10. On the left, the first
four current magnitude and phase columns provide magnitude and
phase-angle values from the EZNEC current table every 5
segments along the antenna's length from the ground upward. The
next two columns provide the analyzed radiation current (Irad)
magnitude and phase angle for each increment of antenna length.
Chapter 32
The final two columns list the analyzed transmission-line current

(Itl) magnitude and phase angle. To the right are the current values
for the linear monopole model.
Except for the 60th segment, the radiation current magnitude

values for both antennas are virtually identical. The slight aberration
in the last value is a function of the folded monopole top connecting
wire. Minimum current occurs at its center. As well, the pattern,
although not the precise values, of radiation current phase is the
same for both antennas. With respect to transmission-line currents
on the folded monopole, the magnitude values increase from the
ground toward the top. However, the phase angle of these currents
is the same all along the antenna and 90 degrees out of phase with
the feedpoint current. We would see similar results from a folded
dipole counting from the feedpoint outward to the antenna end.
Next, let's examine a short folded monopole, perhaps one that is 20

degrees (15.612') long. We may not only perform a similar analysis
(using each of the 12 segments in the model), but as well we may
Chapter 32
again compare it with the corresponding 2.75"-diameter linear

monopole of the same length. The results appear in Table 11.
The linear monopole on the right displays a normal progesssion of

currents from the ground upward, despite the very short length of
the antenna. However, the folded monopole shows something
entirely different, due to the fact that the overall length falls well
below the critical region (between 40- and 60-degree lengths) in
which the antenna transitions from transmission-line-like behavior
to antenna-like behavior. (See Fig. 6, which shows the reactance of
a comparable transmission line, a folded monopole and a linear
monopole to see more vividly the critical transition length region.) At
the very short length of 20 degrees, the folded monopole shows far
higher current magnitude in the transmission-line column than in
the radiation column. Because the two wires, treated as a
transmission line, are almost perfectly out of phase with each other,
the net phase angle is zero and constant along the short length.
The minuscule radiation current magnitude is accompanied by a
minimal phase-angle shift that only becomes apparent at the top of
the folded monopole.
Chapter 32
Applying the current analysis to the modeled current values for the
short folded monopole yields an impression that the short folded
monopole is likely to be a relatively poor radiator. In some
applications, it might actually be superior to the short linear
monopole once we add to the single element the requisite loading
coil at a plausible level of Q. Nevertheless, below the critical
transition length region, the short folded monopole principally acts
like a shorted transmission line rather than like an antenna.
Chapter 32
Chapter 33: Short Folded Monopoles - Extended Applications
I
n Chapter 31, About the Folded Monopole, we briefly explored
the use of a folded monopole that was self-resonant but which
also used a linear extension. The result, as expected, involved
an increase in both the resistance and the inductive reactance at
the source or feedpoint. The folded monopole continued its normal
antenna function in terms of radiation, with a small increase in gain
due to the increased overall length.
In Chapter 32 of this 2-part examination of short folded monopoles,

we explored the properties of complete antennas, that is, folded
monopoles that used equal-length legs. We saw a progression of
properties from physical lengths of 10 through 90 degrees. The
progression may have held some surprises for those not used to
folded monopole behavior. At very short lengths (up to 30 degrees
or so), the source resistance was very low, lower than for an
equivalent linear monopole. In the 50- to 60-degree region, the
source resistance became very high, with an accompanying set of
peaks for inductive and capacitive reactance--and a narrow zero-
crossing length between them. Only as the folded monopole
exceeded about 70 degrees did the source impedance curves
return to behaviors that we associated with linear monopoles
adjusted for the impedance transformation that is inherent to the
folded structure.
In this item, we shall combine the two ideas into a single

exploration. We shall look at short folded monopoles from 10
Chapter 33
through 90 degrees with linear extensions. One goal will be to see

what patterns of antenna performance emerge from the exercises.
One might well wonder whether the exercise might be simply the
satisfaction of idle curiosity. We do not normally hear of extended
short folded monopoles among the many classes of antennas used
by amateur or commercial interests. Just as a rose by any other
name would smell as sweet, so an antenna by any other name
would radiate just as well. There are a number of structures
composed of extended short folded monopoles that we have
managed to re-name--and sometimes, to misunderstand as a result
of the different name. In the vagaries of labels lie many roads to
misconception.
Chapter 33
Consider the left-hand side of Fig. 1. The sketch shows a grounded

tower fed for one of the lower amateur bands by what we call a
shunt feed system. We connect a wire parallel to the tower. The top
end connects to the tower so that the wire length combined with its
distance from the tower allow--so we usually say--coupling to the
tower. We strive for a length that will yield a feedpoint impedance
we can easily match to coaxial cable using the simplest network for
lowest losses. The wire length usually emerges from experience,
and AWG #12 is popular for the purpose. One instruction set that I
read suggests that if the feedpoint impedance is not within a
desirable range, we should change the spacing from the tower until
the impedance is acceptable.
On the right in Fig. 1, we have a short folded monopole that places

its top wire at a height which is not the full length of the return wire.
The return wire happens to be a tower structure that is very wide.
Hence, we obtain a very sizable step-up in impedance relative to a
linear wire of the same length. Since the tower goes on above the
folded monopole, we anticipate that the feedpoint impedance will
show an increase, if we assume that the folded monopole portion is
self resonant. If the folded monopole section is shorter than
resonant, then our work so far leaves the impedance unknown. So
the sketch has added a network at the feedpoint in case we need it
to effect a match with the main feedline.
The perspective that we took in the right portion of the sketch

makes clear that the fed wire is as much a part of the radiating
structure as the tower. Its function is not simply to couple energy to
the tower. Rather, its function is to form with the tower a single
Chapter 33
radiating element. The 2-wire portion of the structure effects an

impedance transformation relative to the tower alone--if base fed--
but the radiating currents are a joint function of both wires. Of
course, between the left and right sides of Fig. 1, we find no
difference in the structure outline. The only differences appear in
the labels for the structure's parts.
The left part of Fig. 2 shows a generalized sketch of a gamma

matched element. The most common occurrence of the match
these days is for the driven elements of Yagi parasitic arrays in
which we wish to connect all elements to the boom. Ordinarily, the
Chapter 33
impedance of the element without the gamma match is lower than

the desired value, usually a value well below the 50-Ohm
impedance of standard coaxial cable. We call the portion of the
gamma match that parallels part of the element length the gamma
rod. We short the end to the element in a position that yields the
desired impedance at the new feedpoint. By a judicious selection of
rod diameter, rod spacing from the element, and rod length to the
connection, we can sometimes produce a purely resistive 50-Ohm
impedance. If there is a reactance, we strive to make it inductive so
that a simple series capacitor will compensate. Note that we
normally consider the gamma rod to be a part of a matching
network and not part of the radiating element.
However, if we redraw the upper half of the sketch and terminate it

at ground (instead of at a grounded boom), we obtain the sketch on
the right of Fig. 2. In this case, the fed wire or rod becomes one
wire of a 2-wire folded monopole, with an extension beyond the
folded structure limit. The ground forms--if we wish to think in these
terms--an image of the upward or physical portion of the antenna.
The gamma match turns out to be a short folded monopole with an
extension. As such, both wires in the folded structure make up the
radiating element until we reach the extension. From our work with
full folded monopoles and their linear equivalents, we know that the
equivalent diameter of the folded portion of the antenna is much
larger than the diameter of the extension along. Hence, every
gamma-matched element contains at least a small imbalance. The
related Tee match overcomes the imbalance by using a folded
monopole on each side of the boom.
Chapter 33
As a consequence of these instances of short folded monopole

applications--whatever the preferred labels--the behavior of short
folded monopoles with extensions becomes more than an idle
exercise. It holds some possibility of improving our understanding
of certain structures that combine impedance transformation and
radiation. As well, we might develop some techniques that would be
applicable to real antenna planning exercises.
Our basic work will follow the pattern on the first episode. We shall
place the folded monopoles over perfect ground and use lossless
wire in order to eliminate some complex variables. In a real
planning situation, we would put into our models the material
conductivity of the proposed wire and the best estimate of real
ground conditions. As well, our models would include the actual
radial system beneath the folded monopole.
Since we need some usable increment

between modeled structures, we shall again
use the physical height of the structures in
degrees, where 360 degrees is 1 wavelength.
We shall explore structures in 10-degree
increments at 3.5 MHz. Table 1 provides a
convenient correlation between the degree
markers and the equivalent height in feet at
the test frequency. At 3.5 MHz, a wavelength
is about 281.02'.
Chapter 33
We shall create models using an implementation of MININEC 3.13.

The program that I am using is Antenna Model, which incorporates
a considerable number of correctives to overcome some MININEC
shortcomings. Since all of our models will use different diameters
for the fed and the return wires, NEC (-2 or -4) will not yield results
that pass Average Gain Test (AGT) muster. As we shall eventually
see, even MININEC's more ready handling of junctions between
wires having dissimilar diameters has limits. However, we shall be
able to create some usable models within those limits. If you try to
replicate the models using a different implementation of MININEC,
expect to find some variance in the output reports. If the
implementation does not include an accessible AGT score, you
may simply have to guess at the reliability of the model that you
create.
The project itself is simple, although long and occasionally tedious.

Fig. 3 shows its general outlines. I shall create a series of short
folded monopoles for each test case. Since the most common
cases of short folded monopole applications involve return wires
that are fatter than the fed wire, the models will all follow this
pattern. The folded monopoles will appear in 10-degrees steps,
from 10 through 90 degrees. To each folded monopole that is
shorter than 90 degrees, I shall attach to the return wire--using the
same wire diameter as the return wire--a series of extensions in 10-
degree steps. Each series will progress until the total element
length is 90 degrees.
Chapter 33
Obviously, one might carry the progressions of folded monopoles

beyond 90 degrees or the extensions beyond 90 degrees.
However, my goal is not to replicate every possible structure we
might use. Rather, it is only to elicit the patterns of behavior of the
resulting antennas, with special attention to the feedpoint
impedance. We would have to end somewhere, and there is little
point in becoming totally lost in a morass of excessive data. Indeed,
the data that we shall observe is complex enough for one episode.
Case 1: D1 = 0.1", D2 = 0.5", Space = 12", 2 Wires
Our study cases can begin with a structure that we examined in the
first episode. It consists of two wires spaced 12" apart. The fed wire
will be 0.1", corresponding to AWG #10 wire that is midway
between the wires used in amateur and commercial practice. The
Chapter 33
return wire is 0.5" in diameter to give us a sense of standard 2-wire

shunt feeding of an existing mast. I have selected this starting point
because the AGT scores are generally very good to excellent within
MININEC despite the difference in element diameter. At a resonant
length of 67.08', the reported source resistance is about 209 Ohms,
compared to a calculated value of 210 Ohms.
In the present context, we shall look at short monopoles in 10-

degree increments. To each of these folded monopoles, we shall
add a 0.5" diameter extension in 10-degree increments. The
minimum length for each portion of the following data will be the
height of the simple folded monopole. However, every folded
monopole will end up with a height of 90 degrees. When we later
speak of the data, we may use expressions such as "20-50" to
designate a line from the table. The first number indicates the
height of the folded monopole portion of the antenna, while the
second number reports the total height that includes the extension,
if any.
The data includes a report of the AGT, the source resistance and
the source reactance. In addition, there are entries for the reported
gain broadside to the pair of wire and edgewise to the wires. In the
latter case, the maximum gain value appears as a measure of the
pattern's circularity or ellipticalness. Table 2 records the data for
the present 2-wire case.
Chapter 33
Chapter 33
Except for the shortest lengths of both the folded monopole and the
extension, the AGT scores are excellent. Therefore, the table
makes no adjustments to the reported values. The gain values
Chapter 33
generally accord with those for folded monopoles that form the
complete structure of the entire length. In the last episode, we
noted the losses that accompany relatively short structures when
we translate our perfect wire into real materials. In practical terms,
there is very little difference in the numerical performance and
virtually no operational real difference in gain performance among
any of the various structures when the total height exceeds perhaps
60 degrees or so.
The impedance progressions that follow on each folded starting

point exhibit interesting patterns. We may use the 10-10 through
10-90 series as a sample. The reactance increases very slowly but
steadily. However, as it crosses the resonant region (between 80
and 90 degrees total height), we do not find resonance. Instead, we
find a reversal of the direction of change of reactance. The
resistance for all total length up to 80 degrees is too low to be
useful. However, the rate of increase climbs so that with a 90-
degree total height, we achieve a matchable pair of resistance and
reactance values. If we track the 20-n and the 30-n series of
models, we find the same pattern for both the resistance and the
reactance, with an adjustment for a new starting value set that
emerges from the new length of the folded section.
Above a folded length of 40 degrees the pattern appears to change.

We seem to find more rapidly changing resistance and reactance
values. However, we are entering the region in which a full folded
monopole would experience rapidly changing impedance values.
The extension portion of each structure has the effect of increasing
the folded monopole length by a small amount with each step.
Chapter 33
Small changes of total length yield large changes in resistance and

reactance. Between 40-80 and 40-90, we see a reversal in the
inductance, suggesting a narrow resonant region. For the 50-
degree folded structure, total height between 70 and 80 degrees
records a similar reactance zero crossing. When the folded
structure is over 60 degrees high, the antenna begins past the
cross-over point and shows predominantly capacitive reactance.
However, the 80-90 case yields another matchable impedance
combination.
One of the oddities of matching practices that uses short folded

monopole structures is the variability of practice. Gamma matches
for Yagi elements tend to use the shortest practical folded section
that will effect the desired impedance transformation. In contrast,
tower shunt feeding tends to use the longer folded structure that will
get the job done. See Chapter 6 of The ARRL Antenna Book, 20th
Edition, and Chapter 9 of ON4UN's Low-Band DXing, 2nd Edition,
for samples of tower shunt feeding. Gamma and related matching
systems appear in Chapter 23 of the ARRL book. In fact, neither
Yagi practice nor tower practice seems to take note of the other
way of achieving the same goal.
Although our case study is not itself very realistic relative to either
HF Yagis or to MF/HF towers, we may use it as a way to explore a
technique for surveying a more complete range of options when
using short folded structures to effect impedance matching. We can
create graphs of the impedance reports. Fig. 4 handles the
reported resistance values for our sample. The X-axis records the
Chapter 33
total height of the structure, while the individual lines represent

different heights for the folded portion of the antenna.
I have cut off the Y-axis for multiple reasons. I arbitrarily set a 500-
Ohm limit to the upper end of the resistance range. The decision in
any real case would rest on an estimate of the highest resistance
value that might be acceptable. The range should be great enough
to show the rate of resistance change from one step to the next.
However, it should not be so high as to obscure how close to an
ideal value of resistance the modeled value comes. In this case, we
Chapter 33
might be concerned with 50 Ohms as a target value. Fig. 5 shows a

similar treatment for the reported reactance values.
The reactance values included in the graph range from -j100 Ohms
to +j500 Ohms. Most short monopole impedance transformation
systems seek an inductive reactance (or zero reactance) at the
feedpoint to allow matching with only high-Q capacitors. The
selected range lets us see both the recorded reactances at the
sampling points and the relevant rates of change to the adjacent
sampling points.
Chapter 33
The combination of the two graphs allows us to select candidates

for implementation. In most cases, we shall find few viable
candidates, since we are likely to be working with an existing
grounded mast or tower. However, in the exercise, we are free to
note any viable combinations. In the present case, the resistance
table offers a number of combinations that show (by the graph's
definition) usable values with modest rates of change in resistance
to the next sampling point. However, the reactance graph reduces
the number of candidates. Within the constraints of the exercise,
total height values between 80 and 90 degrees combined with
folded heights between 10 and 20 degrees offer useable
combinations with relatively low rates of change. The longer folded
structures that showed promise in terms of their resistance reports
in this region tend to disqualify themselves due to the high rate of
reactance change between sampling point. Any network that we
might use to produce a final resistive impedance of 50 Ohms would
likely have at best a very narrow bandwidth.
Our accumulation and exploration of data shows us how we can

use the information as the basis for planning installations. However,
the structure that we used is relatively unrealistic. It appears
because the models are highly reliable as measured by the AGT
values (which are a necessary but not sufficient condition of model
adequacy). Perhaps a more realistic scenario might be useful as a
second exercise.
Chapter 33
Real towers that we might use as a shunt-fed vertical antenna vary

in face size. We may use a modest tower with a 12" face
dimension. For simplicity in the models, we may use the standard
AM BC equivalence and multiple the face by 0.74 to obtain an 8.8"
diameter wire that approximates the tower. (An actual planning
session should model the tower structure as exactly as possible.)
This new diameter forms the return wire for the folded structure and
the extension above and beyond the folded structure. We may
retain the 0.1" diameter fed wire as a realistic value.
The next step is to determine

a workable space between
the wires of the folded
section. Despite MININEC's
superiority in handling
junctions of wires having
different diameters, it will
show limits to its reliability.
We not only have a radical
difference in wire diameters,
but as well, we have two
wires that are fairly closely
spaced. I sampled variety of
spacing values and return-
wire diameters using a 30-
degree folded structure and an 80-degree total height to see what
AGT scores might emerge. The results appear in Table 3.
Chapter 33
For any set spacing, the AGT values degrade as we increase the
diameter of the return and extension wires. In some circles, AGT's
are considered excellent if they fall between 0.995 and 1.005. They
are usable between 0.990 and 1.010. Beyond this latter range, the
data becomes questionable. Even within the usable range, we
should adjust the data by virtue of the AGT score if we are
developing (meaningful and comparative) progressions of values,
especially if the AGT value changes from one sample to the next.
For the projected diameter of the return and extension wires in the
models that we shall run, the minimum spacing for adequate AGT
scores is 36" or 3'. This value tends to coincide with commercial
practice for folded assemblies, so I shall use it in amassing a data
collection.
AGT scores will tend to improve for longer structures and degrade
for shorter total antenna heights. Therefore, we should adjust the
reported values to arrive at the best approximation of a final value.
Because we are dealing with a folded structure, we must reverse
the normal procedure set. Ordinarily, we convert the AGT score into
a "dB' value (=10 log AGT) and subtract it from the reported gain.
(An AGT less than 1 results in a negative dB value, which
increases the reported gain when we do the subtraction.)
Experience with the full folded monopoles in previous studies
suggests that we must add the AGT-dB to the reported value to
obtain gain values that are reasonable in terms of their consistency
with values that emerge from models showing an ideal or very-
nearly ideal AGT value. We normally adjust impedance values by
multiplying the reported number by the AGT itself. However, folded
structures appear to require that we divide the report by the AGT in
Chapter 33
order to obtain values that coincide with calculated impedance

transformations.
With these cautions and conditions, we may proceed to model our

ersatz tower and its shunt wire as a series of 2-wire short folded
monopoles and extensions. We shall use the same increments that
we used for the initial case. Table 4 shows the amassed adjusted
data. However, we must add one more reservation. The standard
impedance transformation equation for resonant folded dipoles
results in an impedance of about 614 Ohms. The resonant version
of our new model is 65.286' high and reports an adjust resistance of
574 Ohms. The difference is about 7%. We cannot be truly
definitive in assigning a source to the difference. However, the end
wire is now 3' long, almost 5% of the total folded monopole height.
As well, the impedance ratio is about 17:1, a very large ratio
indeed.
Chapter 33
Chapter 33
Chapter 33
Within the limits of our case study, the gain values are completely
normal (with the usual reservations about the very shortest folded
and total height values). Broadside gain values at a total height of
height of 90 degrees coincide very closely with corresponding
values for the smaller and narrower model of Case 1. The
edgewise gain values provide a measure of the degree to which the
2-wire structure ovalizes the pattern.
The significant differences between the values in Table 4 vs. those

in Table 2 appear in the resistance and reactance columns.
Between the two tables, we note a large difference in both the
return/extension wire diameter and the spacing between wires.
Therefore, we shall expect to see differences in the impedance
values for each increment of folded structure.
Between folded structures heights of 10 and 30 degrees, perhaps

the biggest surprise may be that we see the same general pattern
of values that we experienced with Case 1. The resistance values
for Case 2 are consistently about 1/3 the level of those for Case 1.
However, the reactance values for each increase in the return wire
length are comparable between the two cases. The parallel extends
even to the decrease in inductive reactance that occurs in the total
height interval between 80 and 90 degrees. The amount of
decrease for the 8.8" example is somewhat less than for the 0.5"
return wire, but it is equally distinct.
As we increase the height of the folded structure, the parallels

remain, but with reservations. For the narrower structure, the
reactance zero-crossing first appeared with a folded structure that
Chapter 33
was 40 degrees high. With our more widely spaced model, the
folded structure reaches 60 degrees before we encounter the first
zero crossing. If the total height for Case 2 is 90 degrees, we find
nearly identical inductive reactance values for folded heights
between 50 and 90 degrees. The shorter the folded structure within
this range, the lower the resistance value.
The general trends give us a picture of both consistency with the

earlier models and of adjustment for the new values of
return/extension wire diameter and spacing. In order to translate
those general trends into a more adequate planning venue, we may
graph both the resistance and the reactance values. Since we do
not have a real tower around which to plan, we may use that same
limits applied to the graphs for the earlier case. Fig. 6 shows the
relevant resistance values.
Chapter 33
With respect to resistance, the 20- and 30-degree folded structures

show perhaps the most promise when used with total heights
between 80 and 90 degrees. Especially notable is the relatively
slow rate of change in resistance with changes in height over this
region. These conditions suggest that variables in the physical
structure relative to the model will be manageable in terms of field
adjustments toward the final antenna.
Chapter 33
Fig. 7 shows the corresponding reactance graph. It tends to

confirm the initial judgment of the promise offered by the 20- and
30-degree folded structures with higher extensions. With the
shorter folded structures, the inductive reactance increases with the
length of the folded monopole for any given total height. However, it
remains manageable. Long folded structures might yield workable
values of inductive reactance at a total height of 90 degrees, but
the rate of change between 80 and 90 degrees total height is very
steep.
Chapter 33
We may note in passing that the resistance curves for all lengths of
folded structure show a downward trend between 80 and 90
degrees. Although we are working with a simple straight tower
(equivalence) in these models, it is fairly easy for typical amateur
towers to exceed 90 degrees in electrical length without passing the
70' mark of our 90-degree tower. Most shunt-fed amateur towers
hold one or more Yagi antennas at the top. These structures tend
to form rudimentary and imperfect top hats that increase the
effective electrical length of the tower considerably. Hence, a 90-
degree physical tower may easily become 100 or 110 degrees
electrically. The consequences lie outside the limits set on this
exercise, but we can expect to find lower resistances and variations
in the reactance. If the tower is too tall electrically, we may easily
find a shift back into capacitive reactance. Within modest electrical
height increases relative to the study limits, we need to study all
potential folded structure heights (or shunt lengths). The longest
folded structures in conjunction with the longest total heights tend
to show very high rates of change in resistance and reactance with
small changes in total height. The result is normally a very narrow
operating bandwidth for any given set of matching components.
One possibility, rarely explored by amateur installations but

regularly used commercially, is to provide multiple fed wires. Fig. 8
shows the configurations that we have and shall explore here. Both
of our earlier sample cases used 2-wire construction. However, 3-,
4-, and 5-wire construction is feasible, assuming that it might
Chapter 33
increase the number of options available for shunt feeding a tower

or other applications of folded monopole structures.
In these notes, we shall take up just two of the added options: 3-

wire and 5-wire configurations. The 3-wire models will use a pair of
0.1" diameter fed wires on opposite sides of the central 8.8" tower-
equivalent return wire and extension. To ensure reasonable AGT
values, we shall retain the 36" spacing between the fed wires and
the return wire. All of the procedures for adjusting values according
to the AGT scores will also apply to this case. The only modeling
Chapter 33
difference lies in the need for having 2 sources which are, in effect,
in parallel relative to the overall structure. If we bring these sources
together, the composite source impedance will be composed of
resistive and reactive components that are each half the value of
the values for the individual fed wires.
The data for the 3-wire assemblies appear in Table 5. Although the
simple folded monopoles in the earlier study ("What is a Folded
Monopole") all produced circular azimuth patterns at resonant
lengths, we find elongated edgewise patterns in these models,
likely due to the very large diameter difference between the fed
wires and the return wire. However, as we increase the total height
of the antenna to 80 degrees for any height of folded structure, the
azimuth pattern becomes circular. The more symmetrical structure
of the 3-wire models also shows a slight improvement in AGT
values relative to the 2 wire models.
Chapter 33
Chapter 33
As we have done for the other cases, we shall break the discussion
of impedances values into 2 parts for Case 3. The first set of values
involves folded structures that are 10 through 30 degrees high. The
general trend that we discovered for the earlier cases applies to the
3-wire models. For any folded structure, as we increase the height
Chapter 33
of the extension, the source resistance slowly rises until we pass

the 80-degree total height mark. Then the resistance declines.
Interestingly, over this range of folded structures, we do not find
very significant differences in the source resistance between 2-wire
and 3-wire models. The reactance also slowly increases as we
increase the total length of models with 10- through 30-degree
folded structures. However, the inductive reactance is considerably
lower for the 3-wire models than it was for the 2-wire counterparts.
When the folded structure exceeds about 40 degrees, the

resistance and the reactance change more rapidly in 2-wire models
than with shorter folded structures. In corresponding 3-wire models,
the resistance tends to change more slowly than in 2-wire models
with the same height characteristics. The first reactance zero-
crossing occurs in 50-80, with the 60-n series of models showing
capacitive reactance at all total heights except 90 degrees. When
the folded structure exceeds about 70 degrees, the 2-wire and 3-
wire source resistance values become very comparable. However,
the reactance values of the 3-wire models appear to be almost
uniformly shifted in the direction of inductive reactance.
The pattern of similarities and differences between 2-wire and 3-

wire structures can naturalize us to the performance behaviors, but
we require more detailed analysis to see if we have developed any
manageable matching potential. Fig. 9 provides the resistance data
for source values up to 500 Ohms. If we can manage source
resistance values up to about 250 Ohms, then the graph suggests
that we may use some taller folded structures (70 and 80 degrees)
in addition to the shorter 20- and 30-degrees structure suggested
Chapter 33
as possibilities for 2-wire structures--so long as we use a relatively

high total structure (>70 degrees).
Whether any of these potentials has a usable or manageable

reactance requires that we examine the graph in Fig. 10. The
reactance values at a height of 90 degrees for the tower are well
within the range of most networks. However, the rate of change for
the reactances associated with the taller folded structures is
somewhat steep as the value shift from capacitive to inductive
between total heights of 80 and 90 degrees. The more rapid the
Chapter 33
change in reactance with smaller height changes, the narrower the

bandwidth will be for any particular set of matching component
values in basic networks. Nevertheless, the slow rate of resistance
changes may effect an improvement in operating bandwidth for the
taller folded structures.
Since 3-wire systems are not very much more difficult to install than
2-wire systems, they may prove useful in a number of applications.
Chapter 33
Developing a data set for 5-wire structures with 4 fed 0.1" wires and
an 8.8" diameter return/extension wire requires far more effort to
model than to explain. It does not matter what name we give to the
structures: caged monopoles, skirted monopoles, or multi-wire
folded monopoles. Central to the modeling is providing a
symmetrical arrangement of the fed wires and providing each fed
wire with a source. The net or parallel source impedance of the
antenna will be 1/4 the value that appears on any single leg.
MININEC lacks any facility for paralleling sources, so the
calculation must be external to the program.
Table 6 provides the adjusted data gathered for the series of 5-wire
models. Increasing the symmetry of the structure provides another
slight improvement of AGT values. As well, the 4 wires circularize
the azimuth patterns, as indicated by the identical gain values in
both the broadside and the edgewise columns. Once we pass very
low levels of total height, the gain for the system does not vary
significantly regardless of the number of wires.
Chapter 33
Chapter 33
Adding two wires to the 3-wire assembly does not greatly change
the net source resistance values relative to 3-wire models when
both use shorter (10- through 30-degree) folded structures. Indeed,
for shorter folded sections, all three series of model (cases 2
through 4) show similar resistance values. The more significant
Chapter 33
change for shorter folded structures occurs in the level of inductive

reactance. The 5-wire models show between 2/3 and 3/4 the level
of inductive reactance for each step of total height for any of the
shorter folded structures.
As we increase the folded structure to a height of 40 degrees or

more, we find a significant difference in antenna impedance
behavior relative to 3-wire models. Model 40-80 shows the first
reactance zero crossing, an event that required a 50-80
combination with 3-wire models. With folded structures between 50
and 60 degrees, all of the reactance values are capacitive,
including the value for a 90-degree total height. Only with folded
structures at least 70 degrees high do we find an inductive
reactance with a 90-degree total antenna height. However, almost
all of the reactance values are quite low, suggesting the potential
for broader bandwidth matching systems.
Chapter 33
In the category of source resistance, Fig. 11 reveals that the 5-wire

assembly offers us new candidates for folded structures in terms of
manageable values. In addition to the 20- and 30-degree folded
structures, we may add 60-90 degrees, with total lengths equal to
or longer than the folded structure.
Chapter 33
With a total height of 90 degrees, Fig. 12 informs us that almost

any of the resistance candidates will suffice in terms of feedpoint
reactance. As we lower the total height to 80 degrees, even the
tallest folded structures show less steep curves than for any of the
preceding models series. Even a 60-degree folded structure will
work well with an 80-degree total height if we can handle a
moderate amount of capacitive reactance.
The 5-wire short folded monopole with extensions expands the

number of options available to the application of folded structures to
Chapter 33
grounded vertical towers. These notes have not explore antenna

lengths greater than 90 degrees, so we cannot say off hand
whether the advantages of multiple feed wires continue with taller
towers.
Matching and Planning
In the course of these notes, I have noted that most amateurs who
feed existing towers for use on one or more of the lower bands
prefer to arrive at an assembly that provides inductive reactance. In
many cases, the operator will sacrifice operating bandwidth for the
inductive reactance. Indeed, they prefer to arrive at a source
resistance that is less than 50 Ohms along with the inductive
reactance. We may fairly ask why this custom prevails.
Chapter 33
Fig. 13 provides part of the answer. Under ideal conditions, we can

effect the required matching to a coaxial cable with a simple series
capacitor, if the antenna feedpoint shows a near-50-Ohm resistive
component along with inductive reactance. The upper left sketch
shows the condition. If the impedance is less than 50 Ohm resistive
and has inductive reactance, then we may use the 2-capacitor
matching scheme at the upper right. Since the scheme is a simple
L-network, it is not clear why the label "omega" match persists,
since nothing in the network resembles the Greek letter in either
upper or lower case forms. For many combinations of impedances
with under 50 Ohms resistance and an inductive reactance, the
system at the lower right will also work. However, it does not
always yield the most desirable values (usually meaning the lowest
values) of capacitance in either the series or the parallel leg.
For general matching with resistances above 50 Ohms, we

normally require the L-network configuration shown at the lower left
in Fig. 13. The sketch shows perhaps the most familiar form of the
L-network, often used with horizontal long-wire antennas. However,
it functions very well with impedance that are near to 50 Ohms with
various levels of inductive or capacitive reactance. Tower shunt
feeders have in the past rejected this configuration--and the
impedances that require its use--because they have preferred to
match entirely with capacitors. One may match a a given
impedance over a narrow operating bandwidth with fixed
weatherproofed capacitors at the base of the antenna. Vacuum
variables can provide fairly weatherproof remote service to vary the
match.
Chapter 33
The desire to remotely match an antenna for upper MF and lower

HF amateur service makes sense if we consider some of the
impedance values that appear in the literature. One set of values
involved a resistance of 17 Ohms and a reactance of 580 Ohms.
Even if we had a resonant condition, the SWR would begin at 3:1
relative to 50 Ohms and become progressive worse with increasing
reactance. With the very high reactive component, the SWR is in
the hundreds.
An alternative offered by multi-wire folded structures is a reduction

in the reactance to levels that may not require remote matching.
Some of the 5-wire models showed resistance values in the 90s as
the total height approached 90 degrees. The inherent SWR relative
to a 50-Ohm line is under 2:1 with those conditions. If we raise the
reactance to perhaps 100 Ohms (capacitive or inductive, the SWR
climbs to about 4.3. If we use 100' of low loss cable (such as
LMR600) and reserve matching for inside the operating room, we
lose about 1/3 dB. A simple L-network at or near the operating
position will effect the required match with little or no concern about
the durability of components as the weather passes through its
many potentially destructive cycles.
The possibly attractive alternative, of course, rests on the numbers

yielded by the sample models, with all of the qualifications and
reservations recorded along the way. Making such a system work
in a specific application requires far better modeling, along with
considerable preliminary field effort. At a minimum, a proper model
should include all of the details shown in Fig. 14. The tower needs
full modeling, as do the specific wires to be used as fed wires in the
Chapter 33
folded portion of the structure. Not only should the model include
any extension of the tower, but as well, it should include any mast
and beam antenna at the top. In addition, the model should specify
the materials for each wire within it, with different values for
material conductivity wherever they occur. Only then will the model
accurately reflect the above-ground structure and its equivalent
electrical length. Of course, the many junctions of wires having
different diameters will force the use of MININEC, since a NEC
model has a very high probability of being unreliable.
Chapter 33
The below-ground structure has equal importance. A buried radial

system deserves below-ground modeling in NEC-4. If we try to
substitute the MININEC ground with a set of specifications for
conductivity and relative permittivity, we would obtain the
impedance for a perfect ground. To resolve the incompleteness of
one system without the other, the modeler can first model the
antenna structure as a whole in MININEC. Then he or she can
develop a simplified substitute model have essentially the same
characteristics, but using throughout a single wire diameter. (In
many cases, this phase of the effort may require the most
ingenuity.) Finally, one may transfer the model to NEC-4 for
completion with the buried radial system to be used in practice.
Of course, Murphy's Law tells us that the individual who bypasses

this process and uses initial estimates will spend long days in the
field adjusting and readjusting the system because reality and the
initial estimates are far apart. Equally, someone who goes through
the entire process will conclude that the initial estimates proved to
capture reality as well as the most detailed model. Such is the lot of
shunt feeders.
The brief account of detailed planning requirements serves to

reinforce my initial notes to the effect that these case studies only
establish the interesting patterns of short folded monopole and
extension behavior. They do not provide precision guides to
building such structures. However, patterns can be useful in
familiarizing us with the territory so that what we encounter has
fewer surprises. Surprise on the battlefield is an effective offensive
weapon. Surprise with antennas is usually simply offensive.
Chapter 33
Chapter 34: A Doublet as a Good Single Choice Antenna
Suppose I wanted to work 40 through 10 meters. And further

suppose that I want to know where my signal is going. Now
what would I choose?
The 40-Meter Dipole Starting Point
The most common answer to the problem I just posed is a 40-meter

#12/#14 copper wire dipole, fed with parallel transmission line for
use as a multi-band doublet. Fig. 1 tells the simple construction
tale.
Placed high enough in the air, the 67' doublet performs very well.
Its bi-directional pattern on 40 has good side QRM rejection, and
with enough altitude, the elevation angle makes DX a real potential
with every band opening. Let's place the antenna at a height of 66'
Chapter 34
(about 20 m) up and see what the models tell us that we can expect
by way of performance.
Freq. Max. Gain TO VBW HBW Feedpoint Z
MHz dBi Deg Deg Deg R +/- jX Ohms
7.15 7.3 28 35 86 70 - j 10
10.1 8.1 20 23 70 275 + j 800
14.15 9.0 15 16 51 4670 - j 345
18.1 10.5 11 12 33 175 - j 860
21.2 8.4 10 10 33 100 - j 115
24.95 9.3 8 9 33 375 + j 730
28.5 9.5 7 8 28 3265 + j 375
The TO angle is the elevation angle of maximum radiation. I have

also provided information on the vertical and horizontal beamwidths
(measured to the -3 dB points from the maximum strength bearing).
We often neglect this information, but the data tell us some
important facts. The vertical beam width is a rough measure of the
range of elevation angles that we can count on for good
communications. The horizontal beamwidth tells us how broad or
narrow our signal is and hence how careful we must be in aiming
the antenna--either when we build it or when we rotate it. (One of
the amusing facets of reading lots of e-mail is discovering how
many beam users demand 1-degree aiming accuracy when their
beamwidths are well over 50 degrees.)
The 67' doublet shows the anticipated lowering of the TO angle as

we increase frequency. As we increase the frequency, the antenna
is increasing in electrical height, that is, its height as a fraction of a
wavelength. So we expect the beam angle to be lower on the upper
bands. The beam widths--both vertical and horizontal--narrow with
rising frequency. Still, the vertical beamwidth is wide enough on all
Chapter 34
bands to catch the main stream of long-range skip. And the

horizontal beamwidth is sufficiently broad to make aiming non-
critical (but not unimportant).
The range of impedances at the antenna suggests that with a

parallel transmission line and an ATU, we should be able to effect a
match on all bands. 20 and 10 meters might present slight
problems, but changing the line length will likely solve them by
presenting the ATU with impedance values it can handle.
However, the gain column presents us with a small problem. The

maximum gain of the 67' doublet occurs on 17-meters, takes a
large dip above that band, and then slowly rises once more. For a
simple wire, the actual gain numbers are not the problem. The
question we want to ask is this: why does the dip occur?
Chapter 34
Fig. 2 can help us figure out what has occurred. The azimuth
patterns overlaid on the plot are for 40, 15, and 10 meters. The 40-
meter pattern is the typical oval. As we increase frequency, the oval
grows narrower (decreasing horizontal beamwidth) while the gain
increases--up to 17 meters. On this band, the 67' doublet is about
1.25 wl long: an extended double Zepp (EDZ). We expect about 3
dB gain over a dipole from an EDZ, and if we compare the 40-
meter and 17-meter gain figures, we can see that we get it.
Chapter 34
Above 17 meters, the antenna is longer than 1.25 wl. At 15 meters,

the antenna is about 1.5 wl long. The main lobe is no longer
broadside to the antenna, but, as shown in Fig. 2, it is broken into 6
distinct lobes. The lobes broadside to the wire are no longer the
strongest. As we move into the 10-meter region, where the antenna
is 2 wl long, we have a pattern composed of 4 lobes at roughly 40-
degree angles to the wire.
For those unfamiliar with pattern development as an antenna

becomes multiple wavelengths long, the following rules of thumb
apply. For an antenna that is N wavelengths long, where N is an
integer (like, 1, 2, 3), the number of lobes is twice the value of N.
So a 2 wl antenna has 4 lobes, and a 1 wl antenna has only 2. For
antenna lengths that are N.5 (like 1.5, 2.5, etc.), the number of
lobes will be the sum of the number of lobes we get at N and at
N+1. At 15 meters, where the antenna is 1.5 wl long, 1 wl gives us
2 lobes and 2 wl gives us 4 lobes, for a total of 6. The higher
number at N.5 wl values arises because the new lobes are growing
and the old ones shrinking--and they are nearly equal strength at
the N.5 wl points.
The 44' Wire Solution
The problem with wire lengths over 1.25 wl is that we are no longer
sure that we have a good signal broadside to our antenna.
Suppose I put up a wire in Tennessee, running it NW to SE. That
makes it broadside to Europe in one direction and to VK/ZL-land in
the other. Not a bad set up. However, the main lobes for
frequencies from 21 MHz up are no longer going where I want them
Chapter 34
to go. (Where they go may result in interesting contacts, but we set

up our problem at the beginning so that it includes the need to keep
our signals where we want them.)
There is a simple solution to this problem, but it may not be the one
some folks would expect. Conventional wisdom tells us always to
make antennas bigger and longer. However, the solution to our
problem is to make our doublet shorter. Let's try a doublet that is
44' long, as in Fig. 3. Again, #14 or #12 AWG copper wire will do
just fine for the antenna.
The first question we may now ask is whether we lose anything with
the shorter wire. Let's find part of the answer in another table,
modeled with the copper wire (#14 AWG) antenna 66' above
average ground.
Chapter 34

7.15 7.0 29 35 94 25 - j 580
10.1 7.6 20 23 83 55 - j 100
14.15 7.7 15 16 72 195 + j 485
18.1 8.6 12 12 60 920 + j1565
21.2 9.0 10 10 51 4160 + j 155
24.95 10.4 8 9 40 520 - j1545
28.5 10.4 7 8 31 140 - j 650
The gain figures begin about a quarter dB below the figures for the
67' doublet on 40 meters and climb steadily. The elevation angles
of maximum radiation and the vertical beamwidths are virtually
identical to those for the longer doublet. The shorter antenna
provides a broader horizontal beamwidth on every band, which
makes aiming less critical. The pattern of impedances offered at the
feedpoint differs in detail from that of the longer doublet, but the
values are manageable. However, see the cautionary note below.
To see one of the major advantages of our short doublet, we should

look at Fig. 4.
Chapter 34
The composite azimuth patterns for the doublet on all of the bands
for which it is intended have their major lobes exactly broadside to
the wire. The 44' length was no accident. On 10 meters, this length
is about 1.25 wl long, the standard EDZ length. The 10-meter
pattern shows the anticipated strong main lobes plus the emerging
"ears," secondary lobes that will become the major lobes at higher
frequencies. On 15 meters, the antenna is about 1 wl long, and on
30 meters it is just under the right length for a half wl dipole (which
Chapter 34
is indicated by the low resistive impedance and the capacitive

reactance for 10.1 MHz). On 40 meters, the antenna is between 1/3
and 3/8 wl long, about the minimum length we should use. Anything
shorter would show very low resistance values and very high
reactance values--a difficult situation for any ATU to handle.
If I wanted to aim at both Europe and at Australia and New Zealand

on all bands from 40 through 10 meters, then the 44' doublet is the
superior antenna to the longer 67' doublet. Of course, larger (102'
or 135') all-band doublets break into fragmented lobe patterns at
lower frequencies than the 40-meter dipole with which we started.
So, if we make aiming one of the criteria for our antenna, the 44'
doublet may be the way to go.
Two side notes. First, if we want to cover the bands only up through
20 meters, but want to include 80 and 75 meters within the
frequencies for which we are well-aimed, then an 88' doublet might
meet our needs. Of course, there is nothing magic in the precise
length numbers chosen, since a length change of a foot or two will
change almost nothing in terms of performance. The bands with the
highest reactances at the feedpoint might show the greatest
change in value as we alter the antenna length, but the feedpoint
values on the other bands would hardly change enough for an ATU
to notice.
For either the 88' or the 44' doublet, the lowest band of use (80/75
and 40, respectively) present challenges in line losses and
matching at the shack end of the line. Hence, the short doublet--
Chapter 34
only 1/3 to 3/8 wavelength on the lowest band--should be

considered as a back-up antenna on those bands.
Second, height is a major consideration with this sort of antenna.

Perhaps 66' is not feasible for everyone. However, every foot of
(safe) height that you can add to the antenna, the better it will work.
This principle goes back to the days of George Grammer of ARRL,
who preferred to add height rather than elements to his antennas.
The idea is no less true today, although there are some limits. Once
we get above 1 wl, there may be some holes in our DX elevation-
angle coverage at certain antenna heights. However, in Grammer's
day, only on 10 meters and VHF did most hams think about heights
above 1 wl or so.
The Aluminum Alternative
There is no good reason why a single element antenna must be

constructed from thin wire--excepting cost and ease of construction.
If there is only one support that is high enough, then we might well
consider constructing an aluminum tubing version of the 44'
doublet. Fig. 5 shows one of many possible schemes for
constructing the element. The wind survival rating for this scheme
is about 70 mph.
Chapter 34
In tabular form, the element structure looks like the following partial
antenna model description.
--------------- WIRES ---------------
Wire Conn.--- End 1 (x,y,z : ft) Conn.--- End 2 (x,y,z : ft) Dia(in) Segs
1 -22.000, 0.000, 66.000 W2E1 -19.000, 0.000, 66.000 5.00E-01 3

2 W1E2 -19.000, 0.000, 66.000 W3E1 -15.250, 0.000, 66.000 6.25E-01 4
3 W2E2 -15.250, 0.000, 66.000 W4E1 -11.500, 0.000, 66.000 7.50E-01 4
4 W3E2 -11.500, 0.000, 66.000 W5E1 -7.750, 0.000, 66.000 8.75E-01 4
5 W4E2 -7.750, 0.000, 66.000 W6E1 -4.000, 0.000, 66.000 1.00E+00 4
6 W5E2 -4.000, 0.000, 66.000 W7E1 4.000, 0.000, 66.000 1.25E+00 9
7 W6E2 4.000, 0.000, 66.000 W8E1 7.750, 0.000, 66.000 1.00E+00 4
8 W7E2 7.750, 0.000, 66.000 W9E1 11.500, 0.000, 66.000 8.75E-01 4
9 W8E2 11.500, 0.000, 66.000 W10E1 15.250, 0.000, 66.000 7.50E-01 4
10 W9E2 15.250, 0.000, 66.000 W11E1 19.000, 0.000, 66.000 6.25E-01 4
11 W10E2 19.000, 0.000, 66.000 22.000, 0.000, 66.000 5.00E-01 3
Chapter 34
The construction shown in the figure should be designated light to

medium for the element length involved. It is based upon
purchasing 8' lengths of aluminum and cutting them in half. Prices
for 6061-T6 aluminum tubing from sources such as Texas Towers
and others are quite reasonable. The tapering scheme allows for
about 3" of tubing overlap at every junction. Any less overlap would
jeopardize physical strength, while too much more overlap will
unnecessarily increase the element weight. The double section of
1.25" and 1.125" diameter tubing at the center provides
reinforcement for most mounting schemes.
Some builders prefer to alternate longer and shorter sections of

tubing for greater overall strength. Whatever scheme one uses,
analyzing it through a program such as YagiStress by Kurt Andress
can go a long way toward ensuring a mechanically sound antenna.
Nonetheless, the 44' doublet is only about 9' longer than most 20-
meter Yagi reflectors and a good bit shorter than the elements used
for 30-meter and 40-meter Yagis. Hence, the use of an aluminum
tubing element is certainly feasible.
Special Note: Carroll Allen, AA2NN, pointed out the taper schedule
suggested would have a wind survival rating of only about 70 mph.
He developed a spread sheet for EXCEL to calculate the stress on
the tubing. For commonly used antenna tubing, such as 6061-T6,
with a wall thickness of 0.058", the maximum stress for each
section should be 40,000 psi or less. He kindly redesigned the
sections for a 100 mph wind survival rating. The following table
presents the revised taper schedule. Like the original schedule, the
Chapter 34
1.125" diameter section is presumed to run all the way through the
1.25" section, but also to have its own exposure length.
44' Aluminum Doublet Half-Element Structure
for 100 MPH Wind Survival
Diameter (") Section L (") Cumulative L (")
1.25 72 72
1.125 19 91
1.0 20.5 111.5
0.875 21.5 133
0.75 23 156
0.625 24 180
0.5 84 264
The use of aluminum tubing provides a small but determinate

increase in electrical performance for the doublet. As we increase
the diameter of an element, the RF resistance decreases
considerably. Although copper wire is certainly efficient enough for
most purposes, the tubing version of the antenna shows an
increase in efficiency, despite the fact that the tubing version user
aluminum, which has a higher resistivity than copper.
Frequency #14 Copper Wire Stepped-Dia. Aluminum
MHz Efficiency (%) Efficiency (%)
7.15 96.98 99.70
28.50 98.35 99.77
Notice that the differential in efficiency grows less as the frequency

increases, that is, as the wire diameter becomes a greater fraction
of a wavelength. Nonetheless, the tubing version shows
systematically higher gain values for each band than the wire
version of the 44' doublet. Compare the following table with the
copper wire table given earlier.
Chapter 34

7.15 7.2 29 35 94 20 - j 410
10.1 7.7 20 23 83 50 - j 85
14.15 7.8 15 16 72 195 + j 295
18.1 8.7 12 12 60 1005 + j 845
21.2 9.1 10 10 51 1700 - j 705
24.95 10.5 8 9 40 285 - j 795
28.5 10.5 7 8 31 100 - j 375
The gain differences are certainly not large enough to make any
kind of operational difference in using the 44' doublet. At most, they
help us better understand some of the variables involved in
antenna structures.
Notice also that the feedpoint impedance figures vary from the wire
values more radically as the frequency increases--and also where
the resistance or reactance values are high to begin with. The
values shown--which will vary considerably as one changes the
precise length of the finished antenna--are nonetheless quite
manageable by most ATUs.
Supporting an aluminum doublet of the size we are suggesting is a

considerable project. Fig. 6 shows the main aspects of the things
we should consider.
Chapter 34
1. We shall need a support tower or pole as tall as we can safely

make it. Although I know some inveterate climbers much older than
I am who regularly scale high towers, my own experience tends to
decrease my tower height by one section for every decade older I
get. Safety comes first; antenna height second.
2. As I doodled the sketch in Fig. 6, I added a rotator for the fun of

the prospect. We only need about 180-degrees rotation for this
doublet, so a side mounting on an existing tower would likely be
Chapter 34
good enough. For sample photos of one version of this antenna

constructed from tubing and able to rotate with a CD44 rotator, see
the web site of Adam, N4EKV. (Today's cheap TV rotators are no
longer able to handle this antenna, but 25 years ago. . .)
3. For elements 40' and longer, consider adding a top-mounted

truss to help support the element. A truss system will require a
longer mast than truss-less elements. The exact position on the
element to place the ends of the truss depends a great deal on the
precise element diameter schedule chosen. Use a weather and UV
resistant material for the truss rope.
4. Since the antenna will use parallel transmission line, stand-off

insulators will be necessary. If the antenna is to rotate, the position
of the line from the feedpoint to the stationary supports on the pole
will require considerable planning. The line should avoid close
proximity to any metal in the mounting region.
5. The parallel transmission line is low loss inherently when the

runs follow standard handbook recommendations. However, for
lowest losses, consider true open-wire line rather than vinyl-coated
lines. Even lines with "windowed" openings in the vinyl between the
wires tend to show higher losses when wet than when dry. Line
losses will generally be low except for the lowest band in the range
shown.
6. When the antenna feedpoint impedance is disproportionately

reactive, you can expect difficulties effecting a match with an
antenna tuner. On the lowest band, the 3/8-wavelength wire has an
Chapter 34
impedance where the reactance is 20 times the resistance. Under

these conditions, the ability of a tuner to effect a match between the
impedance at its terminals and the standard 50-Ohm coax to the rig
will depend both on the line and the tuner configuration. Line
concerns involve an interaction between the line length and the
characteristic impedance. As well, even high-efficiency parallel
lines will add losses to the already lower performance on the lowest
band in the usable antenna range. To avoid tuner damage, tune up
using low power and then raise power to the operating level.
Like all antennas, preventive maintenance at least once per year

(every 6 months is better) will go a long way toward preventing
unexpected catastrophic failures.
However, you decide to construct a single-element antenna, the 44'

doublet has some interesting properties that provide advantages
over other types of multi-band doublets. It is an antenna worth
considering--if you can have only one wire.
The use of the 44' length--or more properly, the approximate length
of a 1.25 wavelength wire at 10 meters--is not new. It dates back to
at least the middle of the 20th century. For example, Gene Fuller,
W2LU, in the July, 1966, issue of CQ wrote an article on using 42'
wires in a set of phased arrays ("Beam Antennas for the H.F.
Range," page 12).
The 44'-wire exercise is designed mainly to encourage antenna

builders to think "outside the box" of antennas that are naturally
resonant at some operating frequency or at the lowest operating
Chapter 34
frequency. Resonance is not a requirement of good performance.

Hence, we are free to set some one or more other goals for our
antenna. In this case, we have set the goal of having radiation
patterns that are broadside to the wire. For 40-10 meters, the 44'
wire fulfills this goal, however, close to or far from resonance its
length may be on any one of the bands covered.
Chapter 34
Chapter 35: Inverted-Vee Doublet Patterns
A
common all-band doublet for smaller ham properties is the
inverted-Vee, usually cut to either 135' or to 67' (or
thereabouts since the exact dimensions are not critical for
an antenna fed with parallel line and an antenna tuner). The
question that recurs is whether the Vee works as well as the level
or flat-top doublet of the same length.
The answer depends on 2 factors. One factor is the frequency of

use. The other factor is the angle of the Vee wires relative to the
level doublet. To see what difference each of these factors make,
let’s look at the doublets in Fig. 1.
The basic length of the doublet will be 67' for a fundamental

frequency of 40 meters. We shall look at both azimuth and
elevation patterns for 40, 20, 15, and 10 meters, on frequencies
about mid-band in each case. However, the patterns will not
change from one end of a band to the other.
Chapter 35
We shall look at patterns for the level doublet as a sort of baseline

against which to measure the patterns of the 2 Vee versions. One
version will be a moderate Vee that slopes downward only 30
degrees relative to the level doublet. The other Vee version will
slope downward 45 degrees. Together, the two Vees will give us a
picture of the trend in pattern change for intermediate or even more
radical slopes.
Chapter 35
In some cases, Vee patterns will be weaker, that is, have no lobes
as strong as the strongest lobes of the doublet. Does that make the
Vee a poorer antenna? Not necessarily. In each case, look also at
the shape of the pattern and the strength of the lower angle
elevation lobes. It will be the evaluation of all of the pattern features
that will tell you which version of the doublet is best for your
operation. This assumes, of course, that some kind of doublet is
best in the first place. On 40 meters, there is really very little to
choose between a flat doublet and a Vee. The elevation angle of
maximum radiation climbs upward as we make the Vee slope more
radically. It changes from 38 degrees for the flat-top to 46 degrees
for the 45-degree Vee. However, in all three cases, the elevation
lobe is so vertically broad that the differences are unlikely to make
a detectable difference in performance.
Chapter 35
Chapter 35
Likewise, there is a slight difference in the strength of the main lobe

broadside to the antenna, which is 50' up at the center in these
patterns. A difference of 1-2 dB is not likely to be detected by the
user without high-cost lab equipment. (That is why we rarely detect
reduced performance due to the lack of antenna maintenance until
the antenna falls down.) So on 40 meters, A Vee and a flat-top are
about equal.
Chapter 35
Chapter 35
On 20 meters, where the doublet is about 1 wavelength long, we

begin to see significant differences. The elevation angles of
maximum radiation climb from 19 degrees for the flat-top to 25
degrees for the 45-degree Vee. As well, the gain drops by over 3.5
dB as we increase the slope of the Vee, although most of that drop
occurs in the move from 30 degrees of slope to 45 degrees.
However, notice the shape of the azimuth patterns. The flat-top

shows extremely deep side nulls (off the ends of the wire. The user
can look at these nulls as QRM fighters or as directions in which
almost no QSOs are possible even under the best propagation
conditions. The lesser nulls of the two Vees offer some hope of
contacts, although condition might have to be very good to get
them.
Chapter 35
Chapter 35
On 15 meters, the situation becomes a good bit more complex,

perhaps even more complex than the maze of lobes and nulls in
the combined sets of patterns. The flat-top has the strongest lobes
by far, although the four best are fairly narrow in width. If these
lobes happen to go exactly where you want them to go, then all is
well. If not, then they may radiate where no one lives. Antenna
orientation is important.
The broader patterns of the Vee antennas offer slightly weaker but
more uniform propagation in most directions. However, nothing is
perfect. Note the elevation patterns, which show almost all radiation
to be at higher angles in the 23 to 30-degree range on a dx band
that does best when radiation is at much lower angles. The flat-top
take-off angle is 13 degrees, just about right if the lobe points at a
target.
Chapter 35
Chapter 35
On 10 meters, we have a similar situation to the one on 15 meters.

Everything is just a bit more extreme. The flat-top lobes are
narrower, but at a nice low 10-degree elevation angle for DX.
Unfortunately, the two Vee antennas have lost virtually all of their
low- angle radiation: their lowest elevation lobes correspond to the
secondary lobe from the flat-top. Hence, the DX potential of the
Vees on 10 meters is somewhat dismal.
Indeed, it might be better for 10 meters to erect a simple dipole,

even one that can rotate. At a 16' length, it can be hidden if need
be. It pattern is likely to be superior to either the narrow flat-top
lobes of the high-angle Vee radiation. In other words, the idea of 1
antenna for everything may not be best for everyone.
Chapter 35
Chapter 36: About the All-Band Doublet
T
he all-band doublet horizontal wire antenna has a history
almost as long as amateur radio itself. Despite all the words
and diagrams in handbooks over the years, newcomers still
send me questions about the antenna. I have collected the
questions and boiled them down to 10, all of which have many
variations. The goal in tackling these frequently asked questions is
to help newer hams erect a successful antenna system.
1. What is an all-band doublet? The all-band doublet is actually

an antenna system and not just an antenna alone. Fig. 1 shows the
basic elements of the system. The horizontal center-fed wire forms
the antenna proper, which accounts for the radiation of transmitted
energy and the reception of incoming energy. The parallel
transmission line transfers the energy from the antenna to the
antenna tuner (or antenna-tuning unit, the ATU) or vice versa. We
insert the tuner because the impedance that shows up at its
terminals will vary widely from one band to another. So we need a
way of matching the impedance at the tuner terminals to the
standard 50-Ohm input and output impedance of the transceiver.
Chapter 36
The antenna wire itself can have many lengths, but should be about
½ wavelength at the lowest operating frequency. Table 1 shows
common doublet lengths that have appeared in handbooks since
the 1930s. It also shows the ham bands covered by the antenna.
Note that the 100' wire, while somewhat shorter than ½ wavelength,
can be pressed into service on 80 meters, and the 67' wire might be
used on 60 meters. However, there are limits that we shall explore
as we proceed through the questions.
Chapter 36
2. What's the difference between a doublet and a dipole? This

interesting question has 2 answers: none and a lot.
Conversationally, the term "dipole" often refers to any antenna that
looks like a dipole, that is, a center-fed wire antenna with a feedline
going to the shack. In this context, we also tend to call any end-fed
antenna a Zepp (although there is a center-fed extended double
Zepp) and to refer to any off-center-fed antenna as a Windom
(although the original Windom had only a single feed wire).
In more precise terms, the coax-fed dipole that we sometimes set

up for single-band use is a more complex antenna than its
appearance suggests. It is actually a center-fed resonant ½-
wavelength dipole. The center-feedpoint is obvious from the
position of the feedline. It is resonant since the feedpoint
impedance is (almost) purely resistive, with little or no reactance.
The length is electrically ½ wavelength, which for any real wire or
tubular element turns out to be shorter than a physical half
wavelength. Finally, it is a dipole because, as Fig. 2 shows, the
charge is minimum at the center feedpoint and maximum at the
element ends. As a result, the current is maximum at the center
and minimum at the wire ends. The dipole undergoes only one
transition in charge and in current from the center to the wire end.1
Chapter 36
When we use the antenna on many bands, it becomes electrically

longer, because the length of a wave grows shorter with rising
frequency. Hence, the charge and current patterns do not satisfy
the dipole conditions above the lowest band or two. Fig. 3 shows
the current distribution along a 135' wire at 80 meters and at 10
meters. Since the current does not follow the dipole pattern, the
charge density is also different from a dipole. In this case, there are
many transitions and the current is not maximum at the center
feedpoint.
Chapter 36
The term doublet is more generic and less fully descriptive than
others. However, it also has a history. In the 1930s, it served as a
label for a center fed wire with a special feed system. Later, the
antenna was renamed the delta feed and the term doublet became
a generic term for center-fed antennas of any length.2 Hence, our
antenna is an all-band doublet.
3. Do I need to measure the wire for precise resonance on the

lowest band? In a word, no. When we set up a resonant
monoband dipole, we want it to achieve resonance or the lowest
possible SWR with our coax cable feedline. However, the all-band
doublet antenna system uses (normally) high impedance parallel
transmission line. Small variations in antenna wire length will make
no difference to doublet performance or to our ability to match the
impedance at the shack end of the feedline. We sometimes see
radiation patterns for a Yagi antenna change shape as we move
Chapter 36
from one end of an amateur band to the other. However, the

patterns produced by the all band doublet change very slowly with
frequency. For example, if we only have room for 125' of wire, then
it will do very well and yield the same results as a 135' wire on the
lower HF bands. On the upper HF bands, we might see some
change in the feedpoint impedances between the two wire lengths,
but they normally will not be severe and certainly not large enough
for us to abandon the antenna. The recommended lengths in Table
1 are ballpark figures, not precise lengths.
4. Why do I need parallel transmission line? Why not just

coax? Or coax with a 4:1 balun? To get our hands on this
question, let's consider only one of the possible doublet lengths:
135'. For this version of the doublet, we can look at the numbers in
Table 2. The second column lists the approximate feedpoint
impedances for each HF amateur band. These numbers will vary
with the exact length of the wire and the height above ground.
However, the approximations will serve well for our demonstration.
Chapter 36
Suppose that we connect a typical coaxial cable to the feedpoint

and use 100' of the line to reach the shack. RG-8X is popular these
days because it is light and easy to handle. How much energy will
we lose if we use this cable as a feedline? We can arrive at some
answers by using a program like TLW. This highly useful software,
written by Dean Straw, N6BV, accompanies The ARRL Antenna
Book, which is a worthy long-term investment for any ham. In the
table, columns 3 and 4 show the 50-Ohm SWR for each of the
impedances and the total cable losses. Notice how many of the
loss entries exceed 10 dB. With a 10-dB loss, only 1/10 of the
energy at one end of the line is available for use at the other end of
the line. The reduction applies whether we are transmitting or
receiving.
The last 2 columns show the SWR for a 450-Ohm parallel

transmission line. The type specified uses a vinyl coating with
windows along the way. The vinyl coating is simply a good way to
Chapter 36
keep the wires evenly spaced, but it does introduce losses that are
slightly greater than open or true ladder line (bare wire with periodic
spacers). Note that even with the highest SWR levels, losses do
not exceed 0.5 dB or a little over 10% of the power, even with 100'
of the line.
Parallel lines do have limits however. Remember that we

recommended ½ wavelength at the lowest frequency as the
shortest antenna wire length. We also suggested that we might
press shorter wires into service, but we did not say how much
Chapter 36
shorter. Let's see what happens below 80 meters as we shorten the

wire from a ½-wavelength starting point. Fig. 4 shows the
approximate resistance and reactance. Although the curves appear
to track each other, remember that the downward path of reactance
actually represents increasing capacitive reactance. As we shorten
the doublet or lower the frequency, the feedpoint resistance
decreases steadily, while the capacitive reactance increases
steadily. The result will be a very high 450-Ohm SWR on the
parallel line. It will rise to the point where even the seemingly low-
loss line shows significant power losses along the way. As a
practical matter, try to keep the antenna at least 3/8 wavelength or
longer at the lowest frequency used if you cannot manage ½
wavelength. Remember that you can always zigzag the wire legs or
let the ends droop downward (but always with their ends out of
human reach) in order to lengthen the wire to the full ½ wavelength
at the lowest frequency.
5. What's the most important factor in setting up an all-band

doublet? Or, we put up a low-band doublet for Field Day about
10-15' off the ground. We did not make many contacts? What
was wrong? The question's second form gives us the answer to
the general question. With an all-band doublet, there is no
substitute for height. However, hams must work with real conditions
and not ideals.
Let's continue to use the 135' doublet as our antenna and see what
happens at various antenna heights that hams actually use. 20' is a
typical Field Day height for wire antennas due to the difficulty of
erecting and sustaining higher supports. 40' is a nice round number
Chapter 36
for a backyard doublet supported by mature trees. 60' is out of

reach for amateurs unless they have a tower or two supporting
rotatable beams. Now look at Table 3. It lists for each sample
operating frequency the height above ground as a fraction of a
wavelength.
The height above ground when measured in terms of a wavelength

is the most important factor that determines the elevation angle of a
horizontal antenna's radiation. (Remember that the radiation angle
is also the angle of reception sensitivity.) Fig. 5 provides a catalog
of typical elevation patterns for the doublet. Each pattern uses the
headings for maximum gain as a basis. The missing bands would
show elevation angles of maximum radiation that are part way
between the bands in the illustration.
Chapter 36
Note that at 80 meters, all three heights are so low that we detect
very little elevation pattern difference. The pattern begins to change
significantly as we raise the antenna to 60' when operating on 40
meters. The 20-meter pattern becomes very usable for low-angle
skip radiation when we raise the wire to 40', a little bit more than ½
Chapter 36
wavelength. 20' is a little over ½ wavelength on 10 meters, and so

we obtain reasonable basic performance on that band.
These notes and graphics cannot change your backyard or field

conditions. However, they do provide food for thought. For
example, if you really want to operate on 80 and 40 meters, but
cannot get the horizontal antenna high enough as a fraction of a
wavelength, then you may wish to consider alternative antennas.
You might achieve better performance on the lowest HF bands with
a different wire antenna, such as the inverted-L.3
6. I carefully set up my 135' doublet to be broadside to Europe.

However, on 15 and 10 meters, signals are much stronger to
Africa than to Europe. Is it propagation? Although propagation
affects all HF communications, the most likely source of the weak
European signals is a misunderstanding of the azimuth patterns
produced on the various amateur bands by the 135' doublet.
Chapter 36
As a center-fed wire antenna grows longer in wavelengths, the

number of lobes that it produces increases. Table 4 lists the length
of our 135' doublet as a function of a wavelength on each operating
frequency. It also lists the number of lobes produced in the azimuth
pattern, that is, around the horizon. The numbers may seem odd,
but there is nothing disorderly about them. As the antenna grows
longer (or we increase the operating frequency), lobes emerge,
reach a peak value, and then disappear--to be replaced by other
emerging lobes. Three main points describe the process. Let n be
the length of the antenna rounded to full wavelengths.
1. If the antenna is n wavelengths long, then the number of lobes

will be 2n. On 40 meters, the antenna is about 1 wavelength, so
there will be 2 lobes. At 10 meters, the antenna is about 4
wavelengths, so we shall find 8 lobes.
2. If the antenna is n.5 wavelengths, we shall find the lobes for n

wavelengths and the lobes for n+1 wavelengths at close to equal
strength. So add the number of lobes for n wavelengths and the
number of lobes for n+1 wavelengths to arrive at the total number
of lobes. At 17 meters, the antenna is about 2.5 wavelengths. A 2-
wavelength wire gives us 4 lobes and a 3-wavelength wire yields 6
lobes. So at the intermediate length, we shall find 4+6=10 lobes.
Chapter 36
3. When the wire is close to n or to n.5 wavelengths, the strongest

lobes will be those farthest from broadside to the wire, that is,
closest to in line with the wire. Table 5 provides the modeled
performance data of the 135' doublet at a height of 40' above
ground. Fig. 6 translates those numbers into a gallery of azimuth
patterns. The virtual antenna runs up and down on the graph page.
Because the take-off (TO) angle (or the elevation angle of
maximum radiation) is so high for 80 through 40 meters, the
azimuth patterns use an arbitrary elevation angle of 45°. All other
patterns use the actual TO angle.
Chapter 36
Chapter 36
Except for 30 meters, where the null between the inner lobes is
hard to detect, all of the patterns clearly exhibit the number of lobes
calculated in Table 4. Since all of the lengths are close to either a
full wavelength or the half-wavelength mark between full
wavelengths, the strongest lobes are those nearest to being in line
with the wire. (When the wire is close to n.25 or n.75 wavelengths,
other lobes may dominate.)
Note that when the length is n.5 wavelengths, the large number of
lobes in the pattern forces the strongest lobes to be closer to in line
with the wire than for the next whole number of wavelengths.
Hence, the angle of the lobes away from broadside is greater on 17
meters than on 15 meters--and greater on 12 meters than on 10
meters. Also note that the larger the number of lobes in a pattern,
the narrower the beamwidth of each lobe.
If we had chosen a 67' doublet, the antenna would be ½

wavelengths on 40 meters, 1 wavelength on 20 meters, and 2
wavelengths on 10 meters. Since the azimuth lobes are functions of
the wire length in wavelengths, we would obtain different lobe
patterns than for the 135' wire. In fact, the 67' wire pattern on 40
would resemble the 135' pattern on 80, and the 67' 10-meter
pattern would look very much like the 135' 20-meter pattern.
How you orient a center-fed doublet depends on understanding

both the elevation and the azimuth patterns for the wire. The
azimuth patterns show where your signal is likely to go, while the
elevation patterns tell you whether the energy is likely to fall within
the skip zone. Orient the doublet so that the pattern for the most
Chapter 36
used band (or bands) covers your most desired target(s) with a
strong, low-angle lobe.
So far, we have concentrated on the wire or antenna-proper portion

of the all-band doublet antenna system. We briefly explored the
main reason for needing to use parallel transmission lines to
connect the antenna to the antenna tuner. Hams who are used to
using coax often ask a number of other questions about parallel
feedlines.
7. Can I run the parallel feeders in a PVC tube underground or

under my house? This question is actually a confession by the
newcomer that he or she knows how to handle coaxial cable, but
not parallel feedline. In a coaxial cable, the energy fields exist
between the outer surface of the inner or center conductor and the
inner surface of the outer conductor, also called the braid. Hence, if
the cable has an outer jacket that can handle soil, burying it does
not affect its use or operation. As well, we can run the cable near to
other wires without significant difficulty.
Parallel feedline is also called open-wire transmission line, and for

good reason. Regardless of whether the wires have insulation, they
are open in the sense of having fields that are not confined by the
structure. Although the main portion of the field is between the two
wires, it also extends around the pair of wires for a considerable
distance--up to a few times the spacing between the wires. Nearby
conductive and semi-conductive materials can disturb the balance
between the lines and cause them to radiate--a job we want the
antenna proper to do. As well, we may lose some energy to those
Chapter 36
objects. So, in a nutshell, the answer to the question is no. Do not

run the transmission lines close to or within the ground, even if you
give them the double insulation of a conduit.
Commercially available lines come in 3 general types, each with a

different characteristic impedance, construction, velocity factor
(VF), and loss value. 300-Ohm transmitting twinlead, sometimes
flat and sometimes tubular, has a VF of about 0.80 and a loss of
about 0.17 dB per 100' at 3.5 MHz. Remember that line losses
increase with frequency. 450-Ohm window line, a form of flat
twinlead with cutouts to minimize the vinyl between the wires, has a
velocity factor of about 0.91 and a 0.07-dB loss per 100' at 3.5
MHz, half the loss of 300-Ohm line. 600-Ohm open-wire ladder line
typically has a velocity factor of about 0.92 or higher and a loss of
only about 0.03 dB per 100' at 3.5 MHz. There are also
commercially available ladder lines in the 400-500 Ohm range, and
their VF and loss values would resemble those of the 600-Ohm
line. Of the 3 types, 450-Ohm window line is perhaps the most
popular for all-band doublets.
Parallel feedline has a few simple rules for effective placement to

maximize energy transfer from the tuner to the antenna proper.
Keep line runs as straight and in the clear as possible. Straight,
clear runs are as important indoors as outdoors. Straight is self-
evident. Wherever possible, keep direction changes shallow. Never
let the line fold back upon itself or roll it in a coil. Clear means as far
from other objects as possible, and in no case less than several
times the line spacing away from anything.
Chapter 36
Of course, we must bring the line indoors. We can use a short

through-wall PVC pipe, perhaps with caps that have slots to keep
the line centered. Or we can use a wood or plastic plate with feed-
through insulators. The difference in spacing and bolt size on the
board relative to the line is not important: it may create a small
impedance bump but will minimize losses. Outdoor supports can be
of two general types: rings or clamps. We can suspend non-
conductive rings (slices of PVC or similar) from limbs and posts to
support the line on its way to the antenna. As well, we can create
non-conductive guides or clamps that extend outward from tree
trunks, posts, or walls to route the transmission line. Be sure to use
enough supports.
At the junction with the antenna, use a strain-relief fixture. A simple

insulator may keep the line from being pulled by the antenna wire.
However, over a relative short time, the feedline wires will flex back
and forth until they break. A fixture that minimizes the flexing at the
junction itself will make the connections much more durable.
8. Will the feedpoint impedance in the tables appear at the

antenna tuner terminals? If the feedline is precisely a multiple of
an electrical half-wavelength, then the feedpoint impedance will
reappear at the far end of the line. (The other condition that would
allow the feedpoint impedance to reappear is an exact match
between the feedpoint impedance and the characteristic impedance
of the cable. With 450-Ohm line, Table 2 makes it clear that this
condition will not exist.) When the characteristic impedance of the
line does not match the feedpoint impedance, the line becomes a
continuous impedance transformer and shows a different
Chapter 36
impedance at each step between the feedpoint and each half-

wavelength or 180° point along the line.
Fig. 7 shows one example of the transformation and applies to 450-

Ohm transmission line and a feedpoint impedance of 2000 - j2000
Ohms. This impedance is similar to some values in Table 2. If the
reactance had been inductive instead of capacitive, we would see
similar curves, but the peaks would appear at 10-15° position along
the line (where 0° is the antenna feedpoint and 180° is a half-
wavelength down the line). Note the very low resistance and the
Chapter 36
relatively low reactance that appear over much of the line's length.
For this reason, placing a 4:1 balun in the line may be a poor
choice for converting the balanced line to a single-ended or
unbalanced line. The transformer may end up converting a low
impedance to a very low impedance, regardless of the balun's
ability to handle the reactance at its terminals.
Let's change the feedpoint impedance to 20 + j100 Ohms. Some

impedances for very short doublets (less than about 3/8
wavelength) show a higher reactance (capacitive), but the peaks
Chapter 36
become very high and the resistance becomes very low--a very
difficult situation to graph. Fig. 8 graphs the resistance and
reactance of the selected values along the line. Note that the peaks
occur just before the 90° or halfway position. If the reactance had
been capacitive, the same peaks would appear just past the
halfway point along each half wavelength of transmission line.
Once more, note for how much of the line the resistance and
reactance are low to very low.
The two samples show some of the extremes of impedance

transformation along a 450-O transmission line. The closer the
feedpoint impedance is to the line's characteristic impedance, the
less radical will be the transformation of resistance and reactance.
It pays to have a small calculation program to assist in finding and
visualizing the data. Earlier, I mentioned the N6BV program, TLW.
You can obtain similar graphs on it. In addition, the graphs will
show the effects of line losses.
The graphs show an electrical half wavelength of line. The physical

length of such a line will vary with the operating frequency. Hence,
it is very difficult (although not impossible) to design a feedline
system so that on each band we end up with just about the same
impedance at the shack entry. Most amateurs let the antenna tuner
do the work of transforming whatever impedance appears at the
terminals to the transceiver's required 50 Ohms.
9. What kind of antenna tuner is best for an all-band doublet?

The best type of antenna tuner is one with a configuration that
naturally has an unbalanced or single-ended input--to accept the
Chapter 36
transceiver's coaxial cable--and a balanced output. From the

earliest days of amateur radio, a common tuner meeting these
conditions has been the link-coupled tuner. Fig. 9 provides a
simplified schematic diagram of one version of this tuner. It
received its name because the input side used a small coil or link
that is inductively coupled to the tank or parallel tuned circuit on the
output side. The most effective forms of this tuner used additional
components on the output side to compensate for the reactance at
the terminals. Taps at every turn (or at least at every other turn) of
the tank coil allowed the user to find a setting that came closest to
providing a good match and maximum power transfer at the same
time. The link might also have switched taps with the later addition
of the so-called variable coupling series capacitor. In fact, the
series capacitor compensates for remnant reactance on the input
side, allowing a purely resistive input impedance. Johnson
Matchboxes, with simplified tank tapping, a fixed link, and no series
input-side capacitor, became famous and still appear at hamfests.
Chapter 36
From the late 1960s onward, the single-ended network came to rule
the commercial manufacture of antenna tuners. Fig. 10 shows 4
popular configurations, with the CLC-T being the most common. It
was perhaps the cheapest to produce in a period of rapidly rising
component costs. It would also handle a very wide range of
impedances at the output terminals. However, the CLC-T was a
high-pass filter network and hence provided little harmonic
suppression for older rigs. Like all of the single-ended
configurations, it required a balun on the output to allow for
balanced lines. The standard version of the balun used a 4:1
impedance ratio either though a misunderstanding of the
impedances likely to be present at the terminals or because such
baluns were cheaper to make than 1:1 baluns. The baluns were
Chapter 36
transmission-line transformers that are most efficient when the

reactance is very low. Most balanced lines, however, did not meet
this condition. The average operator did not have multiple tuners to
compare and so remained unaware that on some bands with some
tuners, efficient power transfer might not occur.
Chapter 36
In recent years, interest in antennas that require parallel

transmission lines has surged, spurring the development of new
inherently balanced tuners. Fig. 11 shows three varieties that are
either on the market or in handbooks. The single-ended CLC-T
network is usable with special precautions not to ground any
component except the transceiver side of the 1:1 input balun that is
common to all of the tuner designs. One commercial tuner uses a
balanced CLC-T network, but the most common balanced network
tuner on the market is the reversible-L circuit. Versions exist for
high power use. However, as the operating frequency increases,
the range of impedances that the reversible-L will match with
standard components grows more limited.
Chapter 36
If you will buy a tuner with an all-band doublet in mind, then one of
the balanced network tuners may be the best bet. However, if you
already have a tuner--even a single-ended network with a 4:1 balun
Chapter 36
on the output side--you might as well try it out. Since none of the
tuners comes with a relative output indicator, you will have to
estimate efficiency on each band indirectly. If you obtain a good
match following the maker's suggestions for the best component
settings, check the temperature of the balun after (not during)
operation. If the balun is warm to the touch, it likely is converting
some part of your transmitted energy into heat. In general, the
broader the tuning, the lower the tuner losses, although there are
exceptions to this rule of thumb. In a tuner designed for all of the
HF ham bands, tuning will naturally become sharper with rising
frequency.
If you cannot obtain a match on a given band, then try inserting a

length of transmission line, preferably outdoors. Using knife
switches, relays, or a simple manual changeover, add a few feet of
line between the line ends of a break that you intentionally make in
the feeders. Form the insertion into a single large loop to avoid
unwanted self-coupling, and use standard precautions to prevent
coupling to other objects. Since the transmission line is a
continuous impedance transformer, the new values of resistance
and reactance at the tuner terminals may fall within the tuner's
range. Since every tuner has a limit to the range of resistance-
reactance combinations that it will handle, the potential need for a
revision in the total feedline length may apply to all antenna tuner
designs.
10. My all-band doublet is 50' high and uses open-wire feeders.

It works well, but I get a lot of RF interference at home, and my
rig sometimes locks up in transmit on CW. How can I
Chapter 36
overcome these problems? Unwanted coupling into home

electronics and into the rig itself has almost as many causes as
there are errors that we may make in installing parallel feedlines.
The first step is to ensure that all station equipment is well
grounded to an earth ground as close to the rig as may be feasible.
The second step is to consider rearranging the station so that you
position the antenna tuner at the place where the feedline enters
the building or shack. Well-grounded coax braid is less likely to
couple RF energy to other lines and objects than open-wire
transmission line.
The third step is to check the routing of the transmission line as it

approaches the entry point. Ideally, the line should approach the
entry perpendicular to the wall or window. If the line runs vertically
down a wall, it may couple energy into various power, telephone, or
computer lines. Some of these lines may use shielded cable, but
unless that cable is also well grounded, it may carry RF energy to
sensitive devices with equally poor grounds.
Sensitive devices, including control inputs for the rig, do not require
very much energy to show signs of interference. If all else fails, you
can try the system shown in Fig. 12. At the building or shack entry,
install a 1:1 choke of ferrite beads, following the designs of W2DU.
The choke acts as a balun, converting the balanced line to the
unbalanced coax. From the coax connector shell, run a very short
earth-ground line. Ideally, the choke should go outdoors, but
modern building construction may require immediate indoor
installation at the entry point. Between the choke and a single-
Chapter 36
ended network tuner, run less than 20' of the largest, lowest-loss
coax that you can obtain.
The system shown will generally eliminate most unwanted RF

energy transfers if the feeders have not already coupled into house
wiring due to improper dress. It bypasses the 4:1 balun in the tuner,
avoiding that loss source. However, the system has losses of its
own. The 1:1 choke will show losses with high impedances having
significant reactive components. The coax will also show some
loss. However, if the length is 20', the line losses will usually be
fairly small. For example, at 30 MHz, 20' of RG-213 will show a 1.1-
dB loss with a 10:1 SWR. A shorter run, lower frequency, or lower
SWR will result in lower coax losses. There are also cables with
even lower losses. Do not use thin cables like RG-58 for this run,
regardless of the operating power level. This system is not ideal,
Chapter 36
but simply a measure of last resort for very tough cases of RF

interference. Before employing this or other radical systems, you
should first use the earlier guidelines to optimize the feeder and
tuner installation.
These notes do not answer every question that we can ask about
the all-band doublet. However, I hope the 10 common questions
that we have tackled give you a good start for reasoning out
answers for yourself.
Notes
1. Fig. 2 follows Stutzman and Thiele, Antenna Theory and Design,

2nd Ed. (Wiley & Sons, 1998), p.57, although the treatment of the
dipole in very short or longer forms is similar in Kraus and in
Balanis. (Kraus, especially, (Antennas, 2nd Ed.) is careful not to
label longer center-fed wire antennas as dipoles).
2. Compare, for example, 2 editions of The Radio Amateur's

Handbook: the 13th or 1936 edition, p. 278, and the 24th or 1947
edition, p. 207.
3. The inverted-L is a highly usable antenna for general-purpose

communication, but requires an entirely different discussion. See
"Straightening Out the Inverted-L" in the 2005 Proceedings of
FDIM, produced by QRP ARCI.
Chapter 36
Chapter 37: The Zig-Zag Dipole-Doublet
One over-age myth about wire antennas is that they must be

straight. Ideally, we would like them to be truly linear. However,
even a kinky wire can perform quite well.
Chapter 37
Consider the scenario sketched in Fig. 1. A standard 1/2 wl dipole

for 80 meters--about 135' long when about 50' up--would look like
the upper sketch if we had the room for a 67.5' long wire runs on
each side of the feedpoint. However, suppose that we do not have
the room for the full length of the wires. We can settle for a shorter
wire antenna, but we do have another option if supports are
available: the zig-zag special. What we did with dimension A in the
top drawing, we shall now do with A + B in the lower drawing.
The antenna could have been made into a U, but the loss of gain
would have been slightly higher than with the zig-zag--due to the
partial cancellation of the radiation from the facing end sections.
However, the amount is small enough that, if a U is all that you can
manage on a site, "U"se it.
To see what happens when we zig-zag our traditional dipole I ran a

series of models, each of #12 copper wire over average soil.
Modeling is limited in that it assumes clear, level terrain, and so it
cannot take into account the hills, valleys, and ground clutter of the
typical ham installation. Nonetheless, the trends are quite useful for
comparative purposes.
If the antenna is set 50' up, the typical dipole pattern at an elevation
angle of about 20 degrees is an oval at right angles to the wire.
Let's see what happens as we turn more and more of the antenna
into opposing end pieces. For the example, I used 5% increments
of the half length, thus shortening each side of center by 3 3/8' with
each move. Theoretically, the end piece should grow by that
amount to keep the antenna resonant. Actually, we shall have to
Chapter 37
lengthen the ends slightly with each change in order to compensate

for coupling between the wires near the corners.
The following table lists the wire lengths each side of center (A)
with both the calculated and actual end pieces (B) need to restore
resonance at 3.5 MHz. The feedpoint resistive impedance at
resonance is also shown, along with the maximum gain. The final
figure is the number of degrees off broadside that the pattern tilts
as a result of the zig-zag ends.
End (B) Calc. Act. Length A Gain Pat. Tilt Feed R
% Feet feet feet dBi degrees Ohms
0 0 0 67.5 0.06 0 70.0
5 3.4 3.7 64.2 0.06 0 67.6
10 6.8 7.3 60.8 0.04 1 66.4
15 10.1 11.0 57.4 0.02 1 64.9
20 13.5 14.5 54.0 -0.01 2 62.3
25 16.9 18.2 50.6 -0.05 2 59.7
30 20.3 21.7 47.3 -0.09 5 56.2
The total loss in gain within the situation set up is 0.15 dB for the
entire spread from a linear wire to an antenna with 30% of each
side turned at right angles to the main wire. If the zig-zag happens
to be more open than the right angle used in the example as an
extreme case, the loss will be less. However, it is already so low as
to be undetectable in operation.
Had we bent the ends to form a U, the gain in the most extreme
case would have been very slightly lower than for the zig-zag
dipole, and so too would have been the source resistance at
resonance. Another comparison of note is between the 20% zig-
zag model and a wire 108' long and linear--something close to the
Chapter 37
traditional G5RV length. The G5RV would have shown about 0.1
dB less gain than the zig-zag, which would have been far less
operationally significant than the high capacitive reactance at the
feedpoint. However, if we feed the antenna with parallel feedline
and an antenna tuner, all of these differences fall among the trivial.
The greater the amount of antenna devoted to the zig-zag ends, the
longer the wire must be to restore resonance. Again, a more open
zig-zag will show smaller amounts of required lengthening.
Likewise, the feedpoint resistance goes down more rapidly as the
zig-zag becomes more extreme.
The amount of pattern tilt is very mild, even at the 30% zig-zag
mark. Fig. 2 below sows an overlay of the straight wire and the zig-
zag azimuth patterns for the 20-degree elevation angle. Again, in
real operation, the difference will be unnoticeable. Notice that the
pattern tilt is away from the bent ends.
Chapter 37
As the zig-zag involves more than 30% of the wire on each side of
center, the pattern tilt becomes more extreme, exceeding 10
degrees as the lengths A and B approach each other. We can view
this amount of tilt as a disadvantage, or we can put it to use.
Suppose the main supports we have will place the broadside
pattern some 10 degrees off target for our desired operation.
Making the antenna into a zig-zag dipole can put us back on target.
Chapter 37
The Dipole Becomes a Doublet
If we choose to use the zig-zag on other HF bands, what happens?
The first thing that happens is that the antenna is no longer a

dipole. A dipole is an antenna with a single current maximum at its
center and voltage maxima at its ends. It is a center-fed dipole in
the version with which we are working. However, since its length
will no longer be apt to produce the current and voltage conditions
along its length once we increase the frequency of operation, it will
no longer be a dipole. Typically, a multi-band single (simple) wire is
best termed a "doublet," a term that implies nothing in itself about
the current and voltage distribution along the length.
The second thing that happens is this: the exact length is no longer
of great consequence. Our first tests intentionally strove for
resonance at 3.5 MHz in order to see what happened to the length
of the end pieces. In multi-band use with parallel feeders and an
antenna tuner, the length is no longer critical. The patterns will not
significantly change with up to 5% differences in overall length, and
antenna resonance is no longer a serious consideration.
A third phenomena is the variation of the patterns of lobes and nulls

from those that we are used to associating with a straight-wire
doublet. To see what happens, let's use 40, 20, 15, and 10 meters
as test bands to compare the patterns of a straight-wire doublet and
our 30% zig-zag doublet--both of #12 copper wire 50' up. Of
course, if we use a smaller amount of zig- zagging, then any
Chapter 37
deviations of patterns from the normal doublet pattern will be that

much less.
In each of the patterns shown below, the antenna extends from one
side of the pattern to the other. The zig-zag legs bend downward
(relative to the page) on the left and upward on the right. Hence,
most of the pattern tilting will be to the upper left corner of the page,
at least at lower frequencies.
Chapter 37
At 7.0 MHz, the zig-zag pattern shows a 5-degree tilt relative to the
broadside lobes of the normal doublet. The elevation angle of
maximum radiation is still very high, so a 20-degree elevation angle
has been selected for the comparison to reflect something
approximating normal skip angles. The gain of the zig-zag is slightly
less (by about 0.5 dB) than that of the straight wire and is
accompanied by a broadening of the beamwidth in both directions.
Since the antenna is about 1 wl long, the feedpoint impedance is
very high. The zig-zag side nulls are shallower than those of the
normal doublet. However, none of these differences are likely to
result in any gained or lost contacts.
Chapter 37
Lest we simply presume that the remaining HF bands will show

essentially similar parallels between the straight and the zig-zag
doublets, operation of the antennas at 14.0 MHz serves as a
reminder that difference might emerge at any frequency. The
elevation angle of maximum radiation on 20 meters is 20 degrees
with the antenna at the 50' mark. Hence, the patterns show the
maximum gain of the antenna. The straight-wire doublet shows the
familiar 4-leaf clover pattern typical of a wire 2 wl long.
Chapter 37
In contrast, the zig-zag antenna shows much greater tilt, with the
peaks being about 20 degrees distant from those of the normal
doublet. The nulls are just barely perceptible, but with that improved
coverage comes a price: the lobes are weaker than those of the
normal doublet by about 1.3 dB.
At 21 MHz, the normal and the zig-zag patterns almost oppose

each other, with zig-zag lobes filling normal nulls and vice versa.
Chapter 37
Once more, the normal doublet shows a higher maximum gain (by
about 1 dB), but the zig-zag doublet tends to have shallower nulls.
Part of the reason for the especially strong zig-zag lobes off the
ends of the antenna is that each bent section of the zig-zag is
approximately 1/2 wl long at 15 meters. Had the zig-zag "B" length
been shortened, the end radiation would have decreased rapidly.
When operating the antenna at multiples of its initial frequency, the
current magnitude shows a number of peaks, and the geometric
configuration plays an increasingly significant role on the ultimate
azimuth pattern generated.
Chapter 37
On 10 meters, at 28 MHz, there are so many lobes that the

difference in the two patterns becomes less significant
operationally. In fact, there is no significant difference in the
strength of the largest lobes of the two antennas. However, we may
note the two small lobes off the end of the zig-zag antenna.
Because the end of "B" lengths are no longer close to 1/2 wl long,
they develop lesser lobes. Some of the versions of the zig-zag with
Chapter 37
shorter "B" dimensions might well show stronger radiation off the
antenna ends.
The feedpoint impedance of both antennas at the even harmonics

of the original 1/2 wl frequency of the antenna will be high. The
exact figures will be functions of the antenna's exact length. At
harmonics, effects of the zig-zag will vary slightly from band to
band, and hence the feedpoint impedances will not be identical to
those of the straight wire. Values of resistance in the 1,000 to 4,000
Ohm range and values of reactance from 500 to 1200 Ohms are
likely to be common for both the straight-wire and the zig-zag
doublets. What values appear at the antenna tuner terminals will
depend not only on these load values, but as well on the
characteristics and length of the feedline used. If a tuner cannot
handle the values presented on a certain band, insertion of a short
length of additional feedline will usually correct the situation.
Other Variations
We have already noted that when the ends of the antenna are bent
in the same horizontal direction, the resulting U-shaped antenna is
only a tiny bit lower in gain than the 30% zig-zag. A more common
scenario is to droop both ends downward. At the fundamental
frequency, this configuration tends to lower gain still further, since
the ends are closer to the ground. However, the result is far from
disastrous. At higher frequencies of operation, the ends may show
significant vertically polarized radiation, but the net effect will not be
sufficient to alter the basic horizontally polarized patterns for each
band.
Chapter 37
Perhaps the ultimate utility of the zig-zag doublet is to fit a full 80-
meter length into a fairly restricted yard size, as suggested in Fig.
7. running the antenna diagonally across the yard for the available
space and then tilting the wires back along the yard lines (assuming
supports are available) can make a multi-band doublet available to
almost anyone.
Chapter 37
The principle can also be applied to hidden roof-top or attic

antennas. The ends can be run along the rafters and roof trusses, if
appropriate care is taken to give clearance to conductive materials.
The urban dweller can still operate effectively even if circumstance

seems to dictate undersized antennas. The key is to think in
designer shapes, of which the zig-zag is a perennial winner. The
losses, compared to traditional straight-line designs, may be far
smaller than initially imagined.
Chapter 37
Chapter 38: The Y-Doublet
A
fter presenting some notes on triangles of doublets of
various lengths, I received more than one message recalling
an old Y-configuration from the 30s and 40s. The basic
scheme was designed for a given band and consisted of three 1/4-
wavelength wires at 120-degree angles coming together at a center
point. There, according to recollections, the old timers used a 3-
wire twisted feedline to the shack. At any one time, the operator
hooked up two of the 3 wires to the antenna tuner (or, in more
remote past times, to the rig output terminals). The result was a
steerable doublet.
Essentially, the operator was selecting the pair of feed wires that
created a doublet, with the third antenna wire relatively inert. We
normally think of a doublet as linear, but bending it by 30 degrees
does not especially harm its performance. So that part of the
system is quite sound.
More foreign to current practice is the twisted feedline. In the

1930s, many hams commonly feed their dipoles with low-
impedance parallel line. A 72-Ohm transmitting parallel line used to
be available, but apparently the high power version is no longer
made. 72-Ohm parallel lines made from round wires are not
feasible with open-wire construction, since the required center-to-
center spacing would require contact between the wires. However,
by using a carefully calculated thickness of an insulating material
with a known dielectric constant on each wire, the desired
impedance is achievable.
Chapter 38
Amateur practice tended to rely on two factors. First, the resulting

parallel line resembled ordinary line cord, sometimes called zip
cord. Second, properly configured antenna tuners or even amplifier
output circuits were capable of handling a fairly wide range on
impedances. Therefore, amateurs used to simply twist pairs of
insulated wires together to form a low impedance parallel feedline.
The wires might be line cord or they might be other insulated wires
twisted and taped together.
Adding a third wire to the set and leaving it disconnected from the
RF source was relatively harmless. If the wire was equally spaced
from the other two hot wires, it would have negligible current on it.
The antenna wire would be at essentially right angles to the main
pattern and hence induce a minimum of coupled antenna current
into the inert feeder. Since the currents on the other two feeder
wires would be equal in magnitude but opposite in phase, any
induced currents in the third wire would cancel, leaving no current
in the third wire.
Overall, the system effects a space savings over three doublets in a

triangle. With good solid AWG #12 wire for the elements, one can
use only three corner supports and let the triangle of antenna wires
support the center assembly. (One can always add a center
support, if convenient or necessary.) With that promise and the
potential for having a steerable doublet, the idea is worth further
exploration.
Chapter 38
The Steerable Y-Doublet Array
Let's begin by looking at the antenna wires and their potential

performance. We shall look more closely at the feed system later
on. The basic configuration of the Y-doublet appears in Fig. 1.
Chapter 38
We shall use as our test array a Y cut for 3.6 MHz. My free-space
model used 67' legs for initial checks. Hence, ignoring the
necessary insulating end ropes to the support trees or posts, we
get a triangle about 116' on a side and capable of fitting within a
rectangular back yard that is about 101' by 116'. The figure shows
the three feed wires, of which we shall use only two at a time. For
modeling, that means terminating each leg short of the exact center
point. Then we connect a short wire between 2 of the 3 wires. I
used a separation between inner leg ends of about 3' so that I could
use a 3-segment wire for the source and use segment lengths of
about 1' on the antenna wire legs.
Chapter 38
Fig. 2 shows the overlaid free-space patterns for the Y-doublet at

3.6 MHz using different pairs of legs to form each of 3 doublets.
The patterns indeed promise full horizon coverage as we switch
pairs of feedlines at the shack end of the feeder lines.
I also checked the antenna's performance on each of the bands

above 80 meters. All of the traditional ham bands (40, 20, 15, and
10 meters) yielded very high feedpoint impedances. Since we are
working with a low-impedance feedline, I set these bands aside as
not especially feasible for use with the system. (We shall review
this decision before we are finished.) However, 30, 17, and 12
meters showed feedpoint impedances sufficiently low to potentially
allow use of the antenna on these bands using the low-impedance
feeder system employed in first half of the 20th century. The
following table shows the free space performance potential of the
array.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Y-Doublet Modeled Performance: Free Space
Freq. Gain Feed Z

MHz dBi R+/-jX Ohms
3.6 1.70 59 - j 6
10.125 5.16 106 - j 375
18.118 4.62 134 - j 103
24.94 4.92 171 - j 291
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The free-space patterns are generally only applicable for a real

horizontal antenna over ground if the height is at least 1
wavelength. 80-meter doublets at 270' or more are rare. Therefore,
I remodeled the antenna at a 50' height to reflect a more realistic
Chapter 38
scenario. At that height, the maximum gain of the antenna has an

elevation angle that is nearly straight up. So I chose for that band
an angle of 34 degrees to reflect typical skip angles. The resulting
3.6-MHz patterns, shown in Fig. 3, are a good bit more oval than
their free-pace counterparts.
On the upper bands, I used the take-off (TO) angle for gathering
potential performance data. The antenna promises performance as
shown in the following table, with the leg-length adjusted to 66.5' to
bring the array close to resonance at 3.6 MHz. (Wire doublets tend
Chapter 38
to vary their feedpoint impedances with height in noticeable ways

when the doublet is less than 1 wavelength above ground.)
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Y-Doublet Modeled Performance: 50' Above Average Ground
Freq. Gain TO Angle Feed Z

MHz dBi degrees R+/-jX Ohms
3.6 3.43 34* 56 + j 8
10.125 9.98 26 110 - j 424
18.118 9.59 14 135 - j 153
24.94 10.23 11 182 - j 365
* 80-meter elevation angle arbitrary.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The patterns on the upper bands are not ovals by any means. Fig.
4 shows these patterns, but only one pattern per band for clarity. As
we increase frequency, we find two especially interesting pattern
properties. First, as the legs become longer in terms of
wavelengths, the patterns develop growing side "wings."
Eventually, by 12 meters, the main lobe has split into two forward
lobes. Second, as we increase frequency, the array becomes more
directional, with a growing differential between the forward and the
rearward gain.
Chapter 38
Still, the patterns may be usable for general amateur operations.

The question left is why we get reasonably low impedances at 30,
17, and 12 meters. Fig. 5 shows part of the reason.
Chapter 38
The graphics display the relative current magnitude distribution

along the doublet for each of the 4 bands. On 80 meters, we have a
somewhat typical dipole current distribution, with the current peak
at the feedpoint. On the other bands, we approximate a 3/2-, 5/2-,
and 7/2-wavelength doublet current distribution. Each of these
configurations places a current peak at the doublet center, resulting
in a relatively low feedpoint impedance.
We should also note that the unused wire shows a flat current line.
The current on it models out (in its perfectly spaced geometry) at
Chapter 38
about 4 orders of magnitude less current than on the active wires.

That is, if the current at the source is 1.0, then the current on the
inert wire shows a value of 0.0001 or 1E-4 or less.
The Feeder Question
The original system was designed for use with a twisted trio of
feedline wires, in other words, a twisted pair plus one. Fig. 6 shows
the general hook-up, but without any poor attempt on my part to
sketch a braid of 3 wires.
There are several questions about the feasibility of using such a

system in modern times. The first quandary is whether we can build
such a feeder system.
Chapter 38
Modern insulated wire tends to use higher quality (lower loss)

insulation than did the wire of yore. I would steer away from line
cord, but modern wires use plastics with better RF characteristics,
even if the only intend use is carrying DC. Since the system is
designed for a low characteristic impedance, but with considerable
SWR on the higher bands, I would recommend a heavy gauge wire,
perhaps #12 or so. The actual characteristic impedance will depend
on the thickness of the wire, the dielectric constant of the insulation
material, and how tightly we hold the wires together. Consequently,
I can give no exact figures.
However, you can make up lengths of a proposed feedline and

check the impedance in a number of ways with a variety of dummy
loads and a low-level signal source. Any one of the current crop of
antenna analyzers will give you a fairly accurate reading. Given the
relatively high dielectric constant of the insulating material, expect
to find a significant velocity factor, something in the 0.6 to 0.7
region.
The next inquiry has to do with the effective inertness of the unused
3rd feeder wire. I re-created the model of the Y-doublet using
parallel feedlines. Since twining the leads is not feasible in a
physical model, I simply dropped the three leads straight down from
the 50' level to 1' above ground. At that point, I connected two of
the feeder ends with a 3-segment source wire. Again, all wires used
a 1' segment length.
The resulting feedpoint impedances are not accurate to the low-Z

feeder system. However, that was not the point of the tests using
Chapter 38
the feeders with something over 800 Ohms as the characteristic

impedance. The question was whether the unused antenna and
feeder wires would remain inert relative to the active wires.
As one measure, the following performance table shows the effects

of the added copper losses of the physically modeled feedlines.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Y-Doublet Modeled Performance: 50' With Feeders
Freq. Gain TO Angle

MHz dBi degrees
3.6 3.35 34
10.125 9.74 26
18.118 9.60 14
24.94 9.85 11
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gain remains virtually unchanged. So, too, do the patterns, and the
outlines shown in Fig. 3 and Fig. 4 remain valid for the
reconfigured model.
A second test is to check the current distribution along both the

unused antenna wire and the ostensibly inert feeder. I actually
performed two tests, one with the unused feeder simply left open
and another with the feeder extended 1 foot to touch the ground.
The 12-meter current distribution graphic in Fig. 7 remains valid for
both.
Chapter 38
Note that the current line on the unused feeder and antenna wires
is flat. The relative current magnitude under either test condition on
all of the bands remained less than 1E-4 relative to a source
current of 1.0.
The modeling test, of course, has limitations relative to an actual

twisted trio of wires. In the test, the modeled wires are widely
separated and perfectly spaced along the entire 49' feeder run.
How well the twisted trio performs may turn out to be as much a
careful-construction issue as any other kind of issue.
Chapter 38
However, the tests suggest an alternative feed system that just

might open up the Y-doublet to use on all of the HF bands. Fig. 8
tells the story.
The Y-doublet on the traditional upper ham bands, 40 through 10

meters, can show feedpoint impedances in the thousands of Ohms,
with considerable reactance. Indeed, shrinking or expanding the
basic 80-meter legs may prove useful in reducing the high
reactance levels that accompany lengths that are close to even
numbers of half-wavelengths. Commonly, we try to select for a
Chapter 38
doublet a feedline characteristic impedance that is about the

geometric mean between the feedpoint impedance extremes that
we are likely to encounter. There is a practical limit to this effort,
since lines above 600-800 Ohms are difficult to produce. Hence,
600-Ohm or so open wire becomes typical for such applications.
We can create a trio of pairs by using circular spacers of the type

shown in the figure. For HF work, Plexiglas or polycarbonate
spacers should be satisfactory. We can cut a hole in the center of
each to reduce the weight. The holes can actually be slots if we add
bridge wires to hold the spacers in place. In essence, we are
adapting techniques normally used to create caged elements and
applying them to the feedline. Such lines might permit the use of
the antenna on all bands with a wide-range antenna tuner and will
go a long distance in maintaining something close to the modeled
ideal geometry we used in the test cases.
Of course, should you choose to work with a system of this order,

you can replace the alligator and crocodile clips of yore with an in-
shack switching system to change the orientation of the pattern on
all bands.
The resurrected Y-doublet as some potential of still being

serviceable today. There are many variables beyond the limits of
this initial feasibility check, so success is not assured. However, for
some hams who are restricted to backyard wire systems but who
wish some directional flexibility, the system may be worth a try.
Chapter 38
With the wide spaced feeder system, the system may also be
adaptable to 102', 88', 67', and 44' doublet lengths discussed in
other notes at this site and in mountains of other literature.
However, as with all horizontal doublets, the rule of thumb that calls
for the maximum feasible height remains in play for effective
operation.
Chapter 38
Chapter 39: The Doublet - Trap or Not?
T
he trap antenna, whether a doublet or a one-sided vertical,
was invented mostly to permit the operator to use coaxial
cable as a feedline. It was not invented for maximum
efficiency. A with all antennas, trap antenna adherents claim they
get good results--and indeed they do. Whether they get better
results than they would with other types of antennas of comparable
size is a question few are positioned to answer. The answer would
require that the trap antenna and the alternative be placed in nearly
the same position at the same height, and few of us can afford the
space, time, or money for such side-by-side comparisons.
There are two types of trap antennas, with examples illustrated in

Figure 1. The most common are those with traps, or parallel tuned
circuits, that are resonant at or just below the edge of the higher
frequency band to be covered, with extensions to make up the
length of the lower band. These antennas will be shorter than a full-
size dipole at the lower frequency, since the trap acts like an
inductor at the lower frequency, much like a mid-element loading
coil. However, the inductive reactance is not a product of the coil
alone, but of the tuned circuit making up the trap.
Chapter 39
The second type of trap antenna is one with a parallel tuned circuit
with the components and position selected to permit the antenna to
show a low SWR one several of the ham bands. W8NX, who has
done a great deal of work on these types of antennas, published an
80/40/17/10 meter antenna with only one trap each side of center,
and it was tuned to 5.16 MHz (QST, July, 1996).
Let's look at the more conventional trap antenna first and simplify it
to just 2 bands, like 80/40 or 20/10. A full size #14 copper wire
resonant dipole will have a gain of about 2.1 dBi in free space, but it
has this gain only in one ham band. We may use the gain figure as
Chapter 39
a standard against which to measure trap antennas for two bands.

The first thing we note is that performance of a two band trap
antenna of conventional design is dependent very heavily on the Q
of the trap. There are many trap designs, but here is a table of one
pretty good design with coils of various Qs. The gain is for free
space. Comparisons between dipoles and doublets at the same
height above real ground will show the same differentials.
Q High-Band Gain (dBi) Low-Band Gain (dBi)
50 0.7 1.7
100 1.4 1.8
200 1.8 1.9
400 2.1 2.0
800 2.2 2.0
Avoid low-Q trap coil designs. It is fairly easy to homebrew air-

wound coils with a Q of 200, and common coil stock usually meets
this figure. Even the best series-wound coaxial trap coils will not
have Qs higher than about 400, and most coils with Qs claimed to
be higher than 400 will not retain that Q under the influence of the
our chemistry-lab atmosphere. Nonetheless, a dipole with a gain of
1.8 or so will not yield results noticeably worse than a full size
dipole, since a half dB of lost gain translates into less than a tenth
of an S-unit. (Where these small losses mount up is in multiband
beams with traps in every element, since the losses of each trap
tend to be cumulative. They also add up in antennas with many
traps for many bands.)
The sample conventional 80/40-meter trap dipole in Figure 1 uses

traps tuned to 6.75 MHz. With a Q of 200, the traps equalize
performance on the two bands at just above 1.85 dBi in free space.
Chapter 39
This is only about 0.35 dB down from a full size dipole for each
band.
Well, that's not too bad. What about the other type of antenna, like
the W8NX improved trap antenna? Since the trap is not resonant at
any ham band, the antenna is functional over its entire length at all
advertised frequencies. On the three upper bands, the trap mostly
adjusts the reactance that appears at the feedpoint so that coax
can handle the feed task. On 80 meters, as Al Buxton notes, the
trap does exhibit significant losses--about 0.6 dB relative to the gain
of the wire of the same length (83.6') without the trap. (The 80
meter performance is down by a bit over 1 dB from a full-size dipole
for 80 meters.) Since most of the impedances are close to 100
ohms, replacing the recommended 1:1 balun with a 4:1 balun will
likely create no problems.
Since the W8NX antenna is operative along its entire length, its
patterns are not true dipole patterns on all but 80 and 40. On the
upper bands, they are multi-lobe patterns typical of a wire of the
same length fed with a parallel transmission line and an antenna
tuner--and at the same lobe strengths. So unlike the conventional
trap antenna, the special trap design acts like a simple doublet.
Now we have an additional selection criterion for our decision-

making machine. If we just have to have coax, then a trap design is
desirable, especially if we do not have space for a yard full of
standard dipoles. If we have to have standard dipole figure-8 or (at
low heights) oval patterns, then the conventional trap design is
indicated. If we have to have the coax, but are willing to accept
Chapter 39
patterns that are a function of the antenna length, then the special
trap design may be useful.
But--what if we do not really have to have coax? What if we could

use parallel feedline and an antenna tuner. And--what if the dipole
pattern were not too important to us? Should we still opt for a trap
antenna? Probably not.
First, traps are always a maintenance problem. More than their

losses, their inability to withstand weather without periodical
disassembly and cleaning is a disadvantage to most users. Open
traps are an invitation to big bug nests and closed traps invite little
insects that get into weep holes and eventually clog them.
Second, a doublet with an ATU allows one to put a signal on all the
ham bands. The W8NX antenna, without the traps, is about the
right length for an EDZ on 20 meters, but the high reactance
requires parallel feedline to avoid losses. With the traps and a
coaxial feedline, the band is not accessible without significant
power losses in the line.
Third, in the short run, a trap antenna may be cheaper than an

ATU, but since ATUs are not out in the weather, they tend to last a
lifetime. Hence, you can prorate their costs over many years more
than a trap antenna.
So if you need the exact things a trap antenna offers, then opt for
either the conventional or the special design types. On the other
hand, if you prefer general operating on all bands, then simply put
Chapter 39
up a doublet and feed it with parallel feedline and an ATU. The 121'
of the conventional trap antenna would translate into a good
doublet at 80 meters and up. Even the 83' length of the W8NX
antenna--which is short by G5RV standards--when used as a
doublet without traps, will still give performance every bit as good
as any trap antenna and on more bands. The length of a doublet is
not critical, but a. try to make it at least close to 3/8 wavelengths
long on the lowest frequency needed and b. be ready to change
parallel feedline lengths in case you run into the occasional
impedance condition your tuner cannot handle well.
Remember that there is no magic to any kind of trap or doublet

antenna. For the band in use, the elevation angle of maximum
radiation will be the same as a dipole at the same height above
ground. Therefore, more height is always a key to improved
performance of any trap or doublet antenna.
My object is not to downgrade traps: use them where your system's

specifications demand them. But do not neglect the multiband
doublet, which can be just as good and occasionally better for
many installations. Hopefully, setting the two side-by-side here will
let you make a more reasoned decision for your installation.
Chapter 39
Chapter 40: Some Notes on Triangles of Doublets
O
n occasion, folks have asked me questions of this sort:
My all-band doublet doesn't seem to be doing the job. So
what if I replace it with a longer one, a loop, or whatever? Or
suppose that I add an antenna of a different kind to the existing
doublet? What should I do?
In many cases, I recommend a second or third antenna of the same

type. It struck me that perhaps some short background on why I
make that recommendation on some occasions--but not all--might
be useful to those who do not have much experience with all-band
wire doublets.
The Single Doublet
Every wire antenna has a length (in feet, meters, etc.) and a height.
We shall use a center-fed doublet throughout these notes. Since
our question does not directly involve elevation angles of radiation
for the best DX, etc., I shall use a constant height of 50' above
average ground for this discussion. Those constants will allow us to
make direct comparisons. Anyone with lower or higher wires can
read other notes at this site to make any adjustments in the
comparisons.
A difference of height will make little or no difference in the azimuth

patterns of a doublet as we move from band-to-band (as long as
the antenna is not too close to the ground). The pattern is mostly a
function of the antenna length in terms of wavelength at each
Chapter 40
operating frequency. Small changes of frequency do not materially

affect the pattern, so we can use a single frequency on each band
as a sample that holds true for the whole band.
Let's start with a 135' center-fed doublet. 135' is about (and "about"
is plenty good enough here) 1/2 wavelength long on the 80-75-
meter band. At 40 meters it is 1 wavelength. On 20, it is 2
wavelengths, and on 10 it is 4 wavelengths.
Any center-fed doublet will have only 2 lobes for any length in
wavelengths up to and just beyond 1 wavelength--1 lobe on each
side of the wire, broadside to the wire. When the antenna is 2
wavelengths long, there will be 4 lobes--2 on each side of the
wires--and they will angle away from the wire leaving a null directly
broadside to the wire. A 4-wavelength antenna will have 8 lobes--4
on each side--and the strongest ones will be angled further away
from the broadside directions.
Chapter 40
Fig. 1 provides snapshots of the azimuth patterns that reflect the

notes I just gave. In addition, it shows the 15-meter situation. The
antenna is 3 wavelengths long, so we get 6 lobes. For all of the
patterns, the antenna runs from left to right (or right to left) across
the center line of the plots. So broadside is up and down on the
patterns.
Chapter 40
The first thing we can do with these patterns is explain why a given
doublet gives good results in certain directions on some bands but
not on others. Assume that we set the wire in the U.S. so that
broadside goes to Europe and to Australia. By the time we operate
on 15 to 10 meters, our strongest lobes are no where near the
headings for those two major target areas.
So our first lesson is in wire antenna orientation. By knowing the

antenna length, we can roughly determine the directions of lobes
on our favorite bands and set up the wire to give the strongest
performance in those directions by how we orient the wire. We
cannot obtain a perfect setting on all bands, but we can (assuming
that we do not need to move any supporting trees) obtain good
settings on our favorite bands.
Before we look further at the matter of direction, we have a few

more preliminaries to note. For example, what band should I
choose as the basic one for DX in my favorite directions? Here is
where a small performance table may help for the bands illustrated
in Fig. 1.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
135' Doublet Performance at 50' Up
Freq. MHz Max Gain dBi TO Angle deg

3.75 3.2 30 (arbitrary)
14.175 9.4 19
21.225 9.8 13
28.5 10.6 9
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Chapter 40
The TO or take-off angle is the elevation angle of maximum

radiation. It is correct for 20-10 meters, but is not for 80 meters,
where the height of the antenna is low enough to direct most
radiation at very high angles. So I chose a reasonable short-skip
angle to take the gain reading.
Since the gain on 80 is so low and the radiation angle so high, 80 is

not a good candidate for use in orienting the antenna for DX
directions. The gain on 20, 15, and 10 are similar, so one of those
bands is a better candidate. However, if we prefer local rag
chewing, then the 80-meter broadside should aim at our target
areas. Remember that the 40-meter pattern will also be a 2-lobe
broadside affair, but the strength will be higher and the beamwidth
narrower (since the added gain has to come from somewhere).
There are, of course, a number of other amateur bands on which

the antenna is not close to an integral number of wavelengths long.
What happens when the antenna is X.5 wavelengths long, where X
is any integer? The answer is a function of how lobes appear. They
do not pop into existence, but grow and shrink as we change the
length of the antenna (or as we raise and lower the operating
frequency, which achieves the same change in antenna length
when measured in terms of wavelengths). At X.5-wavelengths, the
azimuth pattern will show two sets of lobes in approximately equal
strength: the set for wavelength X, longer than which we now are,
and the set for wavelength X+1, which we are approaching.
On 17 meters, a wavelength is about 54' long and our 135' doublet

is about 2.5 wavelengths long (give or take a little). (A wavelength
Chapter 40
is about 984/f feet long, where f is the frequency in MHz.) We shall

have 4 lobes for the 2-wavelengths that we passed in length. We
shall have 6 lobes for the 3-wavelengths that we are approaching.
So the pattern will be composed of 10 lobes total. That sounds
good, since we get lobes in so many different directions. However,
for every lobe, there is also a null. So we have 10 blank directions
and each lobe is narrower than its counterpart in a 2-wavelength or
a 3-wavelength wire.
This note on nulls gives us the second lesson concerning wire

doublets and their azimuth patterns. Study the azimuth patterns
twice. First, look at the lobes that tell you where the radiation is
going. Second study the nulls that tell you where performance will
be very weak. Only then should you make decisions about how to
orient the antenna.
What happens when a wire is X.25 or X.75 wavelengths long. Since

lobes grow and shrink, the answer is almost obvious. At X.25
wavelengths, the wire is not long enough for strong lobes from the
X+1 length, but will show some smaller lobes derived from that
length. At X.75 wavelengths, the wire is too long to support full
scale lobes from the X length, but will have smaller lobes (meaning
lower gain) still present. In either case, as some lobes show low
development, the larger ones are that much larger, since the
radiated power is relatively constant (ignoring external variables)
across all frequencies.
Now suppose that you have studied the azimuth patterns and
decide that you cannot place a strong lobe every where that you
Chapter 40
want to communicate. It is precisely here that folks immediate jump

to thoughts of other antennas. But virtually every other horizontal
wire antenna, whether straight or looped, has a pattern of lobes and
nulls. Most of them are more complex to install. So what is a
solution to our quandary?
Now wire is relatively cheap. Antenna supports are not. So any

solution that we come up with should have the desirable property of
involving a minimum of new supports.
The 135' Doublet Triangle
There is no perfect solution to the problem of working everywhere

we want to work with an array of horizontal wire doublets. However,
we can go a good distance toward that solution by adding only one
more antenna support. We shall create a triangle of wire doublets,
something like the idealized sketch in Fig. 2.
Chapter 40
The drawing shows an equilateral triangle, but almost any shape

will do. In fact, a better idea is to angle each antenna so that its
lobes--on your favorite bands--hit your favorite targets. Just give
yourself a little separation at the wire ends--perhaps 10' on the end
of a 135' doublet--to minimize interactions among the wires.
The sketch also shows two different means of feeding the antenna.
You can run 3 feedlines, each the same length as the others, to a
central point where you install a weather-proof relay box to switch
among the wires. A single feedline runs to the shack, along with a
Chapter 40
relay power line and switching lines via an A-B-C switch to activate
one of the 3 antennas at a time.
The alternative is to run three lines, again, all the same length, to
the shack for use with a manual A-B-C switch at that location. Since
the lines will be parallel transmission lines, follow the usual
precautions about keeping them free and clear of anything that
might disrupt their balance.
The point of using equal-length lines in each case is so that you

can switch between antennas and determine by ear the strongest
signal. If they are not the same length, you will have to do some
rapid re-adjustment of the tuner settings for each switch position.
With identical antennas and feedlines, you should be able to pick
out the strongest one and then only do a final tweaking of the
settings on the tuner.
Chapter 40
As Fig. 3 shows, you may not always have a good choice as to

exactly where you position the supports for the antenna. Fitting the
triangle to available yard space is an eternal amateur antenna
problem. However, as we shall see, we can effect some
improvement on our operation, even if we cannot perfect it.
Notice that we were able to add two antennas for the cost of wire
and with only one extra support. We may well trade any remaining
imperfections in the system for that major simplification of structure.
Now the question is simply this: what do we get for our pains? The
easiest way to show what we get is by overlaying azimuth patterns
for each antenna in one massive plot for each band covered in Fig.
1. We shall note both the advantages that we accrue from the new
arrangement and the remaining problems that we could not solve.
Chapter 40
Fig. 4 shows the three antenna patterns at the same 30-degree

angle that we used in Fig. 1. It is immediately apparent that we can
cover more of the horizon with our signal (and reception) than with
a single antenna. However, since the antenna is close to 1/2
wavelength, there is some interaction between the antennas so that
the inactive ones act as reflectors. The 2-dB difference should not
affect performance too much. You will find the same phenomenon
on 40-meters, where the wires are all nearly 1 wavelength.
Chapter 40
The 20-meter situation appears in Fig. 5. We still have some nulls,

but count the major directions that we can cover, letting overlapping
lobes count as 1. We have 6 directions, not just the 4 that a single
wire would give us. As well, the overlaps are not perfect, so that a
signal that is on the fringe of one lobe may be centered in
overlapping one. Of course, by carefully planning of your triangle,
you can minimize the overlap and spread the area of coverage. The
patterns shown simply use our equilateral triangle as their basis.
Chapter 40
In Fig. 6, we find 15-meter quite well covered by the 3 antennas

and their strongest lobes. Indeed, 15 meters is a band that really
benefits from a triangle of 135' doublets.
Chapter 40
If you look at Fig. 5 and Fig. 7, you may get the impression that
when an antenna is an even number of wavelengths, it leaves more
nulls than when it is an odd number of wavelengths, as in the 15-
meter case. In general, this is a correct conclusion, although as we
further increase the antenna length, the number of lobes becomes
high enough to make it difficult to tell the difference. 10-meters is a
good band for which to redesign the triangle to place a lobe in the
direction that you want it.
Chapter 40
The 88' Doublet Triangle
If you lack yard space for 135' doublets in a triangle, you might try
88' doublets. Here is a performance table for a single doublet on
the same bands that we surveyed for the longer doublet.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88' Doublet Performance at 50' Up
Freq. MHz Max Gain dBi TO Angle deg

3.75 2.8 30 (arbitrary)
14.175 10.1 19
21.225 9.1 13
28.5 10.3 9
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The performance figures for each band are not very much different
than for the 135' doublet, but the patterns are considerably
different. Remember that the 88' doublet is only about 1/3
wavelength at 3.75 MHz. Fig. 8 shows the 88' doublet azimuth
patterns when we place the antenna 50' above ground.
Chapter 40
The 80-meter pattern is similar to the one for the longer wire.
However, the 20-meter pattern shows the typical "ears" of an
extended double Zepp, since that is exactly what the antenna is at
14 MHz. It is 1.25 wavelengths, which means that the 2-wavelength
lobes are just beginning to emerge. On 15 meters, the wire is 2
wavelengths long and shows the same sort of pattern that the 135'
Chapter 40
doublet showed on 20 meters. The 10-meter 88' doublet pattern is

an example of a 10-lobe pattern for a 2.5-wavelength antenna.
Besides taking less space, the triangle of 88' doublets also shows
less interaction among the wires. Hence, we can use somewhat
smaller separations of the wire ends in making the triangle.
However, in exchange for spatial economy, we shall encounter
differences in the ability of the triangle to fill in the nulls on the
single-wire patterns.
Chapter 40
On 80 meters, as shown in Fig. 9, the absence of pattern

distortions created by interactions among the wires yields almost
complete horizon coverage. However, remember that this pattern is
at an elevation angle of 30 degrees, and most radiation is upward.
We can improve long-haul performance of the triangle by "merely"
raising the supports to the 90-100 foot level.
Fig. 10 shows us the 20-meter combined patterns. The azimuth

plot provides a good model for a hex symbol to embroider for good
Chapter 40
luck. More importantly, it shows some nulls that may call for careful
design of the triangle to ensure the desired target-area coverage.
Like its counterpart 20-meter pattern, the 15-meter patterns in Fig.

11 add up to fairly complete coverage, but with nulls and
overlapping lobes. Hence, one might wish to design the triangle to
spread the lobes a bit. However, you will discover that with 4 lobes
per wire, every spread in one direction increases an overlap
somewhere else.
Chapter 40
The situation grows both better and worse on 10 meters, as shown

in Fig. 12. The null areas are wider, but often not as deep, since
minor 10-meter lobes fill the null at about 1.5 S-units lower strength.
Once more, designer triangle formation seems the order of the day.
It would be difficult in a general discussion to provide samples of

designer triangles, since each would prove useful for only one
region of the U.S.--and likely be useless outside the U.S. However,
one can experiment most easily with altering the triangle
Chapter 40
orientations with modeling software. There are inexpensive

packages, and even some free MININEC programs. It may pay to
master them well enough to go with your self-study geography
lessons in order to give yourself the bast chance of placing doublet
supports at the correct locations.
In the end, there are no perfect solutions. However, at a cost of one

extra support and a bale of wire--plus feedlines and an A-B-C
switch--you cannot get much better horizontal coverage much more
cheaply than with a triangle of doublets.
An additional caution or two: The doublet lengths used here place

the lowest band at 80 meters. You can use a 70' doublet if you wish
to cover 40-10 meters. In that case, the 80-meter and the 20-meter
patterns shown in Fig. 1 become the patterns for 40 and 10 meters,
respectively. As well, you can use a 44' doublet in place of the 88'
version used in these notes. The same adjustments then apply to
the patterns in Fig. 8. In both cases, the performance data in the
tables would apply with band adjustments to an antenna at about
25' above ground.
Xs and Ls
Some folks ask what happens when we have only two doublets.
They envision crossing or end-to-end arrangements like those in
Fig. 13. However, they often have in mind to use something other
than a right angle.
Chapter 40
The first thing that happens is a requirement for either 1 or 2 more

support posts. If you have plenty of Douglas Firs handy, a 4-post
system is no problem. But if you have to construct or erect your
own supports, then the support work either matches or exceeds the
work required by a triangle. The L-configuration is like the triangle,
but only lacks 1 wire and its associate feedline. Of course, you can
now get away with an A-B switch, rather than having to figure out
how to make an A-B-C switch.
Let's put up 2 135' doublets that cross at the center (with a

separation to keep the wires apart) and see what we can achieve
with only 2 antennas on each of our sampling bands.
Chapter 40
Fig. 14 shows the 3.75-MHz results. Since the deepest null is now
only about -3 dB or about 1/2 S-unit, it is likely that performance will
be satisfactory within the height limitations that we discussed
earlier.
Chapter 40
On 20 meters, as shown in Fig. 15, we have deep nulls and

considerable lobe overlapping. Hence, for this band, adjusting one
antenna by at least 25-30 degrees off a right angle will likely
produce better coverage. For an individual antenna, the lobes are
only about 35 degrees each side of a broadside tangent line
relative to the wire, so some angling to enhance coverage seems in
order.
Chapter 40
Fig. 16 gives us the 15-meter story. Once more, the main lobes
heavily overlap, but each is about 40 degrees off the line tangential
to the antenna wire. Finding a compromise angle for both 20 and
15 meters will require some thought, especially when we add in the
need to be aiming at communications target areas.
Chapter 40
The most thorough coverage occurs on 10 meters, due to the

multiplicity of lobes. See Fig. 17. Despite the gaps or nulls that
remain on each of the bands, the level of coverage with just two
135' doublets is significantly greater than with a single doublet.
Conclusion: if you cannot swing 3 antennas, at least try for two.
The 88' doublet does not fare quite as well in a 2-wire system as
the 135' doublet. Indeed, the 88' doublet seems best suited to a
triangular environment.
Chapter 40
80 meters appears in Fig. 18. As is evident, there is no significant

difference between the 135' and 88' 80-meter situation with respect
to coverage.
Chapter 40
The 20-meter patterns are in Fig. 19. The extended double Zepp
patterns simply give us two different bi-directional options. Hence,
careful broadside aiming of the wires seems the order of the day.
Chapter 40
If we wish to target areas on 20 meters, they will be broadside to

the antenna. However, on 15 meters, as shown in Fig. 20, we
cannot target the same areas, since the lobes on that band angle
away from broadside by about 35 degrees. Unlike the situation with
a triangle of doublets, the 2-wire system of 88' doublets appears to
force us to declare that either 20 or 15 meters is our favorite band,
but not both--at least not into the same parts of the world.
Chapter 40
Fig. 21 gives us the 10-meter picture. At 2.5 wavelengths, the

antenna yields fairly solid coverage, although there likely is room
for wire aiming on this band.
However we construct the doublets, the triangle provides superior

coverage and more versatility than a crossed or L-ed doublet
system. Obviously, 2 wires are better than 1, but 3 is significantly
better than 2 without requiring any further supports.
Chapter 40
I have omitted construction details, since they are so variable with

the circumstances of the individual builder. As well, I have omitted
inverted Vees, which will tend to change the situation on the lower
bands more than on the higher, since the patterns will be broader
ovals.
Nevertheless, the switched doublet triangle offers flexibility that

takes advantage of the strongest lobes--especially on the upper
bands--in the antenna pattern. A switched-doublet system is
cheaper than a rotator and tower system, and repair costs are
reduced usually to the cost of antenna wire and possibly some
parallel transmission line. However, the performance of a wire--
when you can place one of its main lobes on the desired station--
can be surprisingly good.
Nothing here is intended to compare the doublet with other wire

antennas--or even non-wire antennas. These notes are intended
only to show some possibilities that we often overlook when
thinking about wire doublets.
A love triangle is usually a disaster. However, a triangle of doublets

is often a happy marriage of economy and performance.
Chapter 40
Chapter 41: 40-Meter Vertical Arrays
T
he lower HF amateur bands tend to feature vertical antennas.
We find a few horizontal beams, especially on 40 and 30
meters, but they tend to be very large, heavy, and expensive.
In contrast, vertical arrays make use of wire and usually use much
simpler construction. Hence, their maintenance requirements are
also simpler.
At 40 meters (and by extension and scaling 30 meters as well) the

vertical dipole becomes feasible, especially if we find a way to
shorten it somewhat without losing significant performance. From
the vertical dipole, we may create a large number of array types. In
these notes, I want to examine in order of increasing performance
capabilities a collection of vertical dipole beams and arrays. So that
all comparisons will be fair, every vertical array will use AWG #12
copper wire, perhaps the most common material for amateur wire
antennas. As well, all antennas will be over average soil
(conductivity 0.005 S/m, relative permittivity 13). The performance
of vertical antennas will change with the quality of the soil beneath
them and in the far-field reflection zone, but the changes will tend to
be consistent within any given soil type. Therefore, if we know the
performance of an array over average soil and the relative
performance for a simple vertical dipole over both average soil and
the specific soil type for a given installation, we may extrapolate the
performance values for a more complex antenna system.
Chapter 41
Vertical Dipole Basics
The root antenna for what follows might seem to be a vertical

dipole. Table 1 and Fig. 1 summarize the properties of a vertical
dipole that extends from 1' above ground level to a top height of
67.6'. The total wire length is 66.6', but this exact value applies only
to a resonant vertical dipole over average soil. Slight adjustments
are likely for different soil types, especially since the lower end of
the antenna is so close to the ground. As will be the case for all
models used in these comparisons, the performance and
dimension values do not account for the influence of objects in the
immediate area of the antenna installation, that is, the so-called
ground clutter. All vertical antennas require as much clearance from
ground clutter as the installation site will permit.
Chapter 41
For this and all following antennas, the test frequency will be 7.15
MHz. The clutterless radiation patterns show very normal
characteristics--a circular azimuth pattern and a single elevation
lobe with a low take-off (TO) angle. Hence, despite the low gain of
the antenna, it finds general favor for its coverage and for its
insensitivity to high-angle noise and signals. The simple outline
sketch shows the current magnitude distribution along the antenna
wire. One reason why the antenna patterns show a low TO angle
stems from the relatively high position of the feedpoint or the region
of highest current magnitude, just about 1/4-wavelength above
Chapter 41
ground. However, the close proximity to the ground at the lower

end of the dipole does elevate its feedpoint impedance--from an
expected 70-Ohm value up to the 98-Ohm value reported by NEC-4
for the sample antenna.
Many amateurs do not have 70' supports for a 40-meter vertical

dipole in its simplest form. Individual circumstances vary, but let's
suppose that the maximum support height is about 50'. Within this
height restriction, we may still install a modified vertical dipole.
Rather than accept the losses that center-loading or mid-element
loading might create, we shall use end hats. An end hat or cap is a
symmetrical structure at right angles to the main plane of the
antenna. Radiation from the wires of the end cap largely self-
cancels, leaving us with a vertical antenna in terms of the radiation
patterns. Since the self-canceling portion of the radiation occurs in
the low-current regions of the antenna, we preserve most of the
performance that we might obtain from a full-length dipole. Fig. 2
shows the outlines and current distribution on one version of such a
T-capped dipole/
Chapter 41
The end hats on a vertical dipole can use any symmetrical

arrangement. The more radial arms that we create, the shorter
each one must be for a given length of vertical wire in the center.
However, many-spoked hats set up very significant support
requirements. The T-cap requires only two wires at each end of the
dipole, and we may run these wires along the non-conductive ropes
that we often use to support wire vertical dipoles between two posts
or trees. Reducing the number of hat wires to only two per end
does not affect the performance relative to using greater numbers
of hat wires. For example, the azimuth pattern of the T-cap dipole in
Chapter 41
the sketch shows only 0.03-dB of gain variation as we check all 360
degrees of the horizon.
The T-cap dipole that Fig. 2 shows is only 35.5' long, stretched
from 5' to 40.5' above ground. (One might easily raise the antenna
by another 4 to 5 feet and still remain below the 50' ceiling that we
set. The added height would also increase safety by raising the
base wires with their high RF voltages above the level that family,
friends, or even pets might touch.) The vertical section, then, is just
over 1/4-wavelength. For all following 40-meter vertical antennas,
we shall use the vertical section of the T-cap dipole. The two
horizontal wires at the top and bottom of the antenna, various
called arms or legs, are each exactly 10' long in the sample. Hence,
the total width of the antenna is 20'. This width falls well within the
clear area that we should have for any type of vertical dipole.
If we need to make adjustments for obtaining resonance in

subsequent arrays that use the T-cap dipole, we shall adjust the
length of the legs of the T, thereby leaving the center vertical
section intact. The dimensions for the basic T-cap dipole appear in
Table 2, along with some basic performance data. Within that data,
only the first section is immediately relevant for comparison with the
full-length dipole. First, note the 22-degree TO angle. This value is
higher than the value for the full-length dipole, but the T-cap
vertical's feedpoint is considerably lower (about 0.16-wavelength
above ground). Partly as a function of the lower feedpoint height
and partly as a function of the vertical-element shortening, the gain
of our T-cap dipole is about 0.4-dB lower than the gain of the full-
length dipole. Yet the T-cap dipole, as set up in the sample model,
Chapter 41
is nearly 30' lower in overall height, yielding what is for most

amateur operators a much more manageable construction and
maintenance situation.
In the table, we also find some performance values related to

improvements that we may make in the local ground immediately
beneath the vertical dipole. When we think of local ground
improvements together with virtually any vertical antenna, most
amateurs immediately think of radials. However, unlike the radials
of a monopole, the radials beneath a vertical dipole (either full
length or shortened) perform no antenna-completing function.
Chapter 41
Rather, they simply function to raise the conductivity of the soil

immediately beneath the antenna. In cases 2 through 4 in the table,
I created radial systems with the hub directly below the antenna.
The 1/4-wavelength radials use AWG #12 wire buried 1' deep in the
average ground. As the table shows, adding 4 radials amounts to
wholly wasted effort, since the gain increase is only 0.04 dB. 16
radials provide a 0.3-dB gain increase, perhaps a marginal amount,
considering the work involved. If radial installation is easy, we may
increase the field to 64 radials and obtain nearly 0.7 dB gain
increase. Fig. 3 provides a view of the 3 radial field along with the
T-cap dipole above them. You may estimate the installation work
from the sketches.
Note that none of the radial fields changes the TO angle of the
elevation pattern. The TO angle is mostly a function of the far-field
reflection zone, with is mostly well outside the radial limits. For the
exercises, I did not change the antenna dimensions. Therefore, the
Chapter 41
major influence of the fields appears in the feedpoint impedance

listings. As the size of the radial field increases, the resistive
component of the impedance decreases and the reactance
becomes more inductive. However, none of the changes in cases 2
through 4 present any operational concerns.
The last case presents an alternative method of local ground

improvement. Instead of using a radial system, I modeled a wire-
grid system below ground to simulate laying a screen of some sort
below the antenna. The screen is square and 70' on a side. Fig. 4
overlays the 16-radial system on the screen to show the change in
ground coverage in the screen corner region. (I used the 16-radials
so as not to obscure the grid. However, the most relevant
comparison would be between the screen and the 64-radial
system.)
Chapter 41
With the screen in place, we obtain almost 1.1 dB gain

improvement over untreated soil, and about 0.4-dB improvement
over the 64-radial system. The cost of the gain is a 1-degree
increase in the TO angle and further drift in the feedpoint
impedance. For antennas--like vertical dipoles--that do not require
radials to function as the lower half of a dipole, screens may
sometimes be the easiest and most effective means on improving
soil quality immediately below the antenna. However, they do not
provide that same benefits as living over very good soil, since the
treatment does not also apply to the far-field reflection zone. For
example, local area treatment may reduce ground losses below the
antenna, but that results in a higher TO angle, because the
improvement increases local reflection almost straight upward. The
lower TO angles that we often associate with the same antenna
Chapter 41
over very good soil results from improved soil conductivity at

considerable distances from the antenna.
In the end, soil improvement is a matter for the antenna user to

decide after measuring the anticipated performance improvements
against the amount and cost of work involved in local ground
treatment. For the remainder of these notes, we shall use untreated
soil beneath the antenna, so that the performance values in case 1
in Table 2 become the reference points for all that follows.
Before we leave the basic T-cap vertical dipole, we should

introduce one more set of considerations. We are interested not
only in the performance of any antenna at the test frequency, but
also across the entire band. 40 meters is a wider (but not the
widest) amateur allocation, with a 4.2% bandwidth. Some antennas
will handle the entire band; others will not. A basic dipole--either
full-length or T-capped--shows only a show change of gain across
the band. The case-1 T-cap dipole, for example, changes gain by
only 0.07 dB from 7.0 to 7.3 MHz. Note that our performance
concerns include not only the SWR properties, shown in Fig. 5, but
also such matters of forward gain and front-to-back ratio, both of
which will become important as we add directionality to the basic
dipole.
Chapter 41
Although the resonant impedance of the basic T-cap dipole is in the

vicinity of 75 Ohms, even the 50-Ohm SWR is below 2:1 from one
band edge to the other. Hence, we may feed the dipole using either
50-Ohm or 70-Ohm coaxial cable. As usual, one should route the
cable at right angles to the vertical dipole for as far as possible,
providing supports so that the cable weight does not unduly stress
the wire antenna. As well, I recommend the use of 2 common-mode
current attenuators, one at the feedpoint to also serve as a balun,
and the other at the entry to the operating building to attenuate any
currents induced along the cable run.
Chapter 41
A Bi-Directional Pair of T-Cap Dipoles
Because many of the arrays to follow will use a pair of T-cap

dipoles, we should spend a moment two see what happens when
we place a pair of these dipoles at a spacing of 1/2-wavelength
(about 68.8') and feed them in phase. Because the in-phase fed
pair of antennas will interact (or, otherwise put, exhibit mutual
coupling), I extended the T arms to 10.31' each. As a result, the
total array width, from outer T-tip to outer T-tip is about 89.4'. As
shown in Table 3 and the outline portion of Fig. 6, the array height
above ground has not changed relative to using a single T-cap
dipole.
Chapter 41
Perhaps the most surprising aspect of the array is the 4.1-dB gain
improvement over a single T-cap vertical dipole. Of course, the
Chapter 41
added gain comes at a cost in the available beamwidth, now down

to about 62 degrees. Nevertheless, the array is no taller and only a
bit wider than a half-square, but provides more gain. The gain that
is competitive with a bobtail curtain. In fact, the in-phase-fed pair of
vertical dipoles lies at the theoretical core of all SCV (self-contained
vertical) antennas, since the basic versions--ranging from deltas to
rectangles to half-squares, all place two elements in phase, but at a
spacing limited by the single-wire, single-feedpoint configuration.
The independent feedpoint impedances of the two T-cap dipoles is

about 56 Ohms. Therefore, we may equip the phased pair with a
common feedpoint by using equal lengths of 70-Ohm feedline to a
center point. Since the physical distance between each element's
feedpoint and the center point is 1/4-wavelength, and since all 70-
Ohm transmission lines have a velocity factor (VF) of well under
1.0, you may need to use 3/4-wavelength section to arrive at the
required 100-Ohms at the junction, a value that becomes a good
50-Ohm match in a parallel connection. The lower portion of Fig. 6
provides the modeled 50-Ohm SWR curve. The array covers the
entire 40-meter band with an SWR value that is less than 2:1.
We shall not linger over the in-phase-fed pair of T-cap dipoles and
their broadside bi-directional pattern. However, we shall occasion to
mention them once more before we close our screening survey.
Nevertheless, we should call to attention one more time the gain
value produced by the pair of elements. The maximum gain in each
direction will exceed the forward gain value that we may achieve
from some of the more basic beams that we consider when thinking
about a pair of T-cap dipoles and their best use.
Chapter 41
Parasitic Driver-Reflector Beams Using T-Cap Dipoles
For many operating needs, a front-to-back ratio of at least 10 dB

may be more important than additions to the arrays forward gain. In
such cases, we may create an endfire array using standard 2-
element Yagi principles. In a Yagi with full-length elements, the
reflector is normally longer and the driver normally shorter than a
freestanding resonant dipole. In creating the driver for a useful
Yagi-type beam, we may simply reduce the T-legs to an individual
length of 9.6' (for a total element width of 19.2').
However, we shall not stop to develop a reflector for the T-cap

array that uses a larger element. Instead, we shall simply load the
reflector and use the same dimensions as for the driver. Reflector
loading of a 2-element parasitic array normally does not reduce the
forward gain, and it may actually improve the front-to-back
performance over a full size reflector. Table 4 supplies the
dimensions of the array, which spaces the elements 21' apart. Fig.
7 shows the outline and the radiation patterns that we might expect.
Chapter 41
The azimuth plot overlays two patterns, one in each direction. One
of the key advantages of loading the reflector is that--with very little
Chapter 41
effort--we can reverse the beam's direction. The reflector called for
a 55-Ohm load to achieve the desired pattern. Instead of using an
inductor as the loading element, we may use a shorted length of
feedline, in this case, 50-Ohm cable to match the feedpoint
impedance of the driver element. The electrical length would be
about 18.24', but the physical length will be the electrical length
times the cable's velocity factor. Even a solid dielectric cable will
yield a physical length of at least 12', so that cable running from
each element can meet in the center of the 21' element spacing.
Then, with a suitable remote switch, we can short one cable to
make the inductively reactive load for the reflector and connect the
other line to the main feedline. With a flip of the switch, we have
reversed the beam's direction.
When we first encounter vertically polarized parasitic arrays, we are

sometimes surprised by the facts that their gain is lower and their
beamwidth is much greater than the patterns that we find in
horizontal arrays using the same number of elements. The beam's
gain value is about 3.3 dB higher than for a single T-cap dipole, but
less than the bi-directional gain of the phase-fed pair of verticals.
The beamwidth is nearly 140 degrees, providing very good
coverage in each of the beams two possible directions. Although
ground losses have a role to play in setting the gain of all vertical
antenna types that are close to the ground, the vertical array's gain
would not catch up to the gain of a horizontal counterpart until we
reached a height above 20 wavelengths. The key factor in the gain
and beamwidth difference that we get from rotating the beam 90
degrees along its real or virtual boom is a function of geometry. The
arrangement of element tips restricts the beamwidth in the E-plane,
Chapter 41
that is, in the plane of the elements. However, the H-plane has no
such restrictive influence. Indeed, in free-space, H-plane patterns of
a 2-element parasitic array look very much like the pattern of a
single vertical dipole but displaced in the direction of the forward
gain. A Yagi must have many elements (and be very long) before
the H-plane beamwidth narrows significantly.
Nevertheless, the reversible 2-element vertical array with T-cap

elements provides good service across the 40-meter band. As the
50-Ohm SWR curve at the bottom of Fig. 7 shows, the array will
cover most of the band with less than 2:1 SWR. However, we early
on noted that our bandwidth concerns covered more than just the
SWR. We are also interested in how well the antenna performs in
terms of gain and front-to-back ratio. Fig. 8 provides a partial
answer to our questions.
Chapter 41
The gain curve in the figure shows a 0.4-dB range of forward gain
value across the band, a value that we may consider fairly stable
for such a wide amateur band using elements composed of
relatively thin wire. In contrast, the front-to-back ratio remains
above 10 dB only for a small portion of the band. However, it never
drops below 5 dB. In fact, with careful adjustment of the reflector
load, we may be able to center better the front-to-back curve for
relatively similar performance at the band edges. A slightly higher
load reactance--meaning a slightly longer length for the shorted
stub--would likely do the job. Alternatively, we may favor either the
CW-digital end of the band or the phone end of the band, according
to our operating needs.
Chapter 41
One limitation of the 2-element driver-reflector array is that we

leave some portions of the horizon only weakly covered. It might be
useful if we could cover the entire horizon with less than a 2-dB
drop in gain around the 360-degree span--and at the same time
maintain at least the forward gain that we obtained from the
reversible 2-element array. To achieve that goal, we may think in
triangles.
Suppose that we set up an equilateral triangle of T-cap vertical

dipoles. We may designate the corner containing the driven
element as the apex. The remaining corner will contain identical T-
cap dipoles, but loaded to form reflectors. Table 5 provides the
dimensions and test-frequency performance data, while Fig. 9
supplies a sketch and the radiation patterns.
Chapter 41
The T-legs are each 9.5' long for the individual dipoles. Each
triangle side is about 30.6' long, yielding a distance between the
Chapter 41
apex and the center of the triangle base of 26.5'. It does not matter
how we orient the T-legs of the dipoles, so long as the legs form a
straight line for each dipole. In the triangle shown--and other
dimensions are also possible--each reflector dipole requires a j90-
Ohm load. We shall again use shorted transmission-line stubs as
the source of the required inductive reactance. Since the driver
impedance is close to 70 Ohms, we shall use 70-Ohm lines, which
require a 19.9' electrical length, with physical shortening that
depends on the VF of the line used in the assembly. We shall bring
the stubs to a center point within the triangle for switching. At any
time, one of the stubs will actually be an extension of the main
feedline to the driver, while the switch to form the required reflector
loads shorts the other two lines.
The system gain is about 3.2 dBi, about 0.5-dB higher than for a
standard 2-element parasitic array and about 3.7-dB higher than a
single T-cap vertical dipole. The 15-dB front-to-back ratio is also
several dB higher than we found for the 2-element parasitic beam.
With its higher gain, the triangle shows a beamwidth that is about 5
degrees narrower than the beamwidth of the simpler beam.
Perhaps the key advantage of the triangle is its ability to cover the
entire horizon with only a 2-dB gain deficit at the overlap points
between the forward lobes of the beam in each of its positions.
Although the switching may be more complex for the triangle than
for the reversible beam, the electronics are simple and cheaper
than a rotator.
The lower portion of Fig. 9 shows the 50-Ohm and 70-Ohm SWR
curves for the array. A 50-Ohm feedline from the switch junction to
Chapter 41
the equipment will be satisfactory if SWR values close to 2:1 are

satisfactory at the band edges. However, a 70-Ohm line will reduce
the maximum SWR value to about 1.5:1 at the band edges. With
respect to the gain and front-to-back performance across the 40-
meter band, Fig. 10 supplies the appropriate sweep data.
The performance bandwidth of the triangle is generally similar to

the performance bandwidth of the reversible array, but with a
smaller range of value change across the band. The gain varies by
only about 0.2 dB. The front-to-back ratio remains at 10 dB or more
for the entire band. The band-edge values may be equalized by
slight adjustments to the load values, that is, the length of the
shorted stubs. With the smaller range of performance change over
the 300 kHz of amateur allocation (in the U.S.), one may design a
Chapter 41
triangle for whole-band use and expect only small deficits at each
band edge.
A 2-Element Phased Beam Using T-Cap Dipoles
Obtaining the maximum possible front-to-back ratio from 2

elements is difficult if not impossible using parasitic techniques and
2 vertical elements that are close to the ground. However, it is
easily possible to improve the front-to-back performance and to
obtain a nearly cardioidal pattern by phasing 2 T-cap dipoles in an
endfire arrangement. In fact, we may do so using commonly
available feedline materials, although we may have to do some
experimentation to find the optimal cable for the task.
Experimenting with modeling software is more rapid and less costly
than experimenting with lengths of actual cable.
Table 6 profiles the phased array that uses two standard T-cap
dipoles with 9.6' T-legs and a spacing of 21' (the same spacing
used for the reversible parasitic array). Both dipoles use identical
construction. The phase line uses two separate lengths that reach a
junction where we shall connect the main feedline. The section to
elements 1 in Fig. 11 is 084' of RG62, 93-Ohm line with a VF of
0.84. Hence, the electrical length is 1'. The line does NOT undergo
a half twist or reversal. However, the line from the junction to the
rear element--from the same material--does undergo a reversal. It
is 21' physically or 25' electrically.
Chapter 41
Chapter 41
The forward gain of the array is comparable to the gain of the

reversible parasitic array. However, the front-to-back ratio climbs to
Chapter 41
over 26 dB at the design frequency. As well, the beamwidth

increases to about 145 degrees, providing wide coverage in this
essentially mono-directional array. The rearward quadrants of this
array promise to be exceptionally quiet.
The cost of this type of performance lies in the odd feedpoint

position before we add matching components. The low impedance
has a high inductive reactance. However, the values are nearly
optimal for a beta match, so long as we use an open or capacitively
reactive stub across the feedpoint terminals. 14.3' of 50-Ohm VF
0.78 cable or its equivalent provides the necessary shunt
component for the beta L-network that yields an impedance close
to 50 Ohms. With the beta component in place, the SWR curve in
Fig. 11 shows well under a 2:1 SWR value at the band edges.
Chapter 41
Fig. 12 provides sweep data for the phased array relative to

forward gain and front-to-back ratio. The gain rises steadily across
the band over a 0.4-dB range. The front-to-back ratio peaks near
the middle of the band and decreases to about 15 dB at the band
edges. Compared to parasitic arrays, the whole-band front-to-back
performance of the phased array rates very highly. Perhaps the
major drawback to phasing is that the array does not succumb
easily, if at all, to reversing direction.
Screen Reflectors and the T-Cap Driver
Screen or planar reflectors have been used since the 1920s in

major arrays. Indeed, in their early years, some called them
billboard reflectors. A solid surface properly sized will reflect radio
waves in ways that are analogous to the reflection of light from a
flat mirror. We find them in wide use in dipole arrays for short-wave
broadcasting and in a wide variety of UHF antennas. More recently,
I have recommended their use beneath a number of NVIS
antennas to increase gain, essentially by elevating the quality of the
reflective plane beneath the antenna. In fact, the wire-grid ground
improvement technique noted earlier represents a different
application of the same technology.
A wide range of experience shows that planar or screen reflectors

are most effective in improving directional gain when they exceed
the dimensions of the driver element by between 0.45 and 0.55
wavelength both horizontally and vertically. We can achieve this
goal horizontally with our T-cap vertical dipole driver, but any
screen reflector will be vertically challenged. The ground, of course,
Chapter 41
is one limit, preventing the reflector from extending below the driver
by the desirable amount. Vertically, we shall be limited as well by
some of the height restrictions that we set for this project. If we limit
ourselves to a 50' top height, the forward gain will be only a little
greater than the gain of the reversible beam. A top height of 100'
yields perhaps a half-dB more gain than a 70' height. Therefore, for
the comparisons that we shall show, the height of the screens will
run from just above ground level to 70'. One might hang such a
screen between two widely separated towers that support antennas
for higher amateur bands.
The optimum width turns out to be just about 140', which is 1

wavelength at the test frequency (7.15 MHz) and a half-wavelength
on each side of the driving dipole. Screen reflectors do not have to
be solid surfaces to act like solid surfaces in the HF range.
Chicken-wire fencing and open-weave materials will appear solid to
40-meter energy. The model for the screen-reflector array uses a
wire-grid, as shown in Fig. 15. The dimensions and performance
data appear in Table 7.
Chapter 41
The selected distance from the driver to the reflector is 35' for an
array using a single T-cap dipole driver and a 140' by 70' screen
Chapter 41
reflector. You may ground the screen bottom wires with no effect on
performance, but with a considerable affect on safety. The driver
uses 9.66' T-leg lengths for an 80-Ohm resonant impedance. With
a planar reflector, you may vary two items to set a feedpoint
impedance: the dipole dimensions and the spacing from the
reflector. In general, closer spacing produces lower feedpoint
impedance levels, but narrower operating bandwidths. Each space
adjustment will change the coupling between the reflector surface
and the driver, requiring adjustments to the T-leg lengths to return
to resonance.
The chief merits of the planar reflector array are forward gain and
operating bandwidth, when we compare the results to the reversible
parasitic array. The top height of the reflector limits the front-to-
back ratio to about 13 dB. A top height of 100' would have added
another dB to the ratio, while a more ideal (and unrealistic) height
of 140' would add a further dB or two. The array's forward gain is
close to 4.7 dBi, about 1.8 dB higher than the 2-element parasitic
array. The forward gain comes at the expense of the beamwidth,
which is down to 83 degrees, about 50 degrees narrower than the
beamwidth of the parasitic beam.
Despite the 80-Ohm resonant feedpoint impedance at the design

frequency, the 50-Ohm SWR curve in Fig. 13 reveals a very low
rate of impedance change across the 40-meter band. That curve
shows less than 2:1 SWR at the band edges. The lower 80-Ohm
curve shows less than 1.5:1 SWR at the band's upper and lower
limits. The remaining prime operating parameters are equally slow
to change, as shown in the sweep data in Fig. 14.
Chapter 41
The gain across 40 meters changes by only about 0.1 dB, while the
front-to-back ratio varies by just over 0.5 dB. It is possible to design
driver elements with an inherently wider bandwidth and to use the
array to cover both 40 and 30 meters with very little change in
performance.
One common method of constructing a screen is to use a series of

wires polarized as the driver element. To sample that option, I
reconstructed the screen reflector using 29 AWG #12 wires at 5.0'
intervals. The overall horizontal and vertical screen reflector
dimensions remained the same. Doubling the wire size (to AWG
#6) yielded no performance improvements. The only other change
relative to the wire-grid screen was a 3' increase in the spacing of
the driver from the reflector, as shown in the dimensions in Table 8.
Chapter 41
Chapter 41
As the data and the patterns in Fig. 15 show, the vertical-wire

reflector is not quite as effective as the wire-grid version. Gain
Chapter 41
drops by about a half dB, while the front-to-back ratio decreases by

well over 2 dB. Both decreases are indications that the wire version
of the screen requires a taller top height or is perhaps "leakier" than
the wire-grid. As well, the resonant impedance is 5 Ohms higher
and does not produce 50-Ohm SWR values under 2:1 without
further matching. However, the lower 75-Ohm SWR curve in the
graph does track the wire-grid's 80-Ohm SWR curve very well.
The sweep information in Fig. 16 shows the same broad curves for
both the forward gain and the front-to-back ratio. The gain range is
about 0.15 dB across the band, while the front-to-back ratio
changes by only 0.8 dB. Even with somewhat lesser performance
than the wire-grid reflector, the vertical-wire screen still enjoys a
Chapter 41
considerable advantage over a parasitic reflector with respect to

broadband characteristics.
To compensate even if only partially for ground losses, one may tilt
the reflector back as viewed from bottom to top. Using an 11' tilt
(35' at the bottom and 46' at the top for the distance between the
driver and the vertical-wire reflector), it is possible to add a few
tenths of a dB to the forward gain and a similar amount to the front-
to-back ratio. However, the exercise has its own consequences, as
revealed in the overlaid patterns of Fig. 17.
The peak gain lines both forward and rearward are virtually
indistinguishable in the patterns. However, note the increased high
angle radiation, especially in the rearward lobes. As well, radiation
directly overhead to the driver has also increased. In the end, a
vertically oriented screen appears to yield the best combination of
performance and patterns.
Chapter 41
Earlier in these notes, I promised a second look at the in-phase-fed

pair of T-cap dipoles spaced 1/2-wavelength apart. The dipole pair
produced a bi-directional pattern with a maximum gain that was
over 4 dB stronger than a single T-cap dipole. Since we can readily
feed the pair of dipoles from a single feedpoint, using a pair of
equal-length lines, we might wonder what would happen if we place
the broadside array in front of a screen reflector.
Ideally, the screen should be over 200' wide. Understanding that we

lose performance as we shrink a screen below its ideal proportions,
let's continue to use the smaller 70' by 140' wire-grid screen. With
the screen 40' behind the driver, the T-legs of each dipole are
10.16' long, longer than for a single dipole driver, but slightly shorter
than the T-legs on the bi-directional array. As shown by the
dimensions in Table 9 and the outline sketch in Fig. 18, we lose a
good bit of the horizontal screen extension that optimizes
performance. Nevertheless, the performance is noteworthy.
Chapter 41
Chapter 41
The 6.95-dBi forward gain value for the array may require some
perspective. First, the value is almost 2.3-dB higher than the gain
Chapter 41
we obtain from a single driver and the same reflector. A sign that
the screen is less than optimal in size comes from the fact that the
screenless bi-directional array produces a gain value that was 4 dB
greater than a single T-cap dipole. Nevertheless, the phased dipole
array and screen produce almost 7.5-dB gain relative to a single T-
cap omni-directional dipole. The forward beamwidth is down to 58
degrees, dictating careful aiming of the array.
The front-to-back ratio is not outstanding, at about 11.4 dB, another

sign of using a small screen reflector. Because the unmatched
feedpoint impedance of each dipole is in the vicinity of 90 Ohms,
we may run equal lengths of RG-62 (VF 0.84) to a center point
between the dipoles and obtain a net impedance of 48.9 Ohms. As
the 50-Ohm SWR curve shows, the array easily covers 40 meters
with a maximum SWR that is less than 1.5:1.
Chapter 41
As the sweep data in Fig. 19 reveals, the dual-dipole-driver array is

as stable across 40 meters as the other sample screen-reflector
array. Both the gain and the front-to-back ratio show changes of
just over 0.2 dB across the band.
The construction of a screen-reflector is the most difficult portion of

the overall project. Therefore, one may well wish to place drivers on
either side of the screen for reversible service using separate
feedlines all of the way to the equipment location. Whichever driver
is in use (whether single or double), the inactive one will remain
invisible due to the screen's reflection characteristics. Screen
reflector arrays may not be for everyone, but they may serve a few
40-meter operators.
Chapter 41
Conclusion
We have not only surveyed comparatively the performance of

vertical antennas and arrays on 40 meters, but we have added
screen reflector arrays to the list that we normally see. This last
group of arrays presents serious construction challenges, but offers
in return increased gain over the values that we may obtain from
parasitic arrays.
All of the driving elements in our comparisons have used AWG #12
copper wire in a T-cap arrangement that extends from 5' to 40.5'
above average ground. We changed the length of the T-legs to
obtain a resonant feedpoint impedance. The T-cap dipoles provide
a uniform element length to help validate the comparisons.
Although scarcely longer than 1/4-wavelength, the T-cap dipoles
lose very little performance relative to the reference full-length wire
dipole with which we began these notes.
Just because we cannot obtain the gain level of a phase-fed dipole

pair with a screen reflector does not relegate the intermediate
designs to uselessness. Indeed, many operators use T-cap and
similar 40-meter vertical antennas with surprisingly good results.
The parasitic arrays offer reversibility and even--with the triangle--
full horizontal coverage at the flip of a switch. For maximum front-
to-back ratio, the phased array is difficult to surpass, even though it
uses no complex networks to achieve a nearly cardioidal pattern.
By setting the entire range of vertical arrays on common ground at

a common frequency with common construction, you may gather a
Chapter 41
more precise sense of the benefits and costs as you increase the
complexity of a vertical array. Virtually all of the designs will scale
directly to 30 meters, although bandwidth there is not a major
concern beyond the realm of allowing rather casual construction
without incurring performance losses. Scaling to 80 meters is also
possible, although for most amateurs, the resulting element sizes
may prove to be prohibitive. A T-cap dipole of the present design
will extend from about 10' to about 81' at the top, depending on the
precise frequency to which one scales the design. The T-legs can
be trimmed for resonance if you stick with the AWG #12 wire rather
than doubling its diameter. If the amount of trimming is not too
great, adjusting only the lower T-legs will not disturb the centering
of the feedpoint to any significant degree. (The presence of ground
below the bottom of the antenna and essentially free space above
the top already disturbs the balance that we presume when we
physically center the feedpoint on a vertical dipole. Common-mode
current attenuation devices are necessary adjuncts to any of these
arrays.)
These notes have focused on performance comparisons and

provided very few construction notes. Building any of these arrays,
from the simplest to the most complex, will be an exercise in
making use of available supports and in using what nature or prior
antenna construction provides. In virtually all cases, the result will
be far less expensive than a strong tower able to hold the weight of
a horizontal beam that shows higher-angle elevation lobes and
questionable operating bandwidth.
Chapter 41
Chapter 42: Wire Elements in Lower HF Arrays
W
ire beams are commonplace in the lower HF region. They
have some uses at HF as well, for example, in LPDAs. In
both cases, we sometimes fall into the belief that the wire
array has all of the gain and performance of a comparatively similar
array made from fat tubular elements.
Of course, that belief is simply false. A reduction in the diameter of

elements of the proportions of a move from a fat tube to a skinny
wire reduces not only the operating bandwidth (sometimes), but as
well reduces the inter-element coupling that is critical to deriving full
performance from a beam. For example, standard LPDA
calculations make use of the element length-to-diameter ratio in
determining element lengths, but little has been noted about the
lowering of gain in moving from fat to skinny elements. Even a
beam like the LPDA, that is utterly dependent upon phasing
connections, remains equally dependent upon inter-element
(mutual) coupling, and that coupling decreases with decreases in
element diameter for a given Tau and Sigma.
In many cases, tubular elements cannot be used due to their length

and/or weight. Hence, the builder is forced to use wire. However,
he need not be confined necessarily to reduced performance. I
have shown in a couple of places a method for using wire elements
that retain the performance of fatter tubular elements.
Chapter 42
The technique is quite straight forward, as indicated by Fig. 1. For

each element of a design originally created for single fat elements,
create a double wire of the same length. Then, most likely through
modeling (although field testing will also work), adjust the wire
spacing so that the element is self-resonant on the same frequency
as the original tubing element. The required spacing will vary with
the wire size used. If a modeling approach is used to estimate the
spacing, there are a few constraints that we shall look at further on.
It turns out that this arrangement works well for smaller diameter
tubing equivalencies--say, up to 1 inch or 25 mm. For larger tubing,
we encounter some limitations. The average tubing size--even in
elements that start out at the element center as quite sizable--ends
Chapter 42
up quite modest even in lower HF beams. You can check out the
equivalent uniform diameter for almost any tapered-diameter
element on NEC-2 programs having Leeson corrections by looking
at the substitute elements used in the actual NEC calculations.
Access to these substitute elements is available in EZNEC and
similar programs.
However, we sometimes scale up array sizes from proven upper

HF designs. Proper scaling requires that we increase all
dimensions, including the element diameter, if we are to have what
amounts to a true scaling. And only by a true scaling can we assure
that the array will perform at the lower frequency to the level it
promises at the higher frequency.
Chapter 42
I had occasion to go through this exercise with a 3-element 20-

meter beam adapted from a K6STI design. The exercise was
initially theoretical, so I was not the least troubled by the resultant
4" diameter elements that emerged in the 80-meter model. Fig. 2
shows the basic outline of the Yagi, which will become the center of
attention in what follows. Here is a table of dimensions.
Chapter 42
3-Element Yagi: 3.6 MHz: 4" diameter elements

Element Element Length (ft) Spacing from Reflector (ft)
Reflector 136.08 ----
Driver 129.94 41.167
Director 122.26 88.767
Wondering what the antenna might do if made from wire, I

recreated it from 0.1" diameter wire. This diameter is between #12
and #10 AWG. The conversion required some proportional
lengthening of elements to achieve a coincidence of maximum
front-to-back ratio and driver resonance. The element spacing was
retained, and only the element lengths were changed. The resulting
wire beam had these modeling specifications.
3-Element Yagi: 3.6 MHz: 0.1" diameter elements
Driver 132.40 41.167
Director 124.60 88.767
Both antennas were modeled as copper elements in free space for

comparisons. For the design frequency of 3.6 MHz, the results
were interesting, to say the least.
NEC-4 Modeled Performance: 3-Element Yagis: 3.6 MHz
Model Gain F-B Ratio Feedpoint Z Efficiency
dBi dB R +/- jX Ohms %
4" 8.14 27.3 25.5 - j 0.9 99.81
0.1" 7.06 19.9 43.3 - j 2.9 95.87
Using wire reduced gain by over 1 dB and the front-to-back ratio by

about 7 dB. The feedpoint impedance increases nearly 70%, much
more than the increased wire losses would indicate. The 4%
differential in efficiency is not large enough to account for the
Chapter 42
source impedance increase. (NEC recognizes only wire material

losses and spot load and other network losses in calculating
efficiency. The latter will not be relevant to this exercise. Efficiency
figures can be useful reference points, so long as we do not try to
use them to account for every gain or loss in performance when
comparing designs.)
If we use a current source of 1.0 at a phase angle of 0.0 degrees,

we can gain further insight into the comparative performance of
these arrays. We simply need to look at the element center current
magnitudes and phase angles for the reflector and director
elements for both beams.
NEC-4 Modeled Element Currents: 3-Element Yagis: 3.6 MHz
Model Reflector I Director I
Magnitude Phase Magnitude Phase
4" 0.405 143.5 0.601 -134.0
0.1" 0.414 126.1 0.456 -119.2
Since the elements are both parasitic, the differential in current

magnitudes and phase angles is reflective of the difference in
mutual coupling between the elements and the driver. If the current
values for the 4" model are to be considered as more ideal, then
the values that appear on the elements of the 0.1" model are
considerably off target.
In fact, the wire version of the array might require considerable

redesign to peak its performance, including adjustments to element
spacing and length. This fact is revealed in frequency sweeps of
both antennas from 3.55 through 3.65 MHz.
Chapter 42
Fig. 3 shows the modeled free-space gain curves for both

antennas. The 4" model shows the gain curve across the modeled
passband that we have come to expect from high performance 3
element Yagis: a modest but continuous increase of gain with
frequency. In contrast, the wire model shows the lowest gain at the
design frequency. This dip is indicative antenna operation at a
Chapter 42
different portion of the potential gain curve--somewhat lower in

frequency than the corresponding curve positions for each of the 4"
model gain numbers.
Chapter 42
However, as shown in Fig. 4, the front-to-back peak has been

sustained at or just below the design frequency. In this aspect of
design, the 4" model curve encompasses the wire model curve.
The wire model actually has a wider operating VSWR bandwidth

than the 4" model, as shown in Fig. 5. The wider VSWR bandwidth
results in part from the higher resonant source resistance, so that
Chapter 42
equal amounts of reactance have a lesser effect on the SWR in

terms of increasing its value. In fact, the wire model reactance from
one passband limit to the other is only about 37% greater than that
for the 4" model (40.3 vs. 29.5 Ohms), while the source resistance
has climbed by 70 percent (43.3 vs. 25.5 Ohms). Hence, relative to
the individual resonant source resistances, the wire beam will
permit operation over more of 80 meters.
Despite the wider operating bandwidth, my aim was to see if I might

obtain 4" performance from a wire version of the antenna. So I
applied the two-wire technique described at the beginning of the
exercise, only to be disappointed by the results. 2-wire elements,
each the same length as those of the 4" model, but spaced to
resonate at the same self-resonant frequencies, only brought me
half way to the gain goal.
The 2-wire element substitutes required a spacing of 13" between

0.1" wires. Here we must note that my models will not use full
precision in the interests of keeping numbers as rounded and
simple as possible. So 13" spacing became convenient and close
to the mark. However, the rounding of the spacing value was not
sufficient to account for the gain value of 7.77 dBi, nearly 0.4 dB
short of the mark.
Chapter 42
The answer lies in the insufficient coupling provided by the 2-wire

model elements. So I added a third wire exactly between the 2, as
shown in Fig. 6. The half-elements shown are matched by
equivalent mirror images to the unseen right of the figure. The third
wire does not substantially change the resonant frequency of the
resulting element, so I left the outer spacing of 13" and hence
ended up with a spacing of 6.5" between wires.
Chapter 42
Before we look at the results of the wire-beam models in detail, we

can pause a moment to examine the element models, partially
shown in Fig. 7. The rules for NEC note that closely spaced wires
should have all segment junctions parallel to each other. As well,
angular junctions should have segment lengths of approximately
equal lengths. Finally, the source segment should be protected
from multiple wire junctions, which results in a 3-segment center
section for each element. (Although the parasitical elements might
have been made continuous, I preserved the center sections in
each that resulted from initial resonating tests.) Finally, results will
be most accurate for multi-wire elements where the wires are
closely spaced and parallel if the source wire is centered so that it
meets equal wire lengths in both components of the element. The
wires are joined at the outer ends.
Chapter 42
Since the segments length of the wire from the center or source
wire to the outer wires is 6.5" long, I made this value the segment
length for the entire array. The center wire is 19.5" long in 3
segments. The elements beyond the limits of the figure have over
100 segments per wire on each side of center. Hence, the 2-wire
model has over 1400 segment, while the 3-wire model tops 2100
segments. Although for some, this would be overkill, it meets all
NEC guidelines. Only a little patience is needed while NEC grinds
out the results during frequency sweeps.
Part of my interest in the wire models was to determine if each of

these models had a single-wire model of roughly corresponding
performance. That model would consist of a single fat wire per
element, with the diameter chosen to approximate the performance
of the corresponding wire model. Of course, the element lengths
had to be adjusted relative to those for the 4" model in order to
place peak performance at the design frequency.
1. The 2-Wire Model and a 1" Single Wire Model: A 1" diameter
element model yielded a set of performance curves roughly similar
to those of the 2-wire model. We know the physical dimensions of
the 2-wire model from the discussion above. The following table
presents the physical aspects of the 1" model.
3-Element Yagi: 3.6 MHz: 1" diameter elements
Driver 131.50 41.167
Director 123.76 88.767
Chapter 42
The design-frequency (3.6 MHz) performance for the two models is

in this table.
1" 7.77 25.7 32.7 - j 0.3 99.42
2x0.1" 7.77 23.7 30.4 + j 2.6 96.73
The comparable performance at the design frequency is readily

apparent. Note, however, that the efficiency of the single wire 1"
model is significantly higher than that of the 2-wire model, owing to
the higher material losses and smaller surface area available from
the pair of thinner wires. Indeed, the surface area ratio is about 5:1
in favor of the 1" model. Nevertheless, the increased mutual
coupling made possible by the 2 spaced wires is sufficient to
overcome the increased material losses and nets the same gain at
the design frequency as the more efficient fat-wire model.
The comparison can be extended to the current magnitude and

phase found on the parasitic elements with a source current of 1.0
at 0.0 degrees.
1" 0.407 134.9 0.556 -127.8
2x0.1" 0.391 136.7 0.583 -132.1
Note that the parasitic element current values of these two models
are closer to each other than either is to the standard 4" model. The
models might have more closely corresponded had the elements in
Chapter 42
the 2-wire model been spaced more precisely than the 13" used in
the model. In fact, the 2-Ohm reactance value indicates not only a
slight driver over-length for the spacing, but as well a similar
situation for the other elements as well. Hence, the best SWR value
and front-to-back peak occurs below the design frequency. In
addition, for an equal-gain situation, the required mutual coupling
among elements to overcome the higher material losses in the 2-
wire model would also dictate slightly different parasitic element
current values relative to the single wire model.
Chapter 42
The similarities and differences between the two models become

more apparent when we perform frequency sweeps. In this case,
due to the remnant inductive reactance of the 2-wire element
replacements, I have dropped the lower limit of the sweep to 3.525
MHz. In the case of the gain sweep, shown in Fig. 8, the
differences show up in different rates of increase in gain across the
passband--not widely different, but different nevertheless.
Chapter 42
Fig. 9 shows the front-to-back ratio sweep for the two models. The
1" model reaches peak front-to-back ratio just below the design
frequency, while the 2-wire model shows its peak about 25 kHz
below the design frequency. The curve confirms the note above
that some fine tuning of the 2-wire spacing (a slight narrowing) is
necessary to create a true overlap of curves. However, the ultimate
front-to-back peaks of both antennas are quite close, pushing
Chapter 42
above 27 dB. Moreover, extending the sweep scale shows that the
curves are quite congruent, since the lower end differential is about
the same as the upper end differential.
The SWR curve in Fig. 10, tells a similar tale. The passband end
differentials are reasonably close so that the offset between the
Chapter 42
curves does not lead to any misleading conclusions about the 2:1
operating bandwidths of the 2 models. They are essentially the
same.
The end result--despite the small offsets in the curves and number-
-is the conclusion that for practical purposes, the 2 0.1" diameter
wire elements with 13" spacing in the Yagi model provide the same
performance potential as a single 1" tubular element set in the
same model. The higher mutual coupling of the wire model offsets
the higher material losses, resulting in a beam with the same
performance over the range of vital performance parameters as
that of a single fat element model.
1. The 3-Wire Model and a 2.85" Single Wire Model: I would like
to be able to say that the 3-wire model overcame all remaining
differentials with the 4" model of the 3-element Yagi. However,
some differential remained, although it might be considered minor.
A 2.85" diameter element model yielded a set of performance
curves roughly similar to those of the 3-wire model. Once, more, we
know the physical dimensions of the 3-wire model from the
discussion above. The following table presents the physical aspects
of the 2.85" model.
3-Element Yagi: 3.6 MHz: 2.85" diameter elements
Driver 130.30 41.167
Director 122.66 88.767
Chapter 42
The design-frequency (3.6 MHz) performance for the two models is

in this table.
2.85" 8.04 27.6 27.4 - j 1.7 99.75
3x0.1" 8.04 21.7 26.0 + j 6.1 97.32
The comparable performance at the design frequency is readily

apparent. Note, however, that the efficiency of the single wire 2.85"
model is still significantly higher than that of the 3-wire model,
owing to the higher material losses and smaller surface area
available from the trio of thinner wires. Indeed, the surface area
ratio is about 9.5:1 in favor of the 2.85" model. Nevertheless, just as
with the 2-wire model, the increased mutual coupling made
possible by the 3 spaced wires is sufficient to overcome the
increased material losses and net the same gain at the design
frequency as the more efficient fat-wire model.
The comparison can be extended to the current magnitude and

phase found on the parasitic elements with a source current of 1.0
at 0.0 degrees.
2.85" 0.410 141.1 0.588 -132.0
3x0.1" 0.387 141.3 0.623 -137.2
Once more, there are differences between the two models with
respect to current magnitudes and phases on the parasitic element
Chapter 42
centers--but not great ones. The remnant 6-Ohm reactance on the

3-wore driver element is a result of having performed no
adjustments in the spacing to compensate for the addition of the
center wire. In fact, the reactance value on the driver also indicates
not only a slight driver over-length for the spacing, but as well a
similar situation for the other elements as well. Hence, the best
SWR value and front-to-back peak occurs below the design
frequency. As with the 2- wire model, for an equal-gain situation,
the required mutual coupling among elements to overcome the
higher material losses in the 3-wire model would also dictate
slightly different parasitic element current values relative to the
single wire model.
Chapter 42
However, when we make allowances for the offset in self-resonant

frequencies of the individual elements, the curves for the single
2.85" element model and for the 3-wire model remain remarkable
congruent. Fig. 11 shows the gain curves, which show even
smaller differences than those we saw in Fig. 9 across the same
passband for the frequency sweep.
Chapter 42
Likewise, the two front-to-back ratio curves in Fig. 12, display

excellent congruence with a displacement that is almost exactly 25
kHz. Compare not only the offsets at the front-to-back peaks, but as
well the lower frequency 19+ dB point and the higher frequency 17+
dB points.
Chapter 42
Fig. 13 portrays a similar offset with respect to the SWR curve for
both models. The shapes of the curves are virtually identical,
despite the 25 kHz offset between them. In effect, bringing the
entire set of curves for the two models into alignment would likely
be a matter simply of adjusting the spacing of the wires in the 3-
wire elements.
Chapter 42
Alternatively, one might also slightly shorten the elements of the 3-

wire model slightly. In actual construction practice, this procedure
would become the most practical, since it is likely that the wire
elements, including spacers, would be fixed in construction phases
that precede raising the antenna to operating height.
The 3-wire model ends up slightly short of the goal of achieving the
full gain of the initial 4" diameter element model. However, the
deficit is only about 0.1 dB. Front-to-back performance can be as
good as the initial model, and the feedpoint impedance will be
comparable.
In effect, the exercise has established that it is possible to create

thin wire models with multiple wires per element that simulate
effectively the performance of fat tubular elements in beams and
arrays--at least for this 3-element Yagi model. The requisite
spacing and possibly the number of wires required will vary with the
wire size chosen and the diameter of the element being simulated.
It is not sufficient simply to create elements from multiple wires. The

spacing of the wires is of great significant in increasing the mutual
coupling between wires that yields the desired level of
performance, understanding that some excess coupling may be
needed to overcome higher material losses in the wire substitutes.
If nothing else, this exercise has suggested some interesting
relationships between the roles of material losses and mutual
element coupling in beam performance.
Chapter 42
Besides the flat plane, other wire configurations are possible,

including triangles, squares, and the traditional cage hexagon of
wires. However, the extra weight of 6 wires, relative to the 3-wire
plane used in the models presented here, may not be needed to
effectively simulate fat wires. In the final analysis, two pieces of
design work are needed. One is, for any proposed array, a set of
models to establish the relative structures of equivalent multi-wire
substitutes for a given fat single element. The other, of course, is a
mechanical analysis to determine the best compromise among the
physical properties of the resultant element. The properties might
include ease of construction, durability in the proposed
environment, and ease of raising the structure to its operating
height.
Of course, the home antenna builder might well decide that using
single wire elements is the simplest construction method and that
the performance deficits relative to the ideal still fall within an
acceptable level. While it is one thing to tangle at a computer with
complex models of substitute multi-wire element substitutes, it is
quite another to have to wrestle with the real thing, namely, the
tangle-prone wire elements that require raising into operating
position.
Chapter 42
Chapter 43: The Collinear-Broadside-Endfire Array
For many years, The ARRL Antenna Book has shown a

combination collinear, broadside, endfire array, for example on
pages 8-51 to 8-54 of the 19th Edition. The array looks something
like the outline in Fig. 1.
Chapter 43
The dimensions are locked up in 3 key factors:
1. L = Element Length: Length may vary from 1/2 wavelength up to

well over 1 wavelength.
2. SP = Horizontal Element Spacing: Horizontal spacing may range

from 1/4 wavelength to 3/8 wavelength.
3. H = Vertical Element Spacing: Vertical spacing may range from

3/8 wavelength to 3/4 wavelength.
The write-up shows a typical pattern with good bi-directional gain.

However, a myriad of questions remains for anyone contemplating
building such an array. However, we can boil the questions down to
just two.
What do we get? What do we pay?
If we look carefully at the sketch, we can see that the array in Fig. 1
only becomes a collection of collinear elements when the element
lengths approach and pass beyond the 1-wavelength mark, at
which time, we can consider them to be collinear half-wavelength
elements. The vertical dimensions surround those that we
associate with the Lazy-H antenna. If we look carefully at the
sketch, we can see that the two left-side elements are in phase with
each other, since the lines to each left-side element have half-
twists. Likewise, the right-side elements are also in phase with each
other. The 2 top elements are out of phase with each other, the
sign of a W8JK array. Likewise for the bottom 2 elements. Of
Chapter 43
course, all of this assumes that all 4 lines from the central junction
are equal in both length and characteristic impedance.
Therefore, I have dubbed the array the "Lazy-8JK." This name is

very much shorter than calling the assembly a collinear, broadside,
endfire array. The latter expression is descriptive, because the top
and bottom elements form pairs that are endfire arrangements,
while the left and right pairs are broadside arrangements.
Identifying the array as related to both the Lazy-H and the W8JK
gives us a means of answering the latter question, at least in part.
What we pay is something over twice the complexity of either the
8JK or the Lazy-H. Before we are finished, we shall examine some
other costs for the array, but this much is a beginning.
The tougher question is knowing what we get for our effort.

However, identifying the array as a combination of Lazy-H and 8JK
arrays gives us a foundation from which we can work. Let's
transform the question into this one: What are the advantages in
performance of the Lazy-8JK over either the Lazy-H alone or the
W8JK alone? Now we have something with which to work.
A Review of the Lazy-H Broadside Array
In its basic form, the Lazy-H is a very old but well-proven antenna
design with distinct advantages among wire arrays. Originally, it
consisted of two 1-wavelength elements vertically spaced 1/2-
wavelength apart. However, users later discovered that the antenna
Chapter 43
would operate effectively over a wide frequency range using a

parallel transmission line and a wide-range antenna tuner.
In Fig. 2, we see the essential electrical components of the Lazy-H.

The horizontal wires marked L are the elements. PL, the phase-
Chapter 43
line, is broken into two equal parts, PL1 and PL2. As the diagram
indicates, the two elements are fed in-phase with no twists on either
phase line section. The main feedline, attached at the junction of
PL1 and PL2, provides equal power to each element. Essentially,
then, the Lazy-H consists of two doublets, vertically spaced and fed
in-phase, in order to obtain considerable gain over a single doublet
of the same length mounted at the approximate array center.
The array produces bi-directional patterns on all bands within the

operating range. The elements must be no more than about 1.25
wavelengths to achieve a bi-directional pattern. The antenna will
operate at higher frequencies, but the pattern breaks down into
multiple lobes as the electrical length of the elements increases at
higher frequencies. As well, maximum gain occurs when the
elements are about 5/8-wavelength apart. At the highest frequency
of bi-directional operation, the antenna has been called the
extended or expanded Lazy-H.
Chapter 43
Fig. 3 shows one common form of the Lazy-H that is useful for
operation on amateur bands from 10 meters down to 40 meters. On
10 meters, the 44' wires are about 1.25 wavelengths, dropping to 1
wavelength on 15 meters, and becoming progressively electrically
shorter as we reduce frequency. The 22' spacing is 5/8 wavelength
on 10, 1/2 wavelength on 15, and electrically closer on lower
frequencies. We shall use this model as a foundation for this
evaluation of the Lazy-8JK in order to provide a consistent set of
dimension throughout. In addition, we shall place the bottom wire of
Chapter 43
the Lazy-H at 44', with the top wire at 66'. With these dimensions
and heights, we obtain the performance in the following table.
Extended Lazy-H Performance Potential
Freq. Gain TO angle Beamwidth Feed Z

MHz dBi degrees degrees R+/-jX Ohms
28.1 15.3 9 31 40 + j 305
24.95 14.6 10 41 20 + j 100
21.1 12.5 11 52 25 - j 35
18.118 10.9 13 61 50 - j 145
14.1 9.0 17 73 495 - j 145
10.125 8.1 24 85 50 + j 105
7.1 6.3 33 99 10 - j 100
Because the elements grow shorter and the spacing becomes

closer with decreasing frequency, the gain drops with frequency.
That fact becomes clear from the overlaid azimuth patterns in Fig.
4. The 10-meter azimuth pattern shows its extended-double-Zepp
family resemblance, while the 15-meter pattern has the classic
shape for 1-wavelength elements. Although the full usable
operating band-spread for the antenna appears in the table and
pattern graphic, we shall focus only on operation in the 5 highest
HF bands from 20 meters onward.
Chapter 43
A Review of the W8JK Endfire Array
In its basic form, the 8JK is a very old but well-proven antenna
design with distinct advantages among wire arrays. The array name
derives from its inventor, John D. Kraus, W8JK, who wrote on
various forms and facets of the antenna from 1937 to the present.
Originally, the 8JK consisted of two 1/2-wavelength elements
horizontally spaced from 1/8 to 1/2 wavelength. However, users
Chapter 43
later discovered that the antenna would operate effectively over a

wide frequency range using a parallel transmission line and a wide-
range antenna tuner.
In Fig. 5, we see the essential electrical components of the 8JK.

The horizontal wires marked L are the elements. PL, the phase-
line, is broken into two equal parts, PL1 and PL2. As the diagram
indicates, the two elements are fed out-of-phase with a half-twist on
one of the phase line sections. The main feedline, attached at the
junction of PL1 and PL2, provides equal power to each element.
Essentially, then, the 8JK consists of two doublets, horizontally
spaced and fed out-of-phase, in order to obtain considerable gain
over a single doublet of the same length mounted at the
approximate array center.
Chapter 43
The array produces bi-directional patterns on all bands within the

operating range. The elements must be no more than about 1.25
wavelengths to achieve a bi-directional pattern. The antenna will
operate at higher frequencies, but the pattern breaks down into
multiple lobes as the electrical length of the elements increases at
higher frequencies. The element spacing is subject to significant
variation among builders. The version that we shall explore uses a
spacing of 5/8 wavelength at the highest operating frequency. This
spacing becomes electrically smaller as we reduce frequency, but
still is not the spacing for the highest possible gain. In fact, the
array tends to increase gain with closer spacing, although the
impedance at the feedpoint becomes impractically low. The
advantage of the 5/8-wavelength spacing used in this model is that
it yields almost constant gain over the entire operating range of the
antenna.
The version that we shall explore uses the same wire lengths and
spacing as the Lazy-H in Fig. 3. The only differences are the fact
that the antenna is placed in a horizontal position and there is a
half-twist in one (and only one) of the phase lines from the common
junction to the elements. Since the Lazy-H has elements at both 44'
and 66' above ground, we shall sample performance on 20 through
10 meters at each of these heights.
Chapter 43
44' Above Ground

28.1 11.8 11 33 160 + j 505
24.95 11.8 13 41 35 + j 170
21.1 11.1 15 50 25 + j 25
18.118 11.0 17 55 25 - j 75
14.1 10.5 21 61 110 - j 420
66' Above Ground

28.1 11.7 8 33 165 + j 505
24.95 11.9 9 40 35 + j 170
21.1 11.4 10 40 20 + j 25
18.118 11.3 11 54 25 - j 75
14.1 10.8 14 60 110 - j 410
In these tables--indeed, in all of the tables--the impedance figures

are representative values modeled with 450-Ohm transmission line
for the phase lines. The actual feedpoint values will tend to vary
with changes in either the characteristic impedance or the velocity
factor of the line selected. for that reason, I recommend that any
prospective builder of any one of these arrays model the installation
including the characteristic impedance and velocity factor of the
phase lines to be used. From these values, you may calculate the
anticipated impedance at the antenna tuner using any one of a
number of available transmission line programs, such as TLW or
TLD.
At either height, the gain variation is less than 1 dB from 20 through

10 meters. Although the gain is generally lower, on 20 meters the
8JK outperforms the Lazy-H. Fig. 6 provides the overlaid azimuth
Chapter 43
patterns for the W8JK in extended operation to even lower

frequencies. Once more, the extended-double-Zepp origins of the
10-meter pattern are apparent.
The gain differences between the 8JK and the Lazy-H are largely
due to differences in the amount of energy radiated at high
elevation angles. Fig. 7 shows selected elevation patterns for the
Chapter 43
two arrays, with the 8JK patterns taken for a 66' height above
ground.
At 21 MHz, where the Lazy-H spacing is almost a perfect half-

wavelength, the Lazy-H effectively suppresses high-angle radiation,
in contrast to the multiple strong upper-angle lobes of the 8JK. On
10 meters, the Lazy-H does not suppress overhead radiation
perfectly due to the wider spacing, but overall suppression of high-
Chapter 43
angle radiation is excellent. Still, the 8JK shows even more high-
angle lobes. The 8JK begins to exceed the Lazy-H in gain in the
lowest lobe on 20 meters. The Lazy-H spacing is down to about
5/16 wavelength, a value that is too close for effective suppression
of high angle radiation.
The "Lazy-8JK" Array
Against this background, we may now combine a Lazy-H with an

8JK to arrive at the collinear broadside endfire array, that is, at the
Lazy-8JK. The elements of Fig. 1 describe the most common
version of the antenna. For our comparisons, we shall use 44'
element lengths, which is about 1.25 wavelengths at 10 meters.
The vertical spacing will use the Lazy-H value of 22'.
However, we cannot use the 22' element spacing of the 8JK that
we have reviewed. At that spacing, the Lazy-8JK actually shows
less gain than a single Lazy-H for 15 through 10 meters. We must
compress the horizontal spacing to about 8-11 feet to obtain any
usable additional gain from the 4-doublet array. The modeled data
are based upon the 11' spacing, although the differences between
8' and 11' are not operationally significant.
In addition, there is an alternative feed system to the one shown in

Fig. 1, where each of 4 equal-length phase lines combine at a
central feedpoint. As shown in Fig. 8, we may also construct
individual Lazy-H arrays with center feedpoints. We then run equal
lines to a center position, giving one and only one of those lines the
Chapter 43
necessary half twist to place the left-side doublets out of phase with
the right-side doublets.
For the 44' elements at 44' and 66' above ground, 11' left-to-right
spacing requires two 5.5' phase lines. For the data in the tables, I
again used 450-Ohm line for all phase lines.
Chapter 43
Lazy-8JK Performance Potential
Standard Feed System (Fig. 1)

28.1 16.5 9 33 50 + j 300
24.95 15.4 10 37 9 + j 90
21.1 13.8 11 45 6 + j 10
18.118 12.8 13 51 7 - j 40
14.1 11.6 16 58 42 - j 260
Alternative Feed System (Fig. 8)

28.1 16.5 9 33 1590 + j1870
24.95 15.4 10 37 35 + j 465
21.1 13.8 11 45 10 + j 190
18.118 12.8 13 51 9 - j 85
14.1 11.6 16 58 30 - j 135
1. Gain Factors: There is no difference in the gain performance

between the two feed systems for the Lazy-8JK. The Lazy-8JK
provides a variable gain value according to frequency according to
the Lazy-H part of its origins. However, the W8JK portion of its
origins shows up in the increments of gain above each of the
corresponding simpler arrays. The following table tracks the gain
advantage of the Lazy-8JK over the Lazy-H and the W8JK, where
the latter uses values for a 66' height above ground.
Chapter 43
Gain Advantage of the Lazy-8JK

All gain values are in dBi at the elevation angle of maximum radiation.
Freq. Lazy- Lazy- Added W8JK Added

MHz 8JK H Gain Gain
28.1 16.5 15.3 1.2 11.7 4.8
24.95 15.4 14.6 0.8 11.9 3.5
21.1 13.8 12.5 1.3 11.4 2.4
18.12 12.8 10.9 1.9 11.3 1.5
14.1 11.6 9.0 2.6 10.8 0.8
The Lazy-8JK adds only about 1 dB gain to a single Lazy-H in the

upper-most bands. However, the 8JK portion of its origins
increases the gain on 20 meters by about 2.6 dB. In contrast, the
Lazy-8JK shows its highest gain additions over a single 8JK on the
top bands, with reduced gain additions with descending frequency.
Additionally, the beamwidths for the Lazy-8JK are slightly narrower
than for the standard 8JK as a function of the additional gain.
Chapter 43
Fig. 9 shows selected elevation patterns corresponding to those

shown for the Lazy-H and the W8JK. The added gain is largely due
to a continuing progression of upper lobe suppression. A
comparison of 20-meter patterns among all three arrays is
especially notable.
To complete the comparisons, Fig. 10 shows overlaid azimuth

patterns for each band, with each pattern taken at the elevation
angle of maximum radiation.
Chapter 43
From the perspective of gain, then, there are two considerations

facing the builder who may weigh the increase in construction
complexity against the performance advantage. First, the more
complex array shows a better balance of gain than a single Lazy-H,
with improvements especially in the 20-meter performance.
Second, the Lazy-8JK increases gain on the upper bands over the
sampled standard 8JK array. Whether the total improvement is
sufficient to warrant the increased construction problems is
ultimately a user judgment.
2. Impedance Considerations: The differences in the two ways of

feeding the Lazy-8JK lie wholly in the area of the feedpoint
impedances. The standard X-form of the phase lines (Fig. 1) yields
more bands on which the feedpoint impedance is under 10 Ohms
resistive. Under these conditions, every fraction of an Ohm of loss
resistance in the connections transforms a proportionately higher
percentage of the power into heat. Hence, great care must be used
in the assembly of the system, and equal periodic maintenance
care is required.
The alternative configuration for feeding the Lazy-8JK has fewer

instances of very low resistive components to the feedpoint
impedance. However, 10 meters present a high impedance, with
both the resistive and reactive components above 1500 Ohms. As
a result of the high variability in feedpoint impedances, the builder
very likely will have to pay close attention to the line length between
the antenna feedpoint and the tuner terminals to present
impedances within the tuner's matching capabilities. This aspect of
Chapter 43
the Lazy-8JK is the second half of the answer to the original

question of what do we pay.
Conclusion
The Lazy-8JK (or collinear broadside endfire) array offers different

advantages over the standard Lazy-H and the standard 8JK. Since
these advantages change with each upper HF band, there is no
single answer to the question of whether the larger array is worth
the effort of construction.
Nevertheless, by setting up reasonable comparators, these notes

will hopefully provide a better sense of both the advantages and
disadvantages of going to the more complex arrangement. There
are additional dimensions of the final decision. For example, the
Lazy-8JK does not lend itself to the relatively simple triangular
arrangements of Lazy-H arrays that permit switched directional
changes. Nor does the Lazy-8JK allow one to construct an easily
rotated version of the array, as is possible with a single 8JK. Yet,
the lure of added gain is likely to appeal to at least some
aficionados of wire antenna construction. With dimensions of 44' by
22' by 11', it is still a fairly compact upper HF bi-directional array
with outstanding performance potential.
Chapter 43
Chapter 44: Curtains for the Extended Lazy-H
The Basic Expanded/Extended Lazy-H
A
few years ago, I called attention to a wire array that
receives notice only about every fifteen years--despite its
excellent performance as a fixed bi-directional array of
modest proportions. With two 44' long wires spaced 22' apart, the
extended or expanded Lazy-H provides primary service on 20
through 10 meters, with quite adequate service on 30 and even 40
meters. From each element center, we run a parallel feedline--in
phase--to a center position, which then becomes the primary
feedpoint. We may use any feedline from 300 Ohms to 600 Ohms,
although any figures shown in these notes will apply to 450-Ohm
line with a 0.95 velocity factor. These notes also presume AWG
#14 or #12 copper wire for the elements.
The base of the antenna ideally should be about 44' or more above
ground for the lowest elevation angles of maximum radiation (take-
off or TO angles). However, if necessity prevent the top wire from
reaching 66', then a lower height and higher elevation angles are
tolerable.
Chapter 45 next provides a set of representative elevation and

azimuth patterns for the antenna. The following table summarizes
the performance when set at a 44'-66' height. The gain is the
maximum gain at the TO angle. The Beamwidth is the angle
between half-power or -3 dB points away from the bearing for
Chapter 44
maximum gain. The feedpoint impedance (Feed Z) is at the junction

of the two 11' lengths of 450-Ohm parallel line.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

28.5 15.1 8 31 64 + j 425
24.95 14.6 10 41 17 + j 115
21.1 12.5 11 52 22 - j 18
18.118 10.9 13 61 43 - j 125
14.1 9.0 17 73 403 - j 395
10.125 8.1 24 85 49 + j 105
7.1 6.3 33 99 10 - j 100
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The impedances are all quite manageable for a wide-range

balanced antenna tuner, with the possible exception of 40 meters
and 12 meters. However, the impedance at the tuner will also be a
function of the line length in wavelengths from the feedpoint to the
tuner, so adjustment of the values to be matched is feasible.
Planar Reflectors
My reason for returning to the extended/expanded Lazy-H grew out

of an inquiry from Bill Burton, T88BA of Pelau. He had assembled
the pieces for the antenna. However, he had also seen some
commercial arrays that used a curtain-style reflector. Hence, he
asked whether such a reflector was feasible for the big Lazy-H.
I have seen a number of schemes for multiple extended double

Zepps of about 44' and even for the Lazy-H, all designed to use
Chapter 44
parasitic elements or phased elements to convert the bi-directional

array into a directional beam. Their complexity could be daunting,
often with the result that only the builder knew the correct
adjustments for each band. On the other hand, a planar or curtain
reflector is a broad-band passive addition that might be serviceable.
In UHF work, a planar reflector generally exceeds the dimensions

of the active element or array by perhaps 3/8 to 1/2 wavelength on
all four sides. Such a curtain would be somewhat ungainly for the
extended/expanded Lazy-H. So I somewhat arbitrarily selected a
set of dimensions that seemed close to feasible for someone with
sufficient land to erect the basic antenna at the ideal minimum
height. The modeled screen is 50' wide by 30' high. This is only 6'
wider than the element lengths and 8' larger vertically. In many
ways, it is a minimalist planar reflector.
Chapter 44
Fig. 1 shows the parts of the array, although not to scale. The
extended/expanded Lazy-H remains unchanged from independent
use which results in its bi-directional characteristics. The screen is
centered behind the active array. Although the modeled screen
uses a grid-square assembly, it is possible to use a sequence of
Chapter 44
wires parallel to each other. For this horizontal array, the wires
must also be horizontal.
The 10' spacing was my initial trial spacing for the reflector--just a
bit shy of 1/4 wavelength at 15 meters. After looking at numerous
other spacings, I returned to my intuitive selection, since it provides
approximately equal front-to-back ratios at the array limits, namely
10 meters at the high end and 30 meters at the low end.
To see how the curtain or planar reflector changes array

performance, examine the following table and compare various
values to the ones for the array alone.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Curtained Extended Lazy-H Performance Potential
Freq. Gain TO angle Beamwidth F-B Ratio Feed Z

MHz dBi degrees degrees dB R+/-jX Ohms
28.5 18.4 8 30 13.7 112 + j 430
24.95 17.1 10 40 15.5 29 + j 125
21.1 15.5 11 50 17.1 25 - j 3
18.118 14.4 13 56 16.8 33 - j 100
14.1 13.1 17 63 14.2 175 - j 460
10.125 12.2 23 69 13.8 21 + j 140
7.1 10.0* 30 75 8.3 2 - j 100
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The gain increase in the favored direction is between 3 and 4 dB,

depending upon the band, with peak increases in the 17 and 20
meter bands. The gain improvement values are consistent with
those for any casually designed 2-element driver-reflector Yagi.
Peak front-to-back ratio occurs on 15 meters and decreases slowly
above and below that band.
Chapter 44
The impedance values are interesting. No great change occurs in

the reactance values. However, the resistive component shows an
interesting pattern. On 15 meters, its value is about the same as
the value for the array alone. Above 15 meters, where the spacing
is greater than 1/4 wavelength, the resistive component of the
impedance is higher than for the array alone. Below 15 meters, the
resistive component is less than for the array alone, while the
spacing is less than 1/4 wavelength.
Fig. 2 through Fig. 7 provide azimuth and elevation patterns for the
array on each band. They require no individual comment. However,
the trends in the rear lobe formations should be reasonably clear. In
all cases, the patterns are well-behaved, with no spurious lobes--
other than the emergent secondary lobes inherent to the array
when the active element length approaches 1.25 wavelengths.
Chapter 44
Chapter 44
Chapter 44
Before leaving our patterns, we must take special note of 40

meters. The gain value for that band is starred. That star indicates
that the pattern shows maximum gain in the reverse direction
relative to all of the other bands. See Fig. 8 for azimuth and
elevation details.
Chapter 44
The reason for pattern reversal is simple. The horizontal dimension

for the reflector (50') is 1/2 wavelength or more for all bands from
30 to 10 meters. However, on 40 meters, the reflector horizontal
dimension is only about 3/8 wavelength. Hence, despite its vertical
dimension, the screen acts like a director at less than 0.1
wavelength spacing. The very low resistive component of the
feedpoint impedance reflects this condition. It is unlikely that one
would be able to take advantage of the reverse pattern, given the
potential difficulty in achieving a low-loss match.
Physical Realities
The proof-of-principle exercise suggests that a minimalist planar

reflector or curtain behind an expanded/extended Lazy-H will yield
a competent directional beam with a noticeable improvement in
gain and quite usable front-to-back characteristics. However, the
Chapter 44
requirements for the reflector are sufficiently challenging to make

this array an antenna with a somewhat small niche.
Planar reflectors require both vertical and horizontal dimension for

highest effectiveness. Too narrow a vertical dimension will degrade
the array characteristics as much as too short a horizontal
dimension. Since the reflector is untuned, it must be at least 1/2
wavelength at the lowest frequency used, with additional length up
to about 1 wavelength wherever feasible. Vertically, we improve
performance with height greater than those used here, although the
vertical dimension will reach its limits of helpfulness more quickly
than the horizontal dimension.
Assuming that we can erect vertical supports of the needed height,

I recommend some form of halyard assembly to raise and lower the
reflector screen. Even an extremely open "chicken-wire" screen
reflector will show considerable aggregate wind resistance in
violent storms. However, if one desired beaming in both broadside
directions, then a pair of screens--with only one raised at a time--
will yield a reversible beam. If the user has only one primary target
region, lowering a single screen will return the array to its inherent
bi-directional pattern, and that may suffice for other operations.
For raising and lowering, we need an open-weave metal "fabric"

that will resist snagging when it is crumpled on the ground.
However, one might wind the screen around a ground-level cylinder
instead of lowering it into a heap directly on the ground. One might
even use a combination of horizontal wires and vertical ropes to
create a more flexible screen for this purpose. The final product is
Chapter 44
suited to the skills of one who has experience with square-rigged

sails, with the sail and spar arrangement inverted from those we
find at sea. Of course, we shall invert another matter as well: we
shall seek to slip the wind rather than catching it.
The curtained Lazy-H is a fixed position wire array of considerable

mechanical size and requires equal mechanical ingenuity to
implement. However, it holds promise of providing multi-band gain
and front-to-back ratio so that the operator can use the upper HF
band on which propagation is nearest to optimal for a desired path.
The beamwidth on the upper bands is narrow enough to require
careful sitting. Although the values in the charts emphasize the
amateur bands, the general arrangement may be suitable for short-
wave listening in the intervening spectrum--and possibly even for
an economical short-wave broadcast installation.
The curtained Lazy-H is certainly not an antenna for everyone. But

it may be an antenna for someone.
Chapter 44
Chapter 45: The Expanded Lazy-H
V
ery effective antennas do not have to be exceptionally large
or expensive. The latest designs and construction methods
have their advantages--and also their costs. They tend to
obscure some older designs of high merit as we forget to remember
them.
Rotatable antennas are very effective, but for those unwilling or

unable to put a tower, rotator, and sizable aluminum structure in the
air, fixed position wire arrays can provide excellent gain. Most
designs are bi- directional, but the side rejection is often sufficient
to eliminate most QRM. If we have the trees or the poles to support
the ends, and if we take the trouble to align the antenna in the most
favorable directions for our intended operation, a wire array can
work wonders. For example, a great circle drawn through my QTH
in Tennessee with one end in VK-ZL land will have its other end in
Europe. A bi-directional array might be just the ticket for much of
my operating.
Many broadside arrays are flat-tops--that is, they require at least

two wires with considerable horizontal space between them. For
most purposes, I would need 4 supports. However, a design that
has been around since wire became popular is the Lazy-H, a
vertical stack of two wires fed in phase. The standard Lazy-H
consisted of two 1 wl wires spaced 1/2 wl apart and elevated so
that the lower wire was 1/2 wl above ground.
Chapter 45
John Schultz, W2EEY, wrote in the November, 1968, CQ of the

"Expanded Lazy-H Antenna." Bill Orr, W6SAI, recalled this antenna
in one of his many columns during the 1980s. Another 15 years has
gone by, so let's recall this effective array one more time. The key
to expanding the Lazy-H is to increase both the horizontal and
vertical dimensions by just a little bit.
If we increase the wire lengths from 1 wl to 1.25 wl, we have

stacked extended double Zepps in our Lazy-H. The effect is to give
us a bit more gain per wire and a significant amount more from the
pair. Then, if we increase the spacing from 1/2 wl to 5/8 wl, we
achieve approximately the maximum stacked gain possible with two
simple wires.
Now, let's build one of these expanded Lazy-Hs for 10 meters.
Chapter 45
Fig. 1 shows the antenna outline. For 10 meters, a length of 44' per
wire is satisfactory and not critical: 40' to 50' will work, but the
pattern on 10 meter begins to split up as we lengthen the antenna
too far beyond 1.25 WL. Vertical spacing between the two wires
need not be too fussy, but the recommended 22' gives us not only
5/8 wl at 10 meters but a usable spacing at other frequencies.
The recommended minimum of 1/2 wl at 10 meters is a bit low for

optimal performance. I would recommend that a lower-wire height
Chapter 45
of about 44' be used, which places the top wire at 66' up. Lower
heights will reduce the gain and elevate the TO angle from the
figures I shall present as we think about this simple array.
The mechanical beauty of the Lazy-H design is that it requires only

two supports--although fairly tall ones. The electrical beauty of the
antenna is that it provides excellent bi-directional performance from
10 meters through at least 17 meters, with good performance down
to 30 meters. It can also be pressed into service on 40 meters
without much difficulty.
Chapter 45
Fig. 2 shows the azimuth pattern of the antenna on 10 meters at an

elevation angle of 8 degrees. The phase-fed array still retains the
EDZ "ears." These ears are the beginnings of the multi-lobe pattern
that emerges as the antenna wire length grows toward the 1.5 wl
mark.
On all bands below 10 meters, the length of the antenna is under

the EDZ mark, so the pattern is bi-directional with single lobes each
way. In fact, at 15 meters, the antenna becomes a standard Lazy-
H: two 1 wl wires spaced 1/2 wl apart vertically.
Chapter 45
In Fig. 3, we see the 17-meter pattern at its elevation angle of

maximum radiation of 13 degrees. As the tables below will show,
the lobes become wider as we reduce frequency and narrower as
we increase frequency.
For a more systematic view of anticipated performance on all of the

possible bands on which we might use this one Lazy-H, here is a
table of modeled performance over average ground, with the lower
wire 44' up. The table lists the usual gain and TO angle data, but
also adds numbers for the vertical and horizontal beamwidths
between the -3 dB (half-power) points. This date is useful in
determining the azimuth coverage of the antenna in each direction
and in estimating the elevation angles to catch the skip for varying
circumstances.
Freq Max. Gain TO angle Vert. BW Hor. BW
MHz dBi degrees degrees degrees
28.5 15.1 8 9 31
24.9 14.6 10 11 41
21.2 12.5 11 12 52
18.1 10.9 13 14 61
14.15 9.0 17 18 73
10.1 8.1 24 27 85
7.15 6.4 33 44 99
Chapter 45
By the time the antenna's operating frequency is lowered to 40

meters, the pattern becomes a broad oval with a fairly high TO
angle, as shown in Fig. 4. However, sufficient radiation occurs at
lower angles to make it usable for general purpose communications
on that band.
How good is the antenna's performance? I could use any number of

comparators here, but the simplest would be a single 44' wire
Chapter 45
placed 66' up in height, the same height as the top wire of the
array. The usefulness of this comparison is that it helps reveal
something of the array's characteristics.
Freq Max. Gain TO angle Feedpoint Z
MHz dBi degrees R +/- jX Ohms
28.5 10.5 7 150 - j 695
24.9 10.4 8 620 - j1700
21.2 9.0 10 4200 + j 850
18.1 8.6 12 835 + j1560
14.15 7.7 15 190 + j 490
10.1 7.6 20* 56 - j 105
7.15 7.0* 29* 24 - j 600*
The starred gain entry for 40 meters indicates that the single wire at
this frequency shows more gain than the array (by about 0.6 dB). In
the TO angle column, the starred entries indicate that the single
wire shows a significantly lower angle than shown for the array.
Both phenomena are related. The array elevation angle of
maximum radiation is a composite from radiation from both wires,
with the lower wire radiation raising the angle of the final composite
pattern. The difference is slight until the very lowest bands on which
we might press this antenna into service. On 40 meters, the lower
wire is just over 1/4 wl above ground, so that it raises the overall
pattern angle of the array by a goodly amount and provides slightly
less gain than the single wire that is about 1/2 wl up. As well, The
high ratio of reactance to resistance in the feedpoint impedance
suggests that there may be difficulty in obtaining a good low-loss
match.
From 20 meters on up, the Expanded Lazy-H shows good gain

over a single wire. The benefits increase the higher one goes in
Chapter 45
frequency, up to the break-up of the pattern when the wires are

longer than 1.25 wl. It is certainly possible to scale the antenna for
maximum benefits at a lower frequency, but that lower frequency of
maximum gain will become the highest frequency at which one can
use the array and still have a bi-directional pattern with a single
main lobe off each side of the wires.
The single 44' wire also shows a wide variation in feedpoint

impedance according to the length of the wire. The 10-meter value
is typical for an EDZ. The 15-meter value is also typical, but of a 1
wl center-fed wire. Parallel feeders and a highly competent antenna
tuner would be needed for this antenna. However, careful analysis
of the impedance excursions along the chosen feedline can
minimize the chances that the tuner antenna terminals will see
either a resistance or a reactance value outside its range of
adjustment.
So far, I have given no figures for the feedpoint impedance of the

Expanded Lazy-H. The two elements in Fig. 1 are fed in phase by
the simple expedient of using equal lengths (11') of line to a center
point to which we attach the parallel feeders going to the antenna
tuner. There are two controllable variables that will affect the
feedpoint impedance at the junction with the main line to the shack.
One is the length of the lines, which we have set at 11' each. The
other is the characteristic impedance and velocity factor of the
phasing lines. I shall not here explore other phasing line lengths,
but instead shall show some anticipated feedpoint impedances for
each band using three different phasing lines. One will be a 450-
Ohm, 0.95 VF line, typical of windowed vinyl-covered lines. Another
Chapter 45
will be 300-Ohm, 0.8 VF line, typical of good quality TV line. The

third will be 600-Ohm, 1.0 VF line, which might be bought or built
from wire and spacers.
Feedpoint Impedance (R +/- jX Ohms)
Freq 450-Ohm 300-Ohm 600-Ohm
MHz 0.95 VF 0.8 VF 1.0 VF
28.5 65 + j 425 115 + j 570 105 + j 610
24.9 17 + j 115* 11 + j 140* 30 + j 140
21.2 22 - j 15* 10 + j 38* 40 - j 50
18.1 45 - j 125 16 - j 26* 90 - j 230
14.15 385 - j 395 75 - j 150 1050 - j 350
10.1 50 + j 105 40 + j 65 50 + j 155
7.15 10 - j 95* 6 - j 80* 13 - j 90*
Starred entries represent very low resistive components to the

feedpoint impedance which might present larger excursions along
whatever line is chosen as the main feedline to the shack. Note that
the starred entries are fewest with the 600-Ohm phasing line. Once
more, it is worth noting that these numbers are derived for general
guidance from models. Variations will emerge from the actual
construction of the antenna and from conditions and clutter at the
antenna site.
One question that almost always emerges with respect to

comparing the single wire and the array gain figures for 10 meters
is this: how can the array deliver over 4.5 dB gain over the single
wire? The answer is straightforward if we compare elevation
patterns for the two antennas. Fig. 5 tells the tale.
Chapter 45
Like any single-wire antenna, the EDZ at 66' on 10 meters shows

an array of nearly equal-strength vertical lobes: 4 to be exact. In
contrast, the upper lobes of the Expanded Lazy-H are suppressed
leaving a single dominant lobe and a secondary lobe well over 4 dB
weaker. All other lobes are down by 12 dB or more. The array
Chapter 45
tends to waste far less power at very high angles of radiation

compared to the single wire. This comparative pattern, with
variations, tends to hold true down through 20 meters.
Chapter 45
On 20, the effect is less pronounced but still easily measured, as

shown in Fig. 6. The area enclosed by the upper lobes of the single
wire at the top of the figure is distinctly greater by a considerable
margin than the area enclosed by the upper lobe (barely
discernable as a double lobe) of the array. The difference in area
(assuming that the azimuth patterns are comparable, as they
happen to be in this case) is a rough measure of the added power
appearing in the lower lobes. In this case, that additional power
shows up not only in the maximum gain, but as well in the vertical
beamwidth. The phased feeding of vertically stacked horizontal
wires has benefits hard to match in a typical flat-top wire array.
Alongside the benefits come some limitations. The Lazy-H requires

a pair of tall supports and is suited to the antenna farm with more
tall trees than money. It is possible to lay out more than one of
these inexpensive antennas in order to cover additional regions
along the horizon. It is likely that no special treatment will be
needed to detune unused arrays to prevent them from altering the
pattern of the array in use. Either leaving the shack end of the
unused feedline open or shorting it will introduce to the wire
feedpoints sufficient reactance to detune the wires. However, this is
a facet of multiple array installation that the builder should keep in
mind. Sometimes Murphy dictates that nothing will work to prevent
interaction short of greater physical separation of the arrays.
The Expanded Lazy-H is an outstanding bi-directional array for 10

meters in the design given here. Its performance holds up well
down through 20 meters, and we can press it into service on lower
bands. It takes up very little room horizontally in the yard, although
Chapter 45
a couple of optimally spaced tall trees certainly can aid the

installation process. The wires for the elements and the phasing
lines, as well as the feedline to the shack and the UV-resistant
support ropes, are certainly inexpensive compared to the cost of a
tower, rotator, coax, and commercial aluminum antenna. It is a
design worth recollecting every 15 years or so just to make sure
that we do not forget it.
Chapter 45
Chapter 46: Explaining the Modern Dipole Curtain Array
L
ong-wire antennas served primarily the needs of point-to-
point HF communications in the first half of the 20th century.
Although some rhombics remained in service within the short-
wave broadcast (SWBC) industry, other antenna designs generally
took over. SWBC tends to require a broader beamwidth than a
rhombic provides. Although the rhombic had the frequency range
necessary for frequency shifts in accord with changing HF skip
conditions, other antennas could serve as well--or almost as well.
Once aimed, the rhombic had a line of targets; SW broadcasters
preferred a large region. Even if the target did not encompass the
entire region, slewing the antenna's beam pattern could reduce
costs by avoiding the need for second and third large high-gain
arrays or complex turning mechanisms.
Antique and Modern Billboard Antennas
The solution to the needs of many SW broadcasters arrived with

improvements to a very old antenna, once called the billboard. (See
Kraus, Antennas, 2nd Ed., p. 547, for a representation of a billboard
antenna.) The operational principle is simple. Any bi-directional
antenna, such as a dipole, becomes a directional antenna when
placed in front of a planar reflector. Planar reflectors find many
contemporary uses in the VHF and UHF region today. Hence, we
often overlook their continuing service for SW broadcasting.
However, their current use depended upon a number of advances,
standardizations, and combinatory techniques to give them the
relative predominance that they now enjoy.
Chapter 46
Fig. 1 sketches (with many missing details) an antenna about

which I have received many inquiries. Vacationers encounter them
in unexpected places from coast to coast (and well inland) in the
U.S. The sketch is not to scale. The towers are much too fat for the
array between them. The figure also lists other missing details that
would obscure the main function of the antenna. For example, we
would normally find many more guy wires for the towers and many
more support and spacing wires and jigs for the key elements that
form the antenna's radiation pattern.
In return for omitting some details, we can clearly see both the
dipole elements in a 3-by-3 array and the reflective screen behind
them. In many cases, the screen will consist only of horizontal
Chapter 46
wires, similar to the rod-based planar reflectors in my notes on that

subject in past articles. Since the horizontal lines are very long,
periodic vertical spacers are necessary to maintain the reflector
shape.
The 3-by-3 array of dipoles also represents an evolution from some

of the original billboard antennas that used center-fed full-
wavelength elements. Note also that the dipoles are closely
spaced. The overall reduction in billboard size per driver unit
formed one of many reasons why the modern dipole array is a
primary short-wave broadcast antenna today. Even the original
versions effected a major real estate saving over designs such as
the rhombic. Of course, many of the original rhombics used timber
supports. As steel tower structures became more common and less
expensive, they not only replaced existing supports, but also made
taller antennas more feasible. Hence, the billboard antenna traded
vertical space for longitudinal space, cutting both purchase and
maintenance costs. (Never volunteer to mow the lawn beneath a
truly long rhombic.)
The dipole array rarely uses its original "billboard" name, although
many folks call it a dipole curtain antenna. "Curtain" refers to the
planar reflector behind the driven elements. They could move a bit
in the wind. Early designs were not fully appreciated for several
reasons. First, the high steel structures and copper wire were
subject to corrosion. Breakage required more repair effort than
splicing a rhombic leg. However, one of the electrical limitations of
the billboard was its narrow operating bandwidth.
Chapter 46
In the first half of the last (20th) century, almost all antenna
designers strove to produce as much gain as might be feasible
from a given design. This bad habit still infects much of the antenna
design for amateur radio. We accept excessive problems in
feedpoint matching by designing long-boom Yagis with the
minimum number of elements necessary for a certain gain level.
Even if we overcome that problem, we continue to accept relatively
poor sidelobe suppression because we refuse to add a few more
elements to the design. We continue to make excuses for antenna
designs that are difficult to replicate due to their narrow operating
bandwidth. (There are good reasons in certain circumstances for
using a narrow beamwidth, but in general, it is usually a condition
with which we are stuck for lack of design imagination.)
Early billboard antennas suffered from narrow operating bandwidth

for several reasons. First, the driving elements used a spacing from
the reflector screen that yielded maximum gain. Second, they
looked for element-to-reflector spacings that left the feedpoint
impedance unchanged relative to the same driver with no reflector.
Third, they used driven element lengths and spacings that yielded
maximum gain. For example, a collinear pair of 1/2-wavelength
elements (or a center-fed full-wavelength element) yields a little
more gain than a simple 1/2-wavelength dipole. (The high
impedance of this type of element, of course, permitted the use of
wide-spaced transmission line segment for feeding and phasing, a
condition very suitable for high-power SWBC operations.) Although
we knew that we might obtain even more gain with a vertical
spacing of 5/8 wavelength, 1/2-wavelength became the standard
for the ease of feeding a vertical collection of elements in phase.
Chapter 46
Rarely did we have the room to arrange the elements horizontally

at optimal spacing. Initially, we used some very close spacing to
reflector screens, sometimes as low as 1/8 wavelength. When we
discovered that a wider spacing would yield more gain and weaker
rear lobes, we opted to use that spacing despite the fact it still
limited the operating bandwidth.
Modern dipole curtain arrays operate on other principles. Some

common ones, found in Chapter 26 of Johnson's Radio Engineering
Handbook, 3rd Ed., reappear in Fig. 2. We may note in passing
that an engineer for TCI, a leading producer of dipole curtain
arrays, wrote the 3rd edition version of the chapter on HF antennas.
If the volume is not conveniently available, you may find some of
the same data at the TCI website. Look at model 611 for a general
description of their dipole curtain arrays.
Chapter 46
The side view of the antenna shows the vertical heights generally
used: 1/2 wavelength between dipoles of the array. Studies of
planar reflectors strongly suggest that this antenna type achieves
maximum gain for a given driver set when the reflector screen
exceeds the driver assembly by 1/2 wavelength or so in every
direction. Realities, including catenary effects on an all-wire
assembly, usually dictate less reflector extension except perhaps at
corners.
Chapter 46
The face view shows the equally desirable horizontal reflector

extension, although every extra foot of reflector screen adds to
costs for perhaps marginal performance improvements. The most
notable feature of the face view is the arrangement of the driver
dipoles. Since a driven dipole is normally slightly less than a
physical half wavelength, we may place the dipoles on 1/2-
wavelength centers across the reflector. Because designers still
wish to use wide-spaced transmission lines for feeding and
phasing, the driven elements are usually some form (in some
cases, an exotic form) of a folded dipole.
The final element to note from the sketch is the recommended

spacing of the elements from the reflector screen: 0.3 wavelength.
A simple dipole tended to show maximum gain and weakest
rearward lobes with considerably closer spacing, but by accepting a
lower gain per driver, the designer achieves a wider operating
bandwidth. Before we close these notes, we shall look at the
combination of ingredients that go into extending the operating
bandwidth of a dipole array.
We (but not necessarily the designers) might express the overall

goal in this manner: since wire elements are relatively light, we can
obtain more performance by packing more elements within the
available space rather than from seeking out the maximum
performance from the minimum number of elements. Most of the
array weight (but not necessarily stress in adverse weather) lies in
the reflector lines or screen. The element spacings in the sketch--
and any extension of the sketch--provide the most performance for
a given space (side-to-side and vertical) occupied by the array.
Chapter 46
Performance here includes not only gain, but beamwidth, rear

lobes, and feedline SWR for some specified reference impedance.
How the Dipole Array Achieves Its Performance
Let's back up a step and see how the modern dipole array achieves
its performance. That step requires that we first examine dipoles on
their own, that is, with no reflector screen. We shall survey in
tabular form the maximum gain of various combinations of dipoles.
Of course, the listed gain will be for a bi-directional array. We shall
designate each combination by a code of the order mV-nH,
indicating the number of dipoles stacked vertically (m) and
horizontally (n). Each vertical dipole will be 1/2 wavelength from its
neighbor, and horizontal dipole lines will be on 1/2-wavelength
centers. The data include both the gain and the horizontal
beamwidth. More correctly, the beamwidth is in the E-plane, since
all values for this exercise are for free space. All dipoles consist of
folded dipole made from AWG #10 copper wire. The test frequency
is 10 MHz.
Free-Space Performance of Various Dipole Arrays
Array Size 1V-1H 1V-2H 1V-3H

Maximum Gain (dBi) 2.13 3.79 5.30
Beamwidth (degrees) 78.4 48.2 33.2


Chapter 46
The 2V-2H configuration offers the greatest step-gain increase over

versions with one less vertical or one less horizontal dipole. The
gain steps are not smooth for two reasons. First, gain increases
diminish as we add steps in a linear count. As well, the dipoles
interact, so that gain is not strictly additive. Slightly different spacing
values or even horizontal end-to-end distances may alter some of
the numbers. Nevertheless, the overall progression of dipole
maximum gain values is a fair representation of the potentials of
dipole arrays on 1/2-wavelength centers.
One interesting fact about the progressions is that the E-plane

beamwidth does not significantly change as we add dipoles
vertically. Narrower beamwidths result from adding dipoles
horizontally. Since the beamwidths of each level of horizontal
stacking are constant, regardless of the size of the vertical stack,
we can represent the array patterns with samples taken with a
vertical stack of 2. Fig. 3 shows the pattern shapes for 1, 2, and 3
horizontal dipole stacks. The patterns for 1 and 2 horizontal dipoles
are perfectly normal and familiar. The pattern for 3 dipoles
resembles the pattern for a center-fed 1.25-wavelength extended
double Zepp doublet.
Chapter 46
Over ground, the E-plane patterns would not change shape

significantly. The elevation pattern depends upon the height of the
array above ground. Typically, an installation will adjust the
dominant elevation angle for a design frequency by adjusting the
bottom height for the array selected, which might be still larger than
the samples used here. Although literature tends to use the
average height of the array as a calculating point, the arrays
equivalent height tends to be about 2/3 the distance between the
height of the lower dipole and the height of the highest dipole. This
figure does not vary much from the array's average height, but it
does show up in vertically phased arrays (like the lazy-H) where the
lowest height may be a large fraction of a wavelength above
ground. Indeed, the lazy-H is a billboard antenna without the
billboard, although some amateurs have added screen reflectors for
increased directivity.
Adding a screen to the folded-dipole arrays that we have just

surveyed creates a directive beam antenna. As shown in Fig. 2, the
Chapter 46
recommended spacing between the dipole arrays and the screen

reflector is about 0.3 wavelength. The value is not optimal for
maximum possible gain. In fact, designing a dipole array for
maximum possible gain would require customizing each dipole
element and the array spacing for every possible combination. The
0.3-wavelength spacing provides good gain and pattern shaping
without regard to customizing the dipoles to account for their
interaction. As a test, I created a screen for each folded dipole in
the first sequence. Each screen consisted of a wire grid of standard
modeling proportions (0.1-wavelength squares with a wire diameter
that is the square side divided by PI). Each screen exceeds the
dipole array dimensions both vertically and horizontally by 0.5
wavelength. The test frequency remains 10 MHz. To the data in the
first table, I have added the 180-degree front-to-back ratio as a
measure of rearward performance.
Free-Space Performance of Various Dipole Arrays with Screen Reflectors

Front-to-Back Ratio (dB) 19.24 21.18 21.78


Maximum Gain (dBi) 11.72 13.22 14.87
Chapter 46
Because the rear lobe structure changes, the gallery of E-plane

patterns in Fig. 4 includes plots for all of the entries in the table.
Chapter 46
A number of features of the patterns call for note. The beamwidth

values do not change very much from the values without the
reflector, except that they apply only to the single large forward
lobe. The sidelobes of the versions with 3 horizontal dipoles are
better than 20 dB lower than the main lobe regardless of the
vertical stack size. A single bay consisting of 1 to 3 dipoles
arranged either vertically or horizontally has a good front-to-back
ratio. However, as soon as we add a second bay in one or the other
direction, the ratio approaches 30 dB--even for the 2V-2H version
of the array. For values over average ground with a base height of
at least 1/2 wavelength, you may add about 5 dB to the gain for a
ballpark total gain figure. The gain will slowly climb as we increase
the base height of the array, of course, moving the screen upward
with the dipoles.
One advantage that accrues to the dipole array is the ability to shift
or slew the main direction of the beam by up to 30 degrees each
way, depending on array size. Common installations employ "delay
lines" that shift the phase angle of the current for each vertical bay
of dipoles. We may simulate this effect in models simply by using a
current source and adjusting the source phase angle while holding
the current magnitude constant. Fig. 5 shows the patterns for a 1V-
2H array initially with both vertical dipoles in phase. The center
pattern uses a phase angle of 30 degrees for the first dipole and 60
degrees for the second. The final pattern uses 60 degrees for the
first dipole and 120 degrees for the second. The general rule is to
change the phase angle of subsequent vertical dipole bays by a
multiplier on the baseline phase angle according to the position of
Chapter 46
the dipole (or vertical bay of dipoles) relative to the first vertical
dipole or bay.
The first move changed the heading of the main beam by 7

degrees, and the second changed it by 13 degrees. The angles
would remain the same regardless of the size of the vertical bays in
the array. By the correct selection of delay lines, we can achieve a
relatively precise aim at a target of choice within the span of
allowable slew angles. As we increase the angle of the main beam
by these means, some distortion does appear in the form of forward
and rearward sidelobes. At the angles in the sample, the distortions
are not severe enough to void the use of slewing. However, they
show that slewing has limits. Nevertheless, for a SWBC station that
wishes to change its target from one session to the next, the
process allows the change without physically altering the antenna
or its position. Note that, when used within limits, the beam strength
and beamwidth do not change to any noticeable degree.
Chapter 46
The basic capabilities of a fixed position dipole curtain array are

quite impressive, even using the ubiquitous amateur monoband
Yagi as a standard of comparison. A 2V-2H assembly at a
reasonable height above ground would easily match a 5-element
Yagi, and delay-line slewing of the beam would permit coverage of
all of Europe from the eastern U.S without need for a rotator. If we
built equivalent dipole arrays on each side of the reflector, then we
might cover Europe on one side and the Pacific on the other, at
least from my location in the hills of Tennessee.
Broadbanding Techniques
The needs of SWBC stations are quite different from those of the
average amateur station. SWBC stations tend to use very high
power levels, up to 500 kW in some cases. Since we must provide
energy to each dipole, the use of wide-spaced parallel transmission
is fairly standard, indicating as well the use of high-impedance
antenna feedpoints. A folded dipole of conventional construction--
with equal diameter conductors throughout--goes part of the way
toward the high-impedance goal. However, if we wish to raise the
feedpoint impedance beyond about 280 Ohms, we must resort to
more unconventional techniques. For example, if we use a smaller
diameter wire for the line with the feedpoint and a much larger
diameter wire for the other line, we increase the impedance
transformation to almost any desired level within the limits of lines
to match it. We may simulate very wide second wires using pairs or
cages of wires so that the entire assembly remains lighter than it
would be with a single fat wire or tube for the second conductor.
Chapter 46
Obtaining a high impedance feedpoint does not resolve a second

goal of dipole array designers: achieving a wide operating
bandwidth. The gain of a dipole array changes slowly as we change
the operating frequency as a function of the length of the elements
relative to the operating frequency. However, being able to match
the array over an extended bandwidth requires a combination of
techniques. There is no magic to any of them, although amateurs
rarely use them in complex combinations.
The first step is to begin with a wide-band folded dipole. The AWG
#10 folded dipole used in our initial dipole array models has a 2:1
SWR bandwidth that runs from 9.6 to 10.5 MHz, a 0.9-MHz spread
(given our test frequency of 10 MHz). We need to begin with a
folded dipole array that has inherently a broader operating
bandwidth. That is step 1 in the process. Most dipole array
manufacturers have proprietary designs for their driven elements,
designs to which I am not privy. (Even if I had access to one or
more of them, I likely could not violate agreements that gave me
such access.) So I shall begin with a moderately broadbanded
driver of my own design. It will not have the full capability of some
commercial driver elements, but it will be sufficient for our small
demonstration.
Chapter 46
A fan dipole with a 3:1 length-to-height ratio is capable of

increasing the operating bandwidth over a conventional linear
dipole. However, the feedpoint impedance is about 50 Ohms. If we
create a pair of such fans, we only achieve the standard 4:1
impedance increase that is standard for a conventional folded
dipole. However, if we use a single-wire as the fed portion of the
folded dipole, the fan represents a much "fatter" second portion, for
a significant increase in the feedpoint impedance. We connect the
fed wire to the fan at the centers of the vertical sections, since that
is the pair of points on the fan with minimum current. The free-
space performance data for the arrangement is virtually identical to
the standard folded dipole, but the feedpoint impedance at
resonance is over 550 Ohms. A 600-Ohm SWR curves shows
under 2:1 SWR from 9.4 to 10.8 MHz, a 1.4 MHz spread or about
1.56 times the spread of the standard folded dipole. Like the
standard folded dipole, the folded-fan dipole is composed of AWG
#10 copper wire in all of the models that we shall consider. Fig. 7
Chapter 46
overlays the SWR plots for the standard folded dipole and the
folded-fan dipole. Each curve uses its own reference impedance.
Step 2 in the process of broadening the operating bandwidth is to

place the driver assembly ahead of the reflective screen and
determine the best distance between the two. Like previous
screens, the wire-grid structures used in this sample situation
extend about 1/2 wavelength beyond the driver limits in all
directions. Fig. 8 shows side and face views of the folded-fan
dipole and its screen. In the EZNEC graphic, I have retained the
segment and wire junction dots to lend some color differentiation to
the array pieces.
Chapter 46
Fig. 8 shows no spacing value because I examined 2 cases. The

first placed the driver 0.245 wavelength ahead of the screen. The
free-space gain was 8.34 dBi with a front-to-back ratio of 18.77 dB.
The feedpoint impedance with this spacing was close to 800 Ohms.
Increasing the distance between a planar reflector generally has 3
easily noted effects. First, it raises the feedpoint impedance of the
driver. Second, if the distance is greater than the maximum gain
position, performance gradually declines relative to both gain and
front-to-back ratio. Finally, increased spacing between the driver
and screen tends to widen the operating bandwidth of the array. By
increasing the spacing between the driver and screen to 0.3
wavelength, the feedpoint impedance rose toward 1000 Ohms.
Chapter 46
However, the maximum free-space gain dropped by 0.9 dB to 7.45

dBi, while the front-to-back ratio fell to 16.55 dB. Nevertheless, as
shown in Fig. 9, the small increase in spacing widened the 2:1
SWR bandwidth, with each array design using its own reference
impedance.
Although the curves appear similar, note the difference in the

frequency limits of each graph. At a more optimal position for array
gain, the passband runs from 9.2 to 10.9 MHz or 1.7 MHz. By
increasing the spacing, the operating passband now extends from
9.2 to 11.9 MHz or 2.7 MHz. Notice that the SWR in neither case
Chapter 46
reaches a 1:1 value. That goal is often only an amateur fetish (but
is not always a fetish by any stretch of the imagination). By
selecting an acceptable reference impedance--generally one that
reflects a transmission line that we can use with the system--we
can often attain a wider passband within the upper limits of
allowable SWR.
Stretching the operating passband in terms of SWR does not

guarantee that the array pattern will be equally usable everywhere
within the frequency limits. Fig. 10 presents sample E-plane
patterns from the wider passband, using the upper, lower, and mid-
band frequencies. Within the span of the antenna, the gain drops
from 8.04 dBi down to 5.02 dBi as we raise frequency (and the
spacing becomes wider as a function of a wavelength). The 180-
degree front-to-back ratio tends to be stable in the 16-18-dB region.
However, as we raise the operating frequency, the beamwidth
broadens, especially toward the upper passband limit. At 11.9 MHz,
the pattern shows twin peaks, although there is no noticeable null
between them. However, for some applications, the beamwidth
may have become too wide to meet operating criteria.
Chapter 46
Our step-2 exercise has increased the frequency range for

allowable operation. In the process, the exercise has also shown us
that not every frequency that we can use is one that we can use
well.
The third step on the road to broadening the passband of a dipole

array involves what happens when we phase-feed more than 1
driver. For this rudimentary demonstration, I set 3 folded-fan
dipoles (without a screen) at vertical intervals of 1/2 wavelength.
The center dipole serves as the fed driver relative to the main
transmission line. Each outer driver receives energy from a 1/2-
wavelength transmission line connected to the center driver. The
three drivers are now roughly in parallel. Hence, we can expect a
reduction in both the resistance and the reactance at the feedpoint.
However, the drivers interact with each other. Outer drivers

essentially interact with only one other driver, and mutual coupling
shifts the feedpoint impedance of each of them by like amounts.
Chapter 46
However, the center driver mutually couples to both outer drivers

and shows a different shift in impedance from the value it would
have if used in isolation. By judicious sizing of the drivers we can
overcome the impedance difference. However, let's size them in
concert, that is, make them all the same size. The left portion of
Fig. 11 shows the set-up in outline form.
Chapter 46
Because the impedances on the outer drivers will be increasing for

part of the frequency sweep while the center driver impedance
decreases--and vice versa--we obtain an additional increment of
passband broadening. Fig. 12, at the top shows the new passband,
which even without a reflective screen extends from 9.55 to 11.8
MHz or 2.25 MHz. The reference impedance for the curve is 250
Ohms.
Chapter 46
The bi-directional maximum gain of the drivers across the operating

passband increases from 7.17 dBi to 8.03 dBi, with well-behaved
lobes. Therefore, the driver set seems fit for combining steps 2 and
3, that is, using a phased set of 3 drivers ahead of a screen. The
right portion of Fig. 11 shows the ultimate array (at least for our
demonstration) in outline form. The lower portion of Fig. 12 reveals
the 300-Ohm SWR passband, which now extends from 8.35 up to
11.5 MHz or 3.15 MHz. The following brief table samples the
modeled free-space performance values at each ends of the
passband and at the mid-band frequency.
Modeled Free-Space Performance of a Vertical Stack of 3 folded-Fan Dipoles
0.3 WL Ahead of a Screen Reflector
Frequency (MHz) 8.35 9.925 11.5
Maximum Forward Gain (dBi) 11.99 11.86 10.99
180-Degree Front-to-Back Ratio (dB) 19.81 20.15 20.00
E-Plane Beamwidth (degrees) 58.9 68.8 95.8
Except for not showing relative gain values, the patterns in Fig. 13
put a graphic face on the data in the table.
Chapter 46
The array patterns show the same sort of development that we

found in Fig. 10 for a single driver and screen. At the top of the
passband, we find dual peak gain levels, but without a noticeable
null between them. We also find a rapid rise in the beamwidth
above the mid-band frequency--nearly 30 degrees. Although the
same cautions apply to the expanded array as also apply to the
single driver array, we should note that the operating passband is
now about 50% greater.
Our goal has been only to demonstrate that obtaining a wide

operating passband from a dipole curtain array requires a
combination of ingredients or steps, as we have called them. The
folded-fan dipole is far from being the ideal starting point for
effecting the scheme, although it has served well in the
demonstration. Some manufacturers of dipole screen arrays claim
up to a 2:1 frequency range while also listing tighter SWR limits in
their specification sheets. However, all specification sheets are
incomplete performance records, so we have no idea of whether
the patterns associated with frequencies across a given frequency
range are all usable to the same degree as they are at mid-band.
Regardless of the performance obtained by any given maker, we

have seen that through careful design (even at the crude level used
in this exploratory account), a dipole array with a screen reflector
lends itself to broadband service with gains that one might tailor by
selecting the correct array size. A planar reflector, whether solid or
composed of lines or screens, offers considerable flexibility in
system design. If we give up the amateur habit of always seeking
Chapter 46
the highest gain from the fewest elements, we can achieve a

number of other advantageous performance features in our arrays.
Back to the Billboard
Modern screen or curtain arrays employ dipoles. However, in the

dim recesses of past times, the original standard driver billboard
array was one or more sets of center-fed 1-wavelength elements,
also called collinear half-wavelength elements. We neglected to
test this arrangement to see if it offers any advantage over the use
of dipoles. Let's pause before closing to see what happens if we
replace dipole drivers with the 1-wavelength drivers.
To gain some perspective on the question, let's use a 2V-2H dipole

driver set as one comparator. The corresponding billboard array
would employ a 2V-1H driver set. Both driver arrays would have the
same dimensions and require the same size reflector screen. In
fact, we may also use the same spacing between the drivers and
the screen for both antennas. The top portion of Fig. 14 shows the
two outlines.
Chapter 46
Between the two types of driver arrays we find under 0.1-dB

difference in gain and only 1 degree difference in beamwidth. The
front-to-back ratio of both arrays is near 28 dB. As the sample E-
Chapter 46
plane patterns show in Fig. 14, Nothing in either pattern gives one
or the other array an advantage.
We have already explored some ins and outs of dipole impedance

behavior in a planar-reflector array. The older billboard system
shows individual feedpoint impedances that are very high--in
excess of 3000 Ohms resistance. For a single frequency, the high
impedances are no hindrance to the use of 1-wavelength elements.
However, if we wish to change frequencies, we may encounter a
need to retune the system. The center of a 1-wavelength element is
a region of very large and rapid changes of impedance values as
we shift the operating frequency. Unless all wires composing the
system are very taut, storm winds may result in some impedance
oscillations.
The much lower impedances of dipoles--even raised to several

hundred Ohms--provide the array designer with a much more
controllable situation with respect to impedance matching over a
wider frequency range. The conversion to dipoles as drivers has
made the dipole screen array a mainstay of short-wave
broadcasting by providing the necessary gain and beamwidth
(including slewing) with the ability to handle high power levels.
Increased operating bandwidth supplies the final need of the SWBC
industry.
In these notes, my aim has been to examine some of the features

of an antenna type that we often encounter on our travels across
the U.S., but seldom have occasion to use personally. The maze of
wires--whether active antenna parts or support guys--gives these
Chapter 46
antennas an air of mystery--or at least considerable puzzlement.

(Of course, not all mazes of wires strung between 2 tall support
masts are dipole curtain arrays.) Hopefully, this small set of
exercises has taken some of the mystery out of the arrays.
In the course of developing these notes, I have used published

information on the antenna type, abetted by some first-order free-
space modeling. Hence, my slant on certain features may differ
markedly from the perspective brought to bear by an antenna
engineer deeply involved in the design and implementation of such
arrays. Indeed, I may well have overlooked numerous features that
a manufacturer might consider critical and stressed others
considered marginal or even trivial. Still, I hope that these notes
contain enough analytical information to make the antenna type--
the dipole curtain array--more familiar to and understandable by
those who see one for the first time.
Chapter 46
Chapter 47: The 2-Element Vertical Phased Array
I
n July, 2006, Pete Millis, M3KXZ, published his design of an
array of 2 vertical antennas that he calls "'No-counterpoise'
antenna: 2-element phased array.
The antenna is interesting in several respects. First, it uses a very

simple structure and common materials that you can obtain from
Radio Shack and hardware sources. Pete uses speaker wire and
PVC supports for the vertical elements. Perhaps the only
specialized antenna items are the baluns that he winds on ferrite
cores and the encased 4:1 balun he uses at the center of the 2-
element version of his array. However, we shall have occasion to
evaluate the need for these items as we look into the antenna.
The second significant aspect of the antenna is its performance. A

single element length covers a spread of bands, for example, 20
meters to 6 meters using a total length of 25'. The normal limit for
either a 1/4-wavelength monopole or a 1/2-wavelength dipole is
about 2.5:1, which would suggest a cut-off of about 35-36 MHz, if
the original antenna is cut for 14 MHz. Once a monopole exceeds
about 5/8-wavelength or a dipole exceeds 1-1/4-wavelengths, the
main radiation is no longer broadside to the wire. For a vertical
antenna, the long lengths result in very high angle radiation, rather
than the low angle radiation that we normally need. However, the
M3KXZ antenna and array yield very usable patterns from 20
through 6 meters. In addition, the gain of the antenna is close to the
gain available from either vertical dipole/doublets or from elevated
monopoles with radials on all bands.
Chapter 47
For these reasons, it seems that the antenna in both its 1-element
and 2-element versions deserves a closer look, if only to
understand its operation better. As well, if one wanted to replicate
his antenna using different materials, we shall need to look at some
of the pieces in his arrangement.
A Frame of Reference
As a basic for evaluating the behavior and performance of the

M3KXZ antenna and array, let's first catalog comparable data for a
more familiar antenna, the straight vertical wire element. Since we
shall look at the 20-6-meter version of the M3KXZ antenna, we may
cut the wire for 20 meters. We shall use AWG #12 copper wire and
place the lower end 1' above the ground, the height that we shall
use for the other elements. However, a straight wire that is vertical
requires a top height of 34.6'. For our basic work, we shall use
average ground as the soil throughout.
We have 3 choices for feeding our vertical wire. We might select

the center, which would be natural for a 1/2-wavelength dipole. Of
course, the wire becomes a doublet as the length grows longer
than 1/2-wavelength and the current peak and voltage minimum no
longer occur at the center of its length. Alternatively, we might
select a feedpoint based on the M3KXZ design, that is, a position
2/3 of the distance down from the antenna top, considering the
12.5' fold back in the M3KXZ design as the lower 1/3 of the
antenna. Finally, we might place the feedpoint at the lower end of
the antenna.
Chapter 47
Table 1 summaries some of the key performance data for each of

the three versions of the vertical wire at the center of all amateur
bands from 20 through 6 meters.
As we would expect, the feedpoint impedance values (Feed R and

Feed X) differ widely among the antenna versions. More significant
for eventual comparative purposes is the performance of each
version. None of the three can sustain a low elevation angle for the
main radiation lobe across the range of bands covered by the
survey. Moreover, we find differences within each band depending
on the feedpoint position, and the differences involve more than
small changes in the maximum gain. For an example, we may use
24.94 MHz. The center-fed version produces a main lobe at 13
degrees elevation. The off-center-fed version's main lobe is at 40
degrees, while the end or bottom-fed version main lobe is at 39
degrees elevation.
The radiation pattern differences result from differences in the

current magnitude distribution along the wire on this band--and on
any other band where the wire is longer than 1/2 wavelength. Fig. 1
shows the differences that occur on 12 meters. Most evident is that
Chapter 47
the current minimum occurs ever lower on the structure as we

move from center feeding to bottom-end feeding. As well, but to a
lesser degree, the differences in the current curve between the off-
center-fed and the end-fed version play a role in the ultimate shape
and strength of the pattern lobes for the vertical wire.
Only the center-fed version of the straight wire doublet manages to

cover 20 through 10 meters with the man lobe at a low elevation
angle. The other versions give way to having their main lobes at
Chapter 47
higher and generally undesired elevation angles well under 10

meters. Despite our interest in the radiation patterns, we shall also
discover that the impedance columns of Table 1 will hold
importance as we attempt to see what lies behind the behavior of
the M3KXZ antenna.
Some M3KXZ Antenna Basics
The full 2-element array appears in Fig. 2 in outline form. I have

selected the 20-6-meter version for a detailed look. A single-
element version of the antenna would simply omit the second
element and the two phase-lines marked TL1 and TL2.
Chapter 47
In modeling the antenna, I departed from the original, which uses

speaker wire. To yield adequate models, I spaced the AWG #12
copper wires 1" apart in the lower half. Parallel wires at this
separation have a transmission-line impedance of about 400-450
Chapter 47
Ohms. Although I have seen no tests of insulated speaker wires,

the characteristic impedance of a pair is likely to fall into the 75-
100-Ohm range, due to both the spacing and the relative
permittivity (dielectric constant) of the insulation between them. In
addition, Pete twines the wire along a length of PVC for support,
but without introducing any significant inductance. Although we can
view his elements as essentially straight, we should understand at
the outset that all models will be only approximations of his
antenna.
We should also enter a modeling caution to those who may wish to

replicate the models used in this study. The separation between the
long and the short sections of the element is 1". Although one
would normally use 1" segment lengths for the remainder of the
model, there is a different overriding consideration. Very closely
spaced wires in NEC are subject to errors, even when we precisely
align the segment junctions. In order to obtain a fair set of
comparisons between the M3KXZ element and the straight wire
element, it is necessary to adjust the segmentation to obtain an
average gain test (AGT) score that is as close to 1.00 as may be
feasible. For the models used here, 120 segments in the long
section and 60 segments in the short section produced an AGT
score of 1.004, indicating that gain and impedance values will be
very much on a par with those drawn from the straight wire element
with its essentially perfect AGT score. AGT values below near-
perfect will yield low gain and high impedance reports, while AGT
scores above near perfect will yield values that are too high for the
gain and too low for the impedance. For very closely spaced wires,
the segmentation density alone is enough to yield gain values up to
Chapter 47
1.5 dB off the mark. Hence, close attention of the model's AGT
score is essential, especially when comparing the performance of
models have different geometries.
Before we turn to the full phased array, let's see what we might
obtain from a single M3KXZ element. Table 2 lists the NEC-4
reports from the model, which places the 25' element at a height of
1' above ground. I placed the antenna over a range of soils from
very good to very poor in order to determine if the soil quality had a
significant bearing on performance, given the close proximity of
ground to the lower end of the element.
If we examine the gain and TO angle columns of Table 2 under

average soil, we discover that on all bands through 10 meters, the
M3KXZ element yields competitive gain values and TO angles that
are only slightly worse than those we gather from the center-fed
straight wire. The M3KXZ element TO values are slightly higher
largely because the top height is about 30% lower than the top
height of the center-fed doublet. Nonetheless, all of the TO angles
Chapter 47
shown in the table are suitably low, although they do vary with the
ground quality.
There is an incidental but interesting pattern to note. We normally

think of ground losses as increasing as soil quality decreases so
that as we move toward bad soils, the gain of a vertical element
decreases. However, this thinking has a frequency limit. The old
thinking applies in small amounts from 20 through 15 meters.
However, on 12 meters, maximum gain occurs over average soil.
Above 12 meters, maximum gain occurs over very poor soil. The
trend reversal is accompanied by a shrinkage in the differential in
gain as we change soil, but the reversal is quite real.
Soil quality changes do not make a large difference in the feedpoint

impedance at the antenna base, where the short and the long wires
meet. However, the range of feedpoint impedances is considerable.
Hence, the use of a coaxial cable between the antenna base and
the antenna tuner may prove a considerable loss source unless the
length is very short. With the possible exception of 10 meters, all of
the impedances fall within the easily matched range of a remote
antenna tuner place at the element feedpoint. Otherwise, the use of
parallel feedline--suitably elevated from the ground to prevent
unwanted coupling--may be necessary. However, the very low
impedance on 20 meters may incur some losses even with parallel
feedlines.
Chapter 47
Chapter 47
Fig. 3 provides a gallery of elevation and azimuth patterns for the

bands covered by the array. Although the azimuth patterns all
appear to be quite circular, note the shifting angle of the line that
indicates the bearing of maximum gain. Any antenna with a fold-
back--including the well-known J-pole--will exhibit at least a slight
pattern distortion due to radiation from the two wires in the fold-
back region. The closer that we space the wires, the less will be the
distortion, and with the 1" spacing, the differential is never more
than about 0.03 dB. However, as the shifting line bearings show,
the distortion will change a bit from one band to the next. The fact is
not operationally significant, but will prompt some further
investigation.
The elevation patterns reveal one of the most essential aspects of

the antenna's performance, the well-behaved radiation at low
angles with very little higher-angle radiation until we reach 52 MHz.
At the upper limit of the operating spectrum, the second elevation
lobe nearly equals the gain of the lower lobe over average ground,
and over very good soil, the second lobe at an angle of 50 degrees
is actually stronger. However, the low-angle gain remains
serviceable. For comparison, a straight vertical dipole fed 1/3rd of
the way up its length shows a considerable upper-angle lobe on 15
meters, and at 12 and 10 meters, the higher second lobe
dominates the pattern.
The well-behaved patterns are one of the effects of the 12.5' fold-
back. That fold-back is not just a convenient way of feeding the
antenna at a point 1/3 of its total length (25' plus 12.5'). For
example, if we feed a 20-meter vertical dipole at the 1/3rd point, we
Chapter 47
obtain resistance values ranging from 100 to 3100 Ohms, and

reactance values from -600 to +700 Ohms. The values shown in
Table 2 are far tamer than they are for an off-center-fed straight
dipole--or for any of the other versions of the straight-wire element
in Table 1.
In fact, the fold-back forms a transmission line section that is 12.5'

long. Whatever impedance appears at the junction of the single-
wire top section and the beginning of the double-wire section
undergoes a transformation according to the electrical length of the
double section and its characteristic impedance. Note once more
that modeling requirements have dictated a 1" spacing, and that the
characteristic impedance is not the same as it would be for the
speaker wire. Hence, the impedance numbers in Table 2 are only
representative. As well, they do not account for the transmission-
line velocity factor of the insulated speaker wires used in the
original.
Fig. 4 presents a collection of current magnitude distribution graphs

taken from the EZNEC models. At the far right, I have expanded
the 20-meter graph in order to show that the current magnitude
undergoes a small but noticeable shift at the point where the top
single wire meets the double-wire section. The jog indicates that
below the junction, the graph is showing a combination of both
radiation and transmission-line currents.
Chapter 47
The remainder of the current magnitude graphs show that we

should not try to apply a simplistic J-pole or end-fed Zepp model to
the situation. As the patterns show, the junction between the top
and bottom sections does not occur at a maximum voltage,
minimum current point on any band. As well, the double-line length
is not 1/4-wavelength on any band, although it comes close on 10
meters. Therefore, the impedance transformation differs for each
band in terms of both the impedance at the section junction and the
amount of transformation that occurs in the lower section. One
might use a number of means to roughly calculate the start and end
values for the transformation, but given the higher characteristic
impedance in the model relative to the speaker wire used in the
original, such an exercise might prove to be operationally useless.
For any given installation, the most practical effort is to measure the
impedance at the feedpoint for every planned frequency of
operation.
Chapter 47
One consequence of the lower double-wire or transmission-line

section is that the dominant radiation currents do not follow the
patterns that they would in a straight dipole/doublet, whatever the
feedpoint. Hence, M3KXZ has found essentially an antenna
designer's grail or silver bullet: an arrangement of wires that
extends the range of desirable pattern formation beyond its normal
limits while sustaining good gain for an antenna of its type and
leaving quite workable feedpoint impedance values.
The 2-Element M3KXZ Phased Array
The 2-element version of the M3KXZ antenna, shown in Fig. 2,

consists of 2 elements connected by equal lengths of a
transmission line, with a common junction for connection to the
main feedline. For the 25' version of the antenna, intended to cover
20 through 6 meters, the spacing between elements is 10'. The
spacing is not accidental, since at 52 MHz, it represent about 0.53-
wavelength, the maximum that we would wish to space phase-fed
elements.
Since we shall ultimately feed the antenna at a center point

between the two elements, we have two choices of phasing. We
may connect the two lines so that each long-section wire goes to
the same side of the junction for in-phase feeding. Alternatively, we
may give one (and only one) of the two lines a half twist so that
connections to long-section wires go to opposite junction points and
thus end up with out-of-phase feeding. The original design used
plugs and jacks at the center junction box to allow a quick reversal
of the junction connections. A remote switch might achieve the
Chapter 47
same goal with control transferred to the equipment location. Fig. 5

shows the differences in the current distribution curves that result
from the alternative feedpoint connections.
In-phase feeding of the two elements results in a broadside azimuth

pattern relative to the plane of the two elements, comparable to the
patterns that we obtain from converting a lazy-H into a standing-H.
The gain yielded by the pattern over the gain of a single element is
a function of the narrowing beamwidth in the plane of the array. The
elevation pattern is not materially affected by the dual feed. How
Chapter 47
much gain increase and beamwidth reduction we obtain is a

function of the spacing between the elements measured in
wavelengths at the frequency of operation. Gain increases slowly
from virtually single-element performance at very close spacing to
maximum with a spacing that is just over 0.5-wavelength. With an
exact 0.5-wavelength spacing, the azimuth pattern is a perfect
“figure-8.” Gain continues to increase with slightly wider spacing,
but small sidelobes develop in the plane of the elements. Above
about 0.55-wavelength spacing, the sidelobes grow so fast that the
gain broadside to the antenna decreases.
Chapter 47
Chapter 47
Fig. 6 provides a gallery of elevation and azimuth patterns for the

model of the 25' M3KXZ array with the 10' spacing between
elements. Note that the elevation patterns, taken broadside to the
plane of the two elements, do not differ significantly from the single-
element elevation patterns. The azimuth patterns do not show
relative gain values. Instead, each pattern uses the outer ring as
the pattern limit to reveal more clearly the pattern shapes. If we
look at the azimuth pattern for 52 MHz, we can see the beginnings
of the sidelobes that develop as a result of the 0.53-wavelength
spacing between elements. However, at lower frequencies, the
spacing between elements is considerably less than the 0.5-
wavelength ideal. As a results, as we move down in frequency, the
patterns become more circular, indicating both broader beamwidth
values and lower gain values. At 20 meters, where the spacing is
only about 0.14-wavelength, we should expect--and we obtain--very
little gain increase over a single element.
Table 3 provides 2 sets of performance values, both sets taken

over average ground. The left-most columns give the modeled gain
and TO angle for the patterns in Fig. 6. You may compare these
values to the values obtain for a single M3KXZ elements over
average ground in Table 2. At 20 meters, the gain advantage only
about 0.6 dB due to the close spacing of the in-phase-fed elements.
At 10 meters, the gain advantage of the 2-element array increases
to about 1.9 dB as the spacing increases to nearly 0.3-wavelength.
At the more nearly ideal spacing on 6 meters, the gain advantage
jumps to about 4.4 dB, with a commensurate decrease in the
beamwidth of the two broadside lobes.
Chapter 47
The table's center columns provide modeled values for the

performance of the array when fed out-of-phase by giving one of
the two equal-length lines a single half-twist. The method of feeding
is a simple evolution from the W8JK array that we usually see in
horizontal form. The same principles apply. The antenna becomes
a bi-directional endfire array with the main radiation in the plane of
the elements. Fig. 7 provides a gallery of elevation and azimuth
patterns as they apply to out-of-phase feeding of the two elements
with their 10' spacing. The elevation patterns are taken in the plane
of the elements.
Chapter 47
Chapter 47
Except for the 52-MHz plot, the azimuth patterns all show
comparable beamwidths. If we compare the out-of-phase gain
values to the in-phase values, we find a sudden jump and a leveling
off, so that we show only a slow rise in gain as we increase the
operating frequency. The behavior of an end-fire out-of-phase fed
array differs considerably from the in-phase-fed version. Two
general rules apply. First, the closer the element spacing, the
higher the gain will be over a single element. This trend gives
precedence to 20 and 17 meters, where element spacing is closest.
Second, the gain advantage increases as we increase the element
length relative to an initial length. This trend gives precedence to
the higher frequencies, where the M3KXZ elements are electrically
longer. The broadening of the beamwidth at 52 MHz suggests that
at higher frequencies (and therefore longer element electrical
lengths), the pattern will break into 4 lobes, ruining the bi-directional
characteristic of the antenna with out-of-phase feeding. The net
result of combining the two trends is a much tighter grouping of
gain figures for the out-of-phase version relative to the broadside
in-phase version of the array.
The purpose of switching the phase of feeding is to obtain

maximum possible gain from two fixed vertical elements in the
direction of the signal. An in-phase/out-of-phase switch at the
junction of the two lines marked TL1 and TL2 in Fig. 2 provides a
means of switching the axes of the bi-directional beaming. The
beamwidths of the lobes in each version are complementary so that
little or none of the horizon is excluded from performance equal to
or better than the omni-directional patterns of a single element. Of
Chapter 47
course, in the 2-element phased version of the M3KXZ array, the

elevation patterns retain their low TO angles.
Thus far, the M3KXZ array displays considerable ingenuity in

providing low-angle vertically polarized radiation over a wide
passband. However, the array faces one final challenge: feeding
the system and matching the junction impedance for either phase
condition to the equipment. Here, we can use the modeled
elements only with great caution. M3KXZ constructed his entire
system with speaker wire, an inexpensive but dubious choice for
outdoor durability. Speaker wires generally carry no rating for
performance under the summer-to-winter weather extremes, and so
the quality of such wires will vary from one maker to another,
depending upon the quality of the insulation. In addition, such wires
carry no rating for their characteristic impedance at RF frequencies.
However, similar wires with standard "poly-plastic" insulations
usually show an impedance in the 75- to 100-Ohm range. Due to
modeling limitations, we have had to use a lower or double-wire
section composed of bare wires 1" apart, for a characteristic
impedance in the 400- to 450-Ohm range. Therefore, the
impedance transformation that occurs in this section from the
junction with the single-wire upper section and the end feedpoint
will differ from the transformation obtained in the original version.
Adding phase lines to a central junction is another matter. We may

sample a variety of lines, that is, a range of characteristic
impedance values, by using the TL facility within NEC. The line
impedance will not change the antenna radiation pattern, but it will
change the impedance that we obtain at the junction of the two
Chapter 47
lines under each of the phasing conditions. In creating a survey of

values, I have simply used a velocity factor of 1.0 for two reasons.
First, the element feedpoint impedance values are already off their
marks if we use a double-wire section with a difference
characteristic impedance. Second, common feedlines tend to come
in several versions, each with a specific velocity factor.
Nevertheless, we can obtain a view of the type and size of the

feedpoint impedance challenge using two 5' lengths of transmission
line to a central junction. For the survey, I selected characteristic
impedance values of 50, 75, 125, and 300 Ohms. The 50-Ohm
value is for the most common variety of coaxial cable. 125-Ohms is
the value for RG-63, a very useful but often overlooked cable. 75-
Ohm covers both some common coaxial cables and so-called
twisted pairs of insulated wires. 300 Ohms, of course, applies to
common TV-type parallel feedline. Table 4 provides the results of
the survey, where R and X are the impedance components at the
junction of the two cables (TL1 and TL2) under each phasing
condition.
Chapter 47
The number of cases in which we obtain very low impedances,

regardless of the characteristic impedance values for the phasing
lines, raises a strong question about using a 4:1 balun at the
junction of the two lines. Converting an impedance that is already
well below 50 Ohms down to an even lower values seems to make
little sense, especially if one uses the 50-Ohm main feedline shown
in the originator's sketches. 4:1 baluns come in numerous designs,
some of which may prove to suffer losses when used with high
reactive components or when used outside the range of their
Chapter 47
winding's characteristic impedance. The result may be artificially

favorable impedance values at the terminals that may disguise
what is actually occurring within the device. There are few values in
Table 4 that would benefit from even an accurate 4:1 downward
impedance transformation.
In addition, the author uses 1:1 baluns supposedly to force equal

currents on to the short and long sections. Actually, the current
imbalance on each side of the feedpoint is part of what allows the
array to achieve its broadband characteristics. A balun or choke
might be more applicable at each element feedpoint if TL1 and TL2
are both coaxial cables, where transmission-line currents are inside
the cable between the center conductor and the inner side of the
braid, and common-mode currents are on the outside of the braid.
Indeed, parallel transmission lines may not be the most ideal phase
lines for the array.
The author also correctly notes that he obtains good impedance

matches between whatever impedance the array presents at the
junction--after transformation down the 50-Ohm main feedline--via
his antenna tuner. We shall bypass any losses incurred by the
impedance mis-match between the cable's impedance and the load
impedance at the phase-line junction. The low-impedance loads, if
transferred to a tuner, present challenges of their own.
Depending upon the type of network used, low-impedance loads

sometimes result in acceptable but imperfect impedance matches,
where the best obtainable SWR value at the tuner is perhaps 1.3:1
to 1.6:1. These conditions generally indicate a limit to the range of
Chapter 47
the components within the tuner relative to the impedance at the

terminals. In the table, note that many modeled values show much
higher reactive components than resistive components. Although
the match is acceptable, the efficiency of power transfer may not be
as high as we too often presume. If the network has a low loaded or
operating Q (Terman's "delta" term from the 1940s), we may find a
difference in the settings required for resonance (that is, for zero
reactance), for maximum power transfer, and for impedance
matching. As the circuit's delta increases to about 10, these
settings resolve to a single point. However, so long as tuners
continue to lack any form of relative output indicator, we cannot
easily tell if the impedance match that we obtain is also the point of
maximum efficiency.
The junction of the two phasing lines is a balanced feedpoint.

Ideally, the array might be fed at that feedpoint by a remote,
balanced, weatherproof ATU with a very wide range of impedance
matching capabilities. Such tuners are not generally available,
although we can press existing components into service. In
general, placing the tuner at the balanced feedpoint that joins the
two phase-lines allows the use of a 50-Ohm cable to the equipment
with minimum loss. At the input to the tuner we likely should install
a common-mode current attenuator, such as an unun or a ferrite
bead choke. System ground should occur at the equipment side of
the unun or the choke, not either at the tuner output or at the tuner
input. (The tuner input is likely to have a common ground system
with the output, and we would want the tuner to "float." Grounding
the braid of the coax at the equipment side of a ferrite bead choke
would provide for static discharge.)
Chapter 47
Conclusion
The M3KXZ antenna and array constitute an ingenious

arrangement of element parts that achieves low-angle vertically
polarized radiation over an extended operating bandwidth that
common configurations cannot match. The 2-element version of the
array offers some gain and pattern shaping for bi-directional
operating. Even with improved materials designed for both RF
service and durability through seasonal weather cycles, the
antennas are inexpensive. Moreover, they are relatively short for a
given frequency range, adding to their neighborhood acceptability.
The challenges presented by the antenna and the array revolve

around the matching and the feed system. Increased attention to
these details may result in a very serviceable, wide-band vertical
array.
Chapter 47
Chapter 48: The Terminated Vee-Beam and Rhombic
T
he terminated longwire antenna is perhaps the simplest of the
large terminated wire arrays, but it is not the best performer.
In this session, we shall look at two arrays (other arrays to
follow). The long, terminated Vee-beam provides considerably
more gain for the same length legs, but does not have the
longwire's ease of feeding over a large frequency spread. The
rhombic, in contrast, provides gain even over the Vee-beam and
allows multi-band coverage. However, it is perhaps the most
complex of the terminated arrays.
Designing either type of array for maximum performance is not

simply a matter of stringing hundreds of feet of wire. Behind each
type of array are a number of design equations that take into
account the length of the legs, the angles formed by the wires, and
the antenna height (or desired elevation angle of maximum
radiation). These notes only survey some of the potential of these
wire antennas, but are not sufficient for designing an effective array
for your own farm. Our goal is to provide enough information so that
you may decide whether further study into the designs is a
worthwhile project.
The Vee-Beam: Do not confuse the Vee-Beam with some of the

small Vee-shaped antennas whose total wire length is only about
1/2-wavelength. We do not arrive in Vee-beam territory until each
leg is several wavelengths long. Like the longwire antenna, the
Vee-beam depends upon the fact that as we make a wire longer
Chapter 48
and longer in wavelengths, the main radiation lobe moves from

broadside to the wire to a position nearly in line with the wire.
Fig. 1 shows two ways of designing a terminated Vee-beam. One

uses terminating resistors that operate like those in the simple
longwire. The other uses a cross wire that includes the terminating
resistor. Just as in the last episode, the terminating resistor must be
a non-inductive element with the ability to dissipate about half of the
power supplied to the antenna. Both versions of the Vee-beam use
the same principle of operation.
To design a Vee-beam, we must know the planned length of the

element legs in wavelengths. Since the length of the leg determines
the angle of the main lobes relative to the plane of the leg, the
length of the legs also determines the angle that we must use
between the two legs for maximum forward gain. The goal is to
Chapter 48
align the two legs so that a main lobe from each wire combines to
form a single large forward lobe. See Fig. 2.
The plots show the patterns for the left and right legs of a planned
Vee-beam with 7-wavelength legs. The height of the array is 1
wavelength above average ground. When we combine the two
legs, we obtain the pattern at the right. For the leg lengths that we
chose, the main lobes are about 15 degrees off the plane of the
wire. By making a Vee with a 30-degree total apex angle, we
approach the maximum gain of which the antenna is capable.
Obviously, had we selected longer legs, we would have used a
narrower apex angle, while shorter legs would have called for a
wider angle. If we do not match the angles and the length of the
legs, we shall obtain inferior performance.
With 7-wavelength legs, the Vee-beam yields a maximum forward

gain over average ground of about 13.5 dBi. This value is
considerably higher (by about 5 dB) than the simple longwire
Chapter 48
antenna. However, the Vee-beam is not capable of achieving the

high front-to-back ratio that the terminated longwire gave us. If you
examine the two left patterns in Fig. 2, you will see that each has a
rearward lobe that is less than 10 dB down from the favored lobe.
In the Vee configuration, these lobes also add, giving the Vee-
beam a significant rearward lobe that remains only about 10 dB
down from the forward lobe. Like all arrays composed of wires that
are several wavelengths long, the patterns will be filled with
sidelobes.
Like the longwire antenna, the Vee-beam is capable of good use

with or without its terminating resistors. Fig. 3 gives us comparative
patterns for the two versions, using the same legs. Only the
presence or absence of the terminating resistors marks the pattern
differences. Note that the unterminated Vee-beam has (like the
unterminated longwire) about 2 dB more gain. However, it is more
truly bi-directional (than the longwire), since the rearward radiation
is down by only about 2.5 dB.
Chapter 48
The unterminated Vee-beam is also a useful antenna. It does not

require that we obtain suitable terminating resistors (400 Ohms
each in the sample antenna). Therefore, only land, wire, and
supports stand between us and an array of Vee-beams that can
cover the horizon. Fig. 4 shows the required number of Vee-beam
legs needed, if we use the 7-wavelength legs and a 30-degree
apex angle between legs.
Chapter 48
We need outer-end (and perhaps intermediate) wire supports for

each leg. However, we need only a single inner-end support for all
of the legs. By a suitable means of switching--either at the antenna
or near the shack, to which we bring a circular array of feedline
wires--we select the adjoining pair of legs that gives us radiation in
the two directions that we want.
The Vee-beam has some limitations. It is essentially a 1-band

antenna, although we can press it into service on other bands.
Chapter 48
However, as we move away from the design frequency, the legs

change their length as measured in terms of a wavelength. That
change moves the angle of the main lobe on each leg relative to
the plane of the wire. Hence, on these distant frequencies, the
lobes will not match to form as strong a main lobe. To sustain the
higher gain and re-acquire the broadband characteristics, we need
to use a different shape.
The Rhombic: The terminated rhombic antenna employs two sets

of Vee-beams joined at the outer ends so that the second one
forms a distant apex angle at which we install a terminating resistor.
Fig. 5 shows the general outlines of a rhombic. Each leg has a pair
of main lobes, and the combination of the 4 legs produces a very
strong forward lobe.
Chapter 48
The figure also shows in the lower half two ways of building a
rhombic using either single-wire legs or triple-wire legs. Builders
have reported improved performance with the three wires, although
the 1-wire version is satisfactory for most amateur installations.
Chapter 48
The rhombic is a venerable directional array for which design

equations had been developed in the 1930s. The design equations
take into account the elevation angle of radiation, as well as the
proper combination of angles and leg lengths to produce a strong
forward lobe and a good front-to-back ratio. Indeed, there are
alternative equations for developing various compromise designs
that combine antenna height, leg length, and angles in various
ways. See John Kraus, Antennas, 2nd Ed., pp. 503-508, if you are
truly interested in designing a rhombic that will fit your yard.
For many years, The ARRL Antenna Book has featured an

interesting rhombic design suitable for use on the amateur bands
from 20 through 10 meters. It employs a 600-Ohm terminating
resistor and is a good match for a 600-Ohm transmission line.
Hence, all matching can be done at the shack with a system of
impedance-matching transformers or baluns or with an antenna
tuner. Fig. 6 provides a 600-Ohm SWR curve across the operating
span to show the relatively good match between the terminating
resistor and the feedpoint impedance.
Chapter 48
The rhombic is 377.5' long and 184' wide, with a 52-degree angle at
both the feedpoint and the termination end. Of course, we shall
require longer wire, since each side of the array requires about 420'
of wire. If we set the rhombic at 70' above average ground--1
wavelength at 20 meters--we can anticipate the following
performance figures.
Sample Rhombic Modeled Performance
Freq. Length Height Gain TO Angle Front-Back B/W Feed Z 600-Ohm

MHz WL WL dBi degrees Ratio dB degrees R+/-jX Ohms SWR
14.2 5.5 1 16.2 14 17.1 17 810 + j 60 1.4
18.12 7 1.25 17.8 10 15.3 13 1010 - j200 1.8
21.2 8.1 1.5 18.4 9 19.2 11 830 + j 60 1.4
24.95 9.5 1.75 18.3 7 15.2 9 990 - j 80 1.7
28.3 11 2 17.2 6 20.2 7 900 + j 40 1.5
Note that the array is optimized for 15 meters, where it shows the
highest gain. However, performance is high on all of the bands.
However, the array is not without some important limitations.
Chapter 48
First, the array is fixed in position. We cannot re-aim it easily, if at

all. We may combine this restriction with the second limiting factor:
the beamwidth of the rhombic is very narrow compared to most
wire arrays. The horizontal beamwidth, as measured to the half-
power or -3-dB point is narrower than almost any other array. Fig. 7
provides a band-by-band view of the azimuth patterns of the array
to provide a sense of how narrow the beamwidth really is.
The earliest uses of the rhombic involved point-to-point

communications circuits and well-defined broadcast target areas.
Chapter 48
Contemporary use of the array should have some of these

elements as part of the communications goals before deciding to
erect a rhombic.
Rhombics have also seen use in the VHF range as outdoor

television receiving antennas. More recent improvements in design
techniques have produce double-loop versions that further
suppress the side lobes that naturally occur with multi-wavelength
element legs. Some amateurs have used these principles on bands
as high as 1296 MHz.
The rhombic represents perhaps the pinnacle of refinement of large

wire arrays. More recent developments in steerable arrays have
largely supplanted the rhombic. However, the design still has
adherents and users. It is likely to survive as an antenna option for
generations to come.
Chapter 48
Chapter 49: The Controlled Current distribution (CCD) Antenna
T
he controlled current distribution (CCD) antenna has been
around since the late 1970s. Every so often, it arouses a
flurry of questions in my e-mail. So I decided to look into the
CCD to see what we might reasonably expect of it.
Fig. 1 shows the general outline of the center-fed version of the

CCD. It consists of a number of straight wire sections of any
practical number that we can designate as N. Except for the
feedpoint and the wire tips we must separate each section with a
capacitor. Hence, the total number of capacitors will be N-2. The
CCD simulates a continuously capacitively loaded 1-wavelength
element by using equally spaced discrete components.
There is also a vertical version of the CCD. One way to create a

vertical CCD is to simply use the center-fed antenna in a vertical
installation. However, some literature suggests using the antenna
against ground as a monopole. Fig. 2 shows the general layout of
this configuration.
Chapter 49
Beyond the general claim of "improved performance," I have not

encountered a very clear account of what advantages the CCD is
supposed to offer. The wire length will be 1 physical wavelength.
However, the distributed capacitive loading of the element will
electrically shorten the wire. If we select the correct value for all of
the capacitors (they are all equal), we can arrive at resonance, that
is, a feedpoint impedance with negligible reactance. It appears that
the idea behind the CCD is to avoid the very high impedance of a
Chapter 49
1-wavelength center-fed wire while preserving its gain and

directivity.
I sometimes hear the terms "aperture" applied to the antenna. Most

texts reserve the term for use with highly directive antennas, such
as UHF horns and the like. Nevertheless, the calculation of
aperture rests--according to one text--on the wavelength, the
directivity, the polarization match, and the impedance match. Since
a comparison of a CCD with any other antenna will equate the
wavelength, polarization, and impedance match, the remaining
factor is directivity. Hence, if the CCD is an improvement over a
legitimate comparator, then it will show improvements in directivity.
We can examine the directivity of the CCD by looking at its gain
and beamwidth.
To simplify the examination, let's look at the CCD on a common

frequency. 3.5 MHz is widely used in articles, so that will be our
choice. A wavelength at 3.5 MHz is 280.02' long, so we shall make
the antenna length 280'. Throughout, we shall use AWG #14
copper wire to reflect typical amateur building practices. Our initial
tests will use a free-space environment. In this environment, we
must use the center-fed version of the antenna, but we shall not
have to be concerned with vertical vs. horizontal orientations.
I chose to model the antenna in NEC-4, which creates a small

difficulty. If I use a single wire for the entire element, I can place
load capacitors on individual segments of the wire. However, the
feedpoint region will not be quite perfect. Fig. 3 shows the
alternative models that I used. The upper model uses a single
Chapter 49
source segment, but that does not ensure that the wire lengths on
each side of the feedpoint are equal to the other segments between
capacitors. The lower section uses a split source to simulate a
source at the center segment junction. Although this move
improves the segment spacing, it can result in somewhat erroneous
impedance reports for very high impedances, where the impedance
might change significantly with only a small change of feedpoint
position or total wire length.
For all practical purposes, the difference between the models is not
great enough to jeopardize the modeling analysis. The required
values of capacitance tend to coincide closely with values in the
literature. The next task involves the capacitors themselves.
Literature on the CCD shows that we can build the antenna with
almost any number of wire sections and corresponding numbers of
capacitors. Since constructing each section involves wiring in a
capacitor with appropriate strain relief--a considerable task--I
wondered what one might gain by opting for a "large" CCD over a
Chapter 49
"smaller" CCD, where large and small indicate the number of wire
sections and capacitors. Therefore, I created 2 models, one that
used 26 wires section and 24 capacitors and another that used 50
wire sections and 48 capacitors.
The final step involves selecting the capacitor value. Since

modeling allows easy variation of the capacitor value, we might
explore the performance of the antenna in free space using various
capacitor values. Let's start with the larger model. Table 1 shows
the results of this modeling exercise.
Chapter 49
Perhaps the first notable feature of the data is that as we raise the
value of the capacitors in the string, the gain increases. So too
does the resistive component of the feedpoint impedance. The
reactive component shows an initial capacitive value that becomes
inductive as we increase the capacitor values. This feature is
natural enough, since increasing the capacitance value lowers the
capacitive reactance along the wire. Since the wire is long
compared to a dipole, lowering the compensating capacitive
reactance will leave the feedpoint increasingly inductive.
In fact, more is at stake than just the feedpoint reactance. Note the
entries for resonance. At resonance, the feedpoint reactance
disappears, marking a balance between the inductive reactance of
the wire and the capacitive reactance from the string of inserted
components. Let's also examine the data for the "smaller" CCD that
uses 26 wire sections and 24 capacitors. Table 2 provides the
necessary information, but with fewer steps along the way.
Chapter 49
The antenna gain does not depend for its gain--above some
minimum number of capacitors--on the array size in terms of the
number of capacitors in each leg. The gain of the smaller array with
500-pF capacitors is the same as the gain of the larger array with
1000-pF capacitors and likewise for the smaller 1000-pF and the
larger 2000-pF entries. The feedpoint impedance reports also track
each other in the same manner. This result is also very reasonable,
since 24 500-pF capacitors have the same capacitive reactance as
48 1000-pF capacitors if we measure at the same frequency.
The relative balance between these two factors alters the current
distribution along the 1-wavelength wire. Fig. 4 provides some
samples of the distribution curves for the smaller array, but curves
for corresponding large-array values are virtually identical. The
Chapter 49
curve for the array with a low capacitance value that yield a
capacitively reactive feedpoint impedance is very steep. At the
other extreme, where the capacitor value yields an inductively
reactive impedance, the curve shows dual current peaks. Only the
capacitor values that yield resonance produce a curve that we tend
to associate with a dipole.
The resonant impedance of a center-fed CCD is in the

neighborhood of 200 Ohms. Therefore, one may simplify the
matching problem by installing a 4:1 balun at the feedpoint and
using a standard coaxial cable feedline. Of course, one may also
use parallel transmission line to either a tuner or a 4:1 balun at the
shack end of the line. Since 200-Ohm feedlines are rare, one likely
would need a 300-Ohm feedline cut to the nearest half-wavelength
Chapter 49
(allowing for the line's velocity factor) at the operating frequency to

minimize balun losses.
With the right choice of capacitors in the string, the CCD offers bi-
directional performance with a resonant feedpoint impedance. In
exchange for the more complex construction of the antenna, we
obtain a simplification of matching requirements. However, we have
not yet assessed how good that performance is. For that task, we
need an appropriate comparator.
The CCD vs. the Dipole and 1-Wavelength Center-Fed Wire
In CCD literature, the resonant 1/2-wavelength dipole seems to be

the most popular antenna with which to compare the CCD with
respect to performance as a horizontal antenna. An alternative
comparator is the plain and simple 1-wavelength center-fed wire
antenna. The 1-wavelength plain wire is the same length as the
CCD. However, the current distribution of the plain wire differs from
the current distribution of the resonant CCD. Fig. 5 provides a set
of free-space current distribution curves for the resonant dipole and
for the 1-wavelength wire for ready comparison to the CCD curves
in Fig. 4.
Chapter 49
The dipole curve resembles the resonant CCD curve. Of course,

neither is perfectly sinusoidal, since the voltage at the center
feedpoint never goes to zero. I have not explored the degree to
which each departs from that familiar curve. In contrast, the 1-
wavelength wire curve most resembles the CCD when the latter
uses a very high value for the capacitors and thus loads the wire
least.
Chapter 49
Table 3 compares the free-space performance of the dipole and 1-

wavelength wire along with resonant CCDs of the smaller and
larger type. Clearly, the CCD has more gain (by about 1.1 dB) than
the dipole. However, the CCD falls about 0.7 dB short of the plain
1-wavelength wire. We should have expected this result from the
tabular data on the CCDs as we raised the value of the capacitors
and lowered the level of loading. (The capacitors in all models are
lossless.) As the capacitive reactance decreased, the gain
increased. From a certain perspective, we may view the 1-
wavelength wire as a CCD with capacitors having an indefinitely
high capacitance and hence negligible reactance.
Chapter 49
Fig. 6 overlays the free-space E-plane (azimuth) patterns of the

antennas (using only 1 of the resonant CCDs). The CCD pattern
most resembles the E-plane pattern of a center-fed plain wire about
0.85-wavelength long, although the CCD beamwidth is close to 10
degrees wider.
Very few people operate 3.5-MHz antennas in free space.

Therefore, we may usefully transplant all of the antennas in Table 3
Chapter 49
to a height of 1 wavelength above ground. With this adjustment, we

obtain the data in Table 4.
The gain differentials that we found in the free-space models hold

up when we move the antennas over ground. Of course, the CCD
loading technique does not alter the TO angle of the horizontal
antenna relative to either comparator. Table 4 records the
horizontal beamwidth values for all of the antennas, and we can
see that the CCD beamwidth is closer to the value for the dipole
than it is to the value for the 1-wavelength plain wire. Fig. 7 records
the elevation and azimuth patterns for the antennas in this test
group.
Chapter 49
There is no difference among any of the elevation patterns, since

those patterns emerge as a function of the height of the horizontal
wire above ground. The 2 versions of the CCD show no differences
in their azimuth patterns. Hence, the choice among the antennas
with respect to performance largely hinges upon the combination of
gain and beamwidth.
One of the seeming advantages of the CCD is the fact that in

exchange for more complex construction, one obtains a higher gain
with a simple resonant feedpoint. However, the tables have shown
near-50-Ohm values for the plain 1-wavelength wire. The technique
used to obtain such feedpoint values is very simple and very old.
Since the terminal impedance of the 1-wavelength wire is very high,
we can attach a 1/4-wavelength section of parallel transmission
Chapter 49
line. Somewhere very close to the end of the line will be a 50-Ohm
matching point. The matching sections for the horizontal 1-
wavelength wires use 600-Ohm line. The required length for the
impedances shown is 68' (compared to a full 1/4 wavelength of
70.25'). Fig. 8 shows the simple set-up. The antenna environment
and construction variables will determine the exact line length
required, so the examples use a velocity factor of 1.0. Most 600-
Ohm lines may have values between 0.95 and 0.98.
The matching section may be much simpler to build than the CCD
antenna element. However, the 2:1 SWR bandwidth of the plain
wire plus matching section is fairly narrow--perhaps 150 kHz at 3.5
MHz. (The resonant CCD has a 200-Ohm SWR bandwidth of about
300 kHz.) Indeed, the wisest feed system for the plain wire may be
parallel line all the way to the shack antenna tuner. However,
setting the line length to an odd multiple of a quarter wavelength at
the most used frequency may ease the tuner's task by a good
margin.
Chapter 49
The Vertical CCD and the 1/2-Wavelength Base-Fed Plain Wire
One application suggested for the CCD is as a vertical antenna. We

may hang a full center-fed CCD vertically with its feedpoint at any
height above ground that we can manage to support. However, the
more interesting case is to use a half CCD as a monopole, with the
feedpoint at ground level. A monopole CCD simply uses half the
number of wire sections and half the number of capacitors as a
center-fed horizontal CCD. Hence, the wire will be 1/2-wavelength
long. We shall split the 26-section, 24-capacitor CCD and create a
14-section, 12-capacitor CCD.
The appropriate comparator for this antenna is a simple 1/2-

wavelength monopole. The base feedpoint will use a matching
section to arrive at a near-50-Ohm impedance. Because the natural
impedance of a base-fed 1/2-wavelength monopole is lower than
the impedance of a horizontal center-fed wire, we must use a
parallel transmission line with a lower characteristic impedance.
450-Ohm line works well here, and a 68' length allows us to arrive
at the desire impedance level. Again, the models use a velocity
factor of 1.0, but an actual line would use the actual velocity factor
for the selected line. As well, the variables of installation will likely
require some experimentation to find the correct length.
The simplest way to model both antennas is to use a MININEC

ground. The presumed advantage of using this ground system is
that it allows direct contact between the lower end of the vertical
wire and the ground without need for modeling radials. In some
contexts, this system can be useful, but not in this case. In fact, the
Chapter 49
MININEC ground obscures some critical differences between the

operation of a CCD and a simple 1/2-wavelength base-fed wire.
To use a NEC-4 Sommerfeld-Norton ground with the antenna wire

touching the ground requires that we install some kind of wire
below ground. The simplest below ground wire might be a
simulation of an 8' ground rod. Like the MININEC ground, this
treatment is applicable to both antennas. Whether such a treatment
is advisable is something that the data will tell us.
Alternatively, we can install a buried radial system using as few or

as many radials as the analysis might dictate. The data in Table 5
show 4-, 16-, and 32-radial systems, using 70' AWG #14 copper
radials, for the CCD. The 1/2-wavelength wire only uses a 4-radial
system for reasons that become apparent in the data.
Chapter 49
Each antenna entry set begins with the MININEC ground. If we

were to use this data, we would reach the conclusion that both
antennas perform very similarly, with a negligible difference in gain
and only a 2-degree difference in the TO angle. However, if we
change to the S-N ground with an 8' ground rod, we obtain very
different results. With no change in the TO angle, the plain wire
outperforms the CCD by about 2.6 dB, a noticeable amount.
To improve performance of the CCD, we must replace the ground

rod with radials. The table shows the results for radial fields of
various sizes. For the CCD, adding 4 radials increases the gain by
over 1.5 dB. Only when we have 32 radials does the CCD
Chapter 49
challenge the gain of the plain 1/2-wavelength wire. The data for
the plain wire shows that replacing the ground rod with 4 70' radials
only increases gain by 0.1 dB. The small gain increment tends to
suggest that a radial system is optional with a 1/2-wavelength wire.
Fig. 9 overlays the plain wire with ground rod pattern and two CCD
patterns: one for the ground-rod model and the other for the model
with 16 radials. The patterns reveal not only the gain differences,
but also the TO angle differences for the two types of antennas.
The difference in the radial requirements between the plain wire

and the CCD emerges from the different current distributions on the
two types of antennas. The sketch on the left in Fig. 10 shows the
distribution along the 1/2-wavelength simple wire monopole. The
current maximum occurs at the wire center, which is elevated.
Hence, the antenna shows a lower TO angle than the CCD. As
Chapter 49
well, the current reaches a minimum at the base of the antenna.

Essentially, the element is complete at that point and requires only
a good RF ground. The 4 radials provide a better distributed RF
ground than the simple rod, but the performance difference is small.
In contrast, the CCD design places a current maximum at the base

of the monopole installation. Maximum efficiency requires an
effective completion of the antenna in the form of a low-loss radial
system. The more radials that we have, the more efficient will be
the total antenna system, as indicated by the declining resistive
component of the feedpoint impedance as we add more radials.
Note, however, that as we bring the system to a level of maximum
efficiency, the TO angle actually increases. At a high efficiency
level, the feedpoint would need a 2:1 impedance transformation
device or network for compatibility with the standard 50-Ohm
coaxial cable.
Chapter 49
An effective monopole CCD system thus requires two forms of

construction complexity relative to the plain 1/2-wavelength wire:
the installation of capacitors along the monopole and the burying of
a large number of radial wires. In contrast, the plain 1/2-wavelength
wire needs no modification of the monopole wire and provides good
service at a lower TO angle with only a ground rod or the simplest
of radial systems. However, the high impedance of the antenna
may require a dedicated network or a carefully tuned matching
section to arrive at a 50-Ohm impedance.
Chapter 49
How "Small" Can I Make a CCD?"
The terms "large" and "small" in connection with the discussion of

CCDs in these notes refer to the number of wire sections and the
number of capacitors needed to make up a 1-wavelength antenna.
In our initial free-space modeling experiments, we saw that there is
no significant difference in the performance of a 26-section, 24-
capacitor CCD and a 50-section, 48-capacitor version. Since the
smaller of the two systems requires less work than the larger, we
may naturally pose a question: how small can I make the CCD
antenna and still achieve proper performance. Since the goal is a
self-resonant 1-wavelength antenna, we have a criterion for
success.
Models of the free-space center-fed CCD seemed unable to

achieve resonance with 6 or fewer capacitors. However, an 8-
capacitor, 10-section model of the antenna did achieve resonance.
The feedpoint impedance was considerably lower than the 200-
Ohm target that we have used in order to apply a 4:1 balun at the
feedpoint. As well, the gain was down slightly relative to larger
models, but not enough to be operationally significant. Table 6
shows the results obtain from several different models leading up to
the 26-section, 24-capacitor model.
Chapter 49
Note that between the worst and the best of the model set, we have
only a 0.15-dB difference in gain. As well, the beamwidth changes
by only 3 degrees, indicating a stable E-plane pattern. However,
the feedpoint impedance climbs from about 170 Ohms to well over
200 Ohms. The differences in the gain levels and the feedpoint
impedance values result from the fact that the current distribution
curve changes as we move from fewer to more wire sections and
capacitors. Fig. 11 provides a sense of the evolution of the
distribution curve from essentially a triangle to a smooth curve.
Chapter 49
The shapes of the current magnitude distribution curve do not

materially affect the shape of the radiation pattern, which remains
well-behaved and bi-directional. The absence of any significant
Chapter 49
change in the beamwidth is a further indication that even the 8-

capacitor version of the antenna will perform well.
The answer to the section's lead question then is that about 8-

capacitors and 10-wire section form the smallest practical CCD with
full performance and resonance. It is likely that the precision of the
match between capacitors and between wire sections will play a
stronger role for the small CCD than it does for larger versions.
Conclusion
We have taken a short look at the controlled current distribution

(CCD) center-fed 1-wavelength antenna with an eye toward
understanding its operation and assessing its virtues. The center-
fed version of the antenna provides a 200-Ohm resonant feedpoint
impedance with the choice of the correct capacitor size, compared
to the 70-Ohm impedance of a 1/2-wavelength dipole and the very
high impedance of a plain center-fed 1-wavelength wire. The CCD
gain and beamwidth values fall between those of the two antennas
used as comparators. Models suggest that CCD performance
peaks with about 26 wire sections and 24 capacitors. However,
CCDs as "small" as 8 capacitors and 10 wire sections may work
satisfactorily.
The vertical monopole version of the antenna is perhaps more

problematical, since when ground mounted, it requires an extensive
radial system for most efficient operation. Unlike the plain 1/2-
wavelength wire monopole that has an elevated point of maximum
current, the CCD vertical monopole reaches maximum current at
Chapter 49
ground level and thus requires radials to complete the radiating

structure.
Essentially, the CCD simulates with discrete components a

continuously loaded element with 501 pF/meter. At this loading
level, the gain in free space is 3.10 dBi, with a feedpoint impedance
of 245 - j0.2 Ohms. The performance reports are virtually identical
to those that emerge from the 26-section, 24-capacitor antenna that
used 420 pF capacitors.
The CCD is a viable and potentially useful antenna of its type.

Whether the advantages warrant the relatively complex
construction compared to simpler wire antennas is a user judgment.
Chapter 49
Chapter 50: A Universal HF Back-Up Antenna
I
have had occasion in the past to write on the W2EEY
Expanded/Extended Lazy-H array that first appeared in the
1960s as a mono-band wire array. W6SAI later discovered that
the antenna had available gain on frequencies in a range of at least
2:1. In other words, a 10-meter version of the antenna might be still
useful on 20 meters.
My own looks into this antenna suggest an even wider range of

utility, even though performance tapers off steadily as one lowers
frequency. The chief drawback of the antenna has so far been the
fixation on wire construction. I wondered what tubular elements
might do, and the result is this preliminary note.
Large antenna farms for serious DX and contest use often have far
more than one antenna per band. Hence, back-up on many bands
is almost a matter of course. However, there is often only on
antenna for each of the following bands: 40, 30, 17, and 12.
Now suppose there were a single rotatable antenna of relatively

easy maintenance (compared to a 5-element Yagi or similar) that
might provide emergency performance on 40 through 10 meters,
performance that was not stellar, but usable in a crunch. Further
suppose that the antenna had a top height of not more than say 70'
(about 1/2 wl on 40 meters) and has elements no longer than some
of the half-size 40-meter beams that are commercially available.
Finally, suppose that the antenna required no boom, but looked
Chapter 50
more like one of those 40-meter beams flopped over to point

straight up.
Such an antenna is not a mechanically simple installation, since 44'

foot aluminum elements are not for the beginner. However, once
installed, the antenna poses fewer stress problems than
horizontally positioned arrays that require a boom. If it happened
also be to bi-directional, then it might be laid close in to a pole or
tower, since the total required rotation is 180 degrees.
There might be a niche for such an antenna, so the performance

potentials and the installation challenges seem worth exploring, at
least on a preliminary basis.
The W2EEY Expanded/Extended Lazy-H
The basic array is an extension of the Lazy-H: two 1 wl elements

vertically separated by 1/2 wl and fed in phase. The W2EEY
innovation was to extend the elements to 1.25 wl, extended double
Zepp length. He also expanded the separation to 5/8 wl to
maximize in-phase gain.
Chapter 50
Fig. 1 shows the outlines of a typical extended Lazy-H in typical

wire form. For many years, the Editors and Engineers Radio
Handbook (edited by W6SAI) has carried that standard Lazy-H, fed
at the bottom with a stub for coax matching for mono-band use.
However, for multi-band use, a center junction of equal lengths of
feedline is the simplest route to in-phase feeding on many bands.
The dimensions of the array for 10 meters are a modest 44'

element length and 22' vertical separation. The EDZ element
Chapter 50
lengths represent the practical limit for use on HF bands, including

10 meters. Longer elements will yield a multi-lobe pattern on 10
meters.
The antenna height selected for study was a 66' top height. It can
be mounted higher or lower with standard changes in the elevation
angle of maximum radiation. However, a height of at least 66'
seems advisable for reasonable 40-meter use.
From Wire to Tubing
I have elsewhere looked at the performance of the W2EEY version

of the antenna across many bands. The question that came to mind
regarding rotating the array required a change in element material.
So I redesigned the antenna for aluminum tubing having an
aggressive tapering schedule.
Chapter 50
Fig. 2 shows the arrangement used to develop models. There is no

assurance that this particular schedule meets appropriate
standards. Any actual elements that one might contemplate
constructing should be taken through an exercise or two on
YagiStress to determine their mechanical feasibility. Although the
schedule used here has proven very useful in comparing wire
versions of the antenna to the most severe gradations I could think
of, the element design is hypothetical only in mechanical terms.
Chapter 50
Special Note: Carroll Allen, AA2NN, pointed out the taper schedule
suggested would have a wind survival rating of only about 70 mph.
He developed a spread sheet for EXCEL to calculate the stress on
the tubing. For commonly used antenna tubing, such as 6061-T6,
with a wall thickness of 0.058", the maximum stress for each
section should be 40,000 psi or less. He kindly redesigned the
sections for a 100 mph wind survival rating. The following table
presents the revised taper schedule. Like the original schedule, the
1.125" diameter section is presumed to run all the way through the
1.25" section, but also to have its own exposure length.
44' Aluminum Doublet Half-Element Structure
for 100 MPH Wind Survival
Diameter (") Section L (") Cumulative L (")
1.25 72 72
1.125 19 91
1.0 20.5 111.5
0.875 21.5 133
0.75 23 156
0.625 24 180
0.5 84 264
The chief differences between the wire and tubing versions of the
antenna were two. First, the models had many more wires. Second,
the source impedances (as taken at the junction of the two
feedlines from the elements to a center small segment) varied
somewhat from those associated with the wire version.
Chapter 50
For those unfamiliar with the extended Lazy-H, let's run a series of
tabular entries and some patterns to check the potential
performance. As with all models, these assume level, uncluttered
terrain (average Sommerfeld-Norton ground) in NEC-4.1. Any
serious antenna farmer has already run terrain analysis using
N6BV or K6STI software and can therefore adjust the numbers for
gain and elevation angles accordingly.
The figures in the first table focus on performance. The frequencies

are band centers. Gain is maximum at the elevation angle of
maximum radiation (TO angle). Note that gain is bi-directional. The
horizontal beamwidth is to the -3 dB points on the maximum gain
curve.
Frequency Gain TO Angle Horizontal Notes

MHz dBi degrees Beamwidth (deg)
28.5 15.2 8 37 EDZ-type side lobes
24.94 14.7 10 41
21.225 12.6 11 52 Standard Lazy-H
18.118 11.0 13 61
14.175 9.1 17 73
10.125 8.2 24 85
7.15 6.5 33 99
7.15 7.2 29 89 Using top wire only
As the table shows, gain decreases steadily with frequency, since

the elements grow shorter and the spacing narrower. The 10-meter
gain is similar to a long-boom 5-element Yagi (without the front-to-
back ratio). On 15, gain performance is similar to some 3-element
Yagis, dropping to 2- element performance on 17. Below that
Chapter 50
frequency, the antenna becomes the equivalent of a rotatable

dipole.
The patterns to follow, band by band, show the elevation patterns

along the axis of maximum gain. Only spot azimuth patterns (for 10,
15 and 30 meters) are shown, since the evolution of the azimuth
pattern with frequency changes is perfectly normal. One key
advantage for the array is the relative absence of very strong high
angle radiation, especially on the upper bands, thus reducing a
potential noise source.
Chapter 50
10-Meters
Chapter 50
12-Meters
Chapter 50
15-Meters
Chapter 50
17-Meters
20-Meters
Chapter 50
30-Meters
Chapter 50
40-Meters
The 40 meter graphic shows two patterns, one for the use of both
wires in phase, the other for the use of only the top wire. The latter
is harder to implement in practice, but increases gain while lowering
the take-off angle. On 30 meters, the use of only the top wire gives
the appearance of lowering the take-off angle, but in fact, the
bottom lines of the two patterns overlap, with the 2-wire system
showing more gain. The seemingly unexpected result comes from a
difference between vertical beamwidths for the two arrangements.
Hence, on 30, the phased 2-wire system appears to offer superior
overall performance.
The decrease in gain for each reduction in frequency is clearly

apparent. Given the array of antennas that inhabit some of the most
extensive antenna farms, one might judge 20-meter performance to
Chapter 50
be the weakest in comparative terms. WARC band antennas tend

to be simpler, and relative to them, the array is down by only a
couple of dB at most. (If one has designed against these
tendencies, then the differentials will also be different.) From 15
meters upward, the array gain is quite good--indeed, competitive
might be a reasonable term. However, on 20, 5-element Yagis are
fairly common (as are stacks of beams that give similar
performance). The array on 20 yields performance similar to that of
a 1 wl wire, perhaps 2 S-units weaker than the high-performance
main antenna(s).
Nonetheless, the overall performance of the array relative to all of

the antennas in use for every band must be accounted quite good
in view of what the antenna is in this application: it is an emergency
back-up capable of being switch in to replace any antenna system
that goes dead when Murphy dictates. With a size similar to a half-
size 40-meter 2-element Yagi pointing straight up, it is also much
simpler than its most cogent competitor: a log-periodic for 40-10
meters.
Phase Lines
The models developed here used 600-Ohm, VF=1.0 lines for

phasing. Each line was cut to 14' even though the distance to the
center point was 11' only. A 3' buffer was allowed for each line to
account for routing that would clear any metallic structures that
might disrupt standard performance of the lines.
Chapter 50
With these constraints, the models yield the following source

impedance calculations.
Frequency Source Impedance Notes
MHz R +/- jX Ohms
28.5 140 - j 655 Highly variable with change of Zo
24.94 335 + j 655 Somewhat variable with changes of Zo
21.225 110 + j 85 Decreases with Zo decreases
18.118 165 - j 135 Decreases with Zo decreases
14.175 880 - j 136 Highly variable with change of Zo
10.125 65 + j 250 Reasonably stable
7.15 14 + j 10 Low and stable
7.15 14 + j 30 Top-wire only in use, with line.
The number of cases in which the source impedance decreases

with lower characteristic impedance phasing lines suggest that the
highest feasible value of Zo be used for the antenna. Indeed, with
lower values of Zo, the 40-meter source impedance can decrease
below 10 Ohms, rapidly escalating the negative effects of loss
sources in any installation.
High variability with changes in phasing line type tend to indicate

that the builder of any antenna of this type may encounter quite
different source impedances owing to minor local variations.
Although the phased pair of 44' doublets yields lower ratios of
reactance to resistance, relative to a single 44' doublet, the low
impedance at 40 meters may prove difficult to match at the tuner
end of the line. The reactance will vary more widely than the
resistance over a greater section of each half-wavelength of line.
Hence, careful line-length selection may prove necessary, although
this factor will vary with the particular tuner configuration and
internal components.
Chapter 50
On the basis of these results (typical of several different models),

there appear to be three initially plausible feed systems.
1. Parallel feedline and an ATU: The simplest feed system is to

use parallel feedline from the juncture of the phasing lines to the
operating position, with an ATU providing the requisite matching.
for minimum loss in the matching system and for maximum
isolation of equipment from common mode currents, a link-coupled
tuner is recommended (with due construction precautions against
unintended common mode paths). Without careful measurement of
the actual source impedances encountered and equally careful
measurement of the feedline length, there may be cases in which
the impedance presented to the tuner terminals falls outside the
range for which the tuner can effect a match while compensating
for reactances at that point. Ordinarily, changing the line length in
small increments will overcome this problem without incurring
significant losses.
2. Remote switching of matching circuits--Version 1: A box of

fair proportions installed at or very near the junction of the two
phase lines might contain an array of circuits matching the source
impedance on each band to standard 50-Ohm coaxial cable (or 75-
Ohm hardline). The system would require a power source to control
separate input and output sides of each network. The design of
such a system would have to be a custom installation based on the
actual source impedances of the system.
3. Remote switching of matching circuits--Version 2: A remotely

switched matching network box mounted precisely at the junction of
Chapter 50
phase lines may present mechanical problems if the installation has

a rotating mast joining the two elements. Moreover, the
impedances to be matched may not be the most desirable. A length
of open-wire parallel feeder from the phase line junction to some
point further down the support structure may provide more
desirable values and a more convenient mechanical installation.
Since the variables of this modified system are so many, I have not
explored any particular line lengths to check feasibility. Hence,
electrically, this system can only be classed as equivalent to the
first version.
Either remote matching system serves a certain preference among

antenna farmers: the desire to present sensitive rigs and amplifiers
with loads that require no in-shack tuning during intensive
operations. All adjustments are made during installation and
remade during routine maintenance. Those willing to make real-
time adjustments and who can correctly install open-wire feeders
may wish to use the simplest of the systems outlined here.
Of the three systems, only the remote box at the phase-line junction
is amenable to easy switch-over to using only the top wire on 40
meters. Adapting the other systems to such use will be an exercise
in ingenuity.
Chapter 50
Mechanical Considerations
The basic elements of the antenna

structure appear in Fig. 3.
No support structure is shown, since

this aspect of the installation is subject
to so many variations. The antenna
itself consists of two 44' long elements
of considerable weight. The top one
will be 22' above the lower one, which
might be considered a moderate
stacking challenge to those with large
antenna farms. Since the elements are
mounted close to the line of the
rotating support mast, the stresses
they impose upon the assembly may
be less than those of elements at the
ends of a boom. A mast extension to
handle a truss system to add support
farther out along the elements is
certainly feasible and may well be
advisable.
If the lower element is (or both elements are) are side-mounted

relative to a supporting pole or tower, then there will be a dead
zone in the rotation. The size of this zone in degrees will vary
inversely with the distance of the mast from the support structure. If
we assume that there is a direction from the station that can be
Chapter 50
ignored, side-mounting may offer a means of installing the

emergency antenna on an existing tower or pole. However, the
elements may well interfere with existing guy wires.
These mechanical notes are offered on the premise that anyone

thinking about this antenna has considerable experience with tall
installations and can integrate the structure into a considerable
backlog of diverse variables involved in high pole and high tower
work. Those who may be reading these very preliminary notes
without requisite experience should perhaps review some of the
very considerable stack of reprints offer by Champion (K7LXC)
before becoming too much attracted to the ideas noted here.
The aluminum extended Lazy-H is a bi-directional array with

rotational capabilities that may make use of its reasonably narrow
horizontal beamwidth on 40 through 10 meters. As such, it may fill
a special niche, which I have termed the nearly-universal back-up
for giant antenna farms. The feed system challenges are more
electronic than they are mechanical--except for mounting and
water-proofing a box for two of the suggested systems. Otherwise,
the array presents fewer mechanical problems than most very large
Yagis.
Still, the array not a cure-all for whatever ails. Nor is it a magic elixir
for all antenna installations. The notes presented here simply
suggest that for some installations, the antenna may provide that
nearly universal back-up which is ever handy: ready to go on 40 to
10 meters when Murphy strikes down the primary antenna(s). Just
Chapter 50
do not tell Murphy you have installed one of these or he will take
down your antennas two at a time.
Chapter 50
Wide-Band Yagi Notes - Volume 2 546
Other Publications
We hope you’ve enjoyed Volumes 1 & 2 in a series of books of the Antennas Made
from Wires. Watch for Volume 3 to be released soon. You’ll find many other very
fine books and publications by the author L.B. Cebik, W4RNL and others in the
antenneX Online Magazine BookShelf at the web site shown below.
A Publication by
antenneX Online Magazine
http://www.antennex.com/
POB 271229
Corpus Christi, Texas 78427-1229
USA
Copyright © 2010 by antenneX Online Magazine. All rights reserved. No part of

this book may be reproduced or transmitted in any form, by any means (electronic,
photocopying, recording, or otherwise) without the prior written permission of the
publisher.
ISBN: 1-877992-86-0
Other Publications

Wire 2

Uploaded by

Copyright:

Available Formats

Wire 2

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Wire 2

Uploaded by

Copyright:

Available Formats

Antennas

Simple & Effective

A Collection from the Works

Copyright © 2010 by Publisher antenneX Online Magazine. All rights reserved. No

About the Author

L. B. Cebik, W4RNL, passed away in April 2008. He had written extensively

About the Author

Chapter 27: About the Folded Dipole--------------------------------------------------------7

Chapter 27: About the Folded Dipole

Fig. 1 shows the essential elements of a folded dipole. It consists of

We can construct a folded from common materials ranging from

of finicky adjustments. Something about the ease of building an

Using antenna transformer techniques to raise the impedance of an

The Folded Dipole as an Impedance Transformer

Our understanding of the folded dipole has been stunted in part by

Relative to a single-wire dipole, the feedpoint impedance will be

The log of 2S divided by a wire diameter is a complex quantity that

In free space the impedance, the resonant impedance of a single

• 1. If d1<d2, then R is always >4.

However, R must always be >1. That is, 1 is the limit of R as the

So far, we have ignored S. Before taking a log for the numerator

There is one exception to this consequence of spacing. If the two

Modeling the Folded Dipole

than the small rules of thumb formulae we find in many antenna

The trick to modeling folded dipoles is to use many segments. The

As a test case, let's look once more at the question of spacing. We

Let's compare the performance of folded dipoles having 3 different

Spacing Length Gain Feedpoint Z

Nothing in the spacing of the wires of folded dipoles could produce

Private experience might result in other claims. However, private

values will perform equally well when placed in identical antenna

Of course, the actual feedpoint impedance encountered by the

Before we leave these folded dipoles, let's note the differences in

Although there is a change of resonant length as we change wire

12.4:1, comparable ranges. However, the length range was only

With respect to length and performance, a folded dipole acts very

Antenna Currents and What They Tell Us

One very unhelpful conception of a folded dipole is to think of it

At first glance, a folded dipole operates in ways distinctly unlike a

magnitudes will be equal, but the current phases will be opposite,

Currents Wire 1 Wire 2

The modeling convention that runs the transmission line wires in

To illustrate what we actually encounter with a folded dipole, let us

A phase pattern similar to the one shown is necessary if the folded

a mystery: how can we make sense out of the current magnitudes

The Transmission Line Inside the Folded Dipole

The pattern of current magnitudes and phase angle hides a small

The following table provides the modeled values for a bare-wire

Folded Dipole Currents (with Dipole Currents for Comparison)

Folded Dipole Standard Dipole

There is a very good correlation between the folded-dipole radiation

Ideally, the transmission line currents should all show a phase

The key element in the transmission line currents is their relative

The existence of transmission line currents within a folded dipole

The first step in considering antennas made from insulated wire is

insulation would have yielded higher values--or longer resonant

Folded Dipole Standard Dipole

Nothing in the new table distinguishes the currents in the insulated

radiation currents. Allowing for the fact that the technique of