TABLE OF CONTENTS

10.1 Simple Wire Antennas

10.1.1 Random-Wire Antennas
10.1.2 End-Fed Antennas
10.1.3 Center-Fed Antennas
10.1.4 The 135-Foot, 80 to 10 Meter Dipole
10.1.5 The G5RV Multiband Antenna
10.1.6 The Windom and Carolina Windom
10.1.7 Off-Center-Fed (OCF) Dipoles
10.1.8 Multiple-Dipole Antennas
10.1.9 Terminated Folded Dipole
10.1.10 Horizontal Loop “Skywire”

10.2 Trap Antennas

10.2.1 Trap Losses
10.2.2 Five-Band W3DZZ Trap Antenna
10.2.3 W8NX Multiband, Coax-Trap Dipoles

10.3 Multiband Vertical Antennas

10.3.1 Full-Size Vertical Antennas
10.3.2 Short Vertical Antennas
10.3.3 Trap Verticals

10.4 The Open-Sleeve Antenna

10.4.1 Impedance
10.4.2 Bandwidth
10.4.3 Radiation Pattern and Gain
10.4.4 Construction and Evaluation

10.5 The Coupled-Resonator Dipole

10.5.1 The Coupled-Resonator Principle
10.5.2 Characteristics of Coupled Resonator (C-R) Antennas
10.5.3 A C-R Dipole for 30/17/12 Meters

10.6 HF Log Periodic Dipole Arrays

10.6.1 LPDAs For 3.5 or 7 MHz
10.6.2 5-Band Log Periodic Dipole Array

10.7 HF Discone Antennas

10.7.1 Discone Basics
10.7.2 A-Frame 20-10 Meter Discone
10.7.3 40-10 Meter Discone

10.8 Bibliography

Antenna Fundamentals   1-1

 Chapter 10
Multiband HF Antennas

For operation on a number of bands below 30 MHz, it systems that are designed to be used on two or more of the HF
would be impractical for most amateurs to put up a sepa- bands. Separate chapters cover nonresonant Long-Wire and
rate antenna for each band. But this is not necessary — for Traveling Wave Antennas as well as the popular HF Yagi
example, a dipole one half-wavelength long on the lowest and Quad Antennas. See the Transmission Line Coupling
frequency band to be used can be operated readily on higher and Impedance Matching chapter for more information on
frequencies. In fact, most common antennas can be used on using feed lines and impedance matching circuits.
multiple bands through the use of antenna tuners and other
techniques. What is usually referred to as a “multiband an- Harmonic Radiation from Multiband Antennas
tenna,” however, is one for which a method has been devised Since a multiband antenna is intentionally designed for
that allows the antenna to operate on a number of bands operation on a number of different frequencies, any harmon-
while still offering a good match to a transmission line, usu- ics or spurious frequencies that happen to coincide with one
ally coaxial cable. of the antenna resonant frequencies will be radiated with
When a single physical antenna is used on different very little, if any, attenuation. Particular care should be ex-
bands, one must be aware that the changing electrical heights ercised, therefore, to prevent such harmonics from reaching
and lengths lead to changes in the feed point impedance and the antenna.
the azimuth and the elevation patterns of the antenna as de- Multiband antennas using tuned feed lines have a certain
scribed in the chapters Antenna Fundamentals and Dipoles inherent amount of built-in protection against such radiation,
and Monopoles. For example, a horizontal wire antenna at since it is nearly always necessary to use a tuned coupling
an electrical height of l/2 on 20 meters is 2l/3 high on 15 circuit (antenna tuner) between the transmitter and the feed
meters and l/4 on 40 meters, leading to very different eleva- line. This adds considerable selectivity to the system and
tion patterns than if the antenna were at the same electrical helps to discriminate against frequencies other than the de-
height on all bands. Similarly, the elevation pattern and feed sired one.
point impedance of a single vertical antenna will also change Multiple dipoles and trap antennas do not have this fea-
dramatically on different bands. ture, since the objective in design is to make the antenna
In fact, it is usually more effective to consider the in- show as nearly as possible the same resistive impedance in
stallation as a “multiband antenna system” in which the all the amateur bands the antenna is intended to cover. It
antenna, feed line, and any impedance matching devices is advisable to conduct tests with other amateur stations to
are considered together — as a package. By thinking about determine whether harmonics of the transmitting frequency
the performance of the antenna on different bands you can can be heard at a distance of, say, a mile or so. If they can,
select a combination of system elements that result in good more selectivity should be added to the system since a har-
performance on all bands and not just one. monic that is heard locally, even if weak, may be quite strong
This chapter describes a number of antennas and antenna at a distance because of propagation conditions.
Multiband HF Antennas   10-1
 10.1 Simple Wire Antennas
10.1.1 RANDOM-WIRE ANTENNAS feed point impedance at a current maximum occurs at or near
The simplest multiband antenna is a random length of the transmitter. A short connection (several feet or less) with
wire, attached directly to the output of a transmitter or an- heavy wire or strap to a ground rod or metallic water pipe that
tenna tuner. Power can be fed to the wire on practically any runs through ground may be sufficient on the lower bands but
frequency using one or the other of the methods shown in most ground connections are not short enough to minimize
Figure 10.1. If the wire is approximately 67 or 137 feet long RF voltage by themselves. Regardless of how you address
(l/4 or l/2 on 80 meters) the end impedance will be high this issue, begin by connecting all equipment enclosures to-
on the bands that are harmonics of 80 meters and it can be gether to prevent significant voltages from existing between
fed through a tuned circuit, as in Figure 10.2. Many antenna pieces of equipment.
tuners have the option to feed an end-fed random wire in Using an antenna wire length close to l/4 (65 feet at
this way. Use an SWR meter between the transmitter and the
matching network to adjust for minimum SWR.
If you have a rotatable beam antenna, in many cases it
may be possible to use the beam’s coaxial feed line as an
antenna on HF. Connect the shield and center conductor to-
gether at the station end and use them as a random-length
wire as in Figure 10.1. The beam at the far end will serve to
end-load the wire as a capacitance hat.
The primary disadvantage of all such directly-fed sys-
tems is that the antenna system is composed of the random
wire plus all of the station equipment enclosures and the sta-
tion ground connection. The point at which the antenna is
connected can be thought of as a randomly chosen feed point
in an antenna that has one end tied to ground. As such, there
is a good chance that you will have “RF hot spots” in your
station because of the RF current in the antenna system.
RF voltages within the station can often be minimized by Figure 10.2 — If the antenna length is 137 feet, a parallel-
choosing an antenna and ground wire length so that the low tuned coupling circuit (A) can be used on each amateur
band from 3.5 through 30 MHz, with the possible exception
of the 10-, 18- and 24-MHz bands. C1 should be from 500-
1000 pF with plate spacing capable of withstanding several
hundred volts. L1 should be chosen to resonate with
20-80% of C1’s maximum value. If the wire is 67 feet long,
series tuning can be used on 3.5 MHz as shown at the left;
parallel tuning will be required on 7 MHz and higher fre-
quency bands. The L network shown in Figure 10.1B is also
suitable for these antenna lengths.

Figure 10.1 — At A, a random-length wire can be driven

directly from the pi-network output of a vacuum-tube
transmitter. At B, an L network (or antenna tuner) can be
used with solid state transmitters that do not have tun-
able output networks. C1 should have plate spacing suf-
ficient for at least several hundred volts; a maximum
­capacitance of 100 pF is sufficient if L1 is 20 to 25 µH. A
suitable coil would consist of 30 turns of #12 AWG wire,
21⁄2 inches diameter, 6 turns per inch. Bare wire should be Figure 10.3 — An “artificial ground” can be used to tune a
used so the tap can be placed as required for loading the random length of wire to minimize RF voltage on station
transmitter. equipment enclosures.

10-2   Chapter 10
3.6 MHz, 33 feet at 7.1 MHz), or an odd multiple of l/4 The feed line length can be anything convenient, but odd
(3⁄4 l is 195 feet at 3.6 MHz, 100 feet at 7.1 MHz, 50 feet at multiples of l/4 will transform the high feed point impedance
14 MHz, etc) may be helpful. The goal is to place the antenna to a lower value that is likely to be easier to transform to 50 W.
system’s connection to the transmitter or antenna tuner at a (See “Tuned Feeders” below.) The asymmetrical placement
point of low voltage. Obviously, this can be done for only one of the feed line with respect to the antenna often results in
band even in the case of harmonically related bands, since the common-mode current being picked up by the feed line. This
wire length that presents a current maximum at the transmit- results in radiation from the feed line portion of the system.
ter will present a voltage maximum at two (or four) times (See “Feed Line Radiation” below.)
that frequency. If you have room for only a 67-foot flattop and yet want
Another possibility is to attach a counterpoise wire to to operate in the 3.5-MHz band, the two feed line wires can
the transmitter or antenna tuner enclosure. The counterpoise be tied together at the transmitter end and the entire system
length is adjusted so that RF voltage on the station equip- treated as a random-length wire fed directly, as in Figure 10.1.
ment is minimized. The length may or may not be l/4 at the
operating frequency since the impedance at the end of the 10.1.3 CENTER-FED ANTENNAS
antenna wire is unknown. Be prepared to experiment with A center-fed single-wire antenna can be made to accept
different lengths. Different wires can be attached at different power and radiate it with high efficiency on any frequency
frequencies. higher than its fundamental resonant frequency and, with a
Another option is to use an “artificial ground” such as the reduction in efficiency and bandwidth, on frequencies as low
MFJ-931 (www.mfjenterprises.com) as in Figure 10.3 that as one half the fundamental.
tunes the counterpoise on different frequencies. It is also pos- In fact, it is not necessary for an antenna to be a full
sible in many cases to use an ordinary 100-W antenna tuner half-wavelength long at the lowest frequency. An antenna can
to accomplish the same thing — tuning the random-length be considerably shorter than 1⁄2 l, even as short as 1⁄4 l, and
counterpoise to present a low impedance at the transmitter or still be a very efficient radiator. The use of such short anten-
antenna tuner enclosure. nas results in stresses, however, on other parts of the system
If you do use a counterpoise, be sure to insulate the unat- (for example the antenna tuner and the transmission line) as
tached end because like all unconnected ends of antennas, discussed later on in this section.
there will likely be enough RF voltage to cause an RF burn, The simplest and most flexible (and also least expensive)
particularly at 100 W or higher. all-band antennas are those using parallel-wire feed lines to
the center of the antenna, as in Figure 10.5. Because each half
10.1.2 END-FED ANTENNAS of the flattop is the same length, the feed line currents will
Another common antenna system for multiband opera- be balanced at all frequencies unless, of course, imbalance is
tion is the end-fed Zepp antenna shown in Figure 10.4. The introduced by one half of the antenna being closer to ground
antenna length is l/2 long at the lowest operating frequency. (or a grounded object) than the other. To maintain balance of
(This name came about because the first documented use of the current in each antenna leg and minimize common-mode
this sort of antennas was on the Zeppelin airships where the current on the feed line, the feed line should be run at right
antenna was hung by one end and trailed below the airship.) angles to the antenna, preferably for a distance of at least l/4
An antenna tuner with a balanced output can provide from the feed point.
multiband coverage with an end-fed antenna with any length Center feed is not only more desirable than end feed
of open-wire feed line, as shown in Figure 10.4. Open-wire (described above) because of inherently better balance, but
or window line with an impedance of 300 to 600 W is most it generally also results in a lower standing wave ratio on
often used. the transmission line, provided a parallel-wire line having a

Figure 10.5 — A center-fed antenna system for multi-

Figure 10.4 — An end-fed Zepp antenna for multiband use. band use.

Multiband HF Antennas   10-3

characteristic impedance of 300 to 600 W is used. Ladder or to radiate, such energy could actually have a desirable ef-
window line is satisfactory for all but possibly high power fect. Antenna purists may dispute such a premise, but from
installations (over 500 W), where heavier wire and wider a practical standpoint where you are not concerned with a
spacing are desirable to handle the larger currents and volt- directive pattern, much time and labor can be saved by ignor-
ages that may be present at high SWR. ing possible feed line radiation.
The best type of antenna tuner to use in such an instal-
lation is a balanced type designed for coaxial feed line on Tuned Feeders
the input and parallel-wire line on the output. An unbalanced References are often made to “tuned feeders” meaning
tuner can also be used but because one wire of the output feed sections of feed line with a specific electrical length. The
line is connected to the enclosure, RF current and voltages lengths act to transform load (antenna feed point) imped-
are more likely to be present in the station. ances as described in the Transmission Lines chapter. The
It is not recommended that coaxial feed line be used be- most common application of a tuned feeder is with an end-
tween the antenna tuner and antenna. At frequencies where fed antenna. A feed line that is any number of odd quarter-
the SWR is high, feed line loss in coaxial cable runs of more wavelengths long transforms a high impedance into a low
than 50 feet at HF can quickly become very high. (See the impedance and so can be used to connect a 50-W transmitter
Transmission Lines chapter). to a high-impedance end-fed antenna. This only works at
The length of the antenna is not critical, nor is the length frequencies for which the feed line is the required electrical
of the line. As mentioned earlier, the length of the antenna length, thus the term “tuned.” Most tuned feeders are con-
can be considerably less than l/2 and still be very effective. structed from parallel-wire feed line to minimize loss from
If the overall length is at least l/4 at the lowest frequency, a the high SWR in this application.
quite usable system will result. Some experimentation will Tuned feeders can also create problems due to their
likely be necessary to find the length that works best at a length. For example, a feed line some multiple of l/2 long
specific location on the bands required. connected to grounded equipment enclosures at one end also
has a low impedance at the other end. That can cause trouble
Feed Line Radiation for an end-fed antenna with a high feed point impedance.
Feed line radiation results when currents in a parallel- Resonant feed line lengths (some multiple of l/4 long) also
wire feed line are not balanced so that the radiation from tend to be effective at picking up energy from the antenna
each wire no longer cancels. This imbalance most commonly where it creates common-mode currents and re-radiated sig-
occurs when the feed line picks up energy radiated by the an- nals as discussed above.
tenna on both wires at the same time. This creates common-
mode current which re-radiates a signal just as an antenna 10.1.4 THE 135-FOOT, 80 TO
does. (The equivalent situation for coaxial feed line is for the 10 METER DIPOLE
outer surface of the shield to pick up and re-radiate energy.) As mentioned previously, one of the most versatile an-
Feed lines pick up the antenna’s radiated signal when tennas around is a simple dipole, center-fed with open-wire
they are not symmetrically oriented with respect to the an- transmission line and used with an antenna tuner in the shack.
tenna and its radiated field. For example, a feed line that A 135-foot long dipole hung horizontally between two trees
approaches a dipole at anything other than 90° will couple or towers at a height of 50 feet or higher works very well on
more strongly to the closer leg of the antenna. The closer 80 through 10 meters. Such an antenna system has significant
the feed line is to one leg, the more energy it will pick up. gain at the higher frequencies. (Other lengths reported to work
Feed lines to an end-fed Zepp almost always carry common- well are 88 and 105 feet — don’t be afraid to experiment.)
mode current because they are connected at one end and not The antenna can also be used on 1.8 MHz as a l/4 antenna
the middle. Common-mode feed line current and techniques with some reduction in efficiency.
for minimizing it are addressed in the chapter Transmission
Line Coupling and Impedance Matching. Flattop or Inverted V Configuration?
It should be emphasized that any radiation from a feed There is no denying that the inverted V mounting con-
line is not “lost” energy and is not necessarily harmful. figuration is very convenient, since it requires only a single
Whether or not feed line radiation is important depends en- support. The flattop configuration, however, where the di-
tirely on the antenna system being used. For example, feed pole is mounted horizontally, gives more gain at the higher
line radiation is not desirable when a directive array is being frequencies. Figure 10.6 shows the 80 meter azimuth and
used. Such feed line radiation can distort the desired pattern elevation patterns for two 135-foot long dipoles. The first is
of such an array, producing responses in unwanted directions. mounted as a flattop at a height of 50 feet over flat ground
In other words, you want radiation only from the directive with a conductivity of 5 mS/m and a dielectric constant of
array, rather than from the directive array and the feed line. 13, typical for average soil. The second dipole uses the same
If the feed line passes close to appliances or home entertain- length of wire, with the center apex at 50 feet and the ends
ment equipment, the radiated field can also cause RFI. drooped down to be suspended 10 feet off the ground. This
On the other hand, in the case of a multiband dipole height is sufficient so that there is no danger to passersby
where general coverage is desired, if the feed line happens from RF burns.

10-4   Chapter 10
Figure 10.6 — Patterns on 80 meters for 135-foot, center-fed Figure 10.7 — Patterns on 20 meters for two 135-foot
dipole erected as a horizontal flattop dipole at 50 feet, com- dipoles. One is mounted horizontally as a flattop and the
pared with the same dipole installed as an inverted V with other as an inverted V with 120° included angle between the
the apex at 50 feet and the ends at 10 feet. The azimuth pat- legs. The azimuth pattern is shown in A and the elevation
tern is shown at A, where the dipole wire lies in the 90° to pattern is shown in B. The inverted V has about 6 dB less
270° plane. At B, the elevation pattern, the dipole wire gain at the peak azimuths, but has a more uniform, almost
comes out of the paper at a right angle. On 80 meters, the omnidirectional, azimuthal pattern. In the elevation plane,
patterns are not markedly different for either flattop or the inverted V has a large high-angle lobe, making it a some-
inverted V configuration. what better antenna for local communication, but not quite
so good for DX contacts at low elevation angles.

At 3.8 MHz, the flattop dipole has about 4 dB more down some 6 dB from the flattop.
peak gain than its drooping cousin. On the other hand, the The situation gets even worse in terms of peak gain at
inverted V configuration gives a pattern that is more omnidi- 28.4 MHz for the inverted V configuration. Here the peak
rectional than the flattop dipole, which has nulls off the ends gain is down about 8 dB from that produced by the flattop
of the wire. Omnidirectional coverage may be more impor- dipole, which exhibits eight lobes at this frequency with a
tant to net operators, for example, than maximum gain. maximum gain of 10.5 dBi at about 7° elevation. See the
Figure 10.7 shows the azimuth and elevation patterns comparisons in Figure 10.8.
for the same two antenna configurations, but this time at Whatever configuration you choose to mount the 135-
14.2 MHz. The flattop dipole has developed four distinct foot dipole, you will want to feed it with some sort of low-
lobes at a 10° elevation angle, an angle typical for 20 meter loss open-wire feed line. For example, 450-W window line is
skywave communication. The peak elevation angle gain of popular for this application. Be sure to twist the line once or
9.4 dBi occurs at about 17° for a height of 50 feet above flat twice per foot to keep it from twisting excessively in the wind.
ground for the flattop dipole. The inverted V configuration is (Do not twist it so much that the wire spacing is reduced.)
again nominally more omnidirectional, but the peak gain is Make sure also that you provide some mechanical support
Multiband HF Antennas   10-5
 Figure 10.9 — The G5RV multiband antenna covers 3.5
through 30 MHz. Although many amateurs claim it may be
fed directly with 50-W coax on several amateur bands, Louis
Varney, its originator, recommends the use of a matching
network on bands other than 14 MHz.

Figure 10.8 — Patterns on 10 meters for same antenna con-

figurations as in Figs 10.6 and 10.7. Once again, the
inverted V configuration yields a more omnidirectional pat-
tern, but at the expense of almost 8 dB less gain than the Figure 10.10 — Azimuth pattern at a 5° elevation angle for a
flattop configuration at its strongest lobes. 102-foot long, 50-foot high G5RV dipole (solid line). For
comparison, the response for a 132-foot long, center-fed
dipole at 50 feet height (dashed line) and a 33-foot long half-
wave 20 meter dipole at 50 feet (dotted line) are also shown.
The longest antenna exhibits about 0.5 dB more gain than
for the line at the junction with the dipole wires. This will the G5RV, although the response is more omnidirectional
prevent flexing of the transmission-line wire, since excessive for the G5RV — an advantage for a wire antenna that is not
usually rotatable.
flexing will result in breakage. (See the Antenna Materials
and Construction chapter)

10.1.5 THE G5RV MULTIBAND ANTENNA

A variation on the center-fed antenna that does not re- an analysis of the G5RV feed point impedance shows there
quire a lot of space, is simple to construct and low in cost is no length of balanced line of any characteristic impedance
is the G5RV. Designed in England by Louis Varney, G5RV that will transform the terminal impedance to the 50 to 75-W
(SK), some years ago, it has become quite popular in the US. range on all bands. (Low SWR with coax feed and no match-
(The original article by G5RV in the RSGB Bulletin is in- ing network on bands other than 14 MHz may indicate exces-
cluded on this book’s CD-ROM.) The G5RV design is shown sive losses in the coax.)
in Figure 10.9. The antenna may be used from 3.5 through Figure 10.10 shows the 20 meter azimuthal pattern
30 MHz. Although some amateurs claim it may be fed di- for a G5RV at a height of 50 feet over flat ground, at an
rectly with 50-W coax on several amateur bands with a low elevation angle of 5° that is suitable for DXing. For com-
SWR, Varney himself recommended the use of an antenna parison, the response for two other antennas is also shown
tuner on bands other than 14 MHz (see Bibliography). In fact, in Figure 10.10 — a standard half wave 20 meter dipole at
10-6   Chapter 10
50 feet and a 132-foot long center-fed dipole at 50 feet. The
G5RV on 20 meters is, of course, longer than a standard half
wave dipole and it exhibits about 2 dB more gain compared
to that dipole. With four lobes making it look rather like a
four-leaf clover, the azimuth pattern is more omnidirectional
than the two-lobed dipole. The 132-foot center-fed dipole
is longer than the G5RV and it has about 0.5 dB more gain
than the G5RV, also exhibiting four major lobes, along with
two strong minor lobes in the plane of the wire. Overall, the
azimuthal response for the G5RV is more omnidirectional
than the comparison antennas.
The G5RV patterns for other frequencies are similar
Figure 10.11 — The Windom antenna, cut for a fundamental
to those shown for the 135-foot dipole previously for other frequency of 3.75 MHz. The single-wire feed line, connected
frequencies. Incidentally, you may be wondering why the 14% off center, is brought into the station and the system is
pattern for a 132-foot dipole is shown in Figure 10.10, rath- fed against ground. The antenna is also effective on its
harmonics.
er than the 135-foot dipole described earlier. The 132-foot
overall length describes another antenna that we’ll discuss
in the next section on Windom antennas.
The portion of the G5RV antenna shown as horizontal
in Figure 10.9 may also be installed in an inverted V dipole
arrangement, subject to the same loss of peak gain mentioned
above for the 135-foot dipole. Or instead, up to 1⁄6 of the total
length of the antenna at each end may be dropped vertically
or semi-vertically, or bent at a convenient angle to the main
axis of the antenna, to cut down on the requirements for real
estate.

10.1.6 THE WINDOM AND

CAROLINA WINDOM
An antenna that enjoyed popularity in the 1930s and into
the 1940s was what we now call the Windom. It was known
at the time as a “single-feeder Hertz” antenna, after being
described in September 1929 QST by Loren G. Windom,
W8GZ (see Bibliography).
The Windom antenna, shown in Figure 10.11, is fed
with a single wire, attached approximately 14% off center.
Figure 10.12 — Layout for flattop “Carolina Windom”
In theory, this location provides a match for the single-wire antenna.
transmission line, which is driven against an earth ground.
Because the single-wire feed line is not inherently well bal-
anced and because it is brought to the operating position, “RF
in the shack” is a likely result of using this antenna. For that balun) at the bottom end and a 4:1 “matching unit” (imped-
reason, the true single-feed-wire Windom antenna is rarely ance transformer) at the top. The system takes advantage of
used although the name is often given to wires with non- the asymmetry of the horizontal wires to purposely induce
centered feed points as described in the next section. current onto the outer shield surface of the vertical coax
A recent variation is called the “Carolina Windom,” section. Note that the matching unit is a voltage-type balun
apparently because two of the designers, Edgar Lambert, ­transformer, which purposely does not act like a common-
WA4LVB, and Joe Wright, W4UEB, lived in coastal North mode current choke balun. You must use an antenna tuner
Carolina (the third, Jim Wilkie, WY4R, lived in nearby with this system to present a 1:1 SWR to the transmitter on
Norfolk, Virginia). One of the interesting parts about the the amateur bands from 80 through 10 meters.
Carolina Windom is that it turns a potential disadvantage — The radiation resulting from current induced onto the
feed line radiation — into a potential advantage. 22-foot vertical coax section tends to fill in the deep nulls that
Figure 10.12 is a diagram of a flattop Carolina Windom, would be present if the 132 feet of horizontal wire were center
which uses a 50-foot wire joined with an 83-foot wire at fed. Over saltwater, the vertical radiator can give significant
the feed point insulator. This resembles the layout shown in gain at the low elevation angles needed for DX work. Indeed,
Figure 10.11 for the original W8GZ Windom. The “Vertical field reports for the Carolina Windom are most impressive
Radiator” for the Carolina Windom is a 22-foot piece of for stations located near or on saltwater. Over average soil the
RG-8X coax, with a “line Isolator” (current-type choke advantage of the additional vertically polarized component is

Multiband HF Antennas   10-7

 Figure 10.13 — At A, 20 meter azimuth patterns for a 132-foot
long off-center fed Carolina Windom and a 132-foot long cen-
ter-fed flattop dipole on 20 meters, both at a height of 50 feet
above saltwater. The response for the Carolina Windom is
more omnidirectional because the vertically polarized radia-
tion from the 22-foot long vertical RG-8X coax fills in the
deep nulls. At B, 10 meter azimuthal responses for a 132-foot
long, 50-foot high Carolina Windom over saltwater (solid line)
and over average ground (dashed line), compared to that for
a 20 meter half-wave dipole at 50 feet (dotted line).

ham bands, use the minimum length of feed coax possible to

hold down losses in the coax.
Figure 10.13B shows the azimuth responses for a 50-foot
high flattop Carolina Windom on 28.4 MHz over saltwater
and over average soil. The pattern for a 50-foot high, flattop
20 meter dipole operated on 28.4 MHz is also shown, since
this 20 meter dipole can also be used as a multiband antenna,
when fed with open-wire transmission line rather than with
coax. Again, the Carolina Windom exhibits a more omnidi-
rectional pattern, even if the pattern is somewhat lopsided at
the bottom.

10.1.7 OFF-CENTER-FED (OCF) DIPOLES

The usual practice is to feed a l/2 dipole in the center
where the feed point impedance is low and makes a suitable
match to coaxial cables. The dipole will accept energy from
a feed point anywhere along its length, however, assuming
that the source is matched to the higher impedance that is
presented away from the center point. (As discussed in the
Dipoles and Monopoles chapter, if the feed point is moved
away from the center of the dipole, the impedance rises be-
cause current is dropping while voltage is rising.)
The off-center-fed dipole takes advantage of placing the
feed point in a location along the dipole at which the im-
pedance is similar on more than one band, generally in the
neighborhood of 150-300 W. A suitable impedance match-
ing device such as an impedance transformer is then used to
reduce the feed point impedance to something closer to 50 W.
Note that the feed point impedance of the antenna varies with
height above ground and so will SWR.
Figure 10.14 shows an off-center-fed or OCF ­dipole.

not quite so evident. Figure 10.13A compares a 50-foot high

Carolina Windom on 14 MHz over saltwater to a 50-foot
high, 132-foot long, flattop center-fed dipole. The Carolina
Windom has a more omnidirectional azimuthal pattern, a de-
sirable characteristic in a 132-foot long wire antenna that is
not normally rotated to favor different directions.
Another advantage of the Carolina Windom over a
traditional Windom is that the coax feed line between the
transmitter and common-mode choke balun does not radiate,
meaning that there will be less “RF in the shack.” Since the Figure 10.14 — The off-center-fed (OCF) dipole for 3.5, 7
feed line is not always operating at a low SWR on various and 14 MHz. A 1:4 or 1:6 balun is used at the feed point.

10-8   Chapter 10
Because it is similar in appearance to the Windom of
Figure 10.11, this antenna is often mistakenly called a
“Windom,” or sometimes a “coax-fed Windom.” The two an-
tennas are not the same, since the Windom is driven against
an earth ground, while the OCF dipole is fed like a regular
dipole — just not at its center. The extreme case of an OCF is
the end-fed Zepp where the feed point is moved all the way
to the end of the antenna.
The OCF dipole of Figure 10.14, fed 1⁄3 of its length from
one end, may be used on its fundamental and even harmon-
ics. Its free-space antenna-terminal impedance at 3.5, 7 and
14 MHz is on the order of 150 to 200 W. A 4:1 impedance
transformer at the feed point should offer a reasonably good
match to 50- or 75-W line, although some commercially
made OCF dipoles use a 6:1 transformer. The usual caution is
repeated here about height above ground affecting feed point
impedance.
At the 6th harmonic, 21 MHz, the antenna is three wave-
lengths long and fed at a voltage maximum instead of a cur-
rent maximum. The feed point impedance at this frequency
is high, a few thousand ohms, so the antenna is unsuitable for
use on this band.

Balun Requirements Figure 10.15 — At A, multiband antenna using paralleled

Because the OCF dipole is not fed at the center of the ra- dipoles all connected to a common low-impedance trans-
mission line. The half-wave dimensions may be either for
diator, the feed line is not placed symmetrically with respect the centers of the various bands or selected to fit favorite
to the antenna’s radiated field. As a result, common-mode frequencies in each band. Because of interaction among the
current will flow on the feed line, usually a coaxial cable. various dipoles, the builder should expect to adjust lengths
How much current flows depends on the impedance of the for resonance on each band. B shows a method of con-
structing the dipole that offers less interaction between the
coaxial cable’s outer surface which, in turn, depends on the dipoles, making it easy to tune.
orientation of the cable, how long it is, height above ground,
and so forth. (Some of the common-mode current results
from the slightly unequal impedances presented by the OCF
legs but most of the shield current is induced by the asym- the resonant dipole which “ignores” the nonresonant dipoles.
metric location in the antenna’s field.) In theory, the 4-wire antenna of Figure 10.15A can be
Regardless of how the common-mode current is caused used with a coaxial feed line on five bands. The four wires are
to flow on the feed line, it is generally viewed as undesir- prepared as parallel-fed dipoles for 3.5, 7, 14 and 28 MHz.
able and a current or choke balun is used to increase the The 7-MHz dipole is intended to be used on its 3rd harmonic
impedance of coaxial cable’s outer surface. Radiation from for 21-MHz operation to cover a fifth band. However, in prac-
the feed line may not be a problem in your installation and tice it has been found difficult to get a good match to coaxial
may even improve the antenna’s radiation pattern by filling line on all bands.
in nulls. (See “Feed line radiation” above.) In that case, no The l/2 resonant length of any one dipole in the pres-
balun is required. (Choke baluns are discussed in the chapter ence of the others is not the same as for a dipole by itself
Transmission Line Coupling and Impedance Matching.) due to interaction and attempts to optimize all four lengths
can become a frustrating procedure. The problem is com-
10.1.8 MULTIPLE-DIPOLE ANTENNAS pounded because the optimum tuning changes in a different
The antenna system shown in Figure 10.15A consists antenna environment, so what works for one amateur may
of a group of center-fed dipoles, all connected in parallel at not work for another. The builder should start with a single
the point where the transmission line joins them. Each of the dipole longer than resonance as discussed in the Dipoles and
dipole elements is individually constructed to be an electrical Monopoles chapter and be prepared to make repeated adjust-
l/2 at different frequencies. This is often referred to as a “fan ments to the dipole lengths as more dipoles are added to the
dipole,” although that term is also applied to a dipole con- antenna.
structed as a bow-tie to increase operating bandwidth. (See Even if a perfect match cannot be obtained on all bands,
the section “Broadband Dipoles” in the chapter Single-Band many amateurs with limited antenna space are willing to ac-
MF and HF Antennas.) The general idea is that the feed cept the mismatch on some bands just so they can operate on
point impedance of the dipoles far from resonance will be those frequencies using a single coax feed line. The fewer
high enough that nearly all of the signal power is applied to dipoles that are used in parallel, the easier it will be to adjust

Multiband HF Antennas   10-9

them for the desired performance. makes two dipoles. Thus, two lengths, as shown in the sketch,
If an attempt is made to model the multi-wire dipole, serve to make dipoles for four bands. Be sure to use twin lead
take extra care to define the feed point construction carefully. with copper-clad steel conductors because all of the weight,
As noted in the Antenna Modeling chapter, wires that are including that of the feed line, must be supported by the up-
very close to each other or that join at small angles are hard permost wire (450-W window line could also be used).
to model so that the results reflect actual performance. Two pieces of twin lead are first cut to a length suit-
The multiple-dipole antenna can be fed with parallel- able for the two halves of the longest dipole. Then one of
wire feed line and an antenna tuner but that negates the the conductors in each piece is cut to proper length for the
intended advantage of the design over a conventional single- next band higher in frequency. The excess wire and insula-
wire nonresonant dipole — the use of a single coaxial feed tion is stripped away. A second pair of lengths is prepared in
line. The usual feed method is to use a coaxial feed line and the same manner, except that the lengths are appropriate for
a choke balun at the feed point as described in the chapter the next two higher frequency bands. (Note the potential for
Transmission Line Coupling and Impedance Matching. interaction between higher and lower-frequency dipoles that
The separation between the dipoles for the various fre- may alter the tuning of previously adjusted dipoles.)
quencies does not seem to be especially critical. One set of A piece of thick plastic sheet (plexiglass, polycarbonate,
wires can be suspended from the next larger set, using insu- or high-density polyethylene) drilled with holes for anchor-
lating spreaders (of the type used for feed line spreaders) to ing each wire serves as the central insulator. The shorter pair
give a separation of a few inches. Users of this antenna often of dipoles is suspended the width of the ribbon below the
run some of the dipoles at right angles to each other to help longer pair by clamps also made of poly sheet. Intermediate
reduce interaction. Some operators use inverted V-mounted spacers are made by sawing slots in pieces of poly sheet so
dipoles as guy wires for the mast that supports the antenna they will fit the ribbon snugly.
system. The top (and longest) dipole must support the weight The multiple-dipole principle can also be applied to
of the rest of the antenna plus the feed line, so use heavy wire vertical antennas. Parallel or fanned l/4 elements of wire or
(copper-clad steel is the strongest) for the top antenna. tubing can be driven against ground or tuned radials from a
While the separation between dipoles does not seem common feed point.
to be especially critical to final performance, it does affect
the amount of interaction between them that makes tuning Double-L Antenna
each dipole difficult. A method of construction and tuning The Double-L antenna by Don Toman, K2LQ is a
reported by Don Butler, N4UJW (www.hamuniverse.com/ variation of the multi-wire dipole. (www.yccc.org/Articles/
multidipole.html) is shown is Figure 10.15B. For dipoles in double_l.htm) Shown in Figure 10.17, the antenna is
the 2-18 MHz range, separating the dipoles at the feed point
by at least 51⁄2 inches vertically and at the ends by 38 inches
results in a final length closer than ±2% of a single dipole.
An interesting method of construction used successfully
by Louis Richard, ON4UF, is shown in Figure 10.16. The
antenna has four dipoles (for 7, 14, 21 and 28 MHz) con-
structed from 300-W twin lead. A single length of twin lead

Figure 10.17 — The Double-L antenna by K2LQ is a pair of

vertical dipoles with their ends bent to be parallel to the
ground. The bottom horizontal wires should be at least
Figure 10.16 — Sketch showing how the twin-lead multiple- 10 feet above ground. For single-band operation, install only
dipole antenna system is assembled. The excess wire and a single dipole. The antenna works well as either a single-
insulation are stripped away. band or dual-band antenna.

10-10   Chapter 10
basically a vertical dipole with its ends bent to run horizon- omnidirectional antenna for regional communications on its
tally over ground. It can be constructed as a single antenna for fundamental frequency where its radiation is a maximum at
one band or a second dipole can be added to use the antenna high angles. The loop is also useful on higher bands where
on two bands. the pattern begins to divide into multiple lobes at lower eleva-
Construction is not critical. The bottom wires should be tion angles.
at least 10 feet above ground and no radial system is required. While the feed point impedance might be reasonably low
If you do construct the dual-band version, the vertical wires on some bands, using a coax feed line will result in significant
are connected together at the feed point and separated by losses on others. The best way to feed this versatile antenna
about 3 feet where they bend to become horizontal. The two is with parallel-wire window or ladder line using an antenna
horizontal sections are separated by about 30°. If the anten- tuner in the shack.
nas are supported by a metal tower, the vertical section should The Loop Skywire is shown in Figure 10.19. The anten-
be at least 3 feet from the tower. na has one wavelength of wire in its perimeter at the design
The antenna is inherently unbalanced and may be or fundamental frequency. If you choose to calculate Ltotal in
tuned by removing or adding wire to the lower legs without feet, the following equation should be used:
dramatically affecting performance or feed point imped-
Ltotal = 1005 / f
ance. The dimensions given result in an SWR minimum
near 1.83 MHz and 3.75 MHz. where f equals the frequency in MHz.
Loop shapes other than a square are possible, but the
10.1.9 TERMINATED FOLDED DIPOLE larger the area enclosed by the loop, the better its perfor-
A broad-banded variation of the folded dipole called the mance will be. (A circle encloses the maximum area but this
terminated folded dipole (TFD) adds a 600-W terminating is rarely practical.) The Loop Skywire can also be operated
resistor in the top conductor. The antenna is also called the as a vertical antenna with top-hat loading by tying both feed
terminated tilted folded dipole or T2FD. (See Figure 10.18) line conductors together at the antenna tuner. This method
The function of the resistor is to act as a swamping load, requires good station ground as described in the previous sec-
reducing the higher feed point impedances over a wide fre- tion on Random-Wire Antennas.
quency range. A TFD can be constructed to cover the entire Although the loop can be made for any band or frequen-
2 to 30 MHz range with SWR of 3:1 or less. The resistor cy of operation, the following two Loop Skywires are good
dissipates some of the transmitter power (more than 50% at performers. The 10-MHz band can also be used on both.
some frequencies!), but the improvement in SWR allows a
coaxial feed line to be used without an impedance-matching 3.5-MHz Loop Skywire
unit. The increased convenience and installation outweigh (3.5-28 MHz loop and 1.8-MHz vertical)
the reduction in radiated signal. TD antennas are popular for Total loop perimeter: 272 feet
emcomm operations and where only a single HF antenna can Square side length: 68 feet
be installed and high performance is not required. A com- 7-MHz Loop Skywire
mercial version, the BWD-90, is available from B&W. (7-28 MHz loop and 3.5-MHz vertical)
Total loop perimeter: 142 feet
10.1.10 HORIZONTAL LOOP “SKYWIRE” Square side length: 35.5 feet
A horizontal full-wavelength loop is a very effective

Figure 10.18 — The terminated

folded dipole adds a 600-W termi-
nating resistor to smooth out the
feed point impedance variations
over a wide range. The terminat-
ing resistor dissipates some of
the input power in order to pro-
vide a consistent match.

Multiband HF Antennas   10-11

 Figure 10.19 — A complete
view of the Loop Skywire.
The square loop is erected
horizontally.

The actual total length can vary from the above by a few
feet, as the length is not at all critical. Do not worry about
tuning and pruning the loop to resonance as it will not make
a significant difference in performance.
Bare #14 AWG wire is used in the loop. Copper-clad
steel wire is recommended for the 3.5-MHz version. Fig-
ure 10.20 shows the placement of the insulators at the loop
corners. Two common methods are used to attach the insula-
tors. Either lock or tie the insulator in place with a loop wire
tie, as shown in Figure 10.20A, or leave the insulator free
to “float” or slide along the wire, Figure 10.20B. Most loop
users float at least two insulators. This allows pulling the
slack out of the loop once it is in the air, and eliminates
the need to have all the supports exactly placed for proper
tension in each leg. Floating two opposite corners is
recommended.
Figure 10.21A shows the azimuth performance on 7.2
MHz of a 142-foot long, 7-MHz Loop Skywire, 40 feet high
at an elevation angle of 10º, compared to a regular flattop 1⁄2-l
dipole at a height of 30 feet. The loop comes into its own at
higher frequencies. Figure 10.21B shows the response at 14.2
MHz, compared again to a 1⁄2-l 14.2-MHz dipole at a height
of 30 feet. Now the loop has several lobes that are stronger
than the dipole. Figure 10.21C shows the response at 21.2 Figure 10.20 — Two methods of installing the insulators at
the loop corners.
MHz, compared to a dipole. Now the loop has superior gain
compared to the 1⁄2-l dipole at almost any azimuth. In its fa-
vored direction on 21.2 MHz, the loop is 8 dB stronger than
the dipole. freely and keeps the feed line free of the loop support.
The feed point can be positioned anywhere along the loop Generally a minimum of four supports is required. If trees
that you wish. However, most users feed the Skywire at a cor- are used for supports, then at least two of the ropes or guys
ner. The feed line can be attached as described in the Antenna used to support the insulators should be counterweighted and
Materials and Construction chapter. Placing the feed point allowed to move freely. The feed line corner is almost always
a foot or so from one corner allows the feed line to exit more tied down, however. Very little tension is needed to support
10-12   Chapter 10
 Figure 10.21 — At A, azimuth response of
142-foot long, 7-MHz Loop Skywire, 40 feet in the
air at 7.2 MHz, compared with 1/2-l dipole 30 feet
in the air. At B, response of same Loop Skywire
at 14.2 MHz, compared with 1/2-l 14.2-MHz dipole
30 feet in the air. Now the loop has some advan-
tage in certain directions. At C, response of the
same Loop Skywire at 21.2 MHz compared to a
21.2-MHz dipole at 30 feet. Here, the Loop
Skywire has more gain in almost all directions
than the simple dipole. All azimuth patterns were
made at 10° elevation.

the loop (far less than that for a dipole). Thus, counterweights Recommended height for the antenna is 40 feet or more.
are light. Several such loops have been constructed with bun- Higher is better, especially if you wish to use the loop in the
gee cords tied to three of the four insulators. This eliminates vertical mode. However, successful local and DX operation
the need for counterweighting. has been reported in several cases with the antenna at 20 feet.

10.2 Trap Antennas

By using tuned circuits of appropriate design strategi- an explanation of how a trap antenna works can be elusive.
cally placed in a dipole, the antenna can be made to show For some designs, traps are resonated in our amateur bands,
what is essentially fundamental resonance at a number of and for others (especially commercially made antennas) the
different frequencies. The general principle is illustrated traps are resonant far outside any amateur band.
by Figure 10.22. The tuned circuits are also referred to as A trap in an antenna system can perform either of two
“traps” and so an antenna that uses tuned circuits to change functions, depending on whether or not it is resonant at the
its electrical configuration at different frequencies is called a operating frequency. A familiar case is where the trap is
“trap antenna” or a “trapped antenna.” parallel-resonant in an amateur band. For the moment, let
Even though a trap antenna arrangement is a simple one, us assume that dimension A in Figure 10.22 is 32 feet and
Multiband HF Antennas   10-13
 5 µH and the capacitor 100 pF (reactances of approximately
224 W at 7.1 MHz). But the choice of these values will make
a significant difference in the antenna size for resonance at
3.5 MHz. In the first case, where the L/C ratio is 2000, the
necessary length of section B of the antenna for resonance at
3.75 MHz would be approximately 28.25 feet. In the second
case, where the L/C ratio is 50,000, this length need be only
24.0 feet, a difference of more than 15%.
The above example concerns a two-band antenna with
trap resonance at one of the two frequencies of operation. On
each of the two bands, each half of the dipole operates as an
electrical l/4. However, the same band coverage can be ob-
Figure 10.22 — A trap dipole antenna. This antenna may be tained with a trap resonant at, say, 5 MHz, a frequency quite
fed with 50-W coaxial line. Depending on the L/C ratio of the removed from either amateur band. With proper selection of
trap elements and the lengths chosen for dimensions A and
B, the traps may be resonant either in an amateur band or at the L/C ratio and the dimensions for A and B, the trap will act
a frequency far removed from an amateur band for proper to shorten the antenna electrically at 7 MHz and lengthen it
two-band antenna operation. electrically at 3.5 MHz. Thus, an antenna that is intermediate
in physical length between being full size on 3.5 MHz and
full size on 7 MHz can cover both bands, even though the trap
is not resonant at either frequency. Again, the antenna oper-
that each L/C combination is resonant in the 7-MHz band. ates with electrical l/4 sections. Note that such nonresonant
Because of its parallel resonance, the trap presents a high traps have less RF current flowing in the trap components,
impedance at that point in the antenna system. The electrical and hence trap losses are less than for resonant traps.
effect at 7 MHz is that the trap behaves as an open circuit. It Additional traps may be added in an antenna section to
serves to separate the outside ends, the B sections, from the cover three or more bands. Or a judicious choice of dimen-
inner sections of the antenna. The result is easy to visualize sions and the L/C ratio may permit operation on three or more
— we now have an antenna system that is resonant in the bands with just a pair of identical traps in the dipole.
7-MHz band. Each 33-foot section (labeled A in the draw- An important point to remember about traps is this. If the
ing) represents l/4 with the trap acting as an open circuit. We operating frequency is below that of trap resonance, the trap
therefore have a full-size 7-MHz antenna. behaves as an inductor; if above, as a capacitor. The above
The second function of a trap, obtained when the fre- discussion is based on dipoles that operate electrically as l/2
quency of operation is not the resonant frequency of the antennas. This is not a requirement, however. Elements may
trap, is one of electrical loading. If the operating frequency be operated as electrical 3/2 l, or even 5/2 l, and still present
is below the trap’s resonant frequency, the trap behaves as a reasonable impedance to a coaxial feed line. In trap antennas
an inductor; if above, as a capacitor. Inductive loading will covering several HF bands, using electrical lengths that are
electrically lengthen the antenna, and capacitive loading will odd multiples of l/2 is often done at the higher frequencies.
electrically shorten the antenna. To further aid in understanding trap operation, let’s now
Let’s carry our assumption a bit further and try using choose trap L and C components that each have a reactance of
the antenna we just considered at 3.5 MHz. With the traps 20 W at 7 MHz. Inductive reactance is directly proportional to
resonant in the 7-MHz band, they will behave as inductors frequency, and capacitive reactance is inversely proportional.
when operation takes place at 3.5 MHz, electrically lengthen- When we shift operation to the 3.5-MHz band, the induc-
ing the antenna. This means that the total length of sections A tive reactance becomes 10 W, and the capacitive reactance
and B (plus the length of the inductor) may be something less becomes 40 W. At first thought, it may seem that the trap
than a physical l/4 for resonance at 3.5 MHz. Thus, we have would become capacitive at 3.5 MHz with a higher capacitive
a two-band antenna that is shorter than full size on the lower reactance, and that the extra capacitive reactance would make
frequency band. But with the electrical loading provided by the antenna electrically shorter yet. Fortunately, this is not the
the traps, the overall electrical length is l/2. The total antenna case. The inductor and the capacitor are connected in parallel
length needed for resonance in the 3.5-MHz band will de- with each other.
pend on the L/C ratio of the trap elements.
XL XC
The key to trap operation away from resonance is its L/C Z= j (Eq 1)
ratio, the ratio of the value of L to the value of C. At reso- XL + XC
nance, however, within practical limitations the L/C ratio is where j indicates a reactive impedance component, rather
immaterial as far as electrical operation goes. For example, than resistive. A positive result indicates inductive reactance,
in the antenna we’ve been discussing, it would make no and a negative result indicates capacitive. In this 3.5-MHz
difference for 7-MHz operation whether the inductor were case, with 40 W of capacitive reactance and 10 W of induc-
1 µH and the capacitor were 500 pF (the reactances would be tive, the equivalent series reactance is 13.3 W inductive.
just below 45 W at 7.1 MHz), or whether the inductor were This inductive loading lengthens the antenna to an electrical

10-14   Chapter 10
l/2 overall at 3.5 MHz, assuming the B end sections in Fig- one set of traps is used, resonant at 7 MHz to isolate the in-
ure 10.22 are of the proper length. ner (7-MHz) dipole from the outer sections. This causes the
With the above reactance values providing resonance at overall system to be resonant in the 3.5-MHz band. On 14, 21
7 MHz, XL equals XC, and the theoretical series equivalent is and 28 MHz the antenna works on the capacitive-reactance
infinity. This provides the open-switch effect, disconnecting principle just outlined. With a 75-W feed line, the SWR with
the antenna ends. this antenna is under 2:1 throughout the three highest fre-
At 14 MHz, where XL = 40 W and XC = 10 W, the resul- quency bands, and the SWR is comparable with that obtained
tant series equivalent trap reactance is 13.3 W capacitive. If with similarly fed simple dipoles on 3.5 and 7 MHz. (The
the total physical antenna length is slightly longer than 3/2 l complete article is available on this book’s CD-ROM.)
at 14 MHz, this trap reactance at 14 MHz can be used to
shorten the antenna to an electrical 3/2 l. In this way, three- Trap Construction
band operation is obtained for 3.5, 7 and 14 MHz with just Traps frequently are built with coaxial aluminum tubes
one pair of identical traps. The design of such a system is (usually with plastic tubing in-between them for insulation)
not straightforward, however, because any chosen L/C ratio for the capacitor, with the coil either self-supporting or wound
for a given total length affects the resonant frequency of the on a form of larger diameter than the tubular capacitor. The
antenna on both the 3.5 and 14-MHz bands. coil is then mounted coaxially with the capacitor to form a
unit assembly that can be supported at each end by the anten-
10.2.1 TRAP LOSSES na wires. In another type of trap devised by William J. Lattin,
Since the tuned circuits have some inherent losses, the W4JRW (see Bibliography at the end of this chapter), the coil
efficiency of a trap system depends on the unloaded Q values is supported inside an aluminum tube and the trap capacitor
of the tuned circuits. Low-loss (high-Q) coils should be used, is obtained in the form of capacitance between the coil and
and the capacitor losses likewise should be kept as low as the outer tube. This type of trap is inherently weatherproof.
possible. With tuned circuits that are good in this respect — A simpler type of trap can be easily assembled from
comparable with the low-loss components used in transmit- readily available components. A small transmitting-type
ter tank circuits, for example — the reduction in efficiency ceramic “doorknob” capacitor is used, together with a
compared with the efficiency of a simple dipole is small, but length of commercially available coil material, these being
tuned circuits of low unloaded Q can lose an appreciable por- supported by an ordinary ceramic or plastic antenna strain
tion of the power supplied to the antenna. insulator 41⁄4 inches long. The circuit constants and antenna
The commentary above applies to traps assembled from dimensions differ slightly from those of Figure 10.23, in
conventional components. The important function of a trap order to bring the antenna resonance points closer to the
that is resonant in an amateur band is to provide a high isolat- centers of the various phone bands. Construction data are
ing impedance, and this impedance is directly proportional given in Figure 10.24. If a 10-turn length of inductor is
to Q. Unfortunately, high Q restricts the antenna bandwidth, used, a half turn from each end may be used to slip through
because the traps provide maximum isolation only at trap the anchor holes in the insulator to act as leads.
resonance. The components used in these traps are sufficiently
weatherproof in themselves so that no additional weather-
10.2.2 FIVE-BAND W3DZZ TRAP ANTENNA proofing has been found necessary. However, if it is desired
C. L. Buchanan, W3DZZ, created one of the first trap
antennas for the five pre-1979 WARC amateur bands from
3.5 to 30 MHz. Dimensions are given in Figure 10.23. Only

Figure 10.24 — Layout of multiband antenna using traps

constructed as shown in Figure 10.25. The capacitors are
Figure 10.23 — Five-band (3.5, 7, 14, 21 and 28 MHz) trap 100 pF each, transmitting type, 5000-volt dc rating
dipole for operation with 75-W feed line at low SWR (C. L. (Centralab 850SL-100N). Coils are 9 turns of #12 AWG wire,
Buchanan, W3DZZ). The balanced (parallel-conductor) line 2-1/2 inches diameter, 6 turns per inch (B&W 3029) with end
indicated is desirable, but 75-W coax can be substituted turns spread as necessary to resonate the traps to 7.2 MHz.
with a choke balun at the feed point to maintain symmetry. These traps, with the wire dimensions shown, resonate the
Dimensions given are for resonance (lowest SWR) at 3.75, antenna at approximately the following frequencies on each
7.2, 14.15 and 29.5 MHz. Resonance is very broad on the band: 3.9, 7.25, 14.1, 21.5 and 29.9 MHz (based on measure-
21-MHz band, with SWR less than 2:1 throughout the band. ments by W9YJH).

Multiband HF Antennas   10-15

to protect them from the accumulation of snow or ice, a plas- 15 and 10 meter antenna. The radiating elements are made
tic cover can be made by cutting two discs of plastic slightly of #14 AWG stranded copper wire. The element lengths are
larger in diameter than the coil, drilling at the center to pass the wire span lengths in feet. These lengths do not include
the antenna wires, and cementing a plastic cylinder on the the lengths of the pigtails at the balun, traps and insulators.
edges of the discs. The cylinder can be made by wrapping The 32.3-foot-long inner 40 meter segments are measured
two turns or so of 0.02-inch plastic sheet around the discs, if from the eyelet of the input balun to the tension-relief hole
no suitable ready-made tubing is available. Plastic drinking in the trap coil form. The 4.9-foot segment length is mea-
glasses and 2-liter soft-drink plastic bottles are easily adapt- sured from the tension-relief hole in the trap to the 6-foot
able for use as impromptu trap covers. stub. The 16.l-foot outer-segment span is measured from the
stub to the eyelet of the end insulator.
10.2.3 W8NX MULTIBAND, The coaxial-cable traps are wound on PVC pipe coil
COAX-TRAP DIPOLES forms and use the low-impedance output connection. The
Over the last 60 or 70 years, amateurs have used many stubs are 6-foot lengths of 1⁄8-inch stiffened aluminum or
kinds of multiband antennas to cover the traditional HF copper rod hanging perpendicular to the radiating elements.
bands. The availability of the 30, 17 and 12 meter bands has The first inch of their length is bent 90° to permit attachment
expanded our need for multiband antenna coverage. This to the radiating elements by large-diameter copper crimp
section is based on the August 1994 QST article “Two New connectors. Ordinary #14 AWG wire may be used for the
Multiband Trap Dipoles” by Al Buxton, W8NX. This article stubs, but it has a tendency to curl up and may tangle unless
and two others by the same author are included on this book’s weighed down at the end. You should feed the antenna with
CD-ROM, providing designs for trap dipoles operating on all 75-W coaxial cable using a choke balun.
of the amateur bands below 30 MHz. This antenna may be thought of as a modified W3DZZ
Two different antennas are described here. The first cov- antenna due to the addition of the capacitive stubs. The length
ers the traditional 80, 40, 20, 15 and 10 meter bands, and and location of the stub give the antenna designer two extra
the second covers 80, 40, 17 and 12 meters. Each uses the degrees of freedom to place the resonant frequencies within
same type of W8NX trap — connected for different modes the amateur bands. This additional flexibility is particularly
of operation — and a pair of short capacitive stubs to en- helpful to bring the 15 and 10 meter resonant frequencies to
hance coverage. The W8NX coaxial-cable traps have two more desirable locations in these bands. The actual 10 meter
different modes: a high- and a low-impedance mode. The resonant frequency of the original W3DZZ antenna is some-
inner-conductor windings and shield windings of the traps what above 30 MHz, pretty remote from the more desirable
are connected in series for both modes. However, either low frequency end of 10 meters.
the low- or high-impedance point can be used as the trap’s
output terminal. For low-impedance trap operation, only the 80, 40, 17 and 12 meter Dipole
center conductor turns of the trap windings are used. For Figure 10.26 shows the configuration of the 80, 40,
high-impedance operation, all turns are used, in the conven- 17 and 12 meter antenna. Notice that the capacitive stubs
tional manner for a trap. The short stubs on each antenna are are attached immediately outboard after the traps and are
strategically sized and located to permit more flexibility in 6.5 feet long, 1⁄2 foot longer than those used in the other
adjusting the resonant frequencies of the antenna. ­antenna. The traps are the same as those of the other antenna,
but are connected for the high-impedance parallel-resonant
80, 40, 20, 15 and 10 meter Dipole output mode. Since only four bands are covered by this
Figure 10.25 shows the configuration of the 80, 40, 20, antenna, it is easier to fine tune it to precisely the desired

Figure 10.25 — A W8NX multiband dipole

for 80, 40, 20, 15 and 10 meters. The values
shown (123 pF and 4 µH) for the coaxial-
cable traps are for parallel resonance at
7.15 MHz. The low-impedance output of
each trap is used for this antenna.

Figure 10.26 — A W8NX multiband

dipole for 80, 40, 17 and 12 meters. For
this antenna, the high-impedance out-
put is used on each trap. The resonant
frequency of the traps is 7.15 MHz.

10-16   Chapter 10
Figure 10.27 — Schematic for the W8NX coaxial-cable trap.
RG-59 is wound on a 23⁄8-inch OD PVC pipe.

Figure 10.28 — Construction details of the W8NX coax-

ial-cable trap.
frequency on all bands. The 12.4-foot tips can be pruned
to a particular 17 meter frequency with little effect on the
12 meter frequency. The stub lengths can be pruned to a par-
ticular 12 meter frequency with little effect on the 17 meter
frequency. Both such pruning adjustments slightly alter the
80 meter resonant frequency. However, the bandwidths of the
antennas are so broad on 17 and 12 meters that little need
for such pruning exists. The 40 meter frequency is nearly
independent of adjustments to the capacitive stubs and outer
radiating tip elements. Like the first antennas, this dipole is
fed with a balun and 75-W feed line.
Figure 10.27 shows the schematic diagram of the traps. It
explains the difference between the low and high-impedance
modes of the traps. Notice that the high-impedance terminal
is the output configuration used in most conventional trap
applications. The low-impedance connection is made across
only the inner conductor turns, corresponding to one-half of Figure 10.29 — Other views of a W8NX coax-cable trap.
the total turns of the trap. This mode steps
the trap’s impedance down to approximately
one-fourth of that of the high-impedance
level. This is what allows a single trap de-
sign to be used for two different multiband
antennas.
Figure 10.28 is a drawing of a cross-
section of the coax trap shown through the
long axis of the trap. Notice that the traps
are conventional coaxial-cable traps, except
for the added low-impedance output ter-
minal. The traps are 83⁄4 close-spaced turns
of RG-59 (Belden 8241) on a 23⁄8-inch-OD
PVC pipe (schedule 40 pipe with a 2-inch
ID) coil form. The forms are 41⁄8 inches long.
Trap resonant frequency is very sensitive to
the outer diameter of the coil form, so check Figure 10.30 — Additional construction details for the W8NX coaxial-cable trap.
it carefully. Unfortunately, not all PVC pipe
is made with the same wall thickness. The trap frequencies cable. (See the discussion on waterproofing in the Building
should be checked with a dip meter and general-coverage Antenna Systems and Towers chapter.)
receiver and adjusted to within 50 kHz of the 7150 kHz reso- Also, be sure that you connect the 32.3-foot wire element
nant frequency before installation. One inch is left over at at the start of the inner conductor winding of the trap. This
each end of the coil forms to allow for the coax feed-through avoids detuning the antenna by the stray capacitance of the
holes and holes for tension-relief attachment of the antenna coaxial-cable shield. The trap output terminal (which has the
radiating elements to the traps. Be sure to seal the ends of the shield stray capacitance) should be at the outboard side of the
trap coax cable to prevent moisture from entering the coaxial trap. Reversing the input and output terminals of the trap will
Multiband HF Antennas   10-17
lower the 40 meter frequency by approximately 50 kHz, but major end-fire lobes. On 80 and 40 meters, the antenna has
there will be negligible effect on the other bands. the usual figure-8 patterns of a half-wave-length dipole.
Figure 10.29 shows a coaxial-cable trap. Further details Both antennas function as electrical half-wave dipoles
of the trap installation are shown in Figure 10.30. This draw- on 80 and 40 meters with a low SWR. They both function
ing applies specifically to the 80, 40, 20, 15 and l0 meter as odd-harmonic current-fed dipoles on their other operating
antenna, which uses the low-impedance trap connections. frequencies, with higher, but still acceptable, SWR. The pres-
Notice the lengths of the trap pigtails: 3 to 4 inches at each ence of the stubs can either raise or lower the input imped-
terminal of the trap. If you use a different arrangement, you ance of the antenna from those of the usual third and fifth
must modify the span lengths accordingly. All connections harmonic dipoles. Again W8NX recommends that 75-W,
can be made using crimp connectors rather than by soldering. rather than 50-W, feed line be used because of the generally
Access to the trap’s interior is attained more easily with a higher input impedances at the harmonic operating frequen-
crimping tool than with a soldering iron. cies of the antennas.
The SWR curves of both antennas were carefully
Performance measured using a 75 to 50-W transformer from Palomar
The performance of both antennas has been very satisfac- Engineers inserted at the junction of the 75-W coax feed
tory. W8NX uses the 80, 40, 17 and 12 meter version because line and a 50-W SWR bridge. The transformer is required
it covers 17 and 12 meters. (He has a tribander for 20, 15 and for accurate SWR measurement if a 50-W SWR bridge is
10 meters.) The radiation pattern on 17 meters is that of a used with a 75-W line. Most 50-W rigs operate satisfacto-
3
⁄2-wave dipole. On 12 meters, the pattern is that of a 5⁄2-wave rily with a 75-W line, although this requires different tuning
dipole. At his location in Akron, Ohio, the antenna runs es- and load settings in the final output stage of a vacuum tube
sentially east and west. It is installed as an inverted V, 40 feet amplifier or antenna tuner. The author uses the 75 to 50-W
high at the center, with a 120° included angle between the transformer only when making SWR measurements and at
legs. Since the stubs are very short, they radiate little power low power levels. The transformer is rated for 100 W, and
and make only minor contributions to the radiation patterns. when he runs his 1-kW PEP linear amplifier the transformer
In theory, the pattern has four major lobes on 17 meters, with is taken out of the line.
maxima to the northeast, southeast, southwest and northwest. Figure 10.31 gives the SWR curves of the 80, 40, 20,
These provide low-angle radiation into Europe, Africa, South 15 and 10 meter antenna. Minimum SWR is nearly 1:1 on
Pacific, Japan and Alaska. A narrow pair of minor broad- 80 meters, 1.5:1 on 40 meters, 1.6:1 on 20 meters, and 1.5:1
side lobes provides north and south coverage into Central on 10 meters. The minimum SWR is slightly below 3:1 on
America, South America and the polar regions. 15 meters. On 15 meters, the stub capacitive reactance com-
There are four major lobes on 12 meters, giving nearly bines with the inductive reactance of the outer segment of
end-fire radiation and good low-angle east and west cover- the antenna to produce a resonant rise that raises the antenna
age. There are also three pairs of very narrow, nearly broad- input resistance to about 220 W, higher than that of the usual
side, minor lobes on 12 meters, down about 6 dB from the 3
⁄2-wavelength dipole. An antenna tuner may be required on

Figure 10.31 — Measured SWR curves for an 80, 40, 20, 15 Figure 10.32 — Measured SWR curves for an 80, 40, 17 and
and 10 meter antenna, installed as an inverted V with 40-ft 12 meter antenna, installed as an inverted V with 40-ft apex
apex and 120° included angle between legs. and 120° included angle between legs.

10-18   Chapter 10
this band to keep a solid-state final output stage happy under
these load conditions. Table 10-2
Trap Loss Analysis: 80, 40, 20, 15, 10 meter Antenna
Figure 10.32 shows the SWR curves of the 80, 40, 17
and 12 meter antenna. Notice the excellent 80 meter perfor- Frequency (MHz) 3.8 7.15 14.18 21.3 28.6
Radiation Efficiency (%) 96.4 70.8 99.4 99.9 100.0
mance with a nearly unity minimum SWR in the middle of
Trap Losses (dB) 0.16 1.5 0.02 0.01 0.003
the band. The performance approaches that of a full-size 80
meter wire dipole. The short stubs and the low-inductance
traps shorten the antenna somewhat on 80 meters. Also ob-
serve the good 17 meter performance, with the SWR being Table 10-3
only a little above 2:1 across the band. Trap Loss Analysis: 80, 40, 17, 12 meter Antenna
But notice the 12 meter SWR curve of this antenna,
Frequency (MHz) 3.8 7.15 18.1 24.9
which shows 4:1 SWR across the band. The antenna input Radiation Efficiency (%) 89.5 90.5 99.3 99.8
resistance approaches 300 W on this band because the ca- Trap Losses (dB) 0.5 0.4 0.03 0.006
pacitive reactance of the stubs combines with the inductive
reactance of the outer antenna segments to give resonant rises
in impedance. These are reflected back to the input terminals.
These stub-induced resonant impedance rises are similar to increase in magnitude with upward frequency extrapolation.
those on the other antenna on 15 meters, but are even more Results are believed to be accurate within 4% on 80 and
pronounced. 40 meters, but only within l0 to 15% at 10 meters. Trap Q
High SWR in coaxial cables longer than about 100 feet is shown at both the high- and low-impedance trap termi-
can lead to high feed line losses as shown in the Transmission nals. The Q at the low-impedance output terminals is 15
Lines chapter. If you plan on operating this antenna with an to 20% lower than the Q at the high-impedance output
SWR of greater than 3:1, make sure the amount of feed line terminals.
loss is acceptable. W8NX computer-analyzed trap losses for both antennas
High voltages in the feed line should not cause too much in free space. Antenna-input resistances at resonance were
concern. Even if the SWR is as high as 9:1 no destructively first calculated, assuming lossless, infinite-Q traps. They
high voltages will exist on the transmission line. Recall that were again calculated using the Q values in Table 10-1. The
transmission-line voltages increase as the square root of the radiation efficiencies were also converted into equivalent
SWR in the line. Thus, 1 kW of RF power in 75-W line cor- trap losses in decibels. Table 10-2 summarizes the trap-loss
responds to 274 V line voltage for a 1:1 SWR. Raising the analysis for the 80, 40, 20, 15 and 10 meter antenna and
SWR to 9:1 merely triples the maximum voltage that the Table 10-3 for the 80, 40, 17 and 12 meter antenna.
line must withstand to 822 V. This voltage is well below the The loss analysis shows radiation efficiencies of 90%
3700-V rating of RG-11, or the 1700-V rating of RG-59, the or more for both antennas on all bands except for the 80, 40,
two most popular 75-W coax lines. Voltage breakdown in the 20, 15 and 10 meter antenna when used on 40 meters. Here,
traps is also very unlikely. As will be pointed out later, the the radiation efficiency falls to 70.8%. A 1-kW power level
operating power levels of these antennas are limited by RF at 90% radiation efficiency corresponds to 50-W dissipation
power dissipation in the traps, not trap voltage breakdown or per trap. In W8NX’s experience, this is the trap’s survival
feed line SWR. limit for extended key-down operation. SSB power levels of
1 kW PEP would dissipate 25 W or less in each trap. This is
Trap Losses and Power Rating well within the dissipation capability of the traps.
Table 10-1 presents the results of trap Q measurements When the 80, 40, 20, 15 and 10 meter antenna is operated
and extrapolation by a two-frequency method to higher on 40 meters, the radiation efficiency of 70.8% corresponds
frequencies above resonance. W8NX employed a Boonton to a dissipation of 146 W in each trap when 1 kW is deliv-
Q meter for the measurements. Extrapolation to higher- ered to the antenna. This is sure to burn out the traps — even
frequency bands assumes that trap resistance losses rise if sustained for only a short time. Thus, the power should
with skin effect according to the square root of frequency, be limited to less than 300 W when this antenna is operated
and that trap dielectric loses rise directly with frequency. on 40 meters under prolonged key-down conditions such as
Systematic measurement errors are not increased by fre- RTTY. A 50% CW duty cycle would correspond to a 600-W
quency extrapolation. However, random measurement errors power limit for normal 40 meter CW operation. Likewise,
a 50% duty cycle for 40 meter SSB corre-
sponds to a 600-W PEP power limit for the
Table 10-1 antenna.
Trap Q The author knows of no analysis where
Frequency (MHz) 3.8 7.15 14.18 18.1 21.3 24.9 28.6
the burnout wattage rating of traps has been
High Z out (W) 101 124 139 165 73 179 186 rigorously determined. Operating experi-
Low Z out (W) 83 103 125 137 44 149 155 ence seems to be the best way to determine
trap burn-out ratings. In his own experience

Multiband HF Antennas   10-19

with these antennas, he’s had no traps burn out, even though of RG-59 coax seems rather high. However, W8NX has found
he operated the 80, 40, 20, 15 and 10 meter antenna on the no coax other than RG-59 that has the necessary inductance-
critical 40 meter band using his AL-80A linear amplifier at to-capacitance ratio to create the trap characteristic reac-
the 600-W PEP output level. He did not make a continuous, tance required for the 80, 40, 20, 15 and 10 meter antenna.
key-down, CW operating test at full power purposely trying Conventional traps with wide-spaced, open-air inductors and
to destroy the traps! appropriate fixed-value capacitors could be substituted for the
Some hams may suggest using a different type of coaxial coax traps, but the convenience, weatherproof configuration
cable for the traps. The dc resistance of 40.7 W per 1000 feet and ease of fabrication of coaxial-cable traps is hard to beat.

10.3 Multiband Vertical Antennas

There are two basic types of vertical antennas; either will be more than adequate for the higher frequency bands.
type can be used in multiband configurations. The first is the
ground-mounted vertical and the second, the ground plane. 10.3.1 FULL-SIZE VERTICAL ANTENNAS
These antennas are described in detail in the chapter Dipoles A vertical antenna should not be longer than about 3⁄4 l at
and Monopoles. the highest frequency to be used, however, if low-angle radia-
The efficiency of any ground-mounted vertical depends tion is wanted. You can see why from reviewing the radiation
a great deal on near-field earth losses. As pointed out in patterns for dipoles in the chapter Dipoles and Monopoles.
the chapter Effects of Ground, these near-field losses can As the antenna lengthens, the pattern breaks up into lobes
be reduced or eliminated with an adequate radial system. that are at high elevation angles for a vertical antenna.
Considerable experimentation has been conducted on this Nevertheless, an antenna that is l/4 on the lower frequency
subject by Jerry Sevick, W2FMI (SK), and several important of operation can still be useful over a 3:1 frequency range or
results were obtained. It was determined that a radial system even more if the high-angle radiation can be tolerated. For
consisting of 40 to 50 radials, 0.2 l long, would reduce the example, an 80 meter l/4 vertical around 66 feet high is use-
earth losses to about 2 W when a l/4 radiator was being used. ful through the 30 meter band and a 25-foot vertical would be
These radials should be on the earth’s surface, or if buried, useful from about 10 MHz through the 28 MHz band.
placed not more than an inch or so below ground. Otherwise, In recent years, the 43-foot ground-mounted vertical an-
the RF current would have to travel through the lossy earth tenna with an automatic antenna tuner mounted at the base
before reaching the radials. In a multiband vertical system, of the antenna has become very popular as an all-band HF
the radials should be 0.2 l long for the lowest band, that is, vertical, including 160 meters with the appropriate tuner. See
55 feet long for 3.5-MHz operation. Any wire size may be Figure 10.33. While the elevation angle of maximum radia-
used for the radials. The radials should fan out in a circle, tion begins to increase significantly above the 20 meter band,
radiating from the base of the antenna. A metal plate, such as the combination of simplicity and clean appearance make up
a piece of sheet copper, can be used at the center connection. for the compromise. (A variation on this idea is the “flag-
The other common type of vertical is the ground-plane pole” antenna discussed in the Stealth and Limited Space
antenna. Normally, this antenna is mounted above ground Antennas chapter.) If the lower bands are not required, a
with the radials fanning out from the base of the antenna. 22-foot vertical is quite effective at and above 40 meters.
The vertical portion of the antenna is usually an electrical l/4, The antennas can be constructed from aluminum tubing or as
as is each of the radials. In this type of antenna, the system of a fiberglass mast with wires inside or taped along the out-
radials acts somewhat like an RF choke, to prevent RF cur- side of the mast.
rents from flowing in the supporting structure, so the number In lieu of using an automatic antenna tuner at the base
of radials is not as important a factor as it is with a ground- of the vertical, several QST articles listed in the Bibliography
mounted vertical system. From a practical standpoint, the serve as examples of how a single vertical antenna can be
customary number of radials is four or five. In a multiband put to work on several bands. The referenced articles by Phil
configuration, l/4 radials are required for each band of opera- Salas, AD5X discuss matching the antenna’s impedance on
tion with the ground-plane antenna. 160 and 80 meters.
This is not so with the ground-mounted vertical antenna,
where the ground plane is relied upon to provide an image of 10.3.2 SHORT VERTICAL ANTENNAS
the radiating section. Note that even quarter-wave-long radi- A short vertical antenna (one less than l/4 at the oper-
als are greatly detuned by their proximity to ground — radial ating frequency) can be operated on several bands by load-
resonance is not necessary or even possible. In the ground- ing it at the base, the general arrangement being similar to
mounted case, so long as the ground-screen radials are ap- Figures 10.1 and 10.2. That is, for multiband operation the
proximately 0.2 l long at the lowest frequency, the length vertical can be handled by the same methods that are used for

10-20   Chapter 10
 Figure 10.33 — A ver-
tical whip antenna
tuned by an auto-
matic antenna tuner
at its base is an effec-
tive multi-band
antenna. A 43-foot
whip is a popular Figure 10.34 — Multiband vertical antenna system using base
choice, presenting loading for resonating on 3.5 to 28 MHz. L1 should be wound
reasonable feed point with bare wire so it can be tapped at every turn, using #12
impedances from 80 AWG wire. A convenient size is 21⁄2 inches in diameter, 6 turns
through 10 meters. per inch (such as B&W 3029). Number of turns required
depends on antenna and ground lead length, more turns
being required as the antenna and ground lead are made
shorter. For a 25-foot antenna and a ground lead of the order
of 5 feet, L1 should have about 30 turns. The use of C1 is
explained in the text. The smallest capacitance that will per-
mit matching the coax cable should be used; a maximum
capacitance of 100 to 150 pF will be sufficient in any case.

be run from the antenna base to the nearest convenient loca-

tion for mounting L1 and C1. The extra wire will of course
be a part of the antenna, and since it may have to run through
unfavorable surroundings it is best to avoid using it if at all
possible. (Use the shortest possible ground connection both
for efficiency and to avoid creating an unintended radiating
element of the antenna system.)
This system is best adjusted with the help of an SWR
indicator. Connect the coax line across a few turns of L1 and
take trial positions of the shorting tap until the SWR reaches
its lowest value. Then vary the feed line tap similarly; this
should bring the SWR down to a low value. Small adjust-
ments of both taps then should reduce the SWR to close to
1:1. If not, try adding C1 and go through the same procedure,
varying C1 each time a tap position is changed.
random-length wires.
Another method of feeding is shown in Figure 10.34. L1 10.3.3 TRAP VERTICALS
is a loading coil, tapped to resonate the antenna on the desired The trap principle described in Figure 10.21 for center-
band. A second tap permits using the coil as a transformer for fed dipoles also can be used for vertical antennas. There are
matching a coax line to the transmitter. C1 is not strictly nec- two principal differences. Only one half of the dipole is used,
essary, but may be helpful on the lower frequencies, 3.5 and the ground connection taking the place of the missing half,
7 MHz, if the antenna is quite short. In that case C1 makes it and the feed point impedance is one half the feed point imped-
possible to tune the system to resonance with a coil of reason- ance of a dipole. Thus it is in the vicinity of 30 W (plus the
able dimensions at L1. C1 may also be useful on other bands ground-connection resistance), so 52-W cable should be used
as well, if the system cannot be matched to the feed line with since it is the commonly available type that comes closest to
a coil alone. (This is similar to the techniques described in the matching.
chapter Mobile and Maritime HF Antennas.) Commercial multiband trap verticals such as the Hustler
The coil and capacitor should preferably be installed at 4/5/6BTV series and the Hy-Gain AVQ series have been
the base of the antenna, but if this cannot be done a wire can widely used for many years and provide effective performance
Multiband HF Antennas   10-21
as ground-mounted antennas when used with a good radial reduced to 50 W with a matching network at the base of the
system. antenna. These are particularly useful antennas for tempo-
Verticals advertised as “ground-independent” are in- rary stations and when restrictions prevent the installation of
tended to be mounted above ground. Models such as the ground systems.
Cushcraft R8 and R6000 and the Hy-Gain Patriot are end-fed Most amateurs prefer to purchase multiband trap ver-
systems that are electrically longer than l/4 at the frequency ticals because of the mechanical complexities and require-
of operation. They have a high feed point impedance that is ments to be self-supporting.

10.4 The Open-Sleeve Antenna

Although only recently adapted for the HF and VHF ground conductivity and number of radials. If two sleeve ele-
amateur bands, the open-sleeve antenna has been around ments were added on either side of the central monopole,
since 1946. The antenna was invented by Dr J. T. Bolljahn, of with each approximately half the height of the monopole and
Stanford Research Institute. This section on sleeve antennas at a distance equal to their height, there would be very little
summarizes material by Roger A. Cox, WBØDGF in previ- effect on the antenna mode impedance, ZA, at 14 MHz.
ous editions. The complete article is available on this book’s Also, ZT at 14 MHz would be the end impedance trans-
CD-ROM. formed through a l/8 section of a very high characteristic
The basic form of the open-sleeve monopole is shown in
Figure 10.35. The open-sleeve monopole consists of a base-
fed central monopole with two parallel closely spaced para-
sitics, one on each side of the central element, and grounded
at each base. The lengths of the parasitics are roughly one
half that of the central monopole.

10.4.1 IMPEDANCE
The operation of the open sleeve can be divided into two
modes, an antenna-mode and a transmission-line mode. This
is shown in Figure 10.36.
The antenna-mode impedance, ZA, is determined by
the length and diameter of the central monopole. For sleeve
lengths less than that of the monopole, this impedance is es-
sentially independent of the sleeve dimensions.
The transmission-line mode impedance, ZT, is deter-
mined by the characteristic impedance, end impedance, and
length of the 3- wire transmission line formed by the central
monopole and the two sleeve elements. The characteristic im-
pedance, Zc, can be determined by the element diameters and
spacing if all element diameters are equal, and is found from

Zc = 207 log 1.59 (D/d) (Eq 2)

where
D = spacing between the center of each sleeve ele-
ment and the center of the driven element Figure 10.35 — Diagram of an open-sleeve monopole.
d = diameter of each element

This is shown graphically in Figure 10.37. However,

since the end impedance is usually unknown, there is little
need to know the characteristic impedance. The transmis-
sion-line mode impedance, ZT, is usually determined by an
educated guess and experimentation.
As an example, let us consider the case where the central
monopole is l/4 at 14 MHz. It would have an antenna mode
impedance, ZA, of approximately 52 W, depending upon the Figure 10.36 — Equivalent circuit of an open-sleeve antenna.

10-22   Chapter 10
 Figure 10.38 — Impedance of an open-sleeve monopole for
the frequency range 13.5-15 MHz. Curve A is for a 14 MHz
monopole alone. For curves B, C and D, the respective
spacings from the central monopole to the sleeve elements
are 8, 6 and 4 inches. See text for other dimensions.
Figure 10.37 — Characteristic impedance of transmission-
line mode in an open-sleeve antenna.

impedance transmission line. Therefore, ZT would be on the

order of 500-2000 W resistive plus a large capacitive reac-
tance component. This high impedance in parallel with 52 W
would still give a resulting impedance close to 52 W.
At a frequency of 28 MHz, however, ZA is that of
an end-fed half-wave antenna, and is on the order of 1000-
5000 W resistive. Also, ZT at 28 MHz would be on the
order of 1000 to 5000 W resistive, since it is the end im-
pedance of the sleeve elements transformed through a
quarter-wave section of a very high characteristic imped-
ance three-wire transmission line. Therefore, the parallel
combination of ZA and ZT would still be on the order of 500
to 2500 W resistive.
The actual impedance plots of a 14/28-MHz open-sleeve
monopole appear in Figures 10.38 and 10.39. The length
of the central monopole is 195.5 inches, and of the sleeve
elements 89.5 inches. The element diameters range from
1.25 inches at the bases to 0.875 inch at each tip. The mea- Figure 10.39 — Impedance of the open-sleeve monopole for
sured impedance of the 14-MHz monopole alone, curve A the range 25-30 MHz. For curves A, B and C the spacings
from the central monopole to the sleeve elements are 8, 6
of Figure 10.38, is quite high. This is probably because of a and 4 inches, respectively.
very poor ground plane under the antenna. The addition of
the sleeve elements raises this impedance slightly, curves B,
C and D.
As curves A and B in Figure 10.39 show, an 8-inch spacing is found, the lengths of the sleeve elements can be
sleeve spacing gives a resonance near 27.8 MHz at 70 W, tweaked slightly for a choice of resonant frequency.
while a 6-inch spacing gives a resonance near 28.5 MHz at In other frequency combinations such as 10/21, 10/24,
42 W. Closer spacings give lower impedances and higher 14/21 and 14/24-MHz, spacings in the 6 to 10-inch range
resonances. The optimum spacing for this particular antenna work very well with element diameters in the 0.5 to 1.25-inch
would be somewhere between 6 and 8 inches. Once the range.
Multiband HF Antennas   10-23
 10.4.2 BANDWIDTH and spacings are such that the two sleeve elements approach
The open-sleeve antenna, when used as a multiband an- an interelement spacing of l/8, the azimuthal pattern will
tenna, does not exhibit broad SWR bandwidths unless the show directivity typical of two in-phase vertical radiators,
two bands are very close together. For example, Figure 10.40 approximately l/8 apart. If a bidirectional pattern is needed,
shows the return loss and SWR of a single 10-MHz verti- then this is one way to achieve it.
cal antenna. Its 2:1 SWR bandwidth is 1.5 MHz, from 9.8 to Spacings closer than this will produce nearly circular
11.3 MHz. Return loss and SWR are related as given by the azimuthal radiation patterns. Practical designs in the 10 to
following equation. 30 MHz range using 0.5 to 1.5-inch diameter elements will
produce azimuthal patterns that vary less than ±1 dB.
1+ k If the ratio of the length of the central monopole to the
SWR = (Eq 3)
1− k length of the sleeves approaches 2:1, then the elevation pat-
tern of the open-sleeve vertical antenna at the resonant fre-
where RL
quency of the sleeves becomes slightly compressed. This is
k = 10 20
because of the in-phase contribution of radiation from the l/2
RL = return loss, dB
central monopole.
The third, fifth, and seventh-order resonances of the
10.4.3 RADIATION PATTERN AND GAIN sleeve elements and the central monopole element can be
used, but their radiation patterns normally consist of high-
The current distribution of the open-sleeve antenna where
elevation lobes, and the gain on the horizon is less than that
all three elements are nearly equal in length is nearly that of
of a l/4 vertical.
a single monopole antenna. If, at a particular frequency, the
elements are approximately l/4 long, the current distribution 10.4.4 CONSTRUCTION AND EVALUATION
is sinusoidal.
The open-sleeve antenna lends itself very easily to home
If, for this and other length ratios, the chosen diameters
construction. For the open-sleeve vertical antenna, only a feed
point insulator and a good supply of aluminum tubing are
needed. No special traps or matching networks are required.
The open-sleeve vertical can produce up to 3 dB more gain
than a conventional l/4 vertical. Further, there is no reduction
in bandwidth, because there are no loading coils.
The open-sleeve design can also be adapted to horizontal
dipole and beam antennas for HF, VHF and UHF. A good
example of this is Hy-Gain’s Explorer 14 triband beam which
utilizes an open sleeve for the 10/15 meter driven element.
The open-sleeve antenna is also very easy to model in com-
puter programs such as NEC and MININEC, because of the
open tubular construction and lack of traps or other intricate
structures.
Figure 10.40 — Return loss and SWR of a 10 MHz vertical In conclusion, the open-sleeve antenna is an antenna ex-
antenna. A return loss of 0 dB represents an SWR of infinity.
The text contains an equation for converting return loss to perimenter’s delight. It is not difficult to match or construct,
an SWR value. and it makes an ideal broadband or multiband antenna.

10.5 The Coupled-Resonator Dipole

A variation of the open-sleeve system above is the cou- AL7KK, showed us how to make a three-band vertical. Both
pled-resonator system described by Gary Breed, K9AY, in of these antennas achieve multi-frequency operation by plac-
an article in The ARRL Antenna Compendium, Vol 5, entitled ing resonant conductors very close to a driven dipole or verti-
“The Coupled-Resonator Principle: A Flexible Method for cal — with no physical connection.
Multiband Antennas.” The following is condensed from that
article. 10.5.1 THE COUPLED-RESONATOR
In 1995, QST published two antenna designs that use PRINCIPLE
an interesting technique to get multiband coverage in one As we all know, nearby conductors can interact with an
antenna. Rudy Severns, N6LF, described a wideband 80 antenna. Our dipoles, verticals and beams can be affected by
and 75 meter dipole using this technique (see the Single nearby power lines, rain gutters, guy wires and other metallic
Band MF and HF Antennas chapter), and Robert Wilson, materials. The antennas designed by Severns and Wilson use
10-24   Chapter 10
Figure 10.41 — At A, the SWR of a dipole over a wide frequency range. At B, a nearby conductor is just close enough to
interact with the dipole. At C, when the second conductor is at the optimum spacing, the combination is matched at both
frequencies.

this interaction intentionally, to combine the resonances of

several conductors at a single feed point. While other names
have been used, I call the behavior that makes these antennas
work the coupled-resonator (C-R) principle.
Take a look at Figure 10.41, which illustrates the general
idea. Each figure shows the SWR at the feed point of a dipole,
over a range of frequencies. When this dipole is all alone, it
will have a very low SWR at its half-wave resonant frequency
(Figure 10.41A). Next, if we take another wire or tubing con-
ductor and start bringing it close to the dipole, we will see a
“bump” in the dipole’s SWR at the resonant frequency of this
new wire. See Figure 10.41B. We are beginning to the see
the effects of interaction between the two conductors. As we
bring this new conductor closer, we reach a point where the
SWR “bump” has grown to a very deep dip — a low SWR.
We now have a good match at both the original dipole’s reso-
nant frequency and the frequency of the new conductor, as
illustrated in Figure 10.41C.
We can repeat this process for several more conductors
at other frequencies to get a dipole with three, four, five, six, Figure 10.42 — Evolution of coupled-resonator antennas:
or more resonant frequencies. The principle also applies to At A, the coaxial-sleeve dipole; at B, the open-sleeve dipole;
and at C, a coupled-resonator dipole, the most universal
verticals, so any reference to a dipole can be considered to be configuration.
valid for a vertical, as well.
We can write a definition of the C-R principle this way:
Given a dipole (or vertical) at one frequency and an addi-
tional conductor resonant at another frequency, there is an its driven element to obtain resonance in the 10 meter band.
optimum distance between them that results in the resonance Later on, a few antenna developers finally figured out that
of the additional conductor being imposed upon the original these extra conductors did not need to be added in pairs, and
dipole, resulting in a low SWR at both resonant frequencies. that a single conductor at each frequency could add the ex-
tra resonances (Figure 10.42C). This is the method used by
Some History Force 12 in some of their multiband antennas.
In the late 1940s, the coaxial sleeve antenna was devel- This is a perfect example of how science works. A spe-
oped (Figure 10.42), covering two frequencies by surround- cific idea is discovered, with later developments leading to
ing a dipole or monopole with a cylindrical tube resonant at an underlying general principle. The original coaxial-sleeve
the higher of the desired frequencies. In the 1950s, Gonset configuration is the most specific, being limited to two fre-
briefly marketed a two-band antenna based on this design. quencies and requiring a particular construction method. The
Other experimenters soon determined that two conductors open-sleeve antenna is an intermediate step, showing that the
at the second frequency, placed on either side of the main sleeve idea is not limited to one configuration.
dipole or monopole, would make a skeleton representation Finally, we have the coupled-resonator concept, which
of a cylinder (Figure 10.42B). This is called the open-sleeve is the general principle, applicable in many different antenna
antenna. The Hy-Gain Explorer tribander uses this method in configurations, for many different frequency combinations.
Multiband HF Antennas   10-25
Severns’s antenna uses it with a folded dipole, and Wilson where
uses it with a main vertical that is off-center fed. The author, d = distance between conductors, measured in wave-
K9AY used it with conventional dipoles and quarter-wave lengths at the frequency of the chosen additional
verticals. Other designers have used the principle more sub- resonator
tly, like putting the first director in a Yagi very close to the D = the diameter of the conductors, also in wavelengths at
driven element, broadening the SWR bandwidth the same the frequency of the additional resonator.
way Severns’s design does with a dipole.
In the past, most open-sleeve or multiple-open-sleeve Eq 4 assumes they are both the same diameter and that
antennas built with this technique have also been called open- the feed point impedance at both frequencies is the same as a
sleeve (or multiple-open-sleeve) antennas, a term taken from dipole in free space (72 W) or a quarter-wave monopole over
the history of their development. However, the term sleeve perfect ground (36 W).
implies that one conductor must surround another. This is The equation only describes the impedance due to the
not really a physical or electrical description of the antenna’s additional resonator. The main dipole element is always part
operation, therefore, K9AY suggests using the term coupled- of the antenna, and it may have a fairly low impedance at the
resonator, which is the most accurate description of the gen- additional frequency. This is the case when the frequencies
eral principle. are close together, or when the main element is operating at
its third harmonic. At these frequencies, the spacing distance
A Little Math must be adjusted so that the parallel combination of dipole
The interaction that makes the C-R principle work is not and resonator results in the desired feed point impedance.
random. It behaves in a predictable, regular manner. K9AY K9AY worked out two correction factors, one to cover
derived an equation that shows the relationship between the a range of impedances and another for frequencies close to-
driven element and the additional resonators for ordinary gether. These can be included in the basic equation, which
dipoles and verticals: is rearranged below to solve for the distance between the
conductors:
log10 d Z 0 + 35.5 
= 0.54 (Eq 4) d = 10 0.54 log10 (D/4) × × 1 + e −[((((F2 /F1 )−1.1) × 11.3)+ 0.1)] 
log10 (D / 4) 109  
(Eq 5)

C-R Element Spacing

K9AY’s Eq 5 presented in the text does indeed yield On 21 MHz, the system will exhibit a F/B of 1.6 dB, favor-
a good “first-cut” value for the spacing between coupled- ing the forward direction. Of course, there are systems
resonator elements. Figure 10.A shows the spacing, in where gain and F/B due to the C-R configuration may be
inches, plotted against the ratio of frequencies, for two put to good use, such as the multiband Yagis mentioned
coupled resonator elements with different diameters, above. However, if the elements are spaced above/below
again expressed in inches. This is for an upper frequen- the 14-MHz driven element there is no distortion of the
cy of 28.4 MHz. Beyond a frequency ratio of about 1.5:1 dipole patterns.
(28.4:18.1 MHz), the spacing flattens out to a fixed dis-
tance between elements for each element diameter.
For example, if 1/2-inch elements are used at 28.4 and
18.1 MHz, the spacing between the elements is about K9AY Coupled Resonators, 28.4 MHz, 50 Ohms
3.75 inches. Spacing in Inches for Different Diameter Elements
18.00
EZNEC verifies Eq 5’s computations. Note that a
ANT0288

large number of segments are necessary for each ele- 16.00

ment when they are closely spaced from each other, 14.00

and the segments on the elements must be closely 12.00

Spacing, Inches

aligned with each other. Be sure to run the Average Gain 10.00
test, as well as Segmentation tests. The modeler should 8.00
also be aware that if mutually coupled resonators are 6.00
placed along a horizontal boom (as they would be on 4.00
multiband Yagis using coupled resonators), the higher- 2.00
frequency elements will act like retrograde directors,
0.00
producing some gain (or lack of gain, depending on the 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
azimuth being investigated). Frequency Ratio
For example, in the EZNEC file K9AY C-R 28-21- #14 Wire 0.375" 0.5" 0.625
14 MHz 1 In.EZ, using 1-inch diameter elements spaced 0.75" 0.875" 1.000" 2.000"

6 inches apart, if the 28-MHz element is placed 6 inches

behind the 14-MHz driven element (with the 21-MHz Figure 10.A — Graph of the spacing versus frequency ratio
element placed 6 inches ahead), on 28 MHz the system for two Coupled-Resonator elements at 28.4 MHz, for 50-W
will have a F/B of 2.6 dB, favoring the rearward direction. feed point impedance.

10-26   Chapter 10
where spacers. Other multiband antennas have their complexities as
d and D are the same as in Eq 4 above. well (such as traps that need to be mounted and tuned), but
Z0 = the desired feed point impedance at the fre- C-R antennas will usually be bulkier. The larger size gener-
quency of the additional resonator (between ally means greater windload, which is a disadvantage to some
20 and 120 W). For a vertical, multiply the hams.
desired impedance by two to get Z0. If you The other significant disadvantage is narrower band-
want a 50-W feed, use 100 W for Z0. width, particularly at the highest of the operating frequencies.
F1 = the resonant frequency of the We can partially overcome this problem with large conduc-
main dipole or vertical. tors that are naturally broad in bandwidth, and in some cases
F2 = the resonant frequency of the additional con- we might even use an extra conductor to put two resonances
ductor. The ratio F2/F1 is more than 1.1. in one band. It is interesting to note that the pattern is opposite
e = 2.7183, the base of natural logarithms. that of trap antennas. The C-R antenna gets narrower at the
highest frequencies of operation, while trap antennas gener-
Eq 5 does not directly allow for conductors of unequal ally have narrowest bandwidth at their lowest frequencies.
diameters, but it can be used as a starting point if you use the There are two special situations that should be noted.
diameter of the driven dipole or vertical element for D in the First, when the antenna has a resonance near the frequency
equation. where the driven dipole is 3⁄2 l long (3⁄4 l for a vertical), the
dipole has a fairly low impedance. The spacing of the C-R
10.5.2 CHARACTERISTICS OF element needs to be increased to raise its impedance so that
COUPLED-RESONATOR (C-R) ANTENNAS the parallel combination of the main element and C-R ele-
Here’s the important stuff — what’s different about C-R ment equals the desired impedance (usually 50 W). There is
antennas, what are they good for and what are their draw- also significant antenna current in the part of the main dipole
backs? The key points are: extending beyond the C-R section, contributing to the total
 Multiband operation without traps, stubs or tuners radiation pattern. As a result, this particular arrangement
 Flexible impedance matching at each frequency radiates as three l/2 sections in phase, and has about 3 dB
 Independent fine-tuning at each frequency (little gain and a narrower directional pattern compared to a dipole
interaction) (Figure 10.43). This might be an advantage for antennas
 Easily modeled using MININEC or NEC-based programs covering bands with a frequency ratio of about three, such as
 Pruning process same as a simple dipole 3.5 and 10.1 MHz, 7 and 21 MHz, or 144 and 430 MHz.
 Can accommodate many frequencies (seven or more) The other special situation is when we want to add a
 Virtually lossless coupling (high efficiency) new frequency very close to the resonant frequency of the
 Requires a separate wire or tubing conductor at each main dipole. An antenna for 80 and 75 meters would be an
frequency
 Mechanical assembly requires a number of insulated
supports
 Narrower bandwidth than equivalent dipole
 Capacitance requires slight lengthening of conductors
To begin with, the most obvious characteristic is that this
principle can be used to add multiple resonant frequencies to
an ordinary dipole or vertical, using additional conductors
that are not physically connected. This gives us three variable
factors: (1) the diameter of the conductor, (2) its length, and
(3) its position relative to the main element.
Having the freedom to control these factors gives us the
advantage of flexibility; we have a wide range of control over
the impedance at each added frequency. Another advantage
is that the behavior at each frequency is quite independent,
once the basic design is in place. In other words, making
fine-tuning adjustments at one frequency doesn’t change
the resonance or impedance at the other frequencies. A final
advantage is efficiency. With conductors close together, and
with a resonant target conductor, coupling is very efficient.
Traps, stubs, and compensating networks found on other
multiband antennas all introduce lossy reactive components.
There are two main disadvantages of C-R antennas. The Figure 10.43 — Radiation pattern for the special case of a
first is the relative complexity of construction. Several con- C-R antenna with the additional resonance at the third
ductors are needed, installed with some type of insulating ­harmonic of the main dipole resonant frequency.

Multiband HF Antennas   10-27

example of this. Again, the driven dipole has a fairly low
impedance at the new frequency. Add the fact that coupling
is very strong between these similar conductors and we find
that a wide spacing is required to make the antenna work.
A dipole resonant at 3.5 MHz and another wire resonant at
3.8 MHz will need to be 3 or 4 feet apart, while a 3.5 MHz
and 7 MHz combination might only need to be spaced 4 or
5 inches.
Another useful characteristic of C-R antennas is that they Figure 10.44 — Dimensions of a C-R dipole for the 30, 17
are easily and accurately modeled by computer programs and 12 meter bands.
based on either MININEC or NEC, as long as you stay within
each program’s limitations. For example, Severns points out
that MININEC does not handle folded dipoles very well, and
NEC modeling is required. With ease of computer modeling, space, so these spacing distances may not be exact. Plugging
a precise answer isn’t needed for the design equation given these numbers into your favorite antenna-modeling program
above. An approximate solution will provide a starting point will let you optimize the dimensions for installation at the
that can quickly be adjusted for optimum dimensions. height you choose.
The added resonators have an effect on the lengths of For those of you who like to work with real antennas,
all conductors, due to the capacitance between the conduc- not computer-generated ones, the predicted spacing is ac-
tors. Capacitance causes antennas to look electrically shorter, curate enough to build an antenna with minimum trial-and-
so each element needs to be about 1% or 2% longer than error. You should use a nice round number just larger than
a simple dipole at the same frequency. As a rule of thumb, the calculated spacing for 50 W. For this antenna, K9AY
use 477/f (in feet) instead of the usual 468/f when calculating decided that the right spacing for the desired height would
dipole length, and 239/f instead of 234/f for a l/4 vertical. be 2 inches for the 18 MHz resonator and 1.8 inches for the
24.9 MHz resonator. For simplicity of construction, he just
Summary used 2 inches for both, figuring that the worst he would get
The coupled-resonator principle is one more weapon in is a 1.2:1 SWR if the numbers were a little bit off. Like all
the antenna designer’s arsenal. It’s not the perfect method for dipoles, the impedance varies with height above ground, but
all multiband antennas, but what the C-R principle offers is the 2-inch spacing results in an excellent match on the two
an alternative to traps and tuners, in exchange for using more additional bands, at heights of more than 25 feet.
wire or aluminum. Although a C-R antenna requires more The final dimensions of the dipole for 10.1, 18.068 and
complicated construction, its main attraction is in making a 24.89 MHz are shown in Figure 10.44. These are the final
multiband antenna that can be built with no compromise in pruned lengths for a straight dipole installed at a height of
matching or efficiency. about 40 feet. If you put up the antenna as an inverted V,
you will need each wire to be a bit longer. Pruning this type
10.5.3 A C-R DIPOLE FOR 30/17/12 METERS of antenna is just like a dipole — if it’s resonant too low in
To show how a C-R antenna is designed, let’s build a di- frequency, it’s too long and the appropriate wire needs to be
pole to cover 30, 17 and 12 meters. We’ll use #12 AWG wire, shortened. So, you can cut the wires just a little long to start
which has a diameter of 0.08 inches, and the main dipole will with and easily prune them to resonance.
be cut for the 10.1 MHz band. From the equation above, the A final note: if you want to duplicate this antenna de-
spacing between the main dipole and the 18-MHz resonator sign, remember that the 2-inch spacing is just for #12 AWG
should be 2.4 inches for 72 W, or 1.875 inches for 50 W. At wire! The required spacing for a C-R antenna is related to the
24.9 MHz, the spacing to the resonator for that band should conductor diameter. This same antenna built with #14 AWG
be 2.0 inches for 72 W, or 1.62 inches for 50 W. Of course, wire needs under 11⁄2-inch spacing, while a 1-inch aluminum-
this antenna will be installed over real ground, not in free tubing version requires about 7-inch spacing.

10-28   Chapter 10
 10.6 HF Log Periodic Dipole Arrays
The log periodic antenna whose theory is presented in has more information about the design procedure for arriving
the chapter Log-Periodic Dipole Arrays is intended to be at the dimensions and other parameters of these arrays. The
used across a wide frequency range. Designs that cover two or primary differences between these designs and one-octave
more amateur bands are fairly common and rotatable LPDAs upper HF arrays are the narrower frequency ranges and the
(Log Periodic Dipole Arrays) are popular antennas for 20 use of wire, rather than tubing, for the elements. As design
meters through the UHF bands. examples for the LPDA, you may wish to work through the
This section presents a pair of LPDA designs — a fixed step-by-step procedure and check your results against the val-
array of wire dipoles for the 3.5 and 7 MHz bands and a rotat- ues in Tables 10-4 and 10-5. You may also wish to compare
able array for the five amateur bands from 14 through 30 MHz. these results with the output of an LPDA design software
In addition, the QST article “Practical High Performance HF package such as LPCAD.
Log Periodic Antennas” by Bill Jones, K8CU is provided on From the design procedure, the feeder wire spacings
this book’s CD-ROM for additional design information. for the two arrays are slightly different, 0.58 inch for the
3.5-MHz array and 0.66 inch for the 7-MHz version. As
10.6.1 LPDAS FOR 3.5 OR 7 MHZ
These wire log-periodic dipole arrays for the lower
HF bands are simple in design and easy to build. They are
designed to have reasonable gain, be inexpensive and light- Table 10-4
weight, and may be assembled with stock items found in large Design Parameters for the 3.5-MHz Single-Band LPDA
hardware stores. They are also strong — they can withstand f1 = 3.3 MHz Element lengths:
a hurricane! These antennas were first described by John J. fn = 4.1 MHz l1 = 149.091 feet
Uhl, KV5E, in QST for August 1986. Figure 10.45 shows B = 1.2424 l2 = 125.982 feet
one method of installation. You can use the information here t = 0.845 l3 = 106.455 feet
s = 0.06 l4 = 89.954 feet
as a guide and point of reference for building similar LPDAs.
Gain = 5.9 dBi = 3.8 dBd Element spacings:
If space is available, the antennas can be rotated or re- cot a = 1.5484 d12=17.891 feet
positioned in azimuth after they are completed. A 75-foot Bar = 1.3864 d23 = 15.118 feet
tower and a clear turning radius of 120 feet around the base Bs = 1.7225 d34 = 12.775 feet
of the tower are needed. The task is simplified if you use L = 48.42 feet Element diameters
N = 4.23 elements (decrease to 4) All = 0.0641 inches
only three anchor points, instead of the five shown in Figure
Zt = 6-inch short jumper l/diameter ratios:
10.45. Omit the two anchor points on the forward element, R0 = 208 W l/diam4 = 16840
and extend the two nylon strings used for element stays all the ZAV = 897.8 W l/diam3 = 19929
way to the forward stay line. s' = 0.06527 l/diam2 = 23585
Z0 = 319.8 W l/diam1 = 27911
Design of the Log-Periodic Dipole Arrays Antenna feeder: #12 AWG wire spaced 0.58 inches
Balun: 4:1
Design constants for the two arrays are listed in Tables Feed line: 52-W coax
10-4 and 10-5. The Log-Periodic Dipole Arrays chapter

Table 10-5
Design Parameters for the 7-MHz Single-Band LPDA
f1 = 6.9 MHz Element lengths:
fn = 7.5 MHz l1 = 71.304 feet
B = 1.0870 l2 = 60.252 feet
t = 0.845 l3 = 50.913 feet
s = 0.06 l4 = 43.022 feet
Gain = 5.9 dBi = 3.8 dBd Element spacings:
cot a = 1.5484 d12 = 8.557 feet
Bar = 1.3864 d23 = 7.230 feet
Bs = 1.5070 d34 = 6.110 feet
L = 18.57 feet Element diameters:
N = 3.44 elements (increase to 4) All = 0.0641 inches
Zt = 6-inch short jumper l/diameter ratios:
R0 = 208 W l4/diam4 = 8054
ZAV = 809.3 W l3/diam3 = 9531
s' = 0.06527 l2/diam2 = 11280
Z0 = 334.2 W l1/diam1 = 13349
Antenna feeder: #12 AWG wire spaced 0.66 inches
Balun: 4:1
Figure 10.45 — Typical lower-HF wire 4-element log periodic Feed line: 52-W coax
dipole array erected on a tower.

Multiband HF Antennas   10-29

a compromise toward the use of common spacers for both the way shown in the drawings prevent the wire from break-
bands, a spacing of 5⁄8 inch is quite satisfactory. Surprisingly, ing. All of the rope, string, and connectors must be made of
the feeder spacing is not at all critical here from a match- materials that can withstand the effects of tension and weath-
ing standpoint, as may be verified from the equations in the ering. Use nylon rope and strings, the type that yachtsmen
Log-Periodic Dipole Arrays chapter. Increasing the spac- use. Figure 10.45 shows the front stay rope coming down to
ing to as much as 3⁄4 inch results in an R0 SWR of less than ground level at a point 120 feet from the base of a 75-foot
1.1:1 on both bands. tower. Space may not be available for this arrangement in all
cases. An alternative installation technique is to put a pulley
Constructing the Arrays 40 feet up in a tree and run the front stay rope through the
Construction techniques are the same for both the 3.5 pulley and down to ground level at the base of the tree. The
and the 7-MHz versions of the array. Once the designs front stay rope will have to be tightened with a block and
are completed, the next step is to fabricate the fittings; see tackle at ground level.
Figure 10.46 for details. Cut the wire elements and feed Putting an LPDA together is not difficult if it is assem-
lines to the proper sizes and mark them for identification. bled in an orderly manner. It is easier to connect the elements
After the wires are cut and placed aside, it will be difficult to to the feeder lines when the feed line assembly is stretched
remember which is which unless they are marked. When you between two points. Use the tower and a block and tackle.
have finished fabricating the connectors and cutting all of Attaching the rear connector to the tower and assembling the
the wires, the antenna can be assembled. Use your ingenuity LPDA at the base of the tower makes raising the antenna into
when building one of these antennas; it isn’t necessary to place a much simpler task. Tie the rear connector securely
duplicate these LPDAs precisely. to the base of the tower and attach the two feeder lines to
The elements are made of standard #14 AWG stranded it. Then thread the two feed line spacers onto the feed line.
copper wire. The two parallel-wire feed lines are made of The spacers will be loose at this time, but will be positioned
#12 AWG solid copper-clad steel wire, such as Copperweld. properly when the elements are connected. Now connect the
Copperweld will not stretch when placed under tension. The front connector to the feed lines. A word of caution: Measure
front and rear connectors are cut from 1⁄2-inch thick polycar- accurately and carefully! Double-check all measurements
bonate sheeting, and the feed line spacers from 1⁄4-inch acrylic before you make permanent connections.
sheeting. Connect the elements to the feeder lines through their
Study the drawings carefully and be familiar with the respective plastic connectors, beginning with element 1, then
way the wire elements are connected to the two feed lines, element 2, and so on. Keep all of the element wires securely
through the front, rear and spacer connectors. Details are coiled. If they unravel, you will have a tangled mess of kinked
sketched in Figures 10.47 and 10.48. Connections made in wire. Recheck the element-to-feeder connections to ensure

Figure 10.46 — Pieces for the LPDA that require fabrication. At A is the forward connector, made from 1⁄2-inch polycarbonate.
At B is the rear connector, also made from 1⁄2-inch polycarbonate. At C is the pattern for the phase-line spacers, made from
1
⁄4-inch acrylic. Two spacers are required for the array.

10-30   Chapter 10
Figure 10.47 — The generic layout for the lower HF wire
LPDA. Use a 4:1 balun on the forward connector. See
Tables 10-4 and 10-5 for dimensions. Figure 10.48 — Details of the electrical and mechanical
connections of the elements to the phase-line. Knots in
the nylon rope stay line are not shown.
proper and secure junctions. (See Figures 10.47 and 10.48.)
Once you have completed all of the element connections, at-
tach the 4:1 balun to the underside of the front connector.
Connect the feeder lines and the coaxial cable to the balun. Performance
You will need a separate piece of rope and a pulley The reports received from these LPDAs were compared
to raise the completed LPDA into position. First secure with an inverted-V dipole. All of the antennas are fixed; the
the eight element ends with nylon string, referring to Fig- LPDAs radiate to the northeast, and the dipole to the northeast
ures 10.45 and 10.47. The string must be long enough to and southwest. The apex of the dipole is at 70 feet, and the
reach the tie-down points. Connect the front stay rope to 40- and 80-meter LPDAs are at 60 and 50 feet, respectively.
the front connector, and the completed LPDA is now ready Basic array gain was apparent from many of the reports re-
to be raised into position. While raising the antenna, uncoil ceived. During pileups, it was possible to break in with a few
the element wires to prevent their getting away and tangling tries on the LPDAs, yet it was impossible to break the same
up into a mess. Use care! Raise the rear connector to the pileups using the dipole. The gain of the LPDAs is several
proper height and attach it securely to the tower, then pull dB over the dipole. For additional gain, experimenters may
the front stay rope tight and secure it. Move the elements so wish to try a parasitic director about 1⁄8 l ahead of the array.
they form a 60° angle with the feed lines, in the direction of Director length and spacing from the forward LPDA element
the front, and space them properly relative to one another. should be field-adjusted for maximum performance while
By adjusting the end positions of the elements as you walk maintaining the impedance match across each of the bands.
back and forth, you will be able to align all the elements Wire LPDA systems offer many possibilities. They are
properly. Now it is time to hook your rig to the system and easy to design and to construct: real advantages in coun-
make some contacts. tries where commercially built antennas and parts are not
Multiband HF Antennas   10-31
available at reasonable cost. The wire needed can be obtained
in all parts of the world, and cost of construction is low. If
damaged, the LPDAs can be repaired easily with pliers and
solder. For those who travel on DXpeditions where space and
weight are large considerations, LPDAs are lightweight but
sturdy, and they perform well.

10.6.2 5-BAND LOG PERIODIC

DIPOLE ARRAY
A rotatable log periodic array designed to cover the fre-
quency range from 13 to 30 MHz is pictured in Figure 10.49.
This is a large array having a free-space gain that varies from
6.6 to over 6.9 dBi, depending upon the operating portion
of the design spectrum. This antenna system was originally
described by Peter D. Rhodes, K4EWG, in November 1973
QST. A measured radiation pattern for the array appears in
Figure 10.50.
The characteristics of this array are:
1) Half-power beamwidth, 43° (14 MHz)
2) Design parameter t = 0.9
3) Relative element spacing constant s = 0.05
4) Boom length, L = 26 feet
5) Longest element l1 = 37 feet 10 inches. Figure 10.50 — Measured radiation pattern of the 13-30 MHz
6) Total weight, 116 pounds LPDA. The front-to-back ratio is about 14 dB at 14 MHz and
7) Wind-load area, 10.7 square feet increases to 21 dB at 28 MHz.
8) Required input impedance (mean resistance),
    R0 = 72 W, Zt = 6-inch jumper #18 AWG wire
9) Average characteristic dipole impedance, ZAV:
    337.8 W 10) Impedance of the feeder, Z0: 117.1 W
11) Feeder: #12 AWG wire, close spaced
12) With a 1:1 toroid balun at the input terminals
    and a 72-W coax feed line, the maximum
    SWR is 1.4:1.
The mechanical assembly uses materials readily avail-
able from most local hardware stores or aluminum supply
houses. A complete set of tables and assembly drawings
are included in the original article included on this book’s
CD-ROM.
Experimenters may wish to improve the performance
of the array at both the upper and lower frequency ends of
the design spectrum so that it more closely approaches the
performance in the middle of the design frequency range.
The most apt general technique for raising both the gain and
the front-to-back ratio at the frequency extremes would be to
circularize t as described in the chapter Log-Periodic Dipole
Figure 10.49 — The 13-30 MHz log periodic dipole array. Arrays. However, other techniques may also be applied.

10-32   Chapter 10
 10.7 HF Discone Antennas
The material in this section is adapted from an ar-
ticle by Daniel A. Krupp, W8NWF, in The ARRL Antenna
Compendium, Vol 5. (Additional articles on discone antennas
are referenced in the Bibliography or included on this book’s
CD-ROM.) The name “discone” is a contraction of the words
“disc” and “cone.” Although people often describe a discone
by its design-center frequency (for example, a “20 meter dis-
cone”), discones work very well over a wide frequency range,
as much as several octaves. Figure 10.51 shows a typical
discone, constructed of sheet metal for UHF use. On lower
frequencies, the sheet metal may be replaced with closely
spaced wires and/or aluminum tubing.

10.7.1 DISCONE BASICS

The dimensions of a discone are determined by the low-
est frequency of use. The antenna produces a vertically polar-
ized signal at a low-elevation angle and it presents a good Figure 10.51 — Diagram of VHF/UHF discone, using a sheet-
match for 50-W coax over its operating range. One advantage metal disc and cone. It is fed directly with 50-W coax line. The
of the discone is that its maximum current area is near the top dimensions L and D, together with the spacing S between
the disc and cone, determine the frequency characteristics
of the antenna, where it can radiate away from ground clutter, of the antenna. L = 246 / fMHz for the lowest frequency to be
reducing losses. The cone-like skirt of the discone radiates used. Diameter D should be from 0.67 to 0.70 of dimension L.
the signal — radiation from the disc on top is minimal. This The diameter at the bottom of the cone B is equal to L. The
is because the currents flowing in the skirt wires essentially space S between disc and cone can be 2 to 12 inches, with
the wider spacing appropriate for larger antennas.
all go in the same direction, while the currents in the disc
elements oppose each other and cancel out. The discone’s
omnidirectional characteristics make it ideal for roundtable
QSOs or for a net control station. type of “coaxial taper transformer.”
Electrical operation of this antenna is very stable, with no The July 1950 article was by Mack Seybold, W2RYI.
changes due to rain or accumulated ice. It is a self-contained He described an 11-MHz version he built on his garage roof.
antenna — unlike a traditional ground-mounted vertical ra- The mast actually fit through the roof to allow lowering the
diator, the discone does not rely on a ground-radial system antenna for service. Seybold stated that his 11-MHz discone
for efficient operation. However, just like any other vertical would load up on 2 meters but that performance was down
antenna, the quality of the ground in the Fresnel area will 10 dB compared to his 100-MHz Birdcage discone. He com-
affect the discone’s far-field pattern. mented that this was caused by the relatively large spacing
Both the disc and cone are inherently balanced for wind between the disc and cone. Actually, the performance degra-
loading, so torque caused by the wind is minimal. The entire dation he found was caused by the wave angle lifting upward
cone and metal mast or tower can be connected directly to at high frequencies. The cone wires were electrically long,
ground for lightning protection. causing them to act like long wire antennas.
Unlike a trap vertical or a triband beam, discone antennas
are not adjusted to resonate at a particular frequency in a ham 10.7.2 A-FRAME 20-10 METER DISCONE
band or a group of ham bands. Instead, a discone functions as W8NWF’s first discone was designed to cover 20
a sort of high-pass filter, efficiently radiating RF all the way through 10 meters without requiring an antenna tuner. The
from the low-frequency design cutoff to the high-frequency cone assembly uses 18-foot long wires, with a 60° included
limits imposed by the physical design. apex angle and a 12-foot diameter disc assembly. See Fig-
ure 10.52. The antenna was assembled on the ground, with
History of the Discone the feed coax and all guys attached. Then with the aid of
The July 1949 and July 1950 issues of CQ magazine some friends, it was pulled up into position.
both contained excellent articles on discones. The first ar- The author used a 40-foot tall wooden “A-frame” mast,
ticle, by Joseph M. Boyer, W6UYH, said that the discone was made of three 22-foot-long 2×4s. He primed the mast with
developed and used by the military during World War II. (See sealer and then gave it two coats of red barn paint to make
Bibliography.) The exact configuration of the top disc and it look nice and last a long time. The disc hub was a 12-inch
cone was the brainchild of Armig G. Kandonian. Boyer de- length of 3-inch schedule-40 PVC plumbing pipe. The PVC
scribed three VHF models, plus information on how to build is very tough, slightly ductile, and easy to drill and cut. PVC
them, radiation patterns, and most importantly, a detailed is well suited for RF power at the feed point of the antenna.
description of how they work. He referred to the discone as a Three 12-foot by 0.375-inch OD pieces of 6061 alumi­-

Multiband HF Antennas   10-33

 and have 16 radial spreaders, using telescop-
ing aluminum tubing tapering from 5⁄8 to 1⁄2 to
3
⁄8 inches OD. All spreaders were made from
0.058-inch wall thickness 6063-T832 alumi-
num tubing, available from Texas Towers and
other suppliers. A section of 10-inch PVC
plumbing pipe would be used as the hub for
construction of the disc assembly.
On the air tests proved to be very sat-
isfying. Loading up on 40 meters was easy
— the SWR was 1:1 across the entire band.
W8NWF can work all directions very well
and receives excellent signal reports from
DX stations. When he switches to his long
(333 foot) center-fed dipole for comparison,
he finds the dipole is much noisier and that
received signals are weaker. During the day-
time, nearby stations (less than about 300 to
500 miles) can be louder with the dipole, but
the discone can work them just fine also.
The author happily reports that this
antenna even works well on 75 meters. As
you might expect, it doesn’t present a 1:1
match. However, the SWR is between 3.5:1
and 5.5:1 across the band. W8NWF uses an
antenna tuner to operate the discone on 75. It
Figure 10.52 — Detailed drawing of the A-frame discone for 14 to 30 MHz. The seems to get out as well on 75 as it does on
disc assembly at the top of the A-frame is 12 feet in diameter. There are 45 40 meters.
cone wires, each 18 feet long, making a 60° included angle of the cone. The SWR on 30 meters is about 1.1:1.
On 20 meters the SWR runs from 1.05:1 at
14.0 MHz to 1.4:1 at 14.3 MHz. The SWR
num, with 0.058-inch wall thickness, were used for the on the 17, 15, 12 and 10-meter bands varies, going up to a
12-foot diameter top disc. These were cut in half to make high of 3.5:1 on 12 meters.
the center portions of the six telescoping spreaders. Four From modeling using NEC/Wires by K6STI, W8NWF
twelve foot by 0.250-inch OD (0.035-inch wall thick- verified that the low-angle performance for the bigger an-
ness) tubes were cut into 12 pieces, each 40 inches long. This tenna is worse than that for the smaller discone on the upper
gave extension tips for each end of the six spreaders. frequencies. See Figure 10.54 for an elevation-pattern com-
parison on 10 meters for both antennas, with average ground
10.7.3 40-10 METER DISCONE constants. The azimuth patterns are simply circles. Radiation
When an opportunity arose to buy a 64-foot self- patterns produced by antenna modeling programs are very
supporting TV tower, W8NWF jumped at the chance to helpful to determine what to expect from an antenna.
implement a full 7 to 30-MHz discone. His new tower had The smaller discone, which was built by the book, dis-
eight sections, each eight feet long. Counting the overlap plays good, low-angle lobes on 20 through 10 meters. The
between sections, the cone wires would come off the tower frequency range of 14 through 28 MHz is an octave’s worth
at about the 61.5-foot mark. See Figure 10.53. of coverage. It met his expectations in every way by cover-
W8NWF took some liberties with the design of this ing this frequency span with low SWR and a low angle of
larger discone compared to the first one, which he had done radiation.
strictly “by the book.” The first change was to make the cone The bigger discone, with a modified cone suitable for
wires 70 feet long, even though the formula said they should use on 75 meters, presents a little different story. The low-
be 38 feet long. Further, the cone wires would not be con- angle lobe on 40 meters works well, and 75-meter perfor-
nected together at the bottom. With the longer cone wires, mance also is good, although an antenna tuner is necessary
he felt that 75 and 80-meter operation might be a possibility. on this band. The 30-meter band has a good low-angle lobe
The second major change was to widen the apex angle but secondary high-angle lobes are starting to hurt perfor-
out from 60° to about 78°. Modeling said this should produce mance. Note that 30 meters is roughly three times the design
a flatter SWR over the frequency spectrum and would also frequency of the cone. On 20 and 17 meters there still are
give a better guy system for the tower. good low-angle lobes but more and more power is wasted in
The topside disc assembly would be 27 feet in diameter high-angle lobes.

10-34   Chapter 10
 Figure 10.53 — The large W8NWF discone, designed for operation from 7 to
14 MHz, but useable with a tuning network in the shack for 3.8 MHz.

Figure 10.54 — Computed patterns showing elevation Figure 10.55 — Computed elevation-response patterns for
response of small discone at 28.5 MHz compared to that of the larger W8NWF discone for 3.8, 7.2 and 21.2 MHz opera-
the larger discone at 28.5 MHz. The cone wires are clearly tion. Again, as in Figure 10.54, the pattern degrades at
too long for efficient operation on 10 meters, producing 21.2 MHz, although it is still reasonably efficient, if not
unwanted high-angle lobes that rob power from the desir- optimal.
able low-elevation angles.

The operation on 15, 12, and 10 meters continues to A discone antenna built according to formula will work
worsen for the larger discone. The message here is that al- predictably and without any adjustments. One can modify the
though a discone may have a decent SWR as high as 10 times antenna’s cone length and apex angle without fear of render-
the design frequency, its radiation pattern is not necessarily ing it useless. The broadband feature of the discone makes it
good for low-angle communications. See Figure 10.55 for a attractive to use on the HF bands. The low angle of radiation
comparison of elevation patterns for 3.8, 7.2 and 21.2 MHz makes DX a real possibility and the discone is also much less
on the larger discone. noisy on receive than a dipole.
Multiband HF Antennas   10-35
 10.8 Bibliography
Source material and more extended discussion of topics J. Hallas, “Getting On the Air: The Terminated Folded
covered in this chapter can be found in the references given Dipole,” QST, Sep 2010, pp 51-52.
below and in the textbooks listed at the end of Chapter 2, W. Hayward, “Designing Trap Antennas,” Technical
Antenna Fundamentals. Correspondence, QST, Aug 1976, p 38.
H. B. Barkley, The Open-Sleeve As A Broadband Antenna, D. Hollander, “A Big Signal from a Small Lot,” QST,
Technical Report No. 14, U.S. Naval Postgraduate Apr 1979, pp 32-34.
School, Monterey, CA, Jun 1955. R. H. Johns, “Dual-Frequency Antenna Traps,” QST,
W. M. Bell, “A Trap Collinear Antenna,” QST, Aug 1963, Nov 1983, pp 27-30.
pp 30-31. W. Jones, “Practical High Performance HF Log Periodic
J. S. Belrose, “The HF Discone Antenna,” QST, Jul 1975, Antennas,” QST, Sep 2002, pp 31-37.
pp 11-14, 56. A. G. Kandoian, “Three New Antenna Types and Their
J. Belrose and P. Bouliane, “The Off-Center-Fed Dipole Applications,” Proc IRE, Vol 34, Feb 1946,
Revisited: A Broadband, Multiband Antenna,” QST, pp 70W-75W.
Aug 1990, pp 28-34. R. W. P. King, Theory of Linear Antennas (Cambridge, MA:
J. Belrose, “Technical Correspondence: Terminated Folded Harvard Univ Press, 1956), pp 407-427.
Dipole,” QST, May 1994, pp 88-89. W. J. Lattin, “Multiband Antennas Using Decoupling
H. J. Berg, “Multiband Operation with Paralleled Dipoles,” Stubs,” QST, Dec 1960, pp 23-25.
QST, Jul 1956, pp 42-43. W. J. Lattin, “Antenna Traps of Spiral Delay Line,” QST,
E. L. Bock, J. A. Nelson and A. Dorne, “Sleeve Antennas,” Nov 1972, pp 13-15.
Very High Frequency Techniques, H. J. Reich, ed. (New M. A. Logan, “Coaxial-Cable Traps,” Technical
York: McGraw-Hill, 1947), Chap 5. Correspondence, QST, Aug 1985, p 43.
J. T. Bolljahn and J. V. N. Granger, “Omnidirectional VHF J. R. Mathison, “Inexpensive Traps for Wire Antennas,”
and UHF Antennas,” Antenna Engineering Handbook , QST, Feb 1977, p 18.
H. Jasik, ed. (New York: McGraw-Hill, 1961) pp 27-32 L. McCoy, “An Easy-to-Make Coax-Fed Multiband Trap
through 27-34. Dipole,” QST, Dec 1964, pp 28-30.
J. M. Boyer, “Discone — 40 to 500 Mc Skywire,” CQ, M. Mims, “The All-Around 14-mc. Signal Squirter,” QST,
July 1949, p 11. Dec 1935, pp 12-17.
G. A. Breed, “Multi-Frequency Antenna Technique Uses G. E. O’Neil, “Trapping the Mysteries of Trapped
Closely-Coupled Resonators,” RF Design, November Antennas,” Ham Radio, Oct 1981, pp 10-16.
1994. US Patent 5,489,914, “Method of Constructing W. I. Orr, editor, “The Low-Frequency Discone,” Radio
Multiple-Frequency Dipole or Monopole Antenna Handbook, 14th Edition, (Editors and Engineers, 1956),
Elements Using Closely-Coupled Resonators,” Gary A. p 369.
Breed, Feb 6, 1996. W. I. Orr, “Radio FUNdamentals,” The Open-Sleeve Dipole,
G. H. Brown, “The Phase and Magnitude of Earth Currents CQ, Feb 1995, pp 94-96.
Near Radio Transmitting Antennas,” Proc. IRE, Vol 23, E. W. Pappenfus, “The Conical Monopole Antenna,” QST,
No. 2, Feb 1935, pp 168-182. Nov 1966, pp 21-24.
G. H. Brown, R. F. Lewis and J. Epstein, “Ground Systems P. D. Rhodes, “The Log-Periodic Dipole Array,” QST, Nov
as a Factor in Antenna Efficiency,” Proc. IRE, Vol 25, 1973, pp 16-22.
No. 6, Jun 1937, pp 753-787. P. D. Rhodes, “The Log-Periodic V Array,” QST, Oct 1979,
C. L. Buchanan, “The Multimatch Antenna System,” QST, pp 40-43.
Mar 1955. pp 22-23, 155. P. D. Rhodes, “The K4EWG Log Periodic Array,” The ARRL
R. A. Cox, “The Open-Sleeve Antenna,” CQ, Aug 1983, Antenna Compendium, Vol 3, pp 118-123
pp 13-19. L. Richard, “Parallel Dipoles of 300-Ohm Ribbon,” QST,
G. Countryman, “An Experimental All-Band Nondirectional Mar 1957, p 14.
Transmitting Antenna,” QST, Jun 1949, pp 54-55. P. Salas, “160 and 80 Meter Matching Network for Your
D. DeMaw, “Lightweight Trap Antennas — Some 43 Foot Vertical — Part 1 and Part 2,” QST, Dec 2009,
Thoughts,” QST, Jun 1983, pp 15-18. p 30-32, and Jan 2010, pp 34-35.
W. C. Gann, “A Center-Fed ‘Zepp’ for 80 and 40,” QST, W. Sandford, Jr., “A Modest 45-Foot DX Vertical for 160,
May 1966, pp 15-17. 80, 40, and 30 Meters,” QST, Sep 1981, pp 27-31. Also
D. Geiser, “An Inexpensive Multiband VHF Antenna,” QST, see Feedback, Nov 1981, p 50.
Dec 1978, pp 28-29. R. R. Schellenbach, “Try the ‘TJ’,” QST, Jun 1982,
A. Greenberg, “Simple Trap Construction for the Multiband pp 18-19.
Antenna,” QST, Oct 1956, pp 18-19, 120. R. R. Schellenbach, “The JF Array,” QST, Nov 1982,
G. L. Hall, “Trap Antennas,” Technical Correspondence, pp 26-27. Also see Technical Correspondence, QST,
QST, Nov 1981, pp 49-50. Apr 1983, p 39.

10-36   Chapter 10
R. Severns, “A Wideband 80 meter Dipole,” QST, Jul 1995, S. Stearns, “All About the Discone Antenna: Antenna of
pp 27-29. Mysterious Origin and Superb Broadband Performance,”
T. H. Schiller, Force 12, US Patent 5,995,061, “No loss, QEX, Jan/Feb 2007, pp 37-44.
multi-band, adaptable antenna,” Nov 30, 1999. J. J. Uhl, “Construct a Wire Log-Periodic Dipole Array for
H. Scholle and R. Steins, “Eine Doppel-Windom Antenna 80 or 40 Meters,” QST, Aug 1986, pp 21-24.
fur Acht Bander,” cq-DL, Sep 1983, p 427. (In English: L. Varney, “The G5RV Multiband Antenna . . . Up-to-Date,”
QST, Aug 1990, pp 33-34.) The ARRL Antenna Compendium, Vol 1 (Newington:
M. Seybold, “The Low-Frequency Discone,” CQ, July 1950, ARRL, 1985), p 86.
p 13. R. Wilson, “The Offset Multiband Trapless Antenna
D. P. Shafer, “Four-Band Dipole with Traps,” QST, Oct (OMTA),” QST, Oct 1995, pp 30-32. Also see Feedback,
1958, pp 38-40. Dec 1995, p 79.
R. C. Sommer, “Optimizing Coaxial-Cable Traps,” QST, L. G. Windom, “Notes on Ethereal Adornments,” QST,
Dec 1984, pp 37-42. Sep 1929, pp 19-22, 84.

Multiband HF Antennas   10-37

Antenna Fundamentals   1-1
Antenna Fundamentals   1-1