An Improved Double

Extended Zepp
By Rudy Severns, N6LF
32857 Fox Lane
PO Box 589
Cottage Grove, OR 97424

This article was orinally published in the ARRL Antenna Compendium Vol. 4 1995

T
he sunspot cycle is headed down and
the low bands are coming to life for
DXing. The next several years will N6LF revisits the classic double
be the time to make those low-band contacts
for DXCC. As always, the key to low-band extended Zepp to improve pattern
and SWR bandwidth. He also offers
operation is a good antenna. Unfortunately
for 7 MHz and down, good antennas don’t
come in a box ready to assemble. Every lo-
cation will have a unique set of limitations
and opportunities.
some sage advice about computer
One very important difference between this
sunspot minima and all past ones is the avail-
modeling.
ability of inexpensive, easy-to-use and pow-
erful antenna modeling software.1,2 This soft-
ware allows you to design and optimize an
antenna that exactly fits your situation and
pocketbook. While cut-and-try experimenta-
tion is a very slow way to optimize antennas, amount of gain over a dipole of approximately
modeling is so quick that a wide range of so- 3 dB. The radiation pattern for a DEZepp de-
lutions can be investigated easily. The real signed for 7.15 MHz and suspended 80 feet
problem with modeling is generating the will above ground is shown in Fig 1, along with
power to stop fooling with the variations and the pattern for a λ/2 dipole at the same height
go out and build something! for comparison. The elevation angle is 26°, the
The following article uses 40 and 80/75- peak of the main lobe. The current distribu-
meter double extended Zepps (DEZepp) as tion along the antenna is shown in Fig 2.
examples of what you can accomplish. By The DEZepp does indeed provide gain over
adding two small capacitors, made from the dipole, but only over the relatively small
short lengths of RG-8, in just the right place, angle of approximately 40°. The beamwidth
the pattern can be improved and the driving- between 3 dB points is 35°. Unless the antenna
point impedance changed from reactive and is pointed directly toward the receiving sta-
narrowband to resistive and wideband. This tion, the gain is not usable due to the narrow
allows the antenna to be used without a tuner beam width. In addition to the narrow main
and with an SWR < 1.5:1 over the entire 40- lobe, there are significant sidelobes. These are
meter band, or with SWR < 2:1 over the not big enough to be helpful in those direc-
tions, but they will also certainly pick up noise Fig 1—Azimuth pattern of classic
entire 75/80-meter band. double extended Zepp (solid line) at
and interference. The impedance of the an- 7.15 MHz, compared with standard
A Look at the Classical DEZepp tenna is very reactive, and even when matched dipole (dashed line), both 80 feet high
The classical DEZepp is simply a piece of at midband does not allow the entire band to over average ground. Patterns are
wire 1.25 λ long, fed at the center, usually with be covered without retuning. shown at 26° elevation, where the gain
is maximum. The wire runs along the
open-wire transmission line and a tuner at the For this reason, the DEZepp has tradition- 270° to 90° axis on the graph. Note
transmitter. The DEZepp displays a useful ally been used with an antenna tuner. This is significant sidelobes for DEZepp.
 and therefore less loss. With either choice it is
desirable to use as few components as possible.
As an initial trial I decided to use only two
capacitors, one on each side of the antenna. I
varied the value and position of the capacitors
to see what would happen. It quickly became
clear that I could tune out the reactance at the
feedpoint by adjusting the capacitor value,
making the antenna look like a resistor over
the entire band. The value of the feed-point
resistance could be varied from less than
150 Ω to over 1500 Ω by changing the loca-
tion of the capacitors and adjusting their val-
Fig 2—Schematic for classic DEZepp, Fig 3—Schematic for modified N6LF ues to resonate the antenna. The AO 6
showing current distribution along DEZepp, with new current distribution. (Antenna Optimizer) software1 has the nice
antenna. The “bulging out” of the Overall length is 170 feet, with 9.1 pF feature that it will automatically adjust a vari-
current in the opposite direction near capacitors placed 25 feet each side of
the center of the antenna is responsible center. Now current distribution doesn’t able to tune out reactance. Simultaneously, the
for the sidelobes seen in Fig 1. create sidelobes. pattern was also changing in useful ways.
A number of interesting combinations were
created. The one I elected to use is shown in
Fig 3. The antenna is 170 feet in length. That is
a couple of feet shorter than the classic DEZepp,
but that also just happens to be all the distance
I had between my supporting trees! Two 9.1 pF
capacitors are located 25 feet out each side of
the center. The antenna is fed with 450-Ω trans-
mission line and a 9:1 three-core Guanella
balun3 used at the transmitter to convert to 50 Ω.
The transmission line can be any convenient
length and it operates with a very low SWR.
That’s all there is to it. The radiation pat-
tern, overlaid with that for a standard DEZepp
for comparison, is shown in Fig 4. A compari-
son to a standard dipole is shown in Fig 5. The
sidelobes are now reduced to below 20 dB. The
main lobe is now 43° wide at the 3-dB points,
Fig 4—Azimuth pattern for N6LF DEZepp Fig 5—Azimuth pattern for N6LF as opposed to 35° for the original DEZepp. The
(solid line), compared to classic DEZepp DEZepp (solid line), compared to dipole
(dashed line). The main lobe for the (dashed line) at the same height. antenna has gain over a dipole for > 50° now.
modified antenna is slightly broader than The gain of the main lobe has dropped only
that of the classic model, and the 0.2 dB below the original DEZepp.
sidelobes are suppressed better. The reason for the pattern change can be
seen in Fig 3, showing the modified current
distribution. The main current maxima are
not a terrible hardship but it would be nice still pretty much in the same place, but the
if the tuner could be eliminated, at least on current in the center of the antenna now
one band, and a low SWR presented to the flows in the opposite direction. The result-
transmitter over a whole band. ing pattern is much cleaner.
The gain displayed by the DEZepp is due to
the separation between the two current Experimental Results
maxima. The small inverted current in the I managed to pry myself away from the com-
center section subtracts a little from the main puter and actually build the antenna. It was
lobe and contributes to the sidelobes. The made from #14 wire and the capacitors were
DEZepp is essentially two end-fed collinear made from 3.5-inch sections of RG-213, shown
dipoles. The transmission line and the center in Fig 6A. Note that great care should be taken
portion of the antenna are the feed system. to seal out moisture in these capacitors. The
It would be very beneficial to suppress the voltage across the capacitor for 1.5 kW will be
sidelobes and put that energy into a broader about 2000 V so any corona will quickly de-
main lobe, retaining most of the gain if possible. stroy the capacitor. One of the nice features of
modeling software is that it gives the current
A Modified DEZepp amplitude along the antenna, making it easy to
The key to modifying the radiation pattern determine the stresses on any series reactances.
Fig 6—Construction details for series is to modify the current distribution. One of the I used silicon sealant and then covered both
capacitor made from RG-213 coaxial cable.
At A, the method used by N6LF is simplest ways to do this is to insert a ends with coax seal, finally wrapping it with
illustrated. At B, a suggested method to seal reactance(s) in series with the wire. This could plastic tape. The solder balls indicated on the
capacitor better against weather is shown, either be an inductor(s) or a capacitor(s). In drawing are to prevent wicking of moisture
using a section of PVC pipe with end caps. general, a series capacitor will have a higher Q through the braid and the stranded center con-
ductor. This is a small but important point if meters. For the bands above 40 meters, how- very helpful with comments on the design, and
long service out in the weather is expected. An ever, the reactance of the capacitors drops rap- of course he created the new NEC Wires pro-
even better way to protect the capacitor would idly and the behavior is very much the same as gram. An antenna is only as good as the mate-
be to enclose it in a short piece of PVC pipe for the normal DEZepp. The price paid for im- rials that go into it. I used the beautiful “silky”
with end caps, as shown in Fig 6B. proving operation on 40 meters is the loss of 75/ stranded copper-clad steel wire supplied by
Note that all RG-8 type cables do not have 80 meters. Similarly, in the 75/80-meter ver- Press Jones, N8UG, and also some of his 450-
exactly the same capacitance per foot and there sion, performance on 160 meters is sacrificed. Ω transmission line. It was wonderful to have
will also be some end effect adding to the ca- a wire antenna that didn’t try to envelop you
pacitance. I trimmed the capacitor with a capaci- Some Final Thoughts (and every nearby bush) like a boa constrictor
tance meter. It isn’t necessary to be too exact— The antenna shown here represents a very due to a built-in set in the wire.
I checked the effect of varying the capacitance simple modification of an old idea to suit a par-
±10% and the antenna still works fine. ticular situation. There are any number of varia- Notes
1
The results proved to be close to those pre- tions that could have produced similar results. AO 6.1 and NEC Wires, by Brian Beezley,
dicted by the computer model. Fig 7 shows the Two important lessons were learned during this K6STI, 3532 Linda Vista Dr, San Marcos,
measured value for SWR across the band. These effort. First, the modeling software is pretty CA 92069.
2ELNEC 3.0 , by Roy Lewallen, W7EL, PO
measurements were made with a Bird direc- accurate, particularly now that NEC-based soft- Box 6658, Beaverton, OR 97007.
tional wattmeter. The worst SWR is 1.35:1 at ware is available. The results obtained were very 3Jerry Sevick, W2FMI, Transmission Line
the low end of the band! With a little adjustment close to that predicted—and this is not the first Transformers , Second Edition (ARRL,
of the antenna length this could have been low- time I have seen this. Second, the modeling pro- Newington, CT, 1990), p 9-28.
ered a bit more, but I figured why bother? cess is a great teacher. It helps you to learn how
My antenna was oriented to work into Eu- antennas really work and cuts through many
rope. Prior to putting up this antenna I had misconceptions. By viewing the current distri-
been using a dipole. I could hear a few butions, the associated radiation patterns and
Europeans but was unable to work them. driving-point impedances, it becomes much
Three dB may not seem like much gain but easier to understand which way to modify a
after putting up this antenna I immediately design to achieve a desired result. Being able to
heard many more signals and have been regu- get results quickly is very helpful also.
larly working into Europe with 56/57 reports. There is a whole new world of low-band
Dick Ives, W7ISV, was sufficiently im- antennas out there waiting to be created!
pressed by the success of the 40-meter version
of this antenna to ask me to design a 75-meter Acknowledgments
version for him. In his location one end of the I want to express my appreciation to Dick
antenna could only be 60 feet high (< 0.25 λ), Ives for braving the wet and cold to verify the
and I was concerned about the accuracy of the 75/80-meter version of this antenna in time for
modeling program, because MININEC-based me to add it to this article. Brian Beezley was Fig 7—Measured SWR curve across
programs are known to be inaccurate for gain 40-meter band for N6LF DEZepp.
and feed-point impedance at low heights.
Fortunately, Brian Beezley, K6STI, has a
NEC-based program called NEC Wires.1 This
does model ground accurately and is just the
ticket for low antennas. Using this program I
designed a new antenna for W7ISV.
Despite the temperatures in mid-December,
Dick erected the antenna as shown in Fig 8.
The series capacitors are 17 pF, and since he
isn’t interested in CW, Dick adjusted the
length for the lowest SWR at the high end of
the band. The antenna could have been tuned
somewhat lower in frequency and would then
provide an SWR < 2:1 over the entire band, as
indicated by the dashed line in Fig 8.
This antenna provides wide bandwidth
and moderate gain over the entire 75/80-
meter band. Not many antennas will give
you that with a simple wire structure.
Multiband Operation
When operated with an antenna tuner, one of
the advantages of the classical DEZepp is that it
is a multiband antenna. Typically a 40-meter
DEZepp behaves like a dipole on 75/80 meters
and like a long wire on the higher frequency
bands. Adding the two series capacitors Fig 8—75/80-meter N6LF DEZepp, designed using NEC Wires . At A, a schematic is
decouples the ends of the wires on 75/80 meters shown for antenna. At B, SWR curve is shown across 75/80-meter band. Solid line
shows measured curve for W7ISV antenna, which was pruned to place SWR
and a rather poor antenna results. It behaves minimum higher in the band. The dashed curve shows the computed response
more like a 30-meter dipole being used on 75/80 when SWR minimum is set to 3.8 MHz.