To read a sample of this book 974kb PDF click here

To read a sample of this book 974kb PDF click here
20
12
Successful Wire
Antennas
©
R
SG
B
Edited by
Ian Poole, G3YWX
and
Steve Telenius-Lowe, 9M6DXX, KH0UN
Radio Society of Great Britain
1
20
12
Published by the Radio Society of Great Britain, 3 Abbey Court, Fraser Road,
Priory Business Park, Bedford MK44 3WH. Tel: 01234 832700. Web: www.rsgb.org
Published 2012.
© Radio Society of Great Britain, 2012. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form
or by any means, electronic, mechanical, photocopying, recording or otherwise,
without the prior written permission of the Radio Society of Great Britain.
B
ISBN: 9781 9050 8677 1
SG
Cover design: Kim Meyern
Design and layout: Steve Telenius-Lowe, 9M6DXX, KH0UN
Production: Mark Allgar, M1MPA
R
Printed in Great Britain by Latimer Trend of Plymouth
©
Publisher’s Note:
The opinions expressed in this book are those of the author(s) and are not necessarily those of the Radio Society of Great Britain. Whilst the information presented is believed to be correct, the publishers and their agents cannot accept
responsibility for consequences arising from any inaccuracies or omissions.
2
20
12
Contents
Foreword, by John D Heys, G3BDQ .................................. 4
Preface ................................................................................ 5
Editor’s note ........................................................................ 6
Antenna Basics ................................................................... 7
2
Feeders ............................................................................. 27
3
Dipoles .............................................................................. 42
4
Doublets ............................................................................ 86
SG
B
1
Verticals .......................................................................... 115
6
Loops ............................................................................... 164
7
End-fed wires .................................................................. 181
8
Impedance Matching and Baluns ................................. 200
R
5
©
9
Antenna Masts and Rigging ........................................... 227
Appendix ......................................................................... 234
Index ............................................................................... 237
3
SUCCESSFUL WIRE ANTENNAS
20
12
3
Dipoles
T
B
HE WIRE DIPOLE is a very effective antenna that can be constructed and
installed very easily and for only a small cost. The half-wave version of the
dipole has become the standard against which other radiating systems are
judged and it remains as perhaps the most effective, yet simple, single-band antenna, and one which can virtually be guaranteed to perform well even when used in
far-from-ideal situations.
As the name suggests, it contains two legs or ‘poles’. The most common form is
the half-wave dipole, which (not surprisingly) is an electrical half-wavelength long.
The basic format for a half-wave dipole along with the voltage and current waveforms
can be seen in Fig 3.1. The voltage rises to a maximum at either end and falls to a
minimum at the centre, whereas the current is at its minimum at the end and its
maximum in the centre. Its feedpoint in the centre forms a low impedance point
suitable for many sorts of feeder.
A dipole does not have to be a half-wavelength long. A three half-wavelength
version can be seen in Fig 3.2. Again the points of voltage maximum are at either end
and at a minimum in the centre. Likewise the current is at its minimum at either end
and maximum in the centre.
SG
DIPOLE LENGTHS
©
R
A resonant half-wavelength of wire will be somewhat shorter than its name implies.
RF energy in free space (electromagnetic radiation) can travel at the speed of light,
but when moving along a conductor it travels more slowly. At HF (between 3 and
30MHz) wires exhibit skin effect, i.e. most of the RF energy flows along the outer
surface of the conductor. A practical half-wave antenna made from wire needs end
supports; each end usually being terminated with an insulator. The capacitance
between the ends of dipole and its supports, even when the supporting material is
non-metallic, gives rise to end effect. This effect additionally loads the wire capacitively and contributes towards its shortening from the theoretical half-wavelength.
(Left) Fig 3.1: A basic half-wavelength dipole
antenna with the voltage and current waveforms.
(Above) Fig 3.2: A three half-wavelength dipole.
42
CHAPTER 3 DIPOLES
Frequency
1850
1950
3550
3750
7050
10100
14100
14250
18100
21100
21300
24940
28100
28500
29000
29500
Without insulators
(feet)
(metres)
252' 11"
240' 0"
131' 10"
124' 9"
66' 4"
46' 4"
33' 2"
32' 10"
25' 10"
22' 2"
21' 11"
18' 9"
16' 8"
16' 5"
16' 1"
15' 10"
258' 5"
245' 1"
134' 8"
127' 5"
67' 10"
47' 4"
33' 11"
33' 6"
26' 5"
22' 8"
22' 5"
19' 2"
17' 0"
16' 9"
16' 6"
16' 2"
77.29
73.33
40.28
38.13
20.28
14.15
10.14
10.03
7.90
6.77
6.71
5.73
5.08
5.01
4.93
4.84
Table 3.1: Lengths of half-wave dipoles.
78.75
74.71
41.04
38.85
20.66
14.42
10.33
10.22
8.04
6.90
6.84
5.84
5.18
5.11
5.02
4.93
20
12
(kHz)
Length
With insulators
(feet)
(metres)
The theoretical half-wavelength may be calculated from the expression:
B
Theoretical half wavelength (metres) = 150 / f (MHz)
or
Theoretical half-wavelength (feet) = 492 / f (MHz)
SG
To take account of the end effect and the use of insulators, the length may be
calculated by using either:
Antenna length (metres) = 143 / f (MHz)
or
Antenna length (feet) = 468 / f (MHz).
©
R
When using nylon rope it has been suggested that no insulators are required. In
his book HF Antennas for All Locations (published by the RSGB), Les Moxon, G6XN,
suggests that when no insulators are used a half-wavelength can be found by using
either 478 / f (MHz) feet or 145.7 / f (MHz) metres.
A further factor which influences antenna resonant length is the diameter of the
wire used for that antenna. The formulas above are for typical wire dimensions.
Typical antenna lengths for the amateur bands from 160 - 10m, both when using
insulators or nylon rope, are shown in Table 3.1 above.
DIPOLE IMPEDANCES
A half-wave transmitting antenna, when energised and resonant, will have high RF
voltages at its ends with theoretically zero RF currents there. This means that the
ends of a half-wave dipole in free space will have an infinitely high impedance, but in
practice in the real world there will always be some leakage from its ends and into
the supporting insulators. This means that in reality the impedance at the dipole
ends is close to 100,000Ω, a value which depends upon the wire or element thickness. At a distance of approximately one-sixteenth wavelength from either end it is
1000Ω, and at the dipole centre, where the current is greatest and the RF voltage is
low, the impedance is also low.
43
SUCCESSFUL WIRE ANTENNAS
ANTENNA Q
20
12
If it were made from an infinitely thin conductor wire, our theoretical dipole in free
space would have an impedance of about 73Ω at its centre. Such an antenna is
impossible in the material world, and a practical half-wave dipole made from wire
will have an impedance at its centre at resonance close to 65Ω. Antennas fabricated
from tubing have lower values at their centres, of between 55 and 60Ω. These impedance values also depend upon the height of the antenna above ground, as will
be shown later.
The very high values of self-impedance at the ends of a half-wave wire makes
end-feeding difficult, and this is why breaking the wire at its centre and connecting
the inner ends so formed to a low-impedance feedline makes a convenient and
efficient coupling and match. Suitable feeder is available in the form of twin-lead or
coaxial cable, which both have design impedances lying between 50 and 75Ω. These
present a good match to dipole centres.
At exact resonance the impedance at the centre of a half-wave dipole is like a
pure resistance. At any other frequencies the same dipole will have either inductive
or capacitive reactance at its feedpoint. If the dipole is too short to be resonant the
reactance is capacitive and when it is too long the reactance becomes inductive. In
either case there will be problems in matching the 50 or 70Ω feeder to the dipole and
if the reactances are great, there will be a high SWR on the feeder and considerable
power loss.
©
R
SG
B
A half-wave antenna is something like a conventional tuned circuit where the Q, or
‘Quality factor’, is largely determined by the resistance of the coil. Losses in the
capacitor used in the circuit are generally small and are not so significant in the
determination of Q. A high-Q tuned circuit exhibits very sharp tuning (selectivity) and
this is also the case when an antenna has a high Q.
Using thin wires lowers the bandwidth of a half-wave antenna, but not dramatically.
However, short wires that are brought into resonance will exhibit high Q. The shorter the
wire in terms of wavelength, the higher
the Q. Small changes in the transmitting frequency away from the antenna
resonances will give rise to a rapid rise
in the reactance at the feedpoint.
Thicker wire will lower the Q, reduce
resistive loss and make the half-wave
dipole less frequency conscious. It is
therefore best to ensure that such an
antenna is made from the thickest possible wire consistent with such factors
as the pull on the antenna supports,
windage and sag.
Fig 3.3: Radiation resistance of a half-wave dipole as a
function of height above the ground (reprinted with permission of the American Radio Relay League).
44
DIPOLE HEIGHT
The height of a horizontal dipole above
the ground as a ratio of its design frequency is important (see the standard curves of feed impedance against
height in Fig 3.3). When below about
half a wavelength high the radiation
resistance at the feedpoint will be reduced, and down at a height of just
one-tenth of a wavelength it will only
be 25Ω. This means that a dipole fed
CHAPTER 3 DIPOLES
THE SLOPING DIPOLE
©
R
SG
B
20
12
Horizontal half-wave dipoles require two end supports and it is not always possible to provide
these in some awkward locations. In such situations a single support, preferably a non-metallic
mast or a high point on a building, will suffice,
and then the antenna can be arranged to slope
down towards the ground at an angle lying somewhere between 30° and 60° (Fig 3.14). The sloping half-wave dipole should have its lower end at
least one-sixth of a wavelength above ground,
and its feeder should ideally come away from
the radiator at 90° for at least a quarter of a wavelength. If coaxial feeder is used the braid should
connect to the lower half of the antenna.
The performance of a sloping dipole is quite
different from one of the horizontal variety and it
can be good for long distance work. The radiation from a sloping dipole shows slant polarisation with both vertical and horizontal components
according to the amount of slope. Its lower angle
of radiation to the horizon can result in a little low
angle gain over a horizontal dipole. This kind of
gain is difficult to realise on the low bands in
other ways, where for most amateurs multi-element Yagi beams are out of the question.
There is some high-angle radiation from the
Fig 3.14: A half-wave sloping dipole which can be
sides of the sloping dipole but very little radiation
put up in a small space and which will be useful
from its high end. An actual plan of the horizontal
for long-distance working. Most of its low-angle
radiation pattern resembles a heart with a null
radiation is towards the low end of the antenna
between its two upper lobes. This null correbut there is also considerable radiation at high
sponds with the high end of the sloping dipole. A
angles in other directions.
disadvantage is of course that long-distance
working will only be possible towards one direction, but this may be overcome by having a group of three or four ‘slopers’ suspended
from a common central support, each with its individual feedline which may be
selectively switched to the transceiver. (There are designs which involve the unused
dipoles in such arrangements as reflectors to improve forward gain and front-toback ratios, but their correct adjustment can be complicated.)
Slopers are ideal in many applications where a single support is available. Many
people who have beams and towers, mount a sloper on the tower for one of the lower
frequency bands, ensuring that the direction of maximum radiation is arranged towards the areas of the globe they want to contact, sometimes having two or more
around the tower.
THE VERTICAL DIPOLE
A vertical half-wave dipole will radiate vertically-polarised signals all round, and much
of the radiation will be at the low angles favourable for DX working. If a feed impedance of around 70Ω is required, the centre of this antenna must be around 0.45λ
above the ground and so it is usually more convenient to arrange for a vertical quarter-wave antenna to be used, which can then have its feedpoint at or near ground
level. A vertical dipole cannot be hung down from a metal mast or tower, and it should
have its feeder come away from the radiator wire at right angles if the radiation
pattern is to be preserved, which may present some problems. As a result, vertical
55
SUCCESSFUL WIRE ANTENNAS
half-waves are not often used by amateurs, although they can be practical on the
higher-frequency HF bands. A practical design for vertical dipoles is given in the
chapter on vertical antennas.
A 3λ
λ /2 DIPOLE: THE 40M DIPOLE ON 15M
©
R
SG
B
20
12
The dipole, when fed with coaxial cable, is basically a single band antenna. While
this is true, there are a few ways that dipoles can be made to work on more than one
band. One method is to parallel two or more dipoles for different bands together;
another is to use traps. We will look at both of these methods later in this chapter.
The simplest way, though, is to take advantage of the fact that the low impedance at
the centre feedpoint of a dipole occurs not only when it is one half-wave long, but also
when it is three half-waves long (and in fact all odd numbers of half-waves). We can
take advantage of this where amateur bands have this same harmonic relationship,
i.e. where one band has three times the frequency (or five or seven times, etc), and
therefore one third (or one fifth or one seventh) of the wavelength of another band.
On HF, the best example of this is the relationship between 40m and 15m: the
third harmonic of 7MHz is 21MHz. What this means is that a 40m dipole should also
work on 15m. Unfortunately, life isn’t quite that simple. As we have already seen, the
end effect means that a half-wave dipole is physically about 5% shorter than its
theoretical (electrical) half-wave length. However, when the same antenna is operating on its third harmonic, it becomes fifteen per cent shorter than the electrical length
of a three half-waves antenna. What this means in practice is that it is actually resonant quite a bit higher in frequency than you might expect.
The way around this problem is to resonate the 40m dipole at the very bottom of
the band, 7000kHz, or even make it somewhat longer still, so that the minimum SWR
point is actually below the bottom of the band on, say, 6980kHz. On 21MHz you will
find the minimum SWR point is nevertheless at the top of the band, around 21400
or 21450kHz.
Furthermore, QST Technical Editor Joel Hallas, W1ZR, points out (in the July
2009 QST) that the resonant impedance of a 3λ/2 dipole is above 100Ω, so it’s not as
good a match to 50Ω coax as is the λ/2 case.
The good news is that an
external ATU or the internal
automatic ATU in your rig
should be able to reduce the
Fig 3.15:Comparison between the 40m (black)
and 15m (white) EZNEC SWR plots of a 66ft
high, 67.2ft long, 40m dipole (diagrams reprinted with permission of the American Radio
Relay League).
56
Fig 3.16: Comparison between the 40m (black)
and 15m (grey) azimuth patterns of a 40m λ/2
dipole.
CHAPTER 3 DIPOLES
20
12
SWR to close to 1:1 at your operating frequency of choice in both the 40m and 15m
bands.
Fig 3.15 shows a comparison between the 40m and 15m EZNEC SWR plots of
a 66ft high, 67.2ft long, 40m dipole made of 14 gauge bare wire. In Fig 3.16 note that
the azimuth pattern of a 3λ/2 dipole is not the same as the usual λ/2 case. While
different, the pattern can be useful and provides a bit of additional gain in its prime
directions.
On HF there are a few other combinations of bands that have an odd harmonic
relationship, for example an 80m half-wave dipole cut for the lower-frequency end of
the band is five half-waves long on the 17m band, and seven half-waves long on the
12m band. (However, if the 80m dipole is cut for the SSB DX end of the band, around
3800kHz, its five and seven half-wave resonances will be well above the top end of
the 18MHz and 24MHz bands respectively, and the SWR is likely to be very high on
both bands.)
MOUNTING A WIRE DIPOLE ABOVE A ROTATOR
©
R
SG
B
Most amateurs with masts and towers suspend their HF wire dipole or inverted-V
dipole from a point below the rotator, leaving the beam free to rotate above. However,
mounting a dipole on a mast extension above the beam is a much better option,
provided the inverted-V angle can be made shallow enough to clear the beam as it
rotates underneath. Another strong reason for mounting low-band dipoles above the
beam is the extra height above ground, which makes them more effective – undoubtedly for DX, and often for more local QSOs as well. The problem with this ‘over the
top’ approach is that the centre of the dipole must be able to pivot on the top of the
mast, so that the mast and beams can rotate beneath it.
One successful solution to this problem was designed by Jan Fisher, G0IVZ,
and taken up by Ian White, GM3SEK, in his ‘In Practice’ column in the April 2009
RadCom. The rotating mast for the HF beam is made from scaffold tubing, extended
by a 1.5in fibreglass pole (an aluminium pole of that diameter probably couldn’t
handle the bending forces). The top of the extension pole is filled by a close-fitting
80m dipole over the top of a small HF
beam.
Close-up of the rotating dipole centre
with balun box to the rear.
57
SUCCESSFUL WIRE ANTENNAS
20
12
hardwood plug, drilled 8mm through the centre and secured with epoxy.
G0IVZ’s idea was to use a ready-made centre insulator that was originally designed for mounting a tubular dipole on the boom of a Yagi. The plastic moulding is
strong enough to support a much longer wire dipole, simply tied on through the fixing
holes as shown in the photograph. G0IVZ used the built-in terminal box to connect
the dipole to the coax feedline, while GM3SEK, who has adopted a similar set-up,
connected the two wire ends to the terminals on the balun box.
The main mounting hole of the centre insulator is drilled 8mm for an M8x100mm
stainless steel screw which is the pivot pin for the whole assembly. In GM3SEK’s
version, the balun bracket is fixed to the bottom of the insulator with a nut, so those
two parts rotate together. A large washer is added to spread the down-thrust, and the
free end of the screw simply drops into the hole in the wooden plug. A later addition
was a piece of white PVC waste pipe, taped to the bracket to prevent the bottom edge
scratching the fibreglass. Below this fitting there has to be a rotation loop in the coax
and of course there’s the usual loop around the rotator itself.
THE FOLDED DIPOLE
R
SG
B
Another form of dipole is the folded dipole, shown in Fig 3.17. It is often used as a part
of more complex antennas such as Yagis, but it can also be a useful antenna on its
own. It has the advantages that it has a higher impedance and a wider bandwidth
than an ordinary dipole.
The 300Ω feed impedance is an important feature of a folded dipole. The power
supplied to a folded dipole is evenly shared between the two conductors which make
up the antenna, so therefore the RF current, I, in each conductor is reduced to I/2.
This is a half of the current value (assuming that the same power is applied) at the
centre of the common half-wave dipole, so the impedance is raised. By halving the
current at the feedpoint yet still maintaining the same power level, the impedance at
that point will be four times greater. This means that a two-conductor folded dipole
will have a feed impedance of 280Ω, which is close to the impedance of 300Ω twin
feeder. It can therefore be satisfactorily matched and fed with this feeder, and have a
low SWR along the feedline.
If a third conductor is added to the folded dipole (Fig 3.18), the antenna current
will be evenly split three ways and the impedance at the feedpoint will be nine times
greater than the nominal 70Ω impedance of a simple dipole. Such a three-wire
dipole with its feed impedance
of 630Ω will make a good
match to a 600Ω feeder. This
feeder may be made from
18SWG wires which are
spaced at 75mm (3in).
©
λ/2
Fig 3.17: The basic two-wire folded dipole.
58
Fig 3.18: A three-wire folded dipole. If each wire
is of equal diameter the total current will be
shared equally between the three wires and the
impedance at the feedpoint will be nine times that
of a conventional half-wave dipole (9 x 70Ω =
630Ω), a close match to a 600Ω feedline.
SUCCESSFUL WIRE ANTENNAS
O
20
12
4
Doublets
B
NE CLASS OF antenna that is not as widely used as it might be is that of
tuned feeder antennas. Using an open-wire tuned feedline as part of the
overall antenna system enables multi-band operation to be achieved, although such an antenna - often called a doublet - does require the use of an ATU to
ensure that there is a good match to the transceiver.
The key to tuned feedline antennas is naturally the feeder. As discussed in
Chapter 2, these open-wire feedlines have a characteristic impedance which relates
to the diameter of the wire used and the spacing between the feed wires. This
impedance is important in many applications, but note that it is of no consequence
when considering centre-fed antennas which use tuned lines exclusively.
Tuned feedlines operate on the principle that they are really a part of the antenna
and have ‘standing waves’ along their lengths. Standing waves are a feature of most
radiating wires but, if two such wires of equal length are closely spaced (in terms of
wavelength) and fed in anti-phase, in theory they will not radiate (in practice they will
radiate a very small proportion of the RF power applied).
THE BASIC DOUBLET ANTENNA
©
R
SG
The basic doublet (Fig 4.1) is a probably the most useful simple multi-band antenna
for amateur use. It is simple and yet effective, and requires no special earth or
counterpoise arrangements. The only drawbacks are the requirement to use an ATU
Fig 4.1: The basic doublet antenna.
86
CHAPTER 4 DOUBLETS
SG
B
20
12
and that the balanced feeder cannot be routed through
the house.
The doublet is essentially a balanced system and
each half of the top, plus each wire in the feedline,
must be equal in length. The antenna top is not cut to
resonate at any particular frequency (unlike the halfwave dipole), and almost any length may be chosen
to suit an individual location.
The doublet can be used over a wide range of
frequencies although as the frequency changes so
the radiation pattern of the antenna will alter. A halfwavelength antenna has the maximum radiation at
right angles to the axis or line of the antenna. As the
electrical length of the antenna increases the phasing of the radiation from the antenna wire means that
new lobes appear and grow. Examples of polar diagrams of a half-wave and a three half-wave radiator
are shown in Fig 4.2.
When erecting an antenna of this nature there
are no particular precautions to observe except that,
due to possible problems with reactance making a
good match difficult to achieve, certain combinations
of feeder / top leg length should be avoided. These
are summarised in Table 4.1. The table shows that
when using doublet legs of 15.2m (50ft) together with
16.4m (54ft) of feeder there ought to be little difficulty
with reactance on most amateur bands. There are of
course many other combinations of top length and
Band
(MHz)
Fig 4.2: Horizontal polar diagrams for halfwave and three half-wave horizontal wires.
Lengths to be avoided (metres)
(half the total top length plus feeder length)
56.4m
93.7m
131m
3.6
29.26m
48.8m
68.3m
7
15m
25.14m
35.2m
45.26m
10.1
10.5m
17.52m
24.53m
31.54m
14.15
7.5m
12.6m
17.6m
24.2m
27.7m
32.7m
18.1
5.9m
29.7m
9.9m
33.7m
13.86m
17.83m
21.8m
25.8m
21.2
4.9m
24.7m
8.2m
28.0m
11.6m
31.4m
14.9m
34.8m
18.1m
21.5m
24.94
4.3m
21.3m
7.1m
24.1m
10m
27.1m
12.8m
29.9m
15.6m
18.5m
29
3.7m
18.3m
6.1m
20.7m
8.5m
23.2m
11m
25.6m
13.4m
28m
15.8m
30.5m
©
R
1.8 / 1.9
Table 4.1: Lengths to avoid when designing multi-band doublets with tuned feeders.
87
SUCCESSFUL WIRE ANTENNAS
L1 end (2)
L2 centre (2)
L3 total size
L4 stubs
L5 height
Inductor (μH)
Gain (dBi)
Freq (MHz)
80
40
30
168
65
467
48.6
120
25.9
11.4
3.8
89.4
34.7
248
25.8
64
11.7
11.4
7.15
63.2
24.6
176
18.3
45
7.3
11.3
10.1
20
17
45 35.3
17.5 13.7
125
98
13 10.2
32
25
4.9
3.4
11.2 11.1
14.2 18.12
15
12
10
6
2
30 25.6
11.7 9.97
83.3 71.2
8.66
7.4
21
18
2.8
2.2
11.0 11.0
21.3 24.93
22.4
8.72
62.3
6.47
16
1.9
11.0
28.5
12.7
4.95
35.4
3.68
10
0.85
11.4
50.2
4.37
1.70
12.2
1.26
10
0.13
10.9
146
20
12
Band (metres)
Table 4.3: Lengths (in feet) of an HGSW beam for 10 amateur bands.
SG
B
insulators as shown. The lower ends of the two lines should be stripped and bent
over and soldered together. The resultant active line length must be 13ft. The distance from the centre insulator to the ladder line should be 17.5ft. If you have a lot of
wind in your area you might want to tie a 1oz lead fishing sinker to the bottom of each
of the phasing lines. Alternately a string can be attached and tied to some secure
point below the antenna. AL7KK says that he has had no problem with his phasing
lines except that they curl slightly, which is not ordinarily a serious difficulty.
The antenna is completed by winding five turns of coax near the feedpoint into a
6in diameter coil and securing them with tie wraps. This acts as a cheap but effective
choke balun.
EZNEC modelling results indicate that with the antenna at λ/2 high (32.8ft on
20m), the gain will be about 11.2dBi with a peak of the elevation lobe at 29°. Calculated azimuth, elevation and SWR plots at λ/2 height are shown in Figs 4.14, 4.15 and
4.16 respectively. Even more gain is available, and more importantly lower elevation
angles of the main lobe, with greater heights. For example at 3/4-λ, the peak elevation drops to 20°, and to 15° at 1λ.
Table 4.3 shows the lengths necessary to build an HGSW beam for all bands
from 80 to 2 metres. The dimensions were scaled from the 20m model that was built
and tested, while EZNEC was used to calculate the gain and inductor values.
THE G5RV ANTENNA
©
R
Louis Varney, G5RV, designed his famous G5RV antenna in 1946, but it was not until
1958 that he wrote about it in ‘An Effective Multi-band Aerial of Simple Construction’
(RSGB Bulletin, July 1958). He described it in greater detail, again in the RSGB
Bulletin, in November 1966 (‘The G5RV Aerial - Some Notes on Theory and Operation’).
Finally, he wrote a further article, ‘G5RV Multiband Antenna . . .
Up-to-Date’, published in Radio Communication in July 1984.
The G5RV antenna has achieved almost iconic status
during the last half century, so it is perhaps worth looking at it in
some detail. In this chapter we first look at the design and how
it evolved, using Louis Varney’s own words and the original
diagrams which accompanied his articles. Then we take a 21st
century look at the design, using computer analysis, a luxury
that obviously Varney did not have in 1946, or even in 1984.
Louis Varney’s original design, as published in 1958, is
shown in the diagram opposite. Then, he wrote, “The aerial
consists essentially of a 102ft flat-top split in the centre where a
Pyrex type insulator is inserted, a 34ft long open-wire stub (spacing is unimportant) and sufficient length of 72 ohm coax or twin
Louis Varney, G5RV, in 1998.
feeder to reach the transmitter. Alternatively, open-wire feeder
98
CHAPTER 4 DOUBLETS
©
R
SG
B
20
12
may be employed from the centre of the aerial
right back to the transmitter output or ATU.”
These two methods of feeding the antenna
are shown in Fig 4.17 (a) and (b) respectively. It will be noted that what is shown in
Fig 4.17(b) is simply a basic doublet with a
102ft top and open-wire feeder.
Describing the antenna’s performance
on each band, G5RV wrote that on 20m: “...the
aerial really comes into its own. On this band
it functions as a three half-wavelength
(a)
aerial... Since the impedance at the centre is
about 100 ohms, a satisfactory match to the
72 ohm feeder is obtained via the 34ft of halfwave stub. . . By making the height a halfwave or a full wave above ground at 14 Mc/s
and then raising or lowering the aerial a bit
at a time while observing the standing-wave
ratio on the 72 ohm twin-lead or coax feeder
by means of an SWR bridge, an excellent
impedance match may be obtained on this
band.” The technique of matching an antenna by raising or lowering it seems to have
been lost over the years!
By 1966, 300Ω ‘ribbon’ feeder had become more widely available, and G5RV
wrote, “A word about the matching stub is in
order. If this is of open wire feeder construc(b)
tion (preferred because of lower losses, especially on 21 and 28 Mc/s) its length should
Fig 4.17: The two methods of feeding the G5RV
be 34ft... but if 300 ohm ribbon is used, alantenna, as described by Lois Varney in his original
lowance must be made for the velocity factor
1958 article.
of this type of twinlead. Since this is approximately 0.88, the actual physical length of the
300 ohm ribbon stub should be 29ft 6in. It should be born in mind that this matching
stub is intended to resonate as a half-wave impedance transformer at 14 Mc/s, which
was chosen as the design centre frequency for the G5RV aerial, thus giving a very good
impedance match for a 75 to 100 ohm twin-lead or coaxial cable connected to the base
of the stub.” Thus it is clear that, although Louis Varney described the G5RV as a multiband antenna, he optimised it for use on 20m.
G5RV went on to say, “An alternative arrangement to that of the matching stub and
twin-lead or coaxial cable feeder is to use an 83ft length of open-wire feeder measured
from the centre of the flat top to the terminals of the ATU.” The specific length of 83ft
(modified to 84ft in G5RV’s 1984 article) was chosen because it “permits parallel
tuning of the ATU on all bands from 3.5 to 28 Mc/s with very low feeder losses.”
G5RV’s 1966 article gave current distribution diagrams for the antenna on each of
the five HF bands then allocated to amateurs. He also described the ATU designed for
use with the antenna.
The problem of currents flowing on the outer of the coax was recognised by G5RV,
for he wrote: “Although it may be very convenient to use a length of, say, up to 100ft of
coax direct from the transmitter to the base of the matching stub, it must be remembered that such an arrangement will tend to produce currents which will flow in the outer
conductor of the coax, causing unwanted radiation from the coaxial feeder. This may be
avoided by the use of either 75 ohm twin-lead and a suitable ATU or the open-wire
99
SUCCESSFUL WIRE ANTENNAS
SG
B
20
12
feeder and ATU as already mentioned. However, the use of a wide-band balun... would
be preferable if coaxial cable is to be used. Nevertheless, in practice very satisfactory
operation can be achieved by the simple use of coax direct from the transmitter to the
base of the matching stub even though the VSWR may reach 10 to 1 or more on 3.5
Mc/s. This figure may be reduced to about 5 to 1 on 3.5 Mc/s by ‘pruning’ the coax. On the
higher frequency bands the VSWR on the coax lies between 5 to 1 and 1.5 to 1, the latter
figure applying to 14 Mc/s where, as explained above, the matching is very good.”
This suggestion of using a balun was reversed in Louis Varney’s 1984 article, in
which he wrote: “In the original article describing the G5RV antenna, published in the,
then, RSGB Bulletin November 1966 [Varney himself appears to have forgotten about
the earlier 1958 article - Ed], it was suggested that if a coaxial cable feeder was used,
a balun might be employed to provide the necessary unbalanced-to-balanced transformation at the base of the matching section. This was because the antenna and its
matching section constitute a balanced system, whereas a coaxial cable is an unbalanced type of feeder. However, later experiments and a better understanding of the
theory of operation of the balun indicated that such a device was unsuitable because of
the highly reactive load it would ‘see’ at the base of the matching or ‘make-up’ section
on most HF bands.
“It is now known that if a balun is connected to a reactivbe load presenting a VSWR
of more than about 2:1, its internal losses increase, resulting in heating of the windings
and saturation of the core (if used). In extreme cases, with relatively high power operation, the heat generated due to the power dissipated in the device can cause it to burn
out. However, the main reason for not employing a blaun in the case of the G5RV
antenna is that, unlike an astu [ATU] which employs a tuned circuit, the balun cannot
compensate for the reactive load condition presented to it by the antenna on most of the
HF bands, whereas a suitable type of astu can do this most effectively and efficiently.”
(Louis Varney used the term ‘ATU’ in 1958 and 1966, but in the August 1983 Radio
Communication he had had an article published in which he argued the case that the
device ought more properly be called an ‘Antenna System Tuning Unit’, or ‘astu’. More
accurate or not, the name did not catch on.)
Instead, he recommended the use of an ‘HF choke’, a device which these days is
often referred to as a common-mode choke balun: “Under certain conditions, either
due to the inherent ‘unbalanced-to-balanced’ effect caused by the direct connection of
a coaxial feeder to the base of the (balanced) matching section, or to pick-up of energy
Flat top plus about 17ft (5.18m) of the matching section forms a λ/2 dipole partially
folded up at the centre. Reactive load.
R
3.5MHz
7MHz
Flat top plus 16ft (4.87m) of the matching section functiuons as a partially folded-up
collinear array with two half-waves in phase. Reactive load.
Collinear array with two half-waves in phase. Reactive load.
14MHz
3λ/2 centre-fed long wire. Matching section functions as a 1:1 impedance transformer.
Resistive load, approx 90Ω.
©
10MHz
18MHz
Two full-wave antennas, slightly folded up at the centre, fed in phase. High impedance load, slightly reactive.
21MHz
5λ/2 long wire. High impedance load, virtually non-reactive.
24MHz
5λ/2 long wire with low resistive load of approx 90 - 100Ω.
28MHz
Two x 3λ/2 long wires fed in phase. High impedance load, slightly reactive.
Table 4.4: G5RV antenna theory of operation on each of the HF bands (Source: ‘G5RV Multiband
Antenna... Up-to-Date’, by G5RV, July 1984.)
100
CHAPTER 4 DOUBLETS
R
SG
18MHz
14MHz
B
10MHz
7MHz
20
12
3.5MHz
24MHz
21MHz
28MHz
©
Fig 4.18: Current standing-wave distribution on the G5RV antenna and matching section on each of the
HF bands. (Source: ‘G5RV Multiband Antenna... Up-to-Date’, by G5RV, July 1984. )
radiated by the antenna, a current may flow on the outside of the coaxial outer conductor. This is an undesirable condition and may increase chances of TVI to nearby TV
receivers. This effect may be considerably reduced, or eliminated, by winding the coaxial feeder into a coil of 8 to 10 turns about 6in in diameter immediately below the
point of connection of the coaxial cable to the base of the matching section.”
By 1984, radio amateurs had been allocated additional bands at 10.1, 18.0 and
24.8MHz, and in his article ‘G5RV Multiband Antenna . . . Up-to-Date’, Louis Varney
described the theory of operation on each of the HF bands, including the three new
ones (Table 4.4). The current distribution on each band is shown in Fig 4.18.
101
SUCCESSFUL WIRE ANTENNAS
For use in restricted spaces, Louis Varney wrote that, “because the most useful
radiation from a horizontal or inverted-V resonant antenna takes place from the centre two-thirds of its total length, up to one-sixth of this total length at each end of the
antenna may be dropped vertically, semi-vertically, or bent at some convenient angle
to the main body of the antenna without significant loss of effective radiation efficiency.” This would imply that the full-size G5RV could be fitted into a space just 68ft
(20.72m) long, if 17ft of wire were to be dropped vertically at either end of the antenna.
HALF-SIZE G5RV
20
12
51ft (15.5m)
-------
Matching stub
17ft (5.18m) open-wire feeder or
15ft (4.56m) 300Ω ribbon
Common-mode choke balun
(8 - 10 turns coax, 6in diameter)
50Ω coax to ATU
-------
Wide-range
ATU
The Half-Size G5RV is shown in Fig 4.19.
Writing in 1966, Louis Varney, G5RV, stated
that, “It is quite possible to scale all wire
dimensions (including that of the stub) down
to exactly half-size and the resulting aerial
will work from 7 to 28 Mc/s. Optimum performance and impedance matching will
occur on 28 Mc/s, where the operating conditions will be as for the full-size version at
14 Mc/s.”
In 1984 he added that by strapping the
station end of the feeder (either balanced or
coaxial) and feeding it via a suitable ATU
using a good earth connection or a counterpoise wire, the half-size version may also
be used on the 3.5 and 1.8MHz bands.
To transceiver
B
Fig 4.19: The Half-Size G5RV.
©
R
SG
THE G5RV ANALYSED
Writing in the August 2010 QST, Joel Hallas, W1ZR, commented on three questions
often asked about the G5RV antenna:
 What is the function of the usual 34ft section of window line or twinlead between
the antenna and the coax?
 Should there be a balun or choke at the
transition from the balanced line to coax?
 Should the SWR on the coax be a matter
of concern?
G5RV wanted an antenna that would work
well in certain directions on 20m, and that
could also be used on all the HF bands, at
that time just 80, 40, 20, 15 and 10m. EZNEC
computer analysis, which of course was not
available to G5RV when he designed the antenna, allows us to see the radiation pattern
on 20m (Fig 4.20).
The section of balanced line, 34ft of open
wire in Varney’s article, transforms whatever
the antenna impedance is to a different imFig 4.20: Azimuth pattern of G5RV antenna on
pedance at its bottom. To say that it provides a
20m. At a typical height of 40ft, the peak take-off
good match on all bands may be wishful thinkelevation is 24°. Note the sharp lobes perpening. On 20m it is a half-wave long and thus
dicular to the antenna, as well the broad lobes
repeats the antenna impedance, as it does
at other potentially useful angles. The gain at
on 10m. The half-wave window line or twinlead
each is within a dB or two of the dual lobes of a
that he used as a transforming section resulted
half-wave dipole at the same height.
in an impedance at the bottom on 20m of
102
CHAPTER 4 DOUBLETS
SG
B
20
12
around 100Ω with some reactance.
Computer analysis also allows us to plot
the 75Ω SWR from 3.5 to 30MHz (Fig 4.21). It
shows that there are indeed multiple resonances; however, not many of them line up well
with amateur bands.
The dimensions given by G5RV, and with a
height of 40ft above ground, produce a 75Ω SWR
of 6.5:1 on 3.7, 5.6:1 on 7.1, 2.4:1 on 14.2, 4.6:1
on 21.2 and 2.1:1 on 24.9MHz. Other bands are
higher, typically at least 10:1.
A fundamental limitation of the design is that
there are only three adjustments - the flat-top
length, the height and the transforming section
Fig 4.21: 75Ω SWR plot of Varney’s original
length. With those variables, you can probably
G5RV antenna design using a 34ft section of
find dimensions that will work on multiple, but
air-dielectric open wire line as transforming
not all, bands. Unlike trimming a half-wave disection. The resonances almost line up with
pole, the direction to go with each change is not
some amateur bands, with a 1.7:1 SWR at the
obvious. W1ZR says he has never found a set of
top of 20m.
dimensions that resulted in acceptable SWR on
all, or even most, bands.
The question about the importance of the SWR depends on the length and loss
of the coax used, as well as the tuning range of the ATU used. A high SWR on the
higher bands will result in significant loss for typical coax lengths. This makes the
SWR at the radio look better than it really is, since the loss reduces the power that
gets to the antenna and further reduces the reflected signal. This may explain why
many think it has better SWR on multiple bands than it really does.
As with any balanced load to unbalanced line transition, the need for a balun
depends on the amount of current that flows on the outside of the shield. This in turn
depends on the ground impedance and the electrical length of the coax. Considering
its use on multiple bands, it is likely that there will be some bands that have high
shield currents and thus could benefit from a balun. At least one commercial manufacturer just slips multiple ferrite beads on the coax just below the transition with
good results.
THE ZS6BKW ANTENNA
©
R
In 1982 Dr Brian Austin, ZS6BKW (now G0GSF), used some early computer modelling to optimise the G5RV antenna. By then, UK amateurs had access to three additional bands at 10.1, 18 and 24.9MHz, which were
not available when G5RV designed his antenna,
so operation was also considered on these bands.
In 2007 G0GSF re-computed his design and came
up with new dimensions for an antenna that
presents a better than 2:1 SWR without the use of
an ATU in the 40, 20, 17, 12 and 10m bands. It can
also be used with an ATU on 80, 30 and 15m.
He wrote: “The configuration of the ZS6BKW
is shown in Fig 4.22. The dipole radiator is L1, the
series section impedance matching transformer
(to give its formal name), with characteristic impedance of Z2, is L2 spaced twin-wire. The lower
end of L2 presents an impedance, Z3, to the coFig 4.22: The ZS6BKW antenna as computed
axial cable, Z4 (50Ω as is standard practice in all
in 2007 by G0GSF (ex-ZS6BKW).
modern radio systems). A computer-based pre-
103
CHAPTER 5 VERTICALS
20
12
5
V er ticals
A
SG
B
FTER THE DIPOLE, the vertical antenna in its various guises is probably the
second most widely-used HF antenna today. Like the dipole, the basic quarter-wave vertical is simple to make and can almost be guaranteed to work
with minimal ‘pruning’ required, provided it is made well and certain guidelines are
followed. However, while a horizontal dipole is often easy to mount ‘in the clear’, a
vertical, ground mounted in a typical garden for example, is liable to be screened by
nearby objects such as buildings and trees. As a result its performance in typical
urban or suburban locations can sometimes be disappointing. Furthermore, a quarter-wave vertical needs a ground plane, usually in the form of radial wires, to work
properly, and a less than adequate ground connection can also lead to disappointing results. Nevertheless, a simple quarter-wave vertical wire can work well, and in
certain circumstances extremely well, as we shall discuss later in this chapter.
There is a tendency to think that because a vertical wire takes up virtually no
space at all, it is an ideal antenna for those with very limited space. Unfortunately, this
is usually not the case. Because quarter-wave verticals require radial earth wires, a
quarter-wave vertical antenna system can take up at least as much space as a
horizontal dipole for the same frequency band. In the ideal case, quarter-wave long
radials will extend in all directions and the vertical radiator would therefore be in the
centre of a square a half-wavelength long by a half-wavelength wide. Nevertheless,
it is possible to make certain compromises without affecting the performance too
greatly and, provided you are prepared to put in the ground work (literally), verticals
can be very effective antennas, even for those with limited space for antennas.
THE QUARTER-WAVE VERTICAL
©
R
The most basic vertical antenna is the quarter-wave. In this configuration one connection from the feeder is taken to the quarter-wave vertical radiating element, and
the other is taken to ground. In this way the ground provides the ‘image’, or other half
of the antenna, as shown in Fig 5.1(a). As such the ground connection is an integral
part of the antenna system as a whole, and upon its effectiveness rests the efficiency
of the whole antenna. In fact this is true for any antenna of this nature that uses the
ground for one of its connections.
In view of the fact that one of the connections from the feeder is taken to ground,
this type of antenna is an unbalanced antenna. Accordingly it can be fed directly
using unbalanced feeder, such as coax, without the need for a balun.
The impedance at the point where a resonant quarter-wavelength vertical conductor meets the ground is about 36Ω - half of the feed impedance at the centre of a
resonant half-wave dipole. The current along the quarter-wave vertical antenna is at
its maximum at its base and therefore the greatest radiation will take place at this
point - see Fig 5.1(b). The radiation will be vertically polarised and in the example
illustrated will have equal field-strength levels in all directions.
Much of its radiation will be at low angles to the horizon when above a good
ground, and this makes the vertical antenna very attractive for both short-distance
(ground wave) and long-distance communications on the lower-frequency bands.
115
(a)
20
12
SUCCESSFUL WIRE ANTENNAS
(b)
SG
B
Fig 5.1: (a) The basic quarter-wave vertical antenna positioned over perfect ground, showing its
earth image. Most of the earth return currents flow through the ground in the vicinity of the antenna.
(b) A representation of a quarter-wave vertical antenna over perfect ground which is energised by
a signal with a base current of 1A. The RF current at 10° points along its length is shown and also
the impedance at these points. There is a rapid fall in current towards the top of the antenna and
the impedance therefore rises greatly there. It is interesting to note that the fall in current over the
final 30° of this antenna is almost linear.
©
R
The polarisation of an antenna when used for long-distance work does not matter,
for the effects of refraction in the ionosphere etc will inevitably induce changes in
polarisation.
In order to be able to gain the most from a vertical antenna, the ground system
that is used with it must be efficient. One solution is a mat of buried wire extending to
at least a quarter-wavelength and possibly a half-wavelength from the base of the
antenna but for most practical situations this may not be possible. The antenna will
still work with several buried radial (the more the better). Ground systems were
discussed in Chapter 1 and there is more on wire radial systems later in this chapter.
As an alternative to the ground-mounted vertical it is possible to elevate the
antenna and use a ground plane system, in which case the ground plane wires
should be resonant, a quarter-wave long. Raising the antenna in height allows it to
take advantage of the ‘height gain’ available.
Fig 5.1(a) represents a simplified and ‘ideal’ quarter-wave vertical antenna. The
ground is shown to be a perfect conducting medium, a condition which can only be
realised when it is replaced by a sheet of metal which has dimensions that are large
relative to the length of the antenna or by a large body of salt water. The ground, if it is
a perfect conductor, will behave like an electrostatic shield and provide an ‘image’
antenna a quarter-wave below the radiator. This image completes the missing half
of a half-wave antenna, and earth return currents will be induced in the ground.
116
CHAPTER 5 VERTICALS
SHORTENED VERTICALS
©
R
SG
B
20
12
It is seldom possible or convenient to erect a full-sized quarter-wave vertical for the
lower-frequency bands, although such antennas are often used on the higher frequencies. For the lower frequency bands it is often necessary to look at ways of
physically reducing their length. In Fig 5.2(a) the full quarter-wave is in the vertical
plane and is shown to be bottom fed (impedance 36Ω). Figs 5.2(b), (c) and (d) show
reducing lengths of the vertical antenna sections and corresponding increases in
the lengths of the horizontal components. The total height of the antenna is therefore
lowered and in (d), where only 25% of the quarter-wave is vertical, the antenna is only
0.06-wavelength above ground.
The three ‘bent’ quarter-wave antennas shown in (b), (c) and (d) are called ‘inverted-L’ antennas, and they are very popular arrangements when mast height is
limited. As the vertical part of an inverted-L is reduced in length, the proportion of the
radiated power at low angles and in the vertical plane also diminishes. The horizontal top section will then contribute more of the total radiation, this radiation being
horizontally polarised and at high angles to the horizon. This high-angle radiation is
a result of the antenna being close to the ground.
An inverted-L similar to that shown at (c), where the vertical and horizontal portions are equal in length, should give useful vertically-polarised radiation at low
angles for both DX work and also local working within the ground wave range. The
high-angle radiation from its top horizontal half will be effective for short range communications.
In Fig 5.2(e) the top half of the quarter-wave is dropped down towards the ground.
Fig 5.2: The vertical quarter-wave can have a proportion of its length bent horizontally as shown in
(b), (c) and (d). When this is done the antenna is called an 'inverted-L'. As the proportion of the
vertical section falls the vertically polarised radiation at low angles also falls, the horizontal top
giving horizontally polarised high-angle radiation. The example shown at (d) will have most of its
radiation at very high angles and will only be suitable for short to medium distance working. It will
also have a much reduced ground wave. Bending the top of the inverted-L down (e) will mean that
the antenna currents in the two sections will then tend to be out of phase and begin to cancel. At (f)
the sloping wire will behave almost like a length of unterminated open-wire feeder.
117
CHAPTER 5 VERTICALS
 It doesn’t matter if the vertical section is not fully vertical; for a given support
height it may be advantageous to have a longer ‘semi-vertical’ section by
sloping it slightly away from the truly vertical;
 Finally, and perhaps most obviously, ensure the vertical section is as high as
possible.
20
12
The results could not have been more different between the two inverted-Ls and
are perhaps counter-intuitive, with the shorter antenna working better on 160m than
80m and the longer one working better on 80m than 160m.
©
R
SG
B
FISHING FOR DX: G0GBI FISHING ROD INVERTED-L
This antenna idea was described by Glenn Loake, G0GBI, in the July 2010 RadCom.
It builds on the popular use of a fishing rod as a support for a wire vertical or invertedL by utilising fishing reel and line as well, as shown in Fig 5.15. The basic principle
is to use a fishing rod and reel to loft a weighted, non-conductive leader line to the top
of a tree. The weight should cause the line to hook over a branch and head for ground
level. The other end of the line is permanently attached to a 132ft (40.2m) length of
Fig 5.15: General arrangement of the fishing rod antenna system.
135
SUCCESSFUL WIRE ANTENNAS
Left: The rod support: note how the tube protrudes an
inch or so above the support pole.
20
12
Below: Attaching the feed (the antenna was not deployed at this point, which is why the antenna wire is
not visible).
©
R
SG
B
wire that acts as the antenna element. A stand for the
fishing rod plus a couple of guys and some wiring
completes the set-up.
The parts are quite easy to obtain. When the antenna is dismantled it fits easily into a car. The only
long piece is the aluminium support tube, which could
be cut in half and then sleeve joined. It should take
less than 10 minutes to erect once a suitable tree
has been selected.
None of the parts are terribly critical, so the following is just a guide. You will need a beach caster
type fishing rod. G0GBI used a telescopic one, but
any kind about 8 - 10ft long will do (though you should
avoid the conductive carbon-fibre types). You will also
need a centre pin type plastic or wooden fishing reel;
a commercial fly reel will be fine. Make a connection
point in the side of the reel by putting a bolt through it
or remove the reel winder knob and put a bolt in its
place. Use a solder tag to connect one end of a 132ft
piece of thin wire; any wire can be used provided it is
thin enough to fit comfortably on the spool with space
to spare. Next, attach about a 60ft (20m) length of light
string or nylon fishing line to the other end of the wire.
A solder tag at the end of the wire provides a handy
attachment point. Wind the string on to the reel on top
of the wire.
The rod support is constructed from a length of
aluminium tube, approximately 8ft (2.5m) long and
1.5in (37mm) diameter, although the dimensions are
not at all critical. Attached to the support is a piece of
PVC pipe of a suitable diameter to take the bottom of
The completed fishing rod inverted-L,
showing the support pole, sleeve, reel
and rod.
136
CHAPTER 5 VERTICALS
Length
(m)
(ft in)
160m
80m
40m
30m
20m
17m
15m
12m
10m
39.5m
20.5m
10.5m
7.4m
5.3m
4.1m
3.5m
3.0m
2.7m
129'6"
67'5"
34'8"
24'4"
17'4"
13'7"
11'7"
9'10"
8'9"
Table 5.5: Suggested counterpoise lengths.
20
12
Band
©
R
SG
B
the fishing rod. The pipe is about 16in (40cm) long and does not need to be a tight fit
to the rod. The plastic pipe can be attached to the support rod using two jubilee clips
(hose clamps), as shown in the photo. A piece of steel reinforcing bar about 4 - 5ft
long is used as a ground stake.
The final parts to make are the feed and counterpoise. The prototype used a 10ft
(3m) length of RG58 coax with a PL259 plug on one end. The other end had an
alligator clip on the centre conductor to connect to the end of the antenna wire. The
counterpoise length in metres is calculated as 75/frequency (MHz), which allows a
bit of extra length for trimming. Table 5.5 gives suggested values for the mid-point of
the HF bands, though you may well find that trimming these by 5% or so will be better.
A single counterpoise wire per band was used, although more would probably be
better.
To deploy the antenna you need to know how to beach cast a fishing rod. If you
don’t, please find someone to teach you otherwise you could injure yourself or others. Select a suitable tree and make sure that there are no people or animals nearby
that could be hurt when you cast the leader. Trees beside footpaths are particularly
prone to people walking near them, and folk tend to get upset if you hit them with
flying lead. Respect the wildlife that may be in the tree - after all it’s their home!
Thread the leader through the rod loops (just like a fishing line) and attach the
weight to the end of the leader. Let a good bit of slack off the reel, ensuring it doesn't
tangle. Don't try to cast straight off the reel or a ‘bird's nest’ (tangle) will result. Beach
cast towards the top of the tree. With luck the weight will carry the leader over a high
branch and fall to the ground. Pull the leader over the branch so that the end of the
antenna wire is several feet from the leaf canopy. Tie off the leader at the base of the
tree. Go back to the rod and pay out the antenna wire as you walk away from the tree.
When the wire is fully extended, set up the ground stake, slip the rod support over it
and then put the rod in the top of the support. If the antenna wire is a bit saggy then
you can go back to the tree and tighten it by pulling on the leader.
Depending on the stoutness of the ground stake and the weight of the antenna
wire, you may find it necessary to use some guys to keep the rod support upright.
Finally, connect the feed to the bolt on the reel and arrange your counterpoise.
Another method of feeding the antenna is to put an automatic ATU on the ground
at the base of the antenna with a wire connected to the driven element. The ATU earth
can then be connected to the counterpoise or even just to an earth stake. G0GBI says
that if it is windy the SWR will vary alarmingly as the tree sways about, but in practice
he has not had any real problems. He uses a small LDG auto tuner.
137
SUCCESSFUL WIRE ANTENNAS
20
12
REMOTELY TUNED INVERTED-L
If an automatic ATU is available, a practical remotely tuned multiband inverted-L is a
possibility. See Fig 5.16. As with any inverted-L, the vertical section should be as long
as possible. A good overall length to aim for would be about 86ft, although this is not
critical since the antenna system is brought to resonance with the ATU. For multiband
use, a length of around 65ft should be avoided as this would be close to a half-wave
on 40m and thus would present a high impedance to the ATU and might therefore be
difficult to match.
SG
B
ti
Fig 5.16: Remotely tuned multi-band inverted-L.
©
R
160M INVERTED-L PERFORMANCE
Amateurs who wish to operate on the 160m band, but only have a limited space for
antennas, often use an inverted-L. It is arguably the best ‘compromise’ antenna for
160m if you are short of space. But just how well does it work? In the May 2008 QST,
Al Christman, K3LC, described computer simulations of several different configurations of 160m inverted-L antennas. The height of the vertical section of the radiator
and the length of the radials are varied in 20ft intervals from 30 to 90ft. The complete
QST article includes calculated data for input resistance, peak forward gain, SWR
bandwidth, efficiency, front-to-back ratio and front-to-side ratio for each case. Here,
we present an edited summary of this study.
For those who do not have access to a support high enough to hold up a full-size
160m monopole the choice is straightforward - either use a shortened monopole
with base, centre or distributed loading, or use a full-size λ/4 antenna with the vertical
portion going to the top of an available support and the rest extended horizontally to
a second support, as shown in Fig 5.17. While either technique can be used, the
second provides for efficiency and bandwidth approaching that of a full-size monopole.
The loaded antenna has lower radiation resistance, resulting in more of the transmit
power being lost in the resistance of an imperfect ground, and generally has nar-
138
20
12
CHAPTER 5 VERTICALS
Fig 5.17: The configuration of an inverted-L antenna (diagrams reprinted with permission of the American Radio
Relay League).
Fig 5.18: This inverted-L antenna uses
a ground screen composed of 60 buried radials, each of which is 50ft long.
The height of the vertical section of
the radiator is 50ft, and the horizontal
portion has a length of 84.428ft, which
resonates the antenna at 1830kHz.
©
R
SG
B
rower bandwidth (unless the losses are so high that it starts acting like a dummy
load). The only downside of the inverted-L is that it requires a second support and
has some directivity - but perhaps that can be used to advantage.
All of the antennas described here were modeled using EZNEC/4 with a double
precision calculating engine. Unlike other EZNEC versions, this version of the software is capable of simulating vertical antennas with radials buried in real ground.
The soil was assumed to be ‘average’ (conductivity of 0.005 siemens per metre,
dielectric constant of 13). The radiator (both vertical and horizontal portions) is made
from 12 gauge copper wire and the radials modelled with 16 gauge copper wire. The
number of radials was fixed at 60 because it is well known that it is important for a
vertical antenna to have a good ground system. Tapered segment lengths were
used for all wires, in accordance with the most conservative NEC modelling guidelines. The inner segment of each radial is about 1ft long, and slopes downwards
from the base of the vertical element (at exactly H = 0) to its ultimate burial depth of
3in; the remaining length of the radial is completely horizontal.
The height of the vertical section of the inverted-L radiator was initially set at 30ft,
using 60 buried radials that were also 30ft long. The length of the horizontal portion
of the wire was then adjusted in order to resonate the antenna at a frequency of
1830kHz, after which all of the important performance data was collected. Next the
length of the vertical section was progressively increased to 50, 70 and finally 90ft,
with the tuning and measurement process being repeated each time. This entire
sequence was then carried out again, as the length of the buried radials was increased in succession from 30 to 50 to 70 to 90ft.
Fig 5.18 shows what the antenna looks like when the height of the vertical section is 50ft, and the 60 buried radials are also 50ft long. In this case, the horizontal
portion of the inverted-L had to be cut to a length of 84.428ft to achieve resonance
(input reactance close to zero) at 1830kHz.
Modelling results
The resulting elevation-plane radiation pattern, in the plane containing the invertedL, is given in Fig 5.19. Notice that maximum gain is actually directed opposite to that
of the horizontal section of the radiating element; if you want to beam the strongest
139
SUCCESSFUL WIRE ANTENNAS
R
SG
B
20
12
signal to the north-east, the horizontal portion of the
L must extend towards the south-west. However,
the front-to-back and front-to-side ratios are modest for all of the designs studied here, and are always less than 3dB.
Table 5.6 displays the input resistance at resonance for the various inverted-L configurations. We
can see that, if the length of the radials is held constant, making the antenna taller will increase the
magnitude of the input resistance. In contrast, if the
height of the radiator is fixed, making the radials
longer causes the input resistance to fall because
the ground loss resistance is lower.
Table 5.7 shows how the maximum gain of the
antenna and its corresponding take-off angle (TOA)
vary with the length of the radials and the height of
the vertical section. Notice that, for 30 and 50ft
radials, making the antenna taller always yields
Fig 5.19: This is the elevation-plane radiation
more gain, at least for the range of heights dispattern of the antenna shown in Fig 5.18, in the
cussed here. However, with 70 and 90ft radials, an
plane containing the inverted-L wire. The front
element with a height of 70ft is actually a bit better
of the main lobe is directed towards the left,
than one that is 90ft tall.
opposite to the position of the horizontal secWhen it comes to elevation angles, increasing
tion of the radiator element. The peak gain is
the height of the antenna always produces a lower
0.27dBi at 30.1° take-off angle, and the frontTOA, for radials of any particular length. For examto-back ratio is 1.42dB.
ple, an element height of 30ft generates maximum
gain at an elevation angle of 38.4° (on average)
versus a typical TOA of 30.5° for a height of 50ft. Increasing the antenna height to 70ft
drops the TOA to about 27.1°, while a height of 90ft results in a peak elevation angle
of 25°.
The bandwidth capability of each antenna is listed in Table 5.8. These values
were determined by calculating the standing wave ratio as a function of frequency,
with the input resistance at resonance used as the reference impedance. If the radial
length is fixed, making the antenna taller always increases the SWR bandwidth. On
the other hand, for a given radiator height, making the radials longer always reduces
the bandwidth.
Height (feet)
30
©
Radial
Length
(feet)
30
50
70
90
50
70
90
Ω)
Input Resistance (Ω
19.15
14.72
12.29
11.02
25.43
21.35
19.16
17.97
33.52
29.43
27.17
25.86
40.99
36.81
34.43
32.99
Table 5.6: Input resistance for inverted-L antennas at resonance, as a function of
radial length and antenna height. In each case, the ground screen is composed of
60 radials in ‘average’ soil (see text). The horizontal portion of the wire radiator is
trimmed to resonate the antenna at 1830kHz.
140
CHAPTER 5 VERTICALS
Height (ft)
50
70
90
Gain (dBi) and Take-off Angle (Degrees)
–1.77 @ 38.4
–0.82 @ 39.3
–0.21 @ 37.9
+0.16 @ 38.0
–0.41 @ 30.6
+0.27 @ 30.1
+0.66 @ 30.1
+0.89 @ 31.3
–0.01 @ 27.2
+0.51 @ 27.2
+0.82 @ 26.5
+1.01 @ 27.4
+0.11 @ 25.0
+0.55 @ 24.7
+0.81 @ 25.2
+0.99 @ 24.9
20
12
Radial
Length (ft)
30
50
70
90
30
Table 5.7: Peak forward gain and corresponding take-off angle for inverted-L
antennas, as a function of radial length and antenna height. In each case, maximum gain occurs in the plane containing the radiating element, and is oriented
opposite to the direction of the horizontal portion of the L.
R
SG
B
Efficiency and other data
EZNEC has the ability to estimate the average gain of an antenna and compare this
with a theoretical lossless antenna operating in a lossless environment. Its average
gain (over all angles of elevation and azimuth) is exactly 1, or 0dB, and its efficiency
is therefore 100%, while an antenna whose average gain is -3dB must have an
efficiency of 50%. Table 5.9 provides a compilation of the efficiencies of our invertedL antennas, based on the computer-predicted values for their average gain.
If the antenna height is held constant, we can see that making the radials longer
always increases the efficiency. This intuitively makes sense because we expect a
larger ground screen to reduce losses in the system. If the length of the radials is
fixed at either 30 or 50ft, making the antenna taller always improves the efficiency.
The story changes, though, if longer radials are installed. For 70ft radials, maximum
efficiency is achieved when the height is 70ft (a height of 90ft works almost as well,
followed by a height of 50ft). If the length of the radials is 90ft, a 70ft vertical again
performs best, but now the 50ft tall element takes second place, with the 90ft vertical
in third position. Notice that the inverted-L with the highest efficiency of all those
tested (39.4%) uses 90ft radials in combination with a 70ft vertical section. In contrast, the worst antenna (which uses 30ft long radials and a 30ft tall vertical element)
is roughly half as efficient (19.8%).
Fig 5.19 showed that the inverted-L did not have a significant amount of
directionality in the azimuthal plane and this was true for all the antennas analysed.
30
Radial
Length (ft)
30
50
70
90
2:1 SWR Bandwidth (kHz)
©
Height (ft)
57
44
36
32
50
74
62
55
52
70
97
86
78
74
90
119
106
99
94
Table 5.8: 2:1 SWR bandwidth for inverted-L
antennas, as a function of radial length and antenna height. In each case, the reference impedance for the SWR is the input resistance
value given in Table 5.6.
Height (ft)
Radial
Length (ft)
30
50
70
90
30
50
70
90
Efficiency (%)
19.8
25.2
29.5
32.7
28.2
33.3
36.7
39.1
30.8
34.9
37.6
39.4
31.6
35.0
37.2
38.8
Table 5.9: Efficiency of the various inverted-L
antennas, as a function of radial length and antenna height. In each case, the efficiency is calculated from the average gain of the antenna,
as given by EZNEC. (Note that the efficiency of a
full-size λ/4 vertical, calculated by this method,
is only 40.6% - see text.)
141
SUCCESSFUL WIRE ANTENNAS
20
12
For purposes of comparison, a full-size λ/4 vertical was also modelled, using 60
λ/4 (134.368ft) radials buried in average soil. A height of 130.826ft was required to
obtain resonance at 1830kHz. The resulting input resistance was 38.25Ω. The gain
was 1.15dBi at 21.8° take-off angle, the 2:1 SWR bandwidth 109kHz and the efficiency 40.6%. It is interesting to see that the performance of this antenna is not
significantly better than that of the best inverted-L designs in our study, although it is
much taller and has a much larger ground system.
Computer models are imperfect representations of the real world, and cannot
possibly include all of the features that are actually present, such as buildings,
vegetation, other conductive objects, irregularities in the terrain, non-uniformity of the
ground constants and other local parameters. However, it is hoped that the information in this study will be helpful to those who are considering the use of an invertedL on topband.
G3PJT 160M T-VERTICAL
©
Fig 5.21: Modelling the T antenna in EZNEC.
142
Feed
Fig 5.20: Basic T antenna.
R
SG
B
The simple vertical T antenna shown in
Fig 5.20 is often recommended for 160m
use because of its low angle of radiation.
One of its principal advantages is that its
dimensions are more realistic than those
of a full-size 160m vertical. However, most
of the recommendations on making a vertical T are somewhat casual, usually
couched in terms like “for 160m use your
doublet with the feeders strapped together, fed against ground via your ATU”.
Such a nonchalant approach is unlikely to
lead to the best results. An article by Bob
Whelan, G3PJT, in the July 2009 RadCom
suggests a better way.
CHAPTER 5 VERTICALS
©
R
SG
B
20
12
A T antenna erected at a typical height of between
10 and 20m and having a span of 30 to 40m will have
a feed impedance R ± j X of between 4 - j 190Ω and
23 + j 150Ω. These impedances not only mean that
some sort of antenna tuning unit will be required to
match to 50Ω coax but also that losses due to ground
quality and the system as a whole need to be minimised.
To put this in a practical perspective, G3PJT’s
doublet was 17m high with a span of about 30m.
Modelling this using EZNEC+ (Fig 5.21) showed that
at 1820kHz an impedance of around 11+j 16Ω could
Fig 5.22: L matching section.
be expected, to which an estimate of other losses
needed to be added. The TLA program showed that
such an impedance would need a L section shunt C of 2286pF and series L of 1.8μH
to the antenna, as shown in Fig 5.22. (TLA is the ARRL Transmission Line Advanced
program, bundled with the ARRL Antenna Handbook, available from the RSGB
Bookshop and ARRL.) However, when G3PJT matched the antenna with this network, he found that a shunt C of 2850pF alone gave an excellent match at 1820kHz.
Such a simple arrangement is very low loss.
So much for the theoretical calculated impedance, but what was the actual impedance of the antenna? G3PJT used the program l_network.exe (by R J Edwards,
G4FGQ (SK), available at http://zerobeat.net/G4FGQ/index.htm) to perform the inverse calculation from TLA. This showed that his T antenna impedance was more
like 14+ j 22Ω, which of course includes all the losses. This indicated that his ground
loss was only around 3Ω, which seemed very low, but it did raise two questions: (a)
what is the range of antenna impedances that only require a shunt capacitor to
match to 50Ω?, and (b) what would be the dimensions of such a T antenna?
TLA calculated that 5+ j 14Ω, 10+ j 19Ω, 20+ j 24Ω and 30+ j 24Ω would fill the
Fig 5.23: G3PJT’s T antenna elevation radiation pattern.
143
SUCCESSFUL WIRE ANTENNAS
bill. To use these values as the basis in EZNEC you
need to insert a load that is representative of your
earth loss. In G3PJT’s case, based on the above, he
11.0
50.0
estimated this at about 5Ω and so inserted a 5Ω
17.5
34.0
load in EZNEC. Modelling his own antenna in EZNEC
23.0
23.4
(Fig 5.23) showed that its performance was pretty
good, with maximum gain 1.49dBi achieved at an
elevation angle of 24°.
Table 5.10: Suggested dimension for a T
More modelling suggested the T antenna dimenantenna requiring only a shunt capacitor to
sions
given in Table 5.10 for an efficient and simple
match to 50Ω.
match. These dimensions may need to be adjusted
to trim the antenna to the frequency desired. Of course, ground losses will affect
these values too and should be minimised. If you want to have a very efficient and
simple matching arrangement as well as an effective DX antenna on 160m these
dimensions are the ones to go for.
Span (m)
20
12
Height (m)
THE FOLDED VERTICAL ANTENNA
©
R
SG
B
One of the problems with an ordinary vertical antenna is the relatively low radiation
resistance that can lead to low values of efficiency, especially when the earth connection is poor. This can be overcome to a degree by using a folded vertical antenna.
By using a length of 300Ω ribbon as a ‘half-folded dipole’ or ‘folded unipole’, the
radiation resistance of the antenna is raised to a value in the region of 80 -150Ω,
depending on configuration and height. This means that even when used with an
average earth system the antenna will have an efficiency of around 40%.
The version described here uses a shortened 300Ω ribbon section to which is
added a length of single wire. To calculate the length of the ribbon needed, an
electrical quarter-wavelength at the desired frequency must be multiplied by the
velocity factor of the ribbon used (slotted ribbon has a velocity factor of 0.87). This
Fig 5.24: The folded vertical antenna which uses 300Ω ribbon for most of its length. The use of the
‘folded dipole’ principle raises the feed impedance of this antenna from around 15Ω to four times
this figure. A reasonable match can be obtained with 50Ω coaxial feeder.
144
CHAPTER 5 VERTICALS
THE QUARTER-WAVE SLOPER
20
12
length is less than a quarter-wave, and it must have an additional wire connected at
its end to make it up to be an electrical quarter-wavelength. This technique is very
similar to that used when constructing folded half-wave dipoles from 300Ω ribbon
(see Chapter 3).
The folded antenna illustrated in Fig 5.24 is designed for 3.7MHz operation and it
only needs 10.6m (35ft) high supports. The efficiency of the antenna is proportionally
higher than a single wire vertical because the length of its vertical section is increased as a proportion of the total quarter-wavelength. A minimum of six buried
radial wires, each being at least a quarter-wavelength long, are recommended for a
suitable earth system, although with the limitations of many garden plots this may
not be achievable. For a given length of wire, it is better to use many shorter radials
than fewer longer ones. Versions of this antenna may be scaled up or down for use
on other bands.
The step-up of feed impedance brought about by using this folded dipole technique allows the use of a 50Ω coaxial feeder. The greater distance between the
vertical part of this antenna and any buildings etc, the more effective the antenna will
be for low-angle long-distance communication.
©
R
SG
B
Sloper antennas have become popular as low-band antennas that can be erected
alongside an HF beam on a tower. Although sloping dipoles as described in Chapter 3 are very similar, these slopers comprise a single section and are usually
classed as a form of vertical antenna. They provide a little directivity and can be
erected very easily, making them an ideal antenna for many stations.
The quarter-wave sloper or ‘half-sloper’ antenna is really half of an inverted-V
dipole. A quarter-wave antenna is normally arranged to be bottom-fed, which means
that the maximum radiation is at the base of the antenna. By inverting the feedpoint of
a quarter-wave as shown in Fig 5.25(a), the position of the current maximum and
therefore the greatest radiation can be moved to the top of the antenna.
The ‘ground’ against which slopers are fed is usually the metallic mass of the
support tower, although for this to be as efficient as possible, a good low resistance
ground at the foot of the tower remains important. The slope angle, ‘L’, of a halfsloper is usually 45°, and its maximum radiation is in the direction of the wire away
(a)
(b)
Fig 5.25: (a) A quarter-wave sloper antenna used with a metal tower; (b) Using a quarter-wavelength of wire which is almost vertical to replace the metal tower. The feed impedance of the sloper
depends upon several variables, one being the angle ‘L’.
145
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement