The G5RV Antenna System Re-Visited
The original G5RV antenna system consists of a center-fed horizontal 102' wire plus a 34' length of open-wire 525-Ohm feeder. Louis Varney, the antenna system's developer, intended two other features. First, the main feeder that we connect to the base of the open-wire section should be 75-Ohm twinlead or coaxial cable. Second, the main feeder should go to an antenna tuning unit (ATU) and not directly to a transceiver.
In Part 1, we examined some of the basic properties of the G5RV antenna system at its basic design frequency, 14.15 MHz. We explored some of the variations created by varying the height of the antenna above ground and by using different wire diameters. While none of these variations has much of an effect if we use an ATU between the main feeder and the transceiver, they become important if we attempt to use the antenna system without a tuner. With the physical dimension selected by Varney, the system provides only a partial coverage of 20 meters with a 75-Ohm SWR under 2:1, although a tuner would easily permit full band coverage.
Somewhere along the line of time, the G5RV antenna system has acquired a false aura: namely, that it can cover many amateur bands in the HF region without the use of an antenna tuner. Since almost any rudimentary analysis of the antenna system can show this reputation to be false--and not consistent with what Varney wrote about his antenna system--we shall not dwell on that matter. We shall, of course, present some modeling data that confirms the inaccuracy of the reputation. However, there is a much more interesting question to investigate.
If the antenna system will not provide the desired coverage without an antenna tuner, why use the matching section at all? Why not simply run a feedline of one impedance all the way from the antenna wire to an antenna tuner? Varney recognized that this mode of operation is quite feasible. Nevertheless, he believed that his matching section offered some advantages on most amateur bands. Let's see if we can uncover them.
A single center-fed linear element (regardless of the element diameter) will have a pattern that is broadside to the element from a length of about 1/3 wavelength (about the shortest practical doublet length) to a length that is a bit over 1 wavelength. The electrical length of a fixed length physical doublet will increase as we increase the operating frequency. A 3/2 wavelength doublet at 14.15 MHz is 1/2 wavelength at one third that frequency, or about 4.7 MHz. Obviously, the 102' wire is well under 1/2 wavelength in the 80-meter band. At 3.75 MHz, the wire is about 0.39 wavelength.
As we increase the operating frequency, the wire become electrically longer. When it is about 1.25 wavelengths, we obtain the typical extended double Zepp pattern with the strongest broadside main lobes that we can achieve from a single element, but with "ears." The ears are emerging new lobes that are part of the natural process of pattern evolution. As we increase frequency--that is, as we make the wire electrically longer--the lobes will evolve in a regular fashion.
At 1 wavelength, we have 2 lobes--one on each broadside to the wire. At 2 wavelengths, we have 4 lobes, each at quartering angles relative to the wire orientation. At 3 wavelengths, we obtain 6 lobes. In fact, the total number of lobes for any wire that is an integral number of wavelengths will simply be twice the length as measured in wavelengths.
However, lobes do no simply pop into and out of existence. As we pass any integral wavelength marker in making our wire electrically longer, the old lobes will gradually diminish and the new lobes associated with the next integral wavelength marker will emerge and increase in size. At the 1.25 wavelength point of the extended double Zepp, the 1 wavelength broadside lobe have reached their peak and are ready to diminish, while the new lobes--associated with a 2 wavelength wire--have made their appearance. As we move the wire closer to 1.5 wavelengths, the lobes reach a point of roughly equal strength. Since we have both the 1 wavelength and the 2 wavelength lobes, our lobe total is 6. We can apply similar counting methods to any wire that is x.5 wavelengths, where x is any integer.
So for any wire of any electrical length, we can predict the lobe structure. With that fact in mind, let's survey the patterns that we can obtain from a 102' wire. For the sake of brevity, I shall select only one of the 102' wires and one of the heights that we examined in Part 1. Let's use AWG #12 copper wire and place it 20 m or 65.62' above average ground.
The fixed physical height above ground, of course, will have a bearing upon the pattern by changing the take-off (TO) angle, or the elevation angle of maximum radiation as we change frequency. As we increase frequency and shorten the length of a wave, the antenna will be electrically higher. Hence, the TO angle will be lower. As a rule of thumb--although calculation equations exist in the handbooks--the TO angle of an antenna at 1/2 wavelength height is about 25-26 degrees. At 1 wavelength, the TO angle is 14 degrees. At 2 wavelengths, the angle drops to the 7-8-degree mark. One of the benefits of using a single multi-band wire antennas is that the TO angle tends to correlate with skip properties. As we increase frequency, the dominant skip angles decrease, matching our wire antenna TO angles, if we have it high enough in the first place.
Fig. 1 shows the anticipated azimuth patterns of the 102' wire at a height of 20 m above ground--about 1 wavelength high at 20 meters. Unlike the patterns for a long-boom Yagi, which might change across the span of a single amateur band, the patterns of a single wire antenna are stable and change slowly. Hence, there will be no significant difference in the 15-meter patterns from one end to the other of this 450-kHz wide band.
Each pattern in Fig. 1 shows the frequency at which it was taken, along with the TO angle. 102' represent a little over 1 wavelength at 10.125 MHz, and so we see two broadside lobes. The antenna is about 2 wavelengths long at 17 meters, revealing a 4-lobe pattern. At 10 meters, the antenna is close to 3 wavelengths long and shows 6 distinct lobes.
At 20 meters, where the wire is 3/2 wavelengths, we also find 6 lobes, but these are the product of the 1 wavelength and the 2 wavelength lobes, one set enlarging and the other set diminishing. The other bands shows lobes in various states of emergence or disappearance because the 102' wire in somewhere between the convenient marker lengths that we have designated.
With any multi-band single-wire antenna, the user has some decisions to make. If he has some latitude in orienting the antenna, he can choose a favorite band and orient the wire so that a major lobe points in the direction or directions of favored target communications areas. Or he can spend nights of pencil and paper planning trying to figure out the best orientation that will yield the best possible results on all favored bands.
Before we try to feed this wire, let's examine one other feature of the lobe structure of the 102' wire. The following table provides the maximum gain and TO angle of the 102' wire as we installed it at 20 m above ground. Maximum gain is the strength of the most major lobe (of which there may be more than one).
1. 102' AWG #12 Copper Wire Gain and TO Angles Band Freq. Max Gain TO Angle Meters MHz dBi degrees 80 3.75 6.00 60 40 7.1 7.94 29 30 10.125 9.68 20 20 14.15 8.37 14 17 18.118 9.37 11 15 21.1 10.05 10 12 24.94 10.57 8 10 28.1 10.12 7 Note: Antenna height = 20 m. Maximum gain = gain of the strongest lobe. TO angle = elevation angle of maximum radiation.
There is a general trend toward higher gains in the major lobes as we increase the electrical length of the wire by increasing frequency. This property applies to any horizontal wire antenna, regardless of any special name we might give it. However, increase major lobe gain is accompanied by a disadvantage: the width of the major lobes decreases as we electrically lengthen the antenna wire and place more lobes into the pattern. Hence, the higher the frequency of our 102' wire, the more finicky becomes the aim at a target area.
You may also note another trend in the number, most clearly revealed by examining the numbers of 30, 20, and 17 meters. Note that the maximum gain on 20 meters is less than the values for 30 and 17 meters. One of the phenomena of lobe emergence is that, in general, when we are at the x.5 wavelength region, the emerging and diminishing lobes will have a bit less strength, because we are combining two lobe structures.
The final feature that we want to notice is the feedpoint impedance of the 102' wire as taken at the center point of the wire itself. These values will give us some clue as to the rationale behind the G5RV antenna system.
2. 102' AWG #12 Copper Wire Feedpoint Impedances Band Freq. Feedpoint Impedance Notes Meters MHz R +/- j X Ohms 80 3.75 46 - j 339 High relative X 40 7.1 397 + j 1037 High relative X 30 10.125 1220 - j 2522 High Z and relative X 20 14.15 104 - j 49 Low X 17 18.118 2281 + j 1624 High Z 15 21.1 337 - j 1038 High relative X 12 24.94 203 + j 328 Moderate relative X 10 28.1 2669 + j 678 High Z Note: Antenna height = 20 m
Notice the large range of the resistive components of the impedances on the HF bands--all the way from 46 to 2600 Ohms. (The resistive component at 3.5 MHz would be even lower than 46 Ohms.) As well, note how many of the bands present relatively high values of reactance--some inductive, others capacitive.
To feed this antenna with a single transmission line, we would normally select a characteristic impedance somewhere in the vicinity of the geometric mean between the extremes. Something in the 400-600-Ohm vicinity should prove usable. However, the impedance at the antenna tuner terminals depends upon three general factors--ignoring line losses for the moment: the feedpoint impedance, the characteristic impedance of the feedline, and the electrical length of the feedline. Unless there is a perfect match between the antenna feedpoint impedance and the characteristic impedance of the transmission line, the line itself will continuously transform the impedance components along each half wavelength of line at the frequency of operation. It is not at all unusual to encounter values of resistance and/or reactance at the tuner terminals that fall outside the matching range of the tuner. The most ready cure is often to insert an additional length of line to see if we cannot arrive at resistance and reactance values within the tuner's range. If we are lucky, the insertion may allow matching at all used frequencies. If we are not so lucky, then we may need to developing a switching system to insert the added line length on the bands for which we need it.
Now we are ready to understand part of the rationale behind the G5RV antenna system, with its 34' of 525-Ohm transmission line.
3. G5RV's analysis of the system at all HF frequencies
Note: Load Impedance is the impedance at the end of the "matching
section."
Band Analysis Load Impedance
80 meters Wire + Section = shortened Dipole Reactive (R+/-jX)
40 Wire + Section = partially folded
2-half-waves in phase Reactive (R+/-jX)
30 Wire + Section = partially folded
2-half-waves in phase Reactive (R+/-jX)
20 3-half-waves Resistive (ca. 90 Ohms)
17 2-full-waves in phase High Z, slight X
15 5-half-waves High Z, resistive
12 5-half-waves Resistive (90-100 Ohms)
10 2 x 3-half-waves in phase High Z, slight X
This sort of information style makes it difficult for us to directly compare the results with the matching section with the modeled results that we obtained without the matching section. Therefore, let's do some NEC-4 modeling, using the same TL facility matching section construct that we used in Part 1. As we did initially, we shall confine ourselves to a 20-m height for the 102' AWG #12 copper wire.
While we are at the task, we can also examine some slight variations in the G5RV antenna system. All of the variations represent slight modifications in the matching section transmission line.
Version 1: the original G5RV with 34' of 525-Ohm 0.98 VF open wire line.Version 2: the common U.S. implementation of the G5RV using 34' of 450-Ohm 0.91 VF vinyl-covered window line.
Version 3: a second common implementation using 28' of 300-Ohm 0.82 VF TV-type ribbon or solid vinyl covered line, noted in the 1984 article.
Version 4: 300-Ohm 0.9 VF windowed vinyl-covered TV-type ribbon line (in the U.S., available from The Wireman in SC, but check his specification for the VF).
Allowing for the possible confusion of the VF attached to the original open-wire line by those who suggest alternative line for the matching section, the sections are all cut to be about 1/2 wavelength at 14.15 MHz. Hence, we should see about the same impedance values in all version as we obtained for the wire alone.
The following table shows the modeled impedance values at the base of the matching section for each version on each of the test frequencies spread across the HF region. As well, for reference, the tables also provide the 75-Ohm SWR values in keeping with Varney's intent that the remaining transmission line to the ATU be 75-Ohm twinlead or coaxial cable.
4. Impedances at the base of the "Matching Section" for 4 Variations on
the G5RV Antenna System
All Versions use a 102' AWG #12 copper wire at 20 m above average ground.
differences appear in the "Matching Section."
Version 1: 34' (10.36 m) 525-Ohm, VF 0.98 open wire system (G5RV
recommendation)
Version 2: 34' (10,36 m) 450-Ohm, VF 0.91 windowed parallel line (common
implementation)
Version 3: 28.0' (8.53 m) 300-Ohm, VF 0.82 solid TV-type parallel line
Version 4: 30.6' (9.33 m) 300-Ohm, Vf 0.90 windowed TV-type parallel
line
Version 1 Version 2
Band Freq Impedance 75-Ohm Impedance 75-Ohm
meters MHz R+/-jX SWR R+/-jX SWR
80 3.75 35 + j 136 9.6 31 + j 112 8.0
40 7.1 88 - j 230 9.9 60 - j 110 4.5
30 10.125 95 + j 584 50.0 103 + j 682 62.0
20 14.15 104 - j 52 1.9 104 + j 51 1.9
17 18.118 157 - j 517 25.2 73 - j 230 11.6
15 21.1 77 + j 219 10.2 86 + j 376 23.9
12 24.94 144 - j 73 2.5 145 + j 156 4.5
10 28.1 2398 + j 1002 37.6 409 - j 917 33.0
Version 3 Version 4
Band Freq Impedance 75-Ohm Impedance 75-Ohm
meters MHz R+/-jX SWR R+/-jX SWR
80 3.75 20 - j 10 3.8 20 - j 11 3.8
40 7.1 29 - j 83 5.9 29 - j 85 6.1
30 10.125 25 + j 270 41.9 25 + j 266 41.1
20 14.15 106 - j 64 2.2 106 - j 68 2.3
17 18.118 55 - j 315 26.2 57 - j 326 26.9
15 21.1 24 + j 44 4.2 24 + j 38 4.0
12 24.94 83 + j 24 1.4 83 + j 18 1.3
10 28.1 825 + j 1261 36.8 666 + j 1171 36.4
Let's initially look at a couple of bands in the whole range. Although all of the matching sections show similar impedances at 14.15 MHz, we cannot be assured that the 20 meter SWR curves will be identical for all 4 versions. Therefore, Fig. 2 shows the 75-Ohm curves for the 4 versions.
Versions 1, 3, and 4 show similar curves, since they were cut close to a half wavelength for the line used. However, the common US implementation of the G5RV simply replaces one line with another without allowing for the difference in velocity factor. Hence, the impedance transformation undergoes more than 1/2 wavelength, and the resulting impedance away from the design frequency differs from the other versions. The lesson is that if one wishes to replicate the G5RV system at 20 meters with a different matching section line, one must use some care in accounting for differences in the velocity factor.
Of all the bands, 12 meters shows the greatest promise for avoiding the need for an ATU. Fig. 3 presents the SWR curves for this narrow ham band.
As may be evident, the two 300-Ohm systems provide a good 75-Ohm SWR, while the two higher-impedance matching sections do not. The unsuspecting novice builder of a G5RV may wonder why.
The matching section is 1/2 wavelength long at 14.15 MHz. However, it has a different electrical length at every other frequency across the amateur bands. Lines having different characteristic impedances will yield different impedance transformations.
We are likely familiar with the fact by now that a transmission line of any characteristic impedance will replicate the wire feedpoint impedance if the line is electrically 1/2 wavelength. We may also be familiar with the fact that if a line is electrically an odd number of quarter wavelengths, then the impedance at the base or "sending" end will be the square of the line's characteristic impedance divided by the load impedance--in this case the wire feedpoint impedance.
However, these simplified relationships derive from a much more complex equation describing the transformation of the load impedance for any length of line whatsoever. The following equation shows the transformation, but still simplified by omitting the calculation of line losses. As noted in Part 1, the modeling software uses a lossless-line model for its calculations, and the losses in the short parallel line composing the matching section are almost small enough to be negligible.
The terms l and lambda are in the same units, where l is the electrical length of the transmission line, while lambda is a wave length. Zo is the characteristic impedance of the line; ZL is the load impedance, and Zs equals the impedance at the sending end of the line. This particular version of the impedance transformation equation comes from page 186 of Terman's Radio Engineers' Handbook. Of course, ZL may be complex (R +/- jX), and so, too, may be Zs. There are a number of utility computer programs that will calculate the impedance transformation--with or without losses--including the resistive and reactive components.
The message of the equation for this context is that the complex transformation of impedance along a transmission line, when the load impedance and the line's characteristic impedance are not a perfect match, depends on the line length and the line's characteristic impedance. The transformation on all bands for which the line is not a nearly exact multiple of a half wavelength will differ as we change the characteristic impedance of the line. Therefore, as we develop alternative types of transmission line for the matching section of a G5RV, we should not expect to replicate the impedance values of Varney's original version on bands other than 20 meters.
We can see the effect of moving from the 450-to-525-Ohm region down to 300 Ohms by looking at the impedance values for the bands below 20 meters. The higher impedance lines yield resistive components between 35 and 95 Ohms, while the 300-Ohm lines produce values in the 20-30-Ohm range. These values are also a good reason not to run the feedline to the 4:1 balun that inhabits so many network tuners in common use today. We do not need an already low resistive component further reduced.
However, the 300-Ohm line has a small advantage. It yields impedance values on more bands with 75-Ohm SWR values under 10:1. Although there is no guarantee, given the very wide variety of components used in today's tuners, the lower the overall SWR value, the more likely it is that the feedline from the matching section to the tuner will provide values within the tuning range of the ATU.
Indeed, it is now time to perform one more comparison: between the overall impedance values in the table for the 4 versions of the G5RV and the impedance values for the feedpoint of the 102' wire alone. In general, the matching section yields lower values of both resistance and reactance. Therefore, with a 75-Ohm line from the matching section to the ATU, we are likely to be able to effect a match. We would only be able to achieve this goal with parallel transmission line all the way from the wire to the ATU--and might have to insert some line on some bands.
The final question in this series in inquiries is simple: why do the job in the G5RV manner?
102' of strong copper or copperweld wire--along with sundry end rope, insulators, and a center-junction piece.
A length of parallel transmission line cut to 1/2 wavelength at about 14.15 MHz, accounting for the line's velocity factor.
A length of feedline from the matching section to the ATU. For network tuners, we might as well use 75-Ohm or even 50-Ohm cable. However, since the line will be subject to considerable SWR and hence voltage and current excursions along its length, we should use the shortest possible length to minimize losses. As well, we should use the fattest, lowest loss line that we can obtain (RG-213 or better). Because 75-Ohm transmitting twinlead is no longer made in the U.S., we can only implement the G5RV using coaxial cable, unless we are willing to build our own low-impedance parallel line.
A choke to place at the junction of the matching section and the coaxial cable, as noted in the 1984 RADCOM article.
A wide-range network tuner.
Fig. 4 sketches the essential ingredients of the antenna from the wire down to the network tuner.
When used with a wide-range tuner, there is little to choose among the versions of the matching section illustrated in these notes--or among a large lot of other potential sections. Each should be 1/2 wavelength at about 14.15 MHz. Perhaps the only general rule involved is that the higher the characteristic impedance of the matching section transmission line, the higher the impedance that is likely on the bands below 20 meters. However, 300-Ohm line (the transmitting variety, for lowest losses) offers fewer bands with very high SWR values relative to either 50- or 75-Ohm cable.
Perhaps the only other component of the system calling for comment is the choke. Very often we hear such devices being called choke-baluns or simply 1:1 baluns. Such devices have two functions that are inter-related. They provide a transition between balanced line on the one side and unbalanced line on the other. They also tend to attenuate common-mode currents on the braid of the coax. In fact, these two functions are one and the same, for the only reason for needing a transition device where we effect no impedance transformation is to suppress common-mode currents.
Newcomers to antenna work are sometimes confused by calling these current common-mode currents and also saying that they appear on the coax braid. Normal transmission line currents are ideally equal in magnitude but opposite in phase anywhere along a transmission line. Common-mode currents have the same phase on both conductors. On parallel line, such currents are of equal magnitude on each line. However, on coaxial cable, due to the skin effect which tends to cancel currents at the center of a conductor and place all current at the surface, the current is most measurable on the braid.
Louis Varney warned against the use of transformer-wound 1:1 baluns because many designs show considerable losses when the load reactance is significant. Indeed, Jerry Sevick, W2FMI, who has published the most material on transmission-line transformers, recommends that all reactance compensation occur on the load side of the balun.
In place of such baluns to suppress common-mode currents, Varney recommends a 6" diameter coil of about 8 to 10 turns of the feedline coaxial cable at the junction of the matching section and the main feedline. I have found that W2DU-type ferrite bead chokes also perform well in this function.
One recommendation that I have seen from vendors of commercially prepared G5RV kits is to use as long a run of coaxial cable as possible. Coaxial cable is inherently lossier than parallel transmission line. Any SWR factor acts as a multiplier on the basic matched-line loss of a cable at a given frequency. Hence, the only reason that I can think of for using a very long run of coaxial cable--other than one of necessity for extending from the shack to the antenna--is to use the line losses to mask the SWR at the shack end of the line. If the measured SWR at the shack end of the line is very significantly lower than the sorts of figures produced by these models--or models customized to the system proposed by a user--then they result from line losses. And the only purpose for accepting such losses would be to operate the system without a tuner.
With a wide-range tuner, one achieves the lowest feasible loss level with the shortest possible coaxial cable run.
For a good analysis of the losses associated with various ways of employing combinations of parallel line, coaxial cables, and tuners with the basic G5RV wire, see the extensive notes of Owen Duffy, VK1OD, at owenduffy.net/antenna/G5RV/.
The use of coaxial cable for the main feedline has some advantages in the modern home. Contemporary homes have walls, ceilings, and floors that are rampant with wiring and other metallic conductors associated with heating and air conditioning systems. Hence, indoors, the chances of a parallel line encountering environments that would disrupt the line balance have multiplied with time. A coaxial cable main feedline properly immunized from common-mode currents with a suitable choke offers some isolation from the conductive contents of the modern home with only small losses as the cost.
50-Ohm cable has come to rule the field of amateur feedlines. As well, the ATU remains among many folks a suspect device, since it adds to the number of boxes on the operating desk. As a result, after the appearance of the G5RV antenna system, a search ensued for a combination of antenna wire length and matching section that would yield the highest number of amateur bands offering ATU-less operation on a 50-Ohm cable. We shall devote a final part to this series to explore a G5RV variant, perhaps the most successful effort to reach the 50-Ohm cable goal.
Updated 05-21-2003. © L. B. Cebik, W4RNL. Data may be used for personal purposes, but may not be reproduced for publication in print or any other medium without permission of the author.
Go to Part 3: The Almost-No-ATU ZS6BKW
Go to Part 1: The G5RV on 20 meters
Go to Main Index