G8JNJ

Over a period of time, like most amateurs I have experimented with a variety of wire antennas. Any configurations have to fit in my 100ft long by 40ft wide garden and my aim was to try out several different configurations to see if I could improve my signal on 160m and 80m without compromising the performance on the HF bands.

My options were:- 

• Increasing the size of an existing G5RV type antenna to twice size
• A 132ft inverted L
• 70m Diameter loop
• An 80m long Carolina ‘Windom’ or Off Centre Fed Dipole
• An 80m long single wire fed Windom
• A 100ft doublet fed with 450 ohm ladder line and an auto-atu

I constructed models of all the antennas on EZNEC in order to eliminate the any really bad choices before I started stringing up miles of wire. However modeling can only provide a guide to actual performance. Factors such as physical construction and support, Signal to Noise performance on receive and the level of RF induced into nearby objects and property also have to be taken into account.

Twice size G5RV. This seemed like a good starting point; however by doubling the size of the original design the match on several bands became poor and my feed arrangement (coax fed remote auto-atu) resulted in a worse signal than the normal size version. I messed around with this quite a bit but I couldn't find a good length of antenna wire and feeder which made a significant improvement over the 'normal' sized version. In the end I decided to use a 57m (due to the position of supporting structures) long doublet fed with 450 ohm ladder line and connected to an auto-atu at the base of the antenna. This worked much better and gave good results on all bands. It also formed the basis of the antenna I finally decided to use.

132ft inverted L. This worked quite well when fed with a 4:1 balun at the base and the remote auto-atu. Some elevated counterpoise wires running along the fence under the wire helped to improve radiation efficiency. On frequencies below 18MHz the inverted L performed better than the G5RV, but on 18MHz and above it was much worse.

70m diameter Loop. This also worked quite well once the loop size became greater than twice the wavelength. However at low frequencies the results were comparable to the 132ft inverted L and the G5RV’s. It was difficult to obtain a good match on 160m and the gain seemed lower than the 132ft inverted L. I also found it difficult to support one of the sides of the loop as I could not achieve enough tension to stop the wire sagging in the middle.

Carolina Windom 80m long. This is not really a Windom at all, but an Off Centre Fed dipole with a radiating feed line. I started constructing one of these but soon came to realise that the balun was a compromise. One of the main features of the Carolina Windom is that it is coax fed, with a 4:1 balun at the feed point on the flat top portion of the antenna where the impedance should be somewhere in the region of 200 - 400 ohms. A portion of the vertical coax feed is deliberately used as a radiating element. In order to facilitate radiation from the feeder, the balun is a 4:1 voltage balun, which deliberately provides very poor isolation and balance between the input and output ports. In order to isolate the antenna from the feed cable a further 1:1 current balun is fitted between 20 to 40 feet down the coaxial feed. The use of a voltage balun (auto-transformer) struck me as a very poor feed arrangement with a limited bandwidth, which introduced a lot of additional construction effort. I also found that I had to use an extra 10m length of coax in order to present my auto-tuner (which was mounted at the base of the antenna) with an acceptable impedance it could match to on 160m. After a lot of messing around to get the Carolina Windom to work properly. I finally removed the balun and coax feed and replaced it with 450 ohm ladder line, which I connected directly to the auto-atu. This formed an Off Centre Fed Dipole (OCFD) fed with ladder line, which gave approximately 1.5 to 2dB dB improvement in gain on most bands, when compared to the Carolina Windom. My auto-tuner also seemed a lot happier with the load impedance of the 450 ohm ladder line version..

Single wire fed Windom 80m long. I modelled the Carolina Windom but found that a conventional single wire fed Windom appeared to have marginally more gain. It seemed easier to feed the single wire against either an earth mat or counterpoise wires, either with a 4:1 current balun at the base of the antenna and a coax feed back to the auto-atu in the shack, or with an auto-atu mounted directly at the base of the antenna. The performance of this antenna was generally very good with mainly horizontal polarisation on 160m and vertical polarisation on the other bands. However performance on the HF bands was not as good as the G5RV or Carolina Windom / OCFD antenna. I found the gain was not much better than that obtained with a 1/4 wave vertical, especially on 28 and 50MHz. This seemed to be mainly due to the high angle of radiation and rather scattered polar diagram.

Dipole / doublet, Carolina Windom / OCFD or Tee – Why not have a choice ?

Whilst I was trying out these antennas, It stuck me that I could use a very simple technique that would allow me to quickly modify a basic wire antenna, so that it could be used in a variety of different configurations.

As a result of this I now have a wire antenna which can be altered in a matter of minutes to change between a doublet or OCFD in flat top or inverted Vee configuration. In addition I have installed a remote switching box which allows me to select between a balanced feed or one with strapped feeders fed against a counterpoise for operation as a Tee. This is particularly useful for operation on 160 & 80m depending upon the type of communication distance required. When configured as an OCFD but strapped as a Tee. The configuration is then very similar to the G7FEK design which performs very well on the LF bands.

The design evolved after I started using pulleys and weights to keep my antenna under (almost) constant tension as a result of problems I experienced with one of my supports. This is a large fir tree which moves a lot in heavy wind conditions and used to cause one of the end supports to break. I tried using fishing line and elastic bungee (shock) cord, but the weights give the best results.

The diagram below shows the component parts of the design (sorry about the poor quality, this seems to be due to the way the images are formatted for display, if you copy it to your PC and open it with a graphics package it actually looks quite good).

An 80m long wire is fed 1/3 of the way along its length. The short leg is held under tension by a rope passing over a pulley with a suitable weight at the end. A length of nylon fishing line is attached to the end insulator and is loosely tied to a suitable point near the ground. If the insulator is pulled down by the nylon rope, the weight rises to maintain tension. This allows the height at the ends of the wire flat top section to be varied so that the antenna can be configured as an inverted Vee if desired.

The longer wire section of the flat top is broken with an insulator and a pair of wire jumpers with mating connectors. The insulator can be pulled down with the nylon fishing line and the jumper wires connected or disconnected as required, before the nylon fishing line is released to let the insulator return to its original position. This gives a choice of the antenna being fed as either a balanced dipole or Off Centre Fed Dipole as required.

Conclusions

Based on these experiments, I conclude that for multiband (160 to 6m)  operation from an average UK sized garden, for it's size it is very difficult to beat something like a G5RV, especially the ZS6BK version. If you have more money and space then a doublet with a flat top of around 50 to 70m fed with 450 ohm ladder line and a good auto-atu at the bottom of the feeder will give better results on the LF bands.

For operation on 160m and 80m it is useful to be able to strap the feeders together to form a T antenna, and feed it against a good ground system or elevated counterpoise. With an average antenna height of 10  - 15m above ground, a doublet strapped a as T can give up to 6dB improvement in performance for local contacts on 160m and 2 to 3dB improvement on 80m depending upon the distance between stations.

A Windom or OCFD produces similar results but is slightly more noisy on receive due to common mode currents on the feeder. The ability to swap between Doublet or OCFD is useful if you wish to move lobes in the polar diagram on the higher frequency bands, however it is slightly more difficult to implement. I would only consider using an OCFD if you don't have a convenient support structure  for a doublet.

If you are only interested in the LF bands, an inverted L (approx 40ft high and 132ft long) is a good simple to construct antenna which produces very good results, especially if you don’t have a lot of space for radials or elevated counterpoise wires.

Martin – G8JNJ – 22/04/2012 – V2.2