A conversation with an Antenna Master

G2QT, Frank Cooper (SK 2006)

By Bob Whelan G3PJT (written in 2005)

Frank Cooper, G2QT, started experimenting with antennas in the 1930s. In those far off days a licence was granted for the purpose of experimentation and with 10 watt output limits the antenna made a big difference to what you could hear and work. 60 years later, Frank is still experimenting. Frank has kept an antenna work book which describes the results of all of the modifications and variations and these 100 pages provide a valuable insight into what works and what doesn’t. I will attempt to summarise some of the key findings and so continue Frank’s tradition of helping the experimentally minded of us to try something new, make it, get it to work well and learn something in the process.

As everyone knows the location is the location is the location and at Sellinge, Frank is blessed with an open field about 120m by 60m. In this field Frank has five, 60 foot wooden masts arranged along the edges of the field and an 80 foot tower with a 10 foot extension in the centre. It’s an almost ideal location for antenna experimentation.

In trying a very wide range of antennas, from Zepps to quads to deltas to yagis, possibly as many as 40 or 50 variants, it has to be kept in mind that Frank experimented with and changed antennas for two main reasons :-

n that performance in terms of gain or front to back ratio was poorer than alternatives,

n that there were practical difficulties in construction, robustness or reliability.

These are the same real life considerations as you and I use to judge antenna performance and so Frank’s experience is relevant to all of us. To make the results easier to relate I will divide this article into three parts:- HF (10, 15, 20), LF (40, 80, 160) and some general hints for good practice.

HF (10,15 and 20)

Frank started in the early ‘30s with simple wire doublets, fed with open wire feeders, an effective arrangement if up in the clear, Frank noted, considerable DX has been contacted including an R8 report from Siam !. By correlating his contacts with the antenna orientation Frank was able to show that the radiation pattern was as theory expected and was relevant in deciding the orientation of such an antenna for DX.

By November 1936 his ideas were moving to beam arrays in order to improve his DX performance, he noted a beam antenna now being constructed consisting of four half waves fed in phase two above two; the whole being designed to make 14MC working possible (as well as 28MHz), Fig.1.

Frank’s experiments in the late 1940s were with driven doublet arrays for 28MHz, fed with open wire line.

Even at this stage Frank was finding anomalous results such that a 2 element array had no discernible advantage over a single element ( a double extended Zepp). Following the move to his present QTH at the end of 1956 Frank started work with a six bay Sterba Curtain with reflectors for 28MHz. Since then Frank has used cubical quads, up to 8 elements, delta loops, log periodics and yagis of various sorts.

Delta loops. One of Franks all-time favorite antennas is the delta loop in both tubing and wire forms and as single elements or as part of a multi-element array.

In comparison with fixed three bay Sterba and fixed quads a two element triband delta loop comes out ahead on gain and ease of construction. This design, Fig 2, has also been described in QST [Ref. 1]. Boom is 2 ½” diameter 13’ long.

Freq.	Driven element tubing side and wire top	Reflector element tubing side and wire top	Spacing
14MHz	26’7” and 20’	26’3” and 22’4”	12’
21MHz	16’6” and 14’6”	16’9” and 15’	9’
28MHz	12’7” and 10’6”	12’9” and 11’ 4”	6’6”

The sloping element sides are best made of telescoping tubing else the weight will be excessive, 1”, 7/8” and ¾” suffice for 14MHz, ¾” and 5/8” for 21MHz and 5/8” and ½” for 28MHz. For single band deltas a single tubing size will probably be adequate. Later versions have used fibre glass fishing rods for the outer parts of the 14MHz elements. It is quite difficult to drill the boom accurately and to help alignment and for additional strength a reinforcing plate was added to each element where the side arms meet the boom and further at each boom end the 14MHz elements were locked with bolts and an end plate. This stopped the elements rotating on the boom.

The three driven elements were matched to the feedlines with gamma matches, tuning up being ‘easy and broadband’. Frank was able to further improve the performance of the 21 and 28MHz delta by trimming the reflectors and driven elements.

Few of Frank’s antennas could beat the delta. The fact that the entire antenna is above the boom more than compensates for the lower gain of the triangular element over the conventional square quad element. Compared with a commercial triband yagi the delta was superior.

4 element quad

After building a fixed 4 element quad Frank found that it had 1 S unit gain over a two element and indeed showed even more improvement when the match was improved. As a 4 element delta would be too heavy for the tower Frank used a Gem Quad spider and spreaders Frank to construct a 4 element triband quad of conventional design. He used element lengths he had derived from other experiments and individual feedlines to the three driven elements. There are two constructional points worth noting. Frank used the diamond rather than square configuration, and secondly. used a mast extension truss to hold up the boom and a horizontal strut to the apex of the loops. Thus the loops were prevented from moving back and forth relative to the boom. This greatly improved the rigidity of the structure.

Like many others Frank observed performance differences between the three bands, element trimming again was necessary for best performance. Typical F/B ratio was 20db. In making adjustment to multi director antennas Frank found that results could be confusing if directors were adjusted individually. If directors were adjusted as a set then much more consistent results could be achieved.

A 28MHz Rhombic was constructed for a wave angle of 20° and having 113’ legs but the quad is superior.

The other way of improving the gain of quads and delta loops is to use yagi directors and indeed Frank was able to see a dB or so gain by doing this but he comments that, the mechanical problems are difficult and in some tests the gain was inconclusive.

In 1997 after using a 4 element quad for 20 years, Frank changed to a 10 element, Tennadyne log periodic. Apart from covering all the HF bands with low SWR it shows 2 S units gain over a dipole on most of the HF bands. The front to back ratio is about 6 S units, front to side 8 S units. Frank comments that one other advantage is that all the elements are in the plane of the boom, unlike a quad where the elements are partly below the boom. This means I can guy the tower at the top and therefore the system is more robust in the gales on the south coast.

LF (40, 80 and 160)

40m

Frank has probably experimented with more antenna types on 40 than on any other band to try and find the most effective antenna. For example following the success of the delta loop beams for HF Frank constructed a 5 element array of loops. These were configured as two three element beams, back to back, with common reflector and two driven elements.

Although this arrangement had adequate gain, good as a three element yagi, it proved very difficult to tune, there being considerable interaction between the two halves.

In January 1992 Frank started systematic comparisons between various antenna types for 7MHz. He first erected a 33’ vertical with a single 33’ horizontal, a sort of bent dipole configuration. This was to be a reference 7MHz antenna. Comparing this with the 7MHz yagis in use at the time showed that it was better to the US East Coast ! This was a surprising result. The wire yagis were replaced with two delta loops, the vertical was still better. The addition of two director loops, tuned for maximum forward gain, gave the delta loop array a 2 S units advantage. The vertical was then raised off the ground and 4 radials added. Any improvement in the performance of the now elevated vertical was not discernable. The next comparison was with a 4 element wire yagi. This also showed considerable gain over the vertical.

It is well known that antennas require to be tuned for maximum performance. A two element delta loop for 7MHz when erected with book dimensions did not match the performance of the vertical. When however the loops had been adjusted for optimum performance they were 2 S units superior. They were also superior to a dipole. The addition of more radials to the vertical made no difference to these findings.

The delta loops were found to be noisy on receive. Yagis were quieter and had as good gain.

Bobtails were also constructed for 7MHz but were originally discarded because the SWR bandwidth was too narrow and this caused problems with the linear. However it was found that although they were bi-directional they had comparable performance to the loops. A pair were constructed at right angles as they took up less space than the delta loop arrays.

Last year Frank installed an 8 element 7MHz wire yagi which had about the same performance as the simple vertical in terms of gain. Needless to say the yagi was scrapped and replaced with 4 element which when adjusted correctly delivered one S unit over the vertical.

80m

The best antenna Frank used for 80m was the sloping W8JK in a vee configuration and hung from the top of the extension to the tower. The work book comments in August 1980, This antenna sloping NE worked very well, ‘SWR excellent. In BERU several ZLs answered CQs….. The phasing line is not at all critical.

Running this design through ELNEC shows that the main lobe is broadside with about 3db gain and a F/B of about 10-15db. The 90’ metal tower distorts the pattern somewhat and the lower the tower the less effective the antenna becomes. But for a simple antenna for 80m it would be difficult to beat. Frank has a pair of these still in use.The W8JK was also far better than a 3.5MHz vertical.

The other antenna Frank has used on 80m was a wire log periodic. Although this had a good match at 3.8MHz it did not appear to be particularly broad band. In addition it took up too much space and was therefore discarded as it performance was not superior.

In 1988 Frank installed a bobtail curtain for 80m. This antenna required 64’ vertical sections which due to lack of height had to be folded back on themselves. The curtain was easy to tune up at its design frequency of 3.5MHz but not at3.8MHz. this showed that different lengths were needed.

160m

Initially Frank copied the 80m sloping W8JK for 160 but was disappointed by the results. The ELNEC modeling shows why this might be. As the tower height becomes lower in terms of wavelengths the effectiveness of the antenna declines, it becomes less directive and more coupled into the ground. Frank constructed a Marconi L using the tower as the vertical leg, about 90’tall. However the tower sections generated electrical noise due to poor contacts. A wire was therefore used running from the top of the tower to the base, this cured the problem. Although this arrangement was used for many years it has been modified to a half delta. The overall length is 290’ of which 90’ is vertical, 120’ slopes from the tower top to the matching unit which is 80’ from the tower base and connected to it by a buried wire.

Although this was much more noisy on reception than the sloping W8JK, on transmit it was much better and the states were raised with ease.

A few tips

1. Resonance. Frank has always used close-to-resonant radiators. There is nothing magic about resonance from the aspect of matching. But in so far as current and voltage differences on antennas affect radiation then there will always be a optimum length and diameter for elements in multi-element arrays. In the days when computer modeling of antennas was not possible it was much more accepted to trim antennas to give the best performance and even today antenna models are still only an approximation due to the effect of the antennas environment.

2. Matching. After his early work with open wire feeders and the difficulties of controlling mismatch, Frank has always used matching circuits, such as the gamma match, between the line and the antenna feed point. Flat lines give lower loss but more importantly they show up antenna faults due to increased SWR.

3. Measurements. The use of a remote antenna for comparing radiated signals is very useful. But there is very little substitute for DX comparisons. There are a number of ways this can be done. One of the best is to get comparisons from a friend, Frank uses K4MOG who knows Frank’s signal very well and thus can give objective signal reports which he sends as tape recordings for Frank to hear the difference himself. The other way is to judge performance in contests, DX pile ups or just responses to CQ calls. All of these methods depend to a large extent on conditions but it is surprising how much can be learnt about an antennas effectiveness from such simple methods.

4. Interactions. Few of us can space antennas sufficiently to avoid interaction. Even with the area available to Frank considerable trimming was needed to get the best performance. This is one of the advantages of wire loops over wire yagis. The loop dimensions can be adjusted easily from the ground if the loops are flat top up, apex down.

5. Directional antennas don’t always have gain. It is important to realise that just because an antenna is directional does not mean that it has gain. For example Frank’s notes reveal many antennas which were directional but showed little gain in the desired direction or DX location over a simple vertical, loop or dipole. A gain antenna will show the expected performance advantage over the single element types when it is correctly adjusted and matched.

6. Tips. Attaching guys and pulleys to masts should be done rigidly using an angle bracket and U bolt rather than wire or rope. Twisting halyards can be reduced by fixing the pulley rigidly to such a bracket.

7. Instruments. The MFJ Antenna Analyser is one of Frank’s indispensable instruments. Without it, antenna tuning would be a much slower process.

Summary.

At 85 Frank is still experimenting, he does all his antenna work without assistance. He has a windlass, liberated from a sailing ship (?), with which he can raise and lower each of his masts by himself. And as I and many others can testify he puts out an excellent signal into far off places from a very modern station. Frank makes frequent assessments of his amateur radio system and is not afraid to change when necessary, changing from a lifetime with quads to the log periodic when it became clear that the tower wasn’t robust enough. Frank is interested in all sorts of new ideas, he is into E Mail, packet DXcluster, computers and is clearly fascinated by the relentless march of new technology and its use in amateur radio.

References

1. ARRL Antenna Anthology (1978), page 84 and page 86. Available from RSGB.