A Novel Approach to HF Aerial Design
There is a vast library of literature available to the designer who wants to put a yagi on a tall tower. This library covers the siting, terrain and propagation as well as the design of the aerial itself. For most of us, while this is interesting literature, it has no real relevance to putting up a simple wire aerial in the back garden. Until a few years ago, backyard efforts were largely the results of trial and error. This perhaps accounts for the many weird and wonderful aerial designs that no-doubt met the needs of their own designers but had properties more akin to magic than grounded in fact. In recent years things have improved for the backyard designer. Trying to compare a low quad loop put up this week against a dipole that was up last week was never going to give good results but these aerials can now be readily compared using cheaply available, hugely powerful software such as EZNEC.
But there is still a missing link. Aerial design software will tell you a lot about gains and losses but little about the likely performance of an aerial when trying to work DX from the backyard. In thinking about this dilemma, I stumbled upon an interesting source of information. As part of the ARRL Antenna Book, the ARRL provides tables for all the HF bands showing the probability of propagation at wave angles from 0-35 degrees over various DX paths. These paths include the UK to USA, JA, Southern Asia, Oceania, South America and Southern Africa.
Wave Angles a brief explanation
For any sort of long distance communications in the HF frequency range, the ionosphere must be exploited to refract signals back towards the earth. Without the ionosphere, the signals would simply pass into space, making long range communication impossible. Note that the ionosphere refracts signals, it does not reflect them although the end result is rather similar. Once a signal has returned to earth, it is reflected from the earths surface and back up into the ionosphere where the refraction process continues as before. In this manner, signals can be propagated very long distances. The diagram below shows how this happens. To get long distances, signals generally need to be radiated at low wave angles, this ensures that the signal spends long enough in the ionosphere to be refracted back to the earth. It also reduces the losses inherent in the refraction and reflection processes. The wave angle is the angle between a line that is tangential to the earths surface and the path of the ray heading up into (of down from) the ionosphere.
On a long distance path, the ray may take several routes. These may occur at the same time. Typically a path may be spanned with two hops and/or three hops. When both modes occur at the same time, the signals propagated by each path can interfere with each other, causing fading a common phenomena on HF. This is not the only cause of fading at HF, there are many more!
Some software can predict wave angles. These predictions must be viewed with some caution as the models used, although very complex, are nowhere near as complex as the ionosphere. A typical wave angle prediction from ICEPAC is shown below.
When we talk about the gain of an aerial, by convention we choose the point on the aerial pattern that has the highest gain. That is entirely logical for any case where it is possible to move the aerial to ensure that this point or points is directed to where one wants to send a signal. Unfortunately, most backyard aerials are not rotatable. Many back yard designers will have spent much time squeezing the last tenth of a decibel of gain out of wire aerials designed on the computer, but the question that is generally difficult to answer thus far is whether the radiation gained is actually useful? As a result, I have developed a new approach to measuring the gain of a fixed backyard aerial not with reference to its peak gain but rather by reference to propagating signals to real-world DX locations.
In outline, the method is to plot the polar diagram of a test aerial and calculate a new gain figure based on the probabilities of DX propagation at each wave angle at each bearing over each frequency. This would be impossible by hand as it involves manipulating thousands of pieces of information to do the calculation. I realised this at the outset (or at least quite soon afterwards) and wrote some software to do the tedious calculations. For each of the target DX areas, for which the ARRL provide wave angle information, I have defined a short path azimuth range using a great circle map. These values used in the software are listed below. They are specific to the UK and my own interests.
The wave angle information is then used to provide a weighted gain for each of these paths across these azimuth angles for an aerial orientated just as it would be in real life. The weighting is based on the percentage probabilities provided in the ARRL Tables. An example of the wave angle tables is shown below for the path from England to Japan. The figures in the band columns are the percentage probability for propagation at each wave angle.
It is interesting to note the bi-modality of some paths such as that on 15 metres to JA which probably arises due to the possibility of F2 propagation with three or four hops with peaks at 4 degrees and 8 degrees. This is illustrated on the diagram below.
You can see why this happens in the diagram below. This shows a path that can be spanned with two or three hops and so would show bimodality.
The software shows the gain of a real aerial over each path and also calculates an average "DX Gain" for the aerial over all of the specified paths. The software uses an Excel Spreadsheet and has the UK wave angle information built in for each band. It is available with some brief instructions on how to use it. You will also need to have a copy of EZNEC to generate the aerial patterns.
The results from the software make interesting reading. DX Gains of backyard aerials are well below the "normal" gain specified for the aerial types. Depending upon the orientations of the aerial, some aerials are indicated as dramatically better than others over one path and worse over others. I have been using the software to design an aerial for 30 metres. Below is a comparison between a dipole at 28 feet, a quarter-wave vertical and a low fullwave "quad" loop. The first comparison is the conventional way of doing things with the polar diagrams overlaid upon each other. At lower angles, the differences between the aerials do not seem very great. At angles higher than about 15 degrees, the dipole has significantly more gain (in this plane). The second comparison shows how these same three aerials look when run through the software. Note that these are results specific to my own QTH and are only shown to illustrate what can be discerned. Different results would be obtained with different aerial orientations and heights.
When the polar diagrams are processed through my software, the differences between the aerials are far more obvious and would have been impossible to judge by eye from the polar diagrams alone. It is very apparent that a dipole at 28 feet would be a poor choice for a DX aerial. A fullwave loop would be best for South Asia, USA and South America but slightly worse than a quarter-wave vertical for Africa, Japan and Oceania. The DX Gain of these last two aerials is the same so either would be a valid choice for an all-round DX performer.
Thus far, the discussion has been of backyard aerials, but this approach is equally valid for larger fixed aerials. For example, it would be able to answer questions like would a delta-loop at 80 feet for 80 metres be a better DX aerial than a high dipole at a particular location. Again, typical results are shown below. The delta loop is clearly a better choice in my location. Oceania is omitted from the comparison as there is no wave angle data for this path on 80 metres.
This approach to aerial evaluation gives the designer another tool with which to assess comparative performance. It does not give all of the answers but does certainly show up some intriguing differences between aerials in the real world. It must be borne in mind that the results rely on two models, one for the aerials and another for the propagation. Each model will have inaccuracies so there may be some welcome surprises in spite of the predictions. Inevitably, the backyard designer will always have to make some compromises, but with another source of information such as this perhaps better results could be achieved from the same meagre resources.