- Linear-loaded vertical
- Lazy-H vertical
- Linear-loaded quad and delta loops
Although we have made the model simple with respect to the environment the reality is never like that. It is nevertheless critical and a common enough problem in many vertical installations. I therefore will use as one figure of merit the height of the average antenna current: higher is better.
Just to finish off these (important) preliminaries I want to summarize what the antennas do have in common:
- Omnidirectional - Despite asymmetries in all the antennas, they are all effectively omnidirectional. The worst case is a maximum-to-minimum gain differential of 2 to 3 db for the vertically-polarized loops.
- Match sensitivity - Matching and resonant frequency are very sensitive to even small adjustments to antenna height. Environment is also likely to be a significant factor. In all cases it is therefore desirable to make the antennas easy to adjust in place. Some suggestions are in the articles. The loops are hardest to adjust.
Now then, after all of that lead up over two weeks you may be expecting to hear me say something profound about which antenna is best, worst, etc. This is difficult, and did not turn out as I expected. This was a learning experience for me, and perhaps for you as well.
If you were following along you might have noticed one glaring fact: the patterns of all these antennas are almost identical! Sure, we got good matches to 50Ω coax, but that was easy and not especially important. What I want is to find the best possible DX performer for 40 meters. Yet in every case the low-angle gain ends up somewhere in the vicinity of 0.25 dbi at 15° elevation.
The performance difference, such as it is, can only be found in modest differences in the height of the average current. Since all antennas are mounted at the same low height this mostly comes about by how high the antenna reaches. For the two "true" verticals the average is at the centre point, which are 9.5 and 8 meters for the linear-loading vertical and Lazy-H, respectively, since each is mounted 3 meters up. Due to its squat construction the quad loop has its current centre up only 6.3 meters. The more complex delta loop has its average somewhere around 10.4 meters up, which is not too different from the verticals.
On the basis of this and its complex construction I won't hesitate to eliminate the quad from consideration. Of course it can be placed higher above ground without exceeding the top height of the other antennas but it then becomes more unwieldy and fragile.
All this leaves us to ask why these antennas perform so similarly? It turns out the reason is simple enough: the low angle performance is determined by ground, not the antenna! If you want more gain it is necessary to add more elements to create a vertical array. That is not what I want to do, though many hams have followed this route to great success.
What if we raise the vertical higher off the ground? As already mentioned the match will change, but here we will ignore that and focus on the pattern. It turns out you need to add quite a lot of height to get a significant change in the pattern. For illustrative purposes here is the pattern of the Lazy-H raised to 20 meters above ground; its top will be at 33 meters.
The pattern plot is educational. Total losses are cut to -3.6 db, or about 2 db better than when mounted at 3 meters. Unfortunately this helps little since that additional gain goes into higher-angle radiation. That's great is you want to operate a contest like the ARRL Sweepstakes, but helps little with DX. If that's your objective, put up a dipole.
At our target elevation of 15° the gain is actually slightly less than it was before. The lesson here is that you cannot outsmart Mother Earth; if you want better low-angle performance you'll have to move somewhere with better ground. All you've really gained is a clear shot to the horizon, which is helpful but unlikely to justify the construction of a massive tower. If you do have such a tower it will be far better to choose a dipole or yagi.
The antenna would also have to be lengthened since the change in height shifts the resonance upward to 7.175 MHz. We would also have to move the feed point to near the bottom of the vertical element to regain a 50Ω impedance. However these are merely details.
The big factor is ground loss. This is worth attention and so I took the trouble to model the Lazy-H (at its original 3 meters height) over various grounds. While not modelled, the other verticals in this comparison should be similar.
If you use EZNEC I made the results easily reproducible by using the standard set of ground options that come with the application. In the following table the first 2 columns are the ground conductance and dielectric constants, ranging from extremely poor ground at the top to salt water at the bottom. The final 2 columns are the net gain at 15° and total ground loss. The choice of 15° is done to ease comparisons at a typical DX take-off angle, just keep in mind the elevation of maximum gain ranges from about 20° down to 5°.
|Gain at 15°||Total Loss|
|0.001||3||-2.9 dbi||-7.4 db|
|0.001||5||-2.1 dbi||-6.7 db|
|0.002||10||-1.1 dbi||-6.0 db|
|0.002||13||-0.6 dbi||-5.6 db|
|0.005||13||-0.6 dbi||-5.9 db|
|0.006||13||-0.6 dbi||-5.8 db|
|0.0075||12||-0.5 dbi||-5.8 db|
|0.01||14||0 dbi||-5.4 db|
|0.03||20||1.7 dbi||-4.0 db|
|0.001||80||2.3 dbi||-3.3 db|
|5||81||5.0 dbi||-0.7 db|
The differences in gain and loss are substantial across the full range of ground types. However unless you are one of the rare hams sited on a seashore that final row is pretty useless. In my case the second-last row (fresh water) isn't far wrong for a range of useful directions since I am very close to Lac Déschènes, which is a widening of the Ottawa River. I took the following photograph from the peak of my roof with the camera about 9 meters above ground and looking towards Europe (north-east). As you can see the fresh water is a little bit "hard" this time of year.
Apart from the water scenarios the range of gains and losses are quite modest at under 1 S-unit. I would therefore not be inclined to worry too much about ground quality. I do believe it is important to get above obstructions, many examples of which can be seen in the above photograph. Current averages much below the height of the photograph can result in attenuation, perhaps a lot, although I have no easy way to prove it.
Comparing all these short verticals to a full size, vertically-polarized delta loop at the same 3 meters height shows a modest advantage. At 15° there is about 1 db advantage favouring the delta loop, with a comparable average current height of ~10 meters. The gain is -3 db off the ends when compared to the broadside directions.
The final comparison I want to make is to an inverted-vee. This is the only horizontal antenna I will use in this analysis since it has comparable construction requirements. Like the other antennas it is a single, half-wave wire element, and it will be mounted using a similarly sized mast. The legs will be set 90° apart and the apex at 13 meters.
I won't reproduce the patterns of the inverted-vee here since I've shown almost the very same plots previously. In the broadside direction it shows a gain of -0.7 dbi at an elevation of 15°. Despite its low ground loss almost all the radiation goes out at high angles due to its low height, which is not only useless for DX it will increase QRM from domestic stations. I'd rather hear the DX, not the pile-up.
Where does this leave us? The two short verticals options look interesting but not too much. For myself I still favour the full-sized delta loop. This exercise has set my mind at ease about previously-unexplored alternatives.