Wednesday, August 28, 2013

More on Dipole versus Delta Loop

I lied. I'm not done yet with that 20 meters delta loop versus dipole comparison. The reason is that I was not satisfied that I'd properly understood (or explained) the relatively poor DX performance of the delta loop. I'll attempt to do so in this article.

There are a few key factors to understanding DX performance, which I'll address below:
  • Radiation angle (elevation) for the selected propagation path
  • Local terrain
  • Antenna pattern
It is not easy for the individual amateur to determine the radiation angle of a signal to (or from) a particular DX station or region. Thankfully others have done the work for us. While we cannot know the precise figure in every case, there has been enough experimentation to give us a range of angles within which, with high probability, the actual value falls. This is research I've read in the distant past, but since it was so long ago it made sense to do an internet search to find the source material and refresh my memory.

Since the comparison was done for 20 meters, I'll stick with that. Radiation angles depend on many factors, including: frequency; path length; geomagnetic compass direction; season; geomagnetic indices; and more. That's a lot to consider. However, if we stick to paths that do not intersect the polar regions there is ample experimental data that is pretty reliable.

The ARRL summarized some of this data (PDF) for us. Since Ottawa is not too far north of where the measurements were taken this data should be applicable. The ranges of interest for 20 meters are as follows:
  • For short DX paths (Europe) the radiation angles almost all fall between 3° and 20°
  • For long DX paths (Australia and Japan) the radiation angles are always less than 11°
This was the reason why I was so interested in pursuing an antenna that would perform well at the lowest possible radiation angles. This is difficult to achieve for horizontally-polarized antennas up no more than 10 to 15 meters, which is the upper limit for me in my present circumstance.

As already discussed in some detail, the results of the vertically-polarized delta loop did not meet my expectations. Far-field radiation plots are informative but perhaps not ideal for A/B comparisons, unless one is careful. As an alternative I have produced a table of the gains of the two antennas from the EZNEC models. Here we can easily compare their performance versus elevation angle.

20 Meters – Ground(0.002 S, 13 dc) – Gain (dbi) vs. Elevation angle
Angle
Dipole up 10.5 meters
Delta Loop, vertical, up 7-13.3 meters
Broadside
Off Ends
Broadside
Off Ends
5°
-3.2
-17.5
-1.7
-3.7
10°
2.2
-13.6
1.7
-0.4
15°
4.9
-10.9
2.4
0.3
20°
6.4
-7.9
1.8
-0.4
25°
7.1
-5.1
0.5
-1.8
30°
7.2
-2.9
-1.3
-3.4
35°
6.8
-1.4
-2.5
-4.2
40°
6.1
-0.5
-2.7
-3.8

Note that this was modeled using poor ground, which is perhaps typical of the suburban area in which I live. The terrain is mostly flat out to several kilometers in all directions, except for the southeast to southwest quadrant where the land rises to ~20 meters at 0.4 kilometers distance.

For the dipole the broadside compass directions are 70°/250°, and for the delta loop are 40°/220°.

Notice that at prime DX radiation angles, in the broadside direction, the dipole is moderately superior to the delta loop, except at the very lowest angles. Off the ends the delta loop is clearly superior.

This is not what I found, where the dipole outperformed the delta loop in almost all cases, sometimes by a lot. The models are either wrong or something else is going on. My guess is that the ground may be worse that what I modeled, and that metal in the surrounding houses and utilities had more of an effect on the delta loop, the bottom of which is lower and which is also closer to houses and overhead utility lines.

Rather than redo the model with even poor ground I chose to instead try it with good ground, such as one would find in wide-open farm land.

20 Meters – Ground(0.03S, 20 dc) – Gain (dbi) vs. Elevation angle
Angle
Dipole up 10.5 meters
Delta Loop, vertical, up 7-13.3 meters
Broadside
Off Ends
Broadside
Off Ends
5°
-2.9
-18.0
-0.8
-3.0
10°
2.7
-13.0
1.6
-0.7
15°
5.5
-8.9
1.4
-1.1
20°
7.0
-5.5
-0.3
-3.0
25°
7.8
-2.9
-3.0
-6.0
30°
7.9
-1.1
-5.7
-7.7
35°
7.5
0.0
-5.2
-5.7
40°
6.7
0.6
-3.0
-3.2

As expected the dipole does not fare much differently -- it does slightly better. Surprisingly the delta loop is not only not better, it actually does worse at DX-favouring angles above 5°. One reason is that there are 2 vertical lobes in the radiation pattern, and the null between them gets deeper with improved ground (see this previous article for the pattern). The null is high, at about 35° elevation, but still impacts gain at somewhat lower elevations. The dipole has just one vertical lobe, peaking at about 30°.

This has me wondering whether the ground quality is actually better than expected, since the measured
DX results come closer to what this table indicates. Well, in truth I don't know and I probably never will. While it is possible to test the ground under the antenna (and even enhance it with a metal screen, if you're ambitious), the distant ground, which is what contributes to the far-field radiation pattern, is less known. It also cannot be changed. I suppose one could move and buy an house outside the city, but that isn't in my future plans.

I'm glad I did these table comparisons since they do give me a better sense of what I measured on the radio. It also shows how fraught with uncertainties and problems vertically-polarized antennas can be.

This all leads me to my next experiment, which will be a mostly-horizontal antenna. The factor I am now focussing on is height. I intend to find out what a few meters can do for low-angle performance. Getting there is taking me a few days longer than expected because getting that height requires some construction. I have to get it installed by Labour Day due to exterior work to be done on my house. Site B, the house-bracketed mast, will be inaccessible for 2 weeks.

No comments:

Post a Comment

All comments are moderated, and should appear within one day of submission.