Faraday Rotation and Comparing Inverted Vees

In my previous article I mentioned that I barely got the second 40 meter inverted vee up installed in time for the CQ WW contest this past weekend. Since inverted vees are not particularly interesting antennas to discuss I will only briefly describe the process.

First I modelled the vee with EZNEC to ensure I cut the legs to the right length. The resonant frequency of an inverted vee is very sensitive to the angle between the legs when that angle drops below 90°. In my case that angle had to be 70° to be compatible with the available tie points for the chosen orientation: fence post and a tree limb. For 12 AWG insulated wire the modelled length was 10.35 meters per leg.

I could have opened up the vee to 90° but at the cost of a substantially more complex installation. A pole of 5 or 6 meters height attached to the fence post would be needed, and climbing up into the tree while threading the wire through the branches. The small benefit in radiation angle and feed point impedance wasn't worth the effort. Nor did I have the time if the antenna was to be ready for the contest.

I reused the 40 meter loaded sloper antenna -- a failed experiment from last year -- since it had most of what I needed: centre insulator with SO-239 and most of the needed wire already attached. Splicing brought the length to 10.5 meters per leg, adding 15 cm to allow trimming to resonance. It was quick work to run up the tower to install and connect the feed point, run back down and tie down the legs, then off to the shack to check it out. Trimming 10 cm per leg brought it to resonance approximately where I wanted it.

The SWR dipped almost exactly to 1 even though the model showed an impedance of 35 Ω. This may be due to ground and environmental loss or perhaps interaction with other antennas. When perfectly symmetrical the non-resonant support tower should "disappear" since the vee's legs induce equal currents of opposite phase, thus cancelling. Perfection is often unattainable so there may be a small net current.

Another thing worth mentioning is that I designed this antenna to withstand more tension than is typical. Slack reduces the interior angle of the vee, degrading its performance. I didn't want the angle to be any less than the already small 70°.

Modelling, testing and directivity

I describe my inverted vees as oriented east-west and north-south even though that is not quite the case. Broadside for my existing mult-band inverted vee is roughly 100°/280° true bearing, and 30°/210° for the new vee. Although not ideal I am constrained by the availability of supports and lot width.

The pattern shown is for the new vee. The one for the existing, multi-band vee can be seen in a previous article. Two things to notice are that the multi-band vee shows more directivity than the new one and that, due to asymmetry, the former has a deep notch to the south.

The interior angle affects the directivity. The smaller the angle the more omni-directional it becomes. The multi-band inverted vee has an interior angle of about 120°, which is approaching that of a dipole, and therefore has a pattern not too different from a dipole.

The difference in directivity is small in most cases. The new vee does much better to the south, making it easier to work W4, the Caribbean and South America. The multi-band vee performs better for W6 and the Pacific. Towards Europe the new vee does better. From Ottawa, the bearing to central Europe from is 50°.

This came in very handy during the CQ WW contest. The difference to zone 8 (Caribbean) was as much as 2 S-units on the KX3. The differences to Europe and the Pacific were more subtle, never exceeding 1 S-unit. For many hams this might seem inconsequential, yet it makes a noticable difference. In marginal cases switching from one antenna to the other with my QRP SSB signal allowed a QSO to be made compared to getting no response to my calling them. With QRP every decibel counts. That's no exaggeration.

The inverted vee works well on 15 meters. Since it is broadside to the Caribbean it came in handy to work several multipliers and US stations during the day when the yagi was aimed at Europe. The other inverted vee isn't as favourably oriented and so was less useful for the diversity I require during contests.

Faraday rotation

Now I want to back up the Thursday evening before the contest when I was comparing antennas on 40. I was briefly baffled by the relative performance of the antennas which made for confusing comparisons. Eventually I figured out what was going on. That it took me 30 minutes to work it out is a little embarrassing since it should have been immediately obvious. But first some background.

Faraday rotation in the ionosphere is most often responsible for the cycle of QSB on HF signals. HF antennas are almost always linearly polarized, whether vertical, horizontal or, less often, an angle in between. As Faraday rotation alters the polarization of the incoming signal the induced power on our antennas changes. Signal strength is maximum when the signal and antenna polarities align and can dip 20 db or more when they are opposite (90° apart).

Many hams use both vertically and horizontally polarized antennas on the low HF bands to pull through weak signals (like mine!). To do this you switch between antennas during polarization dips, use a signal combiner or simultaneously listen to two receivers.

The phenomenon I encountered was that on many stations the dip due to Faraday rotation on one antenna seemed to exactly coincide with the peak on the other, and vice versa. At first I put it down to bad luck and I kept trying.

I soon realized this behaviour was no accident. Have another look at the azimuth plot up above. There are 3 field strengths plotted: horizontal, vertical and total. Notice that the orientation of the horizontal and vertical polarization pattern lobes and nulls are at right angles to each other. This can be explained by imagining yourself looking at the antenna from a distance.

When you are broadside to the antenna it appears to have a horizontal orientation. This is the direction in which the horizontally polarized field is maximum. Looking at the antenna from the side and its horizontal width is zero. It now appears to be a vertical, with one leg behind the other. This is the direction in which the vertically polarized field is maximum. The opposite is true for the nulls.

Here then is the answer to the puzzle. When the incoming signal has horizontal polarization its signal strength is maximum on the inverted vee that is broadside to the signal direction. The other inverted vee sees minimum signal strength since it favours vertically polarized signals in that same direction. The situation reverses when the signal is vertically polarized. When the polarization is neither vertical nor horizontal both antennas show some attenuation, and it can be difficult to say which receives best. The ambiguity also occurs for signals that are neither broadside nor off the end.

Doing a proper comparison

A proper comparison of antennas therefore requires more effort than instantaneous A-B switching. On 40 meters it can take 5 or 10 seconds of listening on one antenna to find the maximum signal strength. This must then be repeated on the other antenna. Or two receivers can be used. Comparisons are more difficult and can be muddied by other propagation effects that also affect short-term signal strength variation.

Despite this it is recommended that the inverted vee broadside to the desired direction be selected, and left there. Even though the new inverted vee with its acute interior angle is close to omnidirectional, in practice this rule still seems to apply. For example, toward KH6 (west) the east-west favouring vee has better than a one S-unit advantage during our sunrise opening. Part of the reason may be that the vee with an acute angle has poorer low elevation angle gain due to the lower height of the average current.

I'll be keeping this antenna. It works and it is not an encumbrance or eyesore.

Interactions

As already mentioned, the orientation of the new inverted vee keeps interactions quite low. It is still likely that some coupling is affecting the antenna, though with a magnitude that is not material in any way I can see. The impedance is not quite as predicted by the model, yet any affect on its resonant frequency is small in comparison to the wire length itself.

The acute interior angle also ensures minimal coupling with the yagi. In general I've found in models and in practice that wires that have an angle with the tower under 45° are free of interactions with yagis up above.

Where I did encounter a problem was with the 80 meter vertical, which is comprised of the tower, radials and yagi (capacity hat). The effect is not large and did not stop it from being used in the contest. The coupling, from the model, appears to be non-resonant, perhaps acting more as a coupled resonator due to their proximity and acute angle of 35°. I will address this in a future article, along with other items pertaining to the vertical.