An antenna farm is replete with interactions. Most are unwanted and detrimental. It isn't enough to deal with harmonically related bands since non-resonant antennas, guys, coax, control cables, utility lines and building metals will affect antennas when they are an appreciable fraction of a wavelength or longer, or if they are short and close.
Interactions can be eliminated if the offending conductors are within your control. Unfortunately the effort and expense may not always be worth it. Many hams are unaware of interaction problems or, unless it is especially serious, live with it. A few choose to pretend they doesn't exist.
For those of us invested in high performance antennas those interactions can remove a lot of that performance and therefore require attention. It may not be possible to eliminate every problem but that is not an excuse to do nothing. The answer is to focus our efforts on the worst of the lot and, perhaps, deal with the rest later.
This article is about vertical antennas for 80 and 160 meters since the best for DX are verticals and towers are also verticals. That is, towers are inevitably a part of the antenna system unless they are exceptionally distant. Recall that "distant" is with respect to wavelength and a wavelength on these bands is larger than the vast majority of hams' properties.
There are several common methods for dealing with tower interactions:
- Locate the vertical antennas as far as possible from towers.
- Place verticals arrays and towers so that no tower is in the direction a directional vertical array is pointing.
- Have more than one vertical antenna for the low bands so that the directional deficit of one is covered by another.
- Detune towers so that they don't resonate on 80 or 160 meters.
Methods can be combined when helpful. For my antennas I placed the 80 meter array far but not too far from the towers. I wanted to keep the array close for unrelated reasons and the placement only disfavours one direction (southeast) which is not the most productive for DX paths in contests. Other directions are unaffected.
The situation on 160 meters is more problematic. The towers are supports for 160 meter antennas for the foreseeable future so the interactions must be dealt with in a manner that is most advantageous or, failing that, least deleterious. That can be difficult to achieve due to their large electrical length. As we'll see detuning is not always necessary.
You will not find many satisfying conclusions in this article. It's all part of my learning process. Expect food for thought rather than clearly defined antenna ideas.
Site plan
The 80 meter 3-element vertical yagi is in the north field and the two big towers are in the field to the east and south. The northernmost is 150' and the other is 140'. The heights are nominal due to section overlap and masts. Actual heights to the top of the mast are approximately 46 meters and 43.5 meters, respectively. The Trylon tower near the house is 24 meters to the top of the mast.
The 140' tower is marked with a red dot since the Google Maps image predates it. The distance from the 80 meter central tower and driven element to the 150' tower and Trylon is ~60 meters, as is the distant between the big towers.
Model notes
Since the Beverage antennas to the northeast of the image appear to suffer no pattern distortion due to these towers and antennas they will be omitted from the models. That is not always true, especially with vertical receive arrays, and those interactions must also be dealt with.
For the 80 meter case the 140' tower and Trylon will also be omitted, the first due to the distance and the second due its not being directly in one the array's four directions. That is not to say they have no effect but the effect is small enough that I am not unduly concerned. For the 160 meter case the Trylon and 80 meter array will be ignored due to their short heights.
The big towers will be modelled as 40 meter high thick wires with 2 capacitance hat arms at the top representing the top yagis. This is not exact and is in fact not really even very close to reality. The side mount yagis and the mess of control wires and coax are difficult to model and so they aren't.
To see what the impedance might really be I plotted the R and X of both towers with an antenna analyzer, placing it in line with the wire to the ground rod. The galvanized ground rods are ⅝" × 10' and are located just beyond the subsurface reinforced concrete platform. The tower and pier pin are not bonded to the the re-rod so the bases do not behave as Ufer grounds.
As you can see it's a bit of a puzzle. The only recognizable resonance is around 1.2 MHz, which is not out of line with expectations. Overall the impedance is dominated by the ground resistance. All I can say is that the ground resistance (loss) at these frequencies is probably in the range 25 Ω to 40 Ω. The other tower has a similar profile.
At 1.83 MHz the impedance is 79 + j20 Ω, which is equivalent to a reflector with a lot of ground loss. This is not too unlike the model of an electrically long wire on 160 meter, and is an approximation I've used before. However the impedance is about the same at 3.5 MHz. Who knows, had I inserted the analyzer on a feed line or control cable the result could be very different.
Modelling the tower as a simple thick conductor with yagis as capacitance hats and a resistance load (ground loss) at the bottom has no resemblance to the real world measurement. There is no simple model that mimics the measurement and I decided it would be foolhardy to make further attempts.
As a consequence, what a distant antenna "sees" is exceedingly difficult to know. This complexity limits what I can say about interactions, more so on 80 than on 160 meters. At least on 160 meters the measured impedance is likely closer to what a 160 meter vertical antenna would see.
MININEC ground is specified in the model with resistance loads at the bases of the towers and vertical antenna elements. These can be set to values approximating the ground rods and radial systems. We don't need to be very accurate for this overly rough analysis.
From the foregoing model discussion it should be evident that modelling the interaction is quite a challenge! So I won't try. What I did was try various electrical lengths of the 150' tower that is 60 meters distance.
Regardless of tower tuning the SWR of the 80 meter array does not measurably change. This is not surprising. Feed point impedance is the last thing to be affected by interactions. In order, the effects are felt in directivity and gain. It takes a far tighter interaction to alter the SWR.
Depending on tower tuning, with the 80 meter array pointed at the tower (southeast) gain varies by ±1 db and F/B is reduced by 3 to 10 db. Yes, the tower can act as a wide spaced director and increase gain. From my measurements, however, that is unlikely. From on air use I know there is gain southeast and a small degradation in directivity. That is, there is interaction but nothing too concerning.
When the tower is placed to the side of the 80 meter array, as it would be for northeast and southwest directions, the influence of the tower is small. Recall from the pattern charts for the 80 meter array the F/S is less than 10 db so there is just enough radiation towards the tower to have some effect. Pointed northwest, where the tower is behind the array, there is no interaction, and that is expected.
160 vertical centred between the towers
For the model I adjusted the towers to assign them the approximate measured impedance. This may not be representative of the situation at other stations. Despite that there are qualitative results that are broadly applicable.
Assume a rope catenary from tower top to tower top supporting a vertical wire. Were I to build it that is how it would be done. This is not arbitrary since I planned to use this method to build a reversible 3-element vertical yagi by tuning the towers (LC networks) to act as directors and reflectors (see below for references). However in this case we take the towers as they are with their measured resonance and ground rods. This is an antenna I've looked at before but without the tower impedance data known.
Bandwidth is not great and neither is the efficiency. Both towers together have about the same ground loss as the antenna itself. To improve efficiency you'd have to put radials on the towers since adding radials to the vertical won't fix the tower loss. The pattern is a problem if you have no other 160 meter antenna to cover the broadside directions.
This is not a very good antenna. It would help to detune the towers, not only to reduce tower currents and loss, but also to circularize the azimuth pattern.
You may not know you have a problem unless you measure the impedance and resonance of nearby tall towers. Ignorance is not bliss.
This antenna requires just one tower for the upper support. A rope from the bend is anchored on the ground to give the antenna its shape. The configuration of the towers is identical, electrically and physically, to what is described above. The vertical is 20 meters from one tower and 40 meters from the other.
The antenna is a little like an inverted-L but with efficiency and bandwidth closer to that of a straight vertical. The angle of the bend is 45°. The radiation resistance is higher than the T wire vertical I've used for the past few years.
As you can gather from the EZNEC antenna view, including currents, the left tower's induced current is lower than for the centred vertical wire. Loss in that tower's ground is therefore lower. Current and loss in the right tower is about the same as for the centred vertical wire. Overall loss is lower so the antenna is more efficient. Rather than resulting in higher gain the power appears in the broadside directions so that the pattern is more omni-directional.
With the second tower on the scene the deficit in that direction is reduced (on the plot left is northeast). Compared to the centred wire vertical the broadside deficit is reduced and is acceptable. The modelled 2:1 SWR bandwidth is 80 kHz, which while not great is good enough for my operating interests.
Here I have to use the term "good enough" since this is almost certainly the antenna configuration I will use this winter. It's simple and effective despite its imperfections.
Actually the wire will be offset a few meters to the east so that two of the radials don't run into the stone wall separating the yard and the hay field (see satellite view above). That will skew the pattern so that the east (up) and west (down) directions will differ by about 2 db.
Efficiency of the antenna can be improved with a few radials on just the right tower. That is a feature I will consider. The antenna is simple enough that I could put up another to better cover the east and west directions. That is not a feature I am considering for this winter. I have the 160 meter mode on the 80 meter array available to fill pattern holes since it is more omni-directional although it may be as much as -6 db worse than the full size wire vertical.
160 vertical alongside the tower
A quite common style of wire vertical is to run it alongside a tall tower. It is given its own radial system so that there is no direct connection to the tower and lightning ground rod. Of course the wire couples strongly to the tower and its many cables, and ultimately the ground rod as well. This is not a shunt fed tower. Many report low to no deleterious RFI to antennas and connected equipment while others report severe problems. It seems to depend on details of cabling and yagi feed systems and the resonances therein.
I used the previously measured impedance of the tower at 1.830 MHz to model a grounded thick wire model of the tower and put a vertical AWG 14 wire separated by 2 meters, which seems to be a popular choice among hams who have tried this antenna. Again I use a MININEC ground with resistance loads at the base of the wire and vertical to mimic the loss of radials and ground rods.
The feed point impedance is around 25 Ω and with a 2:1 transformer (another popular choice) has an SWR 2:1 bandwidth of 80 kHz (say, from 1810 to 1890). That's pretty good. The pattern is skewed somewhat and the efficiency is poor. Current in the tower is high (as you can see in the plot at right) so even with a good radial system for the vertical wire about 40% of the power is dissipated in the tower's ground resistance via the ground rod.
Despite the gross simplification of the model this appears to be a risky design. Some of the loss can be mitigated by tying the radial system to the ground rod at the further risk of more serious RFI. It may be little better than a shunt fed tower, if at all. But, again, there is a lot of uncertainty in the model due to the unpredictable behaviour of the many cables on the tower and running toward the shack, whether over or under the ground.
Future 160 meter directional antennas
In an earlier article I explored recruiting the (unavoidable) towers as parasitic elements to make a reversible 3-element vertical yagi. Tuning the elements appropriately will be more difficult than in the simple models (again, all those cables) but it does get useful work out of towers interactions. Radials on the towers reduce the loss due to coupled current on the towers and, to a degree, may reduce the impact of the cables on the tower.
That only provides two directions since, as shown earlier, when the towers are somehow detuned or decoupled they are still present as non-resonant elements that will distort the otherwise omni-directional pattern of the driven element alone. Switchable detuning networks are needed to have an omni-directional mode.
In the future I will explore this antenna further and perhaps attempt an experimental tuning of the towers to see how well they can be adjusted to give the desired effect. I am less hopeful now than I was when I wrote the article that it will work well in practice.
Final notes on mitigation
Towers with resonances on bands with nearby antennas is a problem, even if you aren't fully aware of it in your operating. Some mitigation strategies were discussed but not analyzed in depth. For me these can be summarized as follows:
- Detune towers: This can be more difficult than it sounds. The problem lies with all the various cables alongside the tower. Placing an analyzer at the tower base can be misleading. A field strength measurement may be necessary to confirm that the pattern is circularized.
- Improve tower ground: Towers with substantial induced current and a lightning ground will decrease vertical antenna efficiency. Although it won't stop the interaction, lowering the ground resistance will improve efficiency. Several radials attached to the tower base may be all that's needed. The specifics are not described in this article but were confirmed in the model.
- Recruit the tower: Tall towers often act as reflectors quite naturally. Place your vertical antennas to take advantage of that and thereby improve your signal. Unfortunately you will then need at least one more wire verticals to cover all compass directions. With more effort the tower can be tuned as a director or reflector and that can be a great antenna.
Another technique that can work well in many circumstances is to disconnect the tower from the ground rod and replace it with a high value resistor and spark gap. Think of it as a more radical method of detuning the tower. This technique is typically only used on tower verticals that must be isolated from ground. Its value for a tower supporting multiple yagis and cables is less certain and introduces serious safety concerns. To be complete I want to mention it even though I recommend against it.
This is not the final word. My hope is that this limited analysis of potentially destructive interactions for low band vertical antennas will open eyes. Only then can corrective action be taken. It's food for thought that I will consider for my future 160 meter antenna plans.
160 meter antenna for this winter
As I write this I am not able to operate 160 meters. There is a fault in the 160 meter mode of the 80 meter vertical yagi. I have had little time to work on the problem due to more important tasks. The full size wire vertical can't be deployed until work on the 15 and 20 meter stacks is complete.
I will use a bent vertical described above rather than the sloped T used for the past few years. Interactions with the stacked yagis has not yet been modelled but I expect it to be minimized by locating the wire directly ahead of the lower, fixed yagis.
Next year I may try something different. I'm out of energy and time to be creative with a 160 meter antenna for this winter. Unravelling those tower interactions is an intriguing puzzle that I am sure to return to.
No comments:
Post a Comment
All comments are moderated, and should appear within one day of submission.