In this article I present my preliminary survey into what's available, advantages and disadvantages, and setting of criteria to make a final decision. I have little experience with side mounting so there is the possibility of an error or two creeping in. Hopefully my calculations and research kept that from happening.
The esoteric: Ring rotators and rotatable towers
Neither of these options is in my plans due to expense and disadvantages with respect to my situation. I may have only one tall tower -- though I aim for two, eventually -- which has implications for contest operation.
A rotatable tower is especially expensive and would require substantial custom mechanical work to adapt my tower-in-waiting (LR20) to commercial products. But with one ground-mounted rotator (e.g. large prop pitch motor) you can turn all yagis on the tower. For daily operation this can be a perfect solution, but not in contests.
In a contest you want antenna diversity, whereby antennas can be simultaneously pointed in two or more directions. You can do this in contests if you have multiple towers, but I will be more limited. I plan to build a small contest station, not a super-station.
Ring rotators, such as those by Prosistel and AlfaSpid, allow full, independent rotation of each side-mounted yagi. But, again, they are expensive and require custom fitting to LR20. They have the advantages of full rotation (360° or more) and support large arrays. In most situations one ring rotator is required for each rotatable side-mounted yagi.
Another, less obvious problem is working on a tower with ring rotators. Safely climbing past one of these obstructions can be awkward, and unsafe if not done properly.
In the simplest arrangement two brackets are extended out from a tower vertex as platforms for a rotator and mast bearing. The yagi is mounted to the mast. For a trussed boom or stacked yagis the mast can extend further, with additional bearings as required. The rotator and its bracket typically take the full weight of the mast and antenna. The green circle represents an end view of the yagi boom and its mounting position on the mast.
Although bending moment on the rotator is not large you should consider a bearing directly above the rotator. It depends on the loads due to ice and wind falling within the bending load capacity of the rotator. The plates and tower brackets must withstand the dead and live loads. Here is a commercial example of brackets from IIX Equipment. They can also be home brewed by any ham with suitable metalworking tools.
Some choose a thrust bearing for the bottom plate and place the rotator inside the tower. The rotator turns the mast with a chain drive. Whether this is truly necessary is debatable since the rotator still requires above and below the chain to eliminate lateral loading of the rotator.
So we are done. Or are we? Have we truly achieved our objective so easily? Unfortunately, no.
This simple side mount is too simple. As designed it is not able to achieve 300° of rotation (360° - 60° for the projected angle of a triangular tower). In fact, it doesn't come close. Have a look at the top-down view to see what is going on.
The points where the yagi's boom strikes the tower (red lines) mark the limits of the rotation. This angle is approximately 120°, which is only ⅓ of the compass. Grey marks the azimuth angles the yagi cannot point.
Further, the direction a tower vertex points (there are 3 choices) must be the centre of rotation. The yagi can rotate 60° to either side. If the tower is already standing it may be impossible to select the best 120° range for your specific interests. In my case I want Europe (40° to 60°) within the range. I would then choose either to cover up to East Asia (330°) or Caribbean and South America (150° to 180°). I can't have both. A tower vertex must point approximately 30° or 100°, respectively.
So you would think this method of side mounting is a poor one. Yes and no. For some it may suffice. For others, like me, it is a useful first step. The reason is that it can be supplemented with a swing arm that will give 300° rotation. That's the next subject.
The typical swing arm is a rectangular or triangular attachment to the rotating mast. The yagi attaches to the pipe or plate at the outside edge. The length of the swing arm should just exceed the distance from the mast to the far edge of the tower. LR20, which has 20" faces, requires a swing arm of about 30", assuming a tower vertex to mast distance of 6" to 8".
By swinging the yagi around the tower in this fashion allows the desired 300° of rotation, as shown in the lower diagram.
Most hams build their own swing arms to their own specifications. There are commercial products available, such as that seen at the IIX Engineering link provided earlier. The swing arm must be very rigid and strong in the vertical and horizontal planes to survive high wind and ice loads, and protect the antenna, rotator and tower. The longer the swing arm and the larger the antenna the greater the engineering challenge.
With a swing arm there is some flexibility in choosing the yagi's centre position, something which isn't possible in the previous case. This is helpful if the tower is already built and a vertex is not pointing in a suitable direction. You can see this if you imagine the swing arm at the end of its rotation, up against a tower face. The yagi's boom has up to 60° of additional arc in which it can be oriented. Since this works on both ends of rotation the latitude of boom to swing arm orientation is ±60°. There are no unreachable compass points.
The approach is to first select the vertex that points away from the 60° area that is of lowest utility. In my case that area is between about 90° and 150°, which is mostly the south Atlantic Ocean and long path to the north Pacific Ocean. I therefore want a tower vertex pointing approximately 300°. Second, we must orient the yagi to adjust for the actual orientation of the tower. If the vertex points, say, 275°, the yagi should be turned 25° clockwise on the swing arm mast. That's it.
To close off, I'll point you to a couple of photo galleries of swing arms. First up is K5YA. Notice the rotation offset of the yagis on the swing arms, stacking and eschewing of a bearing to protect the rotator. Many rotators can handle the load, or may require more frequent service. Next up is K7EM. Here we see mast extensions for boom trusses and cohabitation with a yagi on a ring rotator. If you look around on the internet you can find many more examples. You may find a novel approach that strikes your fancy.
Standard controller for rotators typically do not support custom stops. That's a problem with side-mounted yagis. We want the controller to limit rotation before the boom or swing arm strikes the tower. It is preferable to use a controller that permits software-defined stops. Failing this I recommend using a low torque rotator such as a Ham IV (800 inch-pounds) and a spring or similar soft stop on the tower where the boom would otherwise strike.
Also take into consideration the mechanical stops of the selected rotator. Those stops ought to fall within the gray zones shown above, otherwise the rotator will prevent selection of all available direction.
Overall, I consider it almost mandatory to use a software-controlled rotator that can be calibrated to the specific needs of side-mounted yagis, including stops. Mechanical soft stops on the tower can also be used as a form of insurance in case of programming error or the yagi twisting on the mast in a wind storm at night. You can't always be looking out the window to see what's happening.
Symmetry is our friend when it comes to managing interactions. When we disturb symmetry we must carefully analyze the situation to assure no interactions disturb the performance of our antenna. Side mounted yagis, with or without a swing arm, are not symmetric with respect to the tower. Hence the concern.
To begin, in EZNEC I added a long wire running through the 5-element 15 meters yagi I propose to side mount and stack. This wire is 18" in diameter, which is approximately equivalent to a 20" face triangular lattice tower. The objective is to test for interactions between yagi and tower as the yagi is rotated around the tower.
The tower model is truncated and made resonant on 15 meters. There is no easy way to fully model the tower, or to arrive at results that are at once both specific and general. Instead I chose to pursue a worst case result by centering the tower segment with respect to the boom and making it long enough and resonant to excite interaction modes.
To this end I adjusted the tower "wire" to resonate ~40 meters, with the third harmonic at 21.2 MHz. That is, it's a 3λ/2 vertical dipole. The model is placed in free space to isolate the interactions from other environmental factors, including guy wires, ground and other antennas. If I can reduce interactions in this model there should be none of this type in the real world. Of course other interactions must also be addressed, but those can be done separately. In antenna models, as in other types of problem solving, it is best to vary one variable at a time.
In this antenna the mid-point of the 380" boom is 190". The positions of the first and second directors are 106.5" and 239", respectively. The second director is 49" from tower centre, but can come as close as 39" when parallel to a tower face. Clearly the design and construction of the yagi enters into the equation. However in this instance I will only focus on this particular antenna, not yagis in general. I need to keep this work focussed so that I don't get too far afield at this time.
With the model built I proceeded to "move" the position of the tower to different points and measure the resulting impedance and far-field pattern changes. These moves were in the X direction (parallel to the boom) and in the Y direction (parallel to the elements), but always staying within the space bracketed by directors 2 and 3. Without being exhaustive or comprehensive I was able to discern a few general trends that I found helpful.
Note that in the referenced diagrams above I truncated the yagi elements (green) to fit the illustration. In reality directors 2 and 3 or the 5-element 15 meter yagi are ~134", which is several times longer than drawn. The swing arm is also not quite to scale.
- When the tower is adjacent to the boom there is negligible interaction, even where the tower gets as close as 6" to one of the directors. This is good news, and ameliorates that particular concern. It also tells me that fixed side-mounted yagis are not in general compromised when one or two elements get close to the tower. This is increasingly concerning at higher frequency bands where yagi elements are closer together and the tower is larger (wider) as expressed in wavelengths. Interaction could become significant on higher bands, but I did not test this.
- The further the boom is from the tower (that is, in the Y direction, parallel to the elements) the greater the distance required from the adjacent elements to manage interactions. For example, when the separation is 24" (60 cm) the tower should be a minimum of 9", and preferably 12", from the directors of this 15 meter yagi. At double this distance (48" or 120 cm) the elements should be kept a minimum of 18" from the tower, and preferably 24". This is unsurprising since (capacitive) coupling becomes more pronounced toward the element tips; this is the same mechanism used in a Moxon rectangle and also confounds the tuning of fan dipoles.
- Tower resonance by itself has a negligible impact on yagi performance. However when there is interaction as described in the previous point the presence of a resonance exacerbates the problem.
- It takes a severe interaction to reduce gain by greater than 1 db. When this occurs the main lobe and rear side lobes become asymmetrical. SWR also increases due to the shift in impedance. From the model it appears that impedance deviations are a strong indicator of interaction; that is, you'll see the SWR increase, especially high in the band, before pattern distortion becomes significant.
If you play with a pen and paper you'll notice that the worst case for tower interaction is with the swing arm since it places the tower further in Y direction with respect to the yagi. Any rotation from the orientation show in the diagram above brings the tower closer to the boom and therefore will tend to reduce any interaction. Except, perhaps, when at the ±60° points where a direction will most closely approach one of the other two tower vertices. Keep that in mind when designing the yagi and side mount.
The girts on the LR20 tower are every 5' (two per 10' section) and are ideal for attachment of hardware to support side mount yagis. I took measurements of the bolt holes on the girts to get an idea of how to proceed. It is also possible to use commercial clamps that attach to triangular tower legs (e.g. Trylon). However my current plan is to build my own supports.
Out of prudence I will undertake some rough calculations of the loads the yagis will impose on a swing arm, rotator, bearings and mast under wind and ice loads. The tower appears up to the challenge (from talking to other LR20 owners) but that still leaves all the side mount hardware. Because of the array's size and weight I may opt for a bottom thrust bearing so that the rotator can be serviced more easily.
Since the stacking height on 15 meters is quite large there will need to be a bearing mid way. An alternative is to use two independent rotator and swing arm assemblies, but I prefer to turn the antennas as a single unit.
To minimize interaction with the tower I may shift the yagi mount off centre. Weight and wind load compensation may be needed to avoid added mechanical stress. The boom truss will help to avoid boom oscillations due to any remaining imbalance.
Depending on time I may first fix the antennas towards Europe, then build the simple mast system, and only later add swing arms.