Lifting the Side Mounted TH6

One of the smaller tasks I set out for myself this year was to increase the separation between the south pointing TH6 tri-bander and the 40 meter XM240 on the rotatable side mount. Their proximity isn't a problem most of the time, but when they are 90° to each other (XM240 pointed west) the TH6 degrades the performance of the XM240. That's because, from the side, the TH6 behaves as a short 40 meter dipole with large capacitance hats.

In the model I developed, the separation was increased from its previous 2.5 meters (8') to 5 meters (16'). That tamed the interaction to my satisfaction. Unfortunately that isn't achievable since there are guys and an 80 meter inverted vee above the yagis. The reflector element tips of the TH6 would get close to touch them and degrade the tri-bander's performance. I opted for a smaller step of about 11' (3.4 m) to preserve the performance of both antennas as well as the situation allowed. 

Even simple tasks like this involve complications, both mechanical and electrical. It took twice as long as expected when the unforeseen occurred. In this article I'll take you through the process, including what went wrong, and finally the results of the move.

I installed a pulley about 20' (6 m) above the TH6. The extra distance was needed to make it possible to lift the yagi and its tower bracket at the same time. The TH6 centre of gravity, like many Hy-Gain yagis, is at the boom-to-mast clamp so that's where I tied the rope. The bracket is much lighter than the antenna so it doesn't unbalance the load by much. To keep it level for the lift I cinched the rope to the upper end of the mast.

On the right you can see the problem. Convenient at the time, I installed the tower bracket for the TH6 at the desired height with the lower strut under the top strut of the rotatable side mount for the XM240. The TH6 had originally been there so it was easy to install the bracket that way and simply the yagi several feet. However, that made it a hassle for climbing since, pointed south, the boom and truss were near the one climbing face of the tower. I've had to squeeze past it many times! It would have been better, but quite a lot more difficult to mount the yagi on the other side of the face

For this lift the challenge was different: to swing the yagi and bracket outward so that the lower strut cleared the XM240 upper bracket. It was easy to do by placing the pulley high on the tower. I used the same technique for the more complicated repair of the XM240 last year.

After the tower was rigged I asked a friend to come over later in the week to help with the lift. In between we had high winds that may have contributed to a problem I discovered too late.

The lower pulley takes the haul rope horizontally to the winch. The pulley is extended out from the tower so that the rope clears the boom of the XM240 which is below the TH6. The hand winch is attached to a convenient guy anchor. The tarp is for tick protection.

The rigging is straight forward. It is more than capable of handling a 5' lift of a load less than 100 lb. Yet when my friend cranked the winch nothing moved. When I unbolted the bracket the rope held the load but it wouldn't budge.

After searching for obvious problems and seeing none, I climbed up to the high pulley and saw what I really didn't want to see. The ½" polypropylene rope had jumped the sheave and jammed between it and the pulley housing. I didn't take a picture at the time so I took one later in the workshop (below right). It is amazing to think that the rope could squeeze into that small gap. Yet it did.

I have several of these cheap pulleys and I really shouldn't use them. This wasn't the first time they've given me trouble. I picked it since it was first my glance fell upon when I was gathering material for the job. Since this wasn't going to be a heavy lift I didn't give it a second thought.

With a bit of jiggling I was able to lower the load slightly and reattach the bracket to the tower. I climbed up to disassemble the pulley to free the rope (left picture). After a critical appraisal of both rope and pulley I put it back together with the rope properly riding the sheave. The second lift attempt went smoothly. In less than 30 seconds the bracket was swung outward and upward and ready for bolting to the tower in its new location. 

The girts on the tower are approximately 5' apart so that was the minimum distance it could be moved. I lowered the TH6 on the stub mast to where I wanted it. The next day I took a few pictures to show how close the reflector is to wires and cables.

I took these pictures the next day after pointing the TH6 south. We are looking down on the TH6 and XM240. To the northwest (left) the rear reflector is several feet from a 105' level guy and one leg of the 80 meter inverted vee. To the northeast (right) the other leg of the inverted vee if not as close but the guy (nearly invisible in the picture) is just a few feet from the opposite reflector tip. The south guy (not shown) is centred on the boom and some distance from the TH6. From previous modelling of interaction scenarios I am confident the south guy has negligible impact on the TH6.

Despite the proximity, I was surprised to note that the SWR on all 3 bands (20, 15, 10) was affected only a little. SWR measurements on the tower were not accurate since I couldn't stand far enough away to avoid detuning the antenna -- the jumper to the analyzer was too short. Body interaction was particularly bad on 15 meters; it's happened before with this antenna. On 20 meters the SWR at the band edges was a little high at 2, however back in the shack it was about 1.7 to 1.8, probably due to some loss in the long run of LDF5 Heliax.

The 15 meter SWR was very good when measured in the shack. The 10 meter SWR was about the same as measured on the tower, with the SWR 2 bandwidth ranging from 28.0 to 28.8 MHz. 

That's good enough. The TH6 isn't used heavily since it is primarily a "multiplier antenna" used in contests to work the multiplier rich and QSO poor Caribbean, South and Central America while the other, larger antennas can stay pointed to Europe and elsewhere.

But what of the XM240? Did this manoeuver help? That's difficult to measure with any confidence. What I can say for certain is that the SWR deviation when the antennas are at right angles is significantly improved. 

The deviation is now small enough that there is little need to re-tune the amps when the XM240 is rotated. I consider that a win. If modelling is any guide, gain and F/B are also better preserved. That, however, is very difficult to measure and confirm.

With this job out of the way I can proceed to finish the 40 meter Moxon and swap it for the XM240. I hope to get that job done in the near future. 

One final and slightly amusing note. After measuring the distance from the TH6 feed point and the Heliax connector I realized it was a perfect fit for the 20 meter yagi delay line that was retired from service last year. It had been hanging on a hook ever since. Further reducing the clutter, the small box for the L-network is also being reused in the 40 meter Moxon project. 

There are a few benefits to being a pack rat who rarely throws anything away.

Trees

Trees are not towers. Living things are engineered by billions of years of selection for survival, not by design. I have some expertise with one but not the other. Since trees grow and surround us, and can be very attractive and desirable, we admire them most of the time. Other times they are a curse. This was one of those times.

When the wind blows I don't worry too much about my many towers and antennas. They've been engineered to survive. Trees may not fare as well. On my 50 acres there are thousands of trees. A few are used to support antennas, particularly several Beverage receive antennas.

Other trees are unremarked until they age, die and fall. That's dangerous when they're near towers and antennas. Tall trees do the most damage. When they break near the roots they measure their full height on the ground. Despite steel being stronger than wood, the momentum can and will destroy engineered steel structures.

This spring I cut down several large trees that threatened the house and garage, and a few smaller ones that were unwanted or that threatened wire antennas. Yesterday we had a wind storm with 100 kph gusts that took down several large trees. One antenna got in the way. 

I knew what to expect when the reversible Beverage worked poorly in its normal direction but not at all in the reverse direction. That's the most obvious symptom of a cut wire or crossed wires in a reversible Beverage. Even lying on the ground the now unterminated Beverage becomes bi-directional. It can't work in the reverse direction because the twinned wires no longer act as a transmission line.

There isn't enough resolution in the picture to see that both wires of the northeast-southwest Beverage are trapped under the fallen tree. Obviously both wires snapped, but at different locations some distance from the tree -- wires break where they are more vulnerable to the sharp spike of tension. It was only the day before that I splices one of those wires when a branch I was cutting from a dead tree at the other end of the antenna, over 100 meters away, fell across the Beverage.

I returned with a chainsaw and material to splice the antenna. In a minute the rotten wood was cut into several lengths and tossed aside. There were more dead trees in the boggy ground (there's a large swamp just beyond the trees in the picture) that I had to leave until I have a partner to pull the tree away from the antenna while it is cut. It is very difficult to perform both actions simultaneously. Never do it solo unless there is nothing of importance in the fall zone.

There were already 4 wire splices in the area due to all those trees and the occasional deer skirting the swamp. Preventive tree clearing doesn't catch every potential problem short of clear cutting. I took the opportunity to replace several spliced sections with new wire. Since tick season has already begun I worked early in the morning when it was cold and the bugs were asleep.

Another challenge with trees is that they grow in diameter. They do it by building new layers on the outside. Over time they envelope attachments such as nails and screws. The simple cleats made from nails were exposed less and less over time until they become unusable. The nails are difficult to extract so it is easier to use new ones. This antenna, originally a uni-directional Beverage, was installed over 8 years ago. There wasn't much nail left (see the lower nails in the above left photo -- the other old nails were hammered flat).

The Beverage is high enough that many repair jobs require a ladder. The Beverage would work at a lower height but that would put it in the path of deer and the occasional hunter (the latter with my permission). 

The picture on the above right gives an idea of how splicing is done. It can be complicated because the wire must first be loosened so that there is enough slack for the wire ends to be twisted together. Then they are once again drawn tight. A lot of walking must be done if the break is far from a termination -- this Beverage is 175 meters long.

At right is another tree that came down in the storm. This one was midway along the north-south RG6 reversible Beverage. The tree was blazed last fall to mark it for removable. It decided not to wait for me and my chainsaw. 

I got lucky in this case as I have before. The tree fell away from the Beverage rather than into it. It was a big tree that dislodged limbs and saplings, some of which landed on the coax. No damage was done. I'll clean up the mess later.

Felling big trees before they cause damage is not a job for the timid or fools. I have quite a few of these trees that I've been methodically removing over the past months. There is one near the power line that crosses my property that I'll leave for the utility crew to deal with.

One large dead maple threatened the garage and also hovered above the two overhead coax cables running from the tree line to the garage and house. Those are above ground to keep them from being a hazard to walkers. This was not an area where they could easily be buried.

I dropped the two coax runs to the ground in early winter since there would be little foot traffic and the snow would provide protection. I then cut several large limbs to make felling it less of a danger. Only then did I attack the trunk. Well-placed chainsaw cuts, cables and a winch directed major limbs to where I wanted them to fall. Falling trees that large are very difficult to direct by muscle and a rope. I don't recommend trying it. It took a long time to clean up the remains, brush and good firewood. 

Although I have a lot to say on this blog about towers, I will not do the same for dealing with large trees. The reason is that I'm an amateur. I am not qualified to give directions or advice. I am leaving out a lot of detail! For example, the huge maple I felled in the above sequence of pictures taken by my valued assistant for this job, Alan VE3KAE. He has more tree sense than I do.

On the left I am making the initial cuts. Two are made to make a wedge opening on the side where I want the tree to fall. The felling cut is then made on the opposite side. It helps that the tree is already leaning in the direction I want it to fall. If that were not true, well, you need an expert not an amateur for the job! As I said earlier in the article, a human (or a few) armed with a rope cannot direct the fall of a tree this large.

The middle picture is the "cut and run" phase of the operation. My spotter is updating me on the state of lean by monitoring the motion of the upper limbs. The cut widens very little at this stage so don't use that as an indicator. There were a couple of false alarms, but it is better to be cautious. 

If you expand the picture you'll see that there are several cuts well above ground where I previously removed large limbs. That was done to have the remaining weight lean the tree towards the falling side and to reduce damage to other trees (those I want to save). A tree this large is both wide and tall.

Finally down it goes. When it's on the ground you can see how tall it was. We were sure to move our vehicles out of the path before starting the job. That dead maple, although far from the buildings, still fell close to the garage and would have either struck parked vehicles or showered them with large limbs that broke up when it hit the ground. Damage to the stone wall surrounding the yard could not be entirely avoided. 

The trunk and large limbs of the trees I felled are being turned into firewood. They are cut into lengths, stacked and dried, and will eventually be split. That's a lot of chainsaw work but it isn't very difficult. The worst is the multitude of small branches that have no use. They can be piled up and (very slowly) composted or piled up and burned. I have two piles like the one on the right. It's almost 10' (3 m) high. I have lots of space deposit the refuse where no one is inconvenienced but no good place to safely burn the brush.

To end this article I'll show one more tree -- it's the large one near the right of the picture that is leaning left. The tree supports the head end of the northeast-southwest Beverage. It's an otherwise unremarkable tree, until it died about a year ago. The lean has since increased a small amount.

Unfortunately there's a guy anchor for my 150' tower not far enough away were the tree to fall directly towards the anchor or, worse, onto the guys. That might or might not bring down the tower. When I did my initial site plan for the station and surveyed the area to place the tower I thought this would be a safe distance. The trees had other ideas and grew higher and higher.

I've lopped off several lower limbs and I am starting of the on the larger ones higher up. The objective is to lessen the area of the fall zone and to reduce its momentum if it does happen to strike the guys when it's cut down. I'll leave about 8' standing since I have no other convenient anchor point for the Beverage.

This work is dangerous if not done correctly. I'll climb a tower alone but I will not fell large trees without at least one knowledgable helper. As I stated at the beginning of the article, I won't go into details since this is not my expertise. If all goes well, that tree will be safely down in the coming days. However, the job will be delayed if there are any doubts as to safety. The tree is not yet decrepit and will last until the fall when I resume tree work. 

Update May 5, 2025:

The tree threatening the guys is down. Alan VE3KAE and I took it down after we completed work on one of the towers. There were other tower jobs to be done that day but this fit better with the weather and the available time. He took the following picture after the felling.

Notice that several large limbs were removed before the main trunk was cut. That lowered the momentum of the falling tree. Cutting it high up shortened its reach on the ground to what we deemed acceptable. The upper branches struck the guys, as I knew they would. They have little strength so they shook the guys but were incapable to doing any damage. The Beverage terminating on the tree was unharmed.

Where I cut the main trunk the diameter was about 9" (23 cm). As before I will demur on details other than to mention that the ladder and I were very well secured so that I could devote both hands to operating the chainsaw. Don't try a job like this on your own. If you don't have an experienced person on hand, please hire a professional.

I can rest easy with this risk defused. The rest of the trunk above the Beverage will be cut in the fall. Other trees have been marked for their potential risk in coming years as they grow higher. In the meantime my pile of firewood is growing.

40 Meter Reversible Moxon: Construction

After disposing of several other tasks I returned to the construction of the 40 meter reversible Moxon. The plan is for a more efficient and versatile replacement for the XM240. In an earlier article I described the computer model using EZNEC with the NEC5 engine. This is an antenna that requires NEC5 for acceptable accuracy.

After an unexpected spring snowstorm and other matters, spring is finally asserting itself and I can comfortably resume work on the antenna. There is an urgency since the hay will get tall in mid-May and more distractions are imminent. I hope to have it working and on the tower by then.

It's a large antenna but not so large as some others that I've built and raised onto the towers. The boom is reasonably short (typical of a Moxon) yet the elements are long and there are large capacitance hats. Due to the constraint of symmetry necessary to make it reversible, the elements are identical and that reduces the design challenge. My experience building most of my yagis allowed me to proceed with confidence with respect to material selection and the joining of tubes, plates and pipes.

This article is about how I mechanically designed and built the antennas. A future article will cover its electrical design, testing and performance. I prefer to split the discussion rather than wait to the end and write one very long article. Impatient readers can read previous articles about the electrical design and modelling with NEC5, and how I am approaching the feed system.

One note before I start is that this is not an XM240 conversion per the W6NL design. I am holding on to that antenna for another use. However, my mechanical design is similar to the conversion project with its relatively light duty elements. My XM240 was improved in line with the conversion specs by its previous owner.

I did not want to make it too heavy (over 100 lb) by specifying stronger elements such as those on my 3-element 40 meter yagi. The elements are therefore trussed, like the XM240 conversion, and other measures were taken to limit element flexing. The antenna will be low enough for service, if necessary, although I of course hope to avoid it. You can judge for yourself how well I've done from the description in this article.  

Boom

The choice was easy. I had a 20' length of 2.375" OD schedule 80 6061-T6 aluminum pipe in my stock (0.218" wall thickness), which was perfect for this antenna. I bought it as used/surplus several years ago with the intent of using it as a mast, and then never did. With it, I can forgo a boom truss since boom droop is quite small. Its hefty 35 lb weight is an acceptable trade off.

This may not be a solution for most antenna builders since these pipes are expensive when new. A lighter duty boom with a truss may be more economical. Just don't get so economical that it is under-built for wind and ice loads. The load on the boom due to those long, end-loaded elements, is substantially more than for the XM240 or similar 40 meter yagis.

The distance between the element centres is about 225" (18.5' or 5.7 m). that leaves room outboard of the boom-to-element clamps for the clamps that support the 5' element truss posts.

Clamps

The boom-to-mast clamp is not easily seen in the picture above. A close up isn't necessary since it is nothing out of the ordinary. Again, I dug into my junk box and found a suitable plate clamp. Galvanized saddle clamps on a ¼" aluminum alloy plate fit the 2-⅜" boom and 1.9" pipe mast. It goes right at boom centre since, due to antenna symmetry, that coincides with the centre-of-gravity. The mast size is pre-determined by the Hy-Gain rotator used in the rotatable side mount where the XM240 currently resides. 

The boom-to-element clamps are more elaborate. When I first designed the elements the halves were entirely supported by the clamp. The large bending moment puts a lot of stress on the polymer insulators (the element must be electrically isolated from the boom) and the clamp plate. 

The polymer tubes were scavenged from an ancient 40 meter coil-loaded yagi that a friend gave me. The plates are 18" × 6" × ¼" aluminum alloy bars. Aluminum alloy strips protect the polymer from the stress points due to the saddles and u-bolts.

The galvanized angles take a lot of the stress but not enough of it. I built ¼" steel backing plates to make up the strength deficit. It worked but I was not satisfied by the complexity and weight of the completed clamps.

I purchased 2' lengths of 1" fibreglass rods from an antenna builder and fit them inside the inner pipes of the element halves. Aluminum flashing was used as a shim since the ID of the pipes is 1.049" -- I wanted a close fit so that more of the rod contacted the pipe to reduce point stresses. The rods are very strong and can support the element on their own. Fibreglass tube is lighter and cheaper and may have sufficed but this joint is critical. The steel backing plates were no longer needed. 

The rods are held in place solely by the hose clamps at the slit ends of the pipes. If there is any "creep" I may need to drive screws through the pipe and into the rod. I prefer not to do that so we'll see how it fares. The inner hose clamps also hold the tabs from the switch boxes.

Elements

I made a deliberate decision to reduce the cost and weight of the elements by making them similar to what is found in XM240 conversions to W6NL Moxon antennas. It helped that I had almost all of the required aluminum on hand.

The capacitance hats, although not heavy, present cumbersome and complex loads since they are placed at the element tips. Heavier element construction might be sufficient to deal with those loads, but at a "heavy" cost. Thus, a truss is the preferred solution.

The details of element construction are as follows:

• 1.315" (0.133" wall) 6061-T6 pipe: 72". There is a 1" gap between the element halves, plus inductance due to the ~2" (5 cm) of conductors into the switch box and up to and through the relays.
• 1"  (0.120" or 0.130" wall) 6061-T6 tube: 115".
• ¾" (0.058" wall) 6061-T6 tube: 48.5", with a ⅝" (0.058" wall) 6063-T832 tube inside
• The same ⅝" tube projects 19" beyond the ¾" tube to form the element tip.
• 4" × 3" × ¼" 6061-T6511 plate with stainless u-bolt to attach the capacitance hats to the end of the ¾" tube.
• ⅞" (0.058" wall) 6061-T6 tube: 7" (half of capacitance hat centre tube).
• ¾" (0.058" wall) 6061-T6 tube: 20.5". A single length runs through the centre tube and also forms the second tube section for the other half of the capacitance hat.
• ⅝" (0.058" wall) 6063-T832 tube: 33" (capacitance hat).
• ½" (0.065" wall) 6061-T6 tube: 46.25" (capacitance hat tip)

Most of the joints are fixed with screw fasteners in a manner I've used for other yagis. That leaves few places to adjust the yagi should it not exactly match the NEC5 model. The only adjustable tubes are the tips of the capacitance hats and the element tip. The latter can only be lengthened. Since the distance between capacitance hat tips should not vary far from its design value, one option is to only adjust the outside arms of the hats. This must be done on all 4 to preserve symmetry. For small adjustments the mechanical imbalance won't be an issue.

The element tips can be lengthened by insertion of a ½" tube fastened with a hose clamp. Since I don't expect to use this method (it is more likely the elements will need to shortened slightly) I have not cut slits in the tips. That can be done in the field if necessary. I'll know more after the antenna is in the air and its electrical behaviour is measured.

The capacitance hats could have been made by stepping down the tubing schedule by ⅛", similar to that for the typical 40 meter Moxon. I opted for a stronger design because of the antenna's symmetry and for stable element coupling, the latter of which has a strong impact on antenna performance.

In the W6NL Moxon, which is unidirectional, the capacitance hats are laterally offset. The coupling between driven element and reflector is therefore less than in a Moxon rectangle, and the tips are less likely to touch. This appears to be a reasonable trade off between off performance and reliability. However, I have never seen this stated so it may not have been one of his design criteria.

The capacitance hat tips in my antenna are aligned and 30 cm (12") apart. It is far easier for the tips to touch. That could cause havoc in the shack when amplifiers fault and go offline or (worst case) cause a hardware failure. There are two approaches to prevent this from happening: insulation barriers to reduce or eliminate metal-on-metal contact, or fix the relative positions of the capacitance hat tips. I chose the latter option, which is described below.

Element truss

My first design didn't work. I used a long ⅜" threaded rod to "reach" over element centre to adjust for the post's boom clamp that positions it about 5" to 6" towards the end of the boom. I went further and removed the saddles from the clamps so that the post would lean inward. The reason this failed to work as intended was that the tension on the truss rope was greater than I expected. 

Although the element halves are not heavy, the geometry is not in my favour. The truss attachment to the post is ~4' above the element and it attaches to the element at the end of the 1" tube, which is 18' from the boom. The angle between the truss rope and the element is therefore only 13°. It doesn't look that acute when I stand next to it but it is. 

The tension to lift the element is on the order of several tens of pounds. There was too much inward bending stress on the post. I had intended to use a non-tempered aluminum tube for the post. Instead I went with a 1.315" pipe. I'm glad I did. I am surprised that the W6NL XM240 conversion specifies a 3' tall PVC pipe, yet I know that many of those antennas are working perfectly well.

I replaced the saddles and attached the truss ropes closer to the post. That worked. Only a small adjustment was required to keep the element straight when the truss rope tension is high enough to level the element. Shorter ⅜" bolts serve to properly align the post and element. Each bolt is covered with a ½" scrap aluminum tube to protect the rope from the sharp threads. 

Rather than turnbuckles, the tension is manually adjusted and the rope secured by a simple cleat made from long #10 screws. The rope can be taped to the post to prevent unravelling. The nylon rope used during setup (seen in the photo) was replaced by low stretch, UV-resistant Dacron. The improvised cleat is on both sides of the post to allow each rope to be independently adjusted.

I tested the truss as well as I could by simulating wind and ice load by bending them in various direction by hand and to simulate oscillation in the wind. The antenna performed well with no sign that the element was prone to buckling or other weakness. However I can't be certain without a full mechanical analysis which will not be done. That's beyond my expertise.

An alternative truss whereby lateral flexing of the elements can be largely eliminated is with a cross bar on the post and attaching ropes to each end of the cross bar. The weight and wind load penalty is small. Although that will not reduce flexing of the element tips and capacitance hats, their flexing will be less prone to amplification by in phase element oscillations.

We'll see how the built truss fares after it is installed. I was also pleased to observe that the capacitance hats resisted moving inward under simulated lateral (wind) load. That won't prevent the tips from touching but it reduces the difficulty to mitigate that problem, as described in the next section.

Capacitance hat tip management

The capacitance hats are collinear, unlike in the W6NL Moxon. That makes them far more prone to touching as they bounce in the wind. Even if they don't touch, the tip spacing has a significant impact on performance since the "critical coupling" that makes a Moxon work depends on that spacing. I considered and rejected a few alternatives:

• Double truss the elements to reduce lateral motion
• Several narrow ring insulators on each inside tip so that when they try to occupy the same space the insulators strike the aluminum and prevent metal-on-metal contact
• Side trusses on the elements using boom extensions as truss posts and a rope between the capacitance hat tips -- the tensioned structure is rigid in the horizontal plane

My preferred solution is simple and should be effective. It works well when the antenna is agitated during testing at ground level. A fibreglass rod connects the capacitance hat tips.

The ½" aluminum tube tips of the hats are slotted for compression with hose clamps. The fibreglass rods are 4' long driveway marker posts that are common in snowy climates and can be found for a few dollars at any hardware store. I pulled off the top caps and inserted them into the tubes. 

Since the rods are about 5/16" (8 mm) diameter, a little smaller than the ID of the tubes I wrapped the rod with a few turns of electrical tape at each end for a tighter fit. The distance between tips is 12" (30 cm) per the EZNEC (NEC5) model. I used the full length of the rods so there are 18" (45 cm) inside each tube, adding to the strength of the tubes.

The distance between the elements is ~18.5' (5.6 m) so there is droop in the capacitance hats and the fibreglass rod bends to fit the curves. Although they may appear flimsy they withstood a lot of abuse when the elements were strongly agitated by hand. The elements moved as a rigid structure. 

We'll see how the hat tips fare once the antenna is installed on the tower. If there is a failure I can always fall back to one of the alternatives listed earlier. But for now I like it: cheap, simple and effective.

Weight and wind load

The antenna is almost ready for its first live test. It'll be trammed to a low height on the 15/20 meter tower where antenna interactions will be minimal. It was assembled at the approximate launch point in the hay field between the big towers since it is an awkward antenna to move once assembled.

Although most hams would classify it as a large antenna, it is a middling large antenna for my station. When assembled it is quite heavy, approximately 105 lb (50 kg). That's at least 15 lb more than I estimated before construction. The element clamps and truss posts are two of the culprits. The weight was measured with a method I've used before: weigh myself and then again holding the antenna. I had to hold the antenna steady for the 10 seconds until it stopped wobbling and the scale reading was stable. 

I'll probably use a vehicle to tram the antenna onto the tower. There will be less grumbling from my friends! We're all getting older and it takes a lot of muscle to pull an antenna this size up the tram line.

One detail I'll have to deal with is that the antenna can't get past the tower because the capacitance hat tips are connected. The forward fibreglass spacer will have to be removed and reinstalled after the hats advance to the far side of the tower. Good communication with the ground crew is critical.

The wind load is less that what you might expect for an antenna of this size. The reason is that there are only two elements and the boom is not very long. 

The wind surface area (cylindrical) is simple to calculate from the tube and pipe dimensions:

• Boom: 4.0 ft²
• Each element (not including the capacitance hat): 3.4 ft²
• Each capacitance hat: 0.7 ft² 
• Element truss: 0.5 ft²

The total is 4 + 6.8 + 2.8 + 1 = 14.6 ft². The wind load experienced is never that high because all components of the antenna are not parallel. The expected wind load would exceed 7.8 ft² by a small amount, which is that of the two elements or, coincidentally, the boom and capacitance hats. Due to droop, there is some wind load due to their vertically projected area. The element trusses always face the wind and their area is included for all wind directions. 

Since the elements are long and will oscillate in high winds, the spikes in the dynamic load suggest that the load is greatest when the antenna is pointed toward or away from the wind direction. You don't want so much tension on the truss that the element is at risk of buckling when it bends due to wind and/or ice load. 

To reduce the risk, the post height can be increased and the element attachment point moved inward. I anchored the rope at the end of the 1" (~0.125" wall), not at the capacitance hat clamp as in the W6NL XM240 conversion design. The element is still strong at that point and the truss rope angle is higher than it would be when anchored at the capacitance hat

The wind load is not really within the capacity of the Ham-M rotator that is used on the tower side mount where the XM240 is currently mounted. I don't anticipate serious problems unless we get a severe wind storm. I'd likely have the same risk with the XM240 since its wind loading is only a little less, the main differences being the capacitance hats and element trusses on the Moxon.

There it is, ready for liftoff. First I need to complete the feed system so that it can be tested. I'll discuss that after the antenna is raised and tested since that's more associated with tuning, not the mechanical construction. 

My hope is to complete this project by mid-May before the hay gets too high for comfortable ground work.

Cookies for Hams

I've been doing tower work for others since I was a teenager. At first for my friends, then for their friends and, well, on it went. I enjoyed the gratification of putting a smile on another ham's face. Besides, I came to love working on towers and since I had just one to play with, helping others with theirs brought me joy.

It has also become a way for me to pay it back -- to give back to a hobby that has given me so much.

Tower work never tempted me as a career. Not only did my interests lie elsewhere, having grown up with little I worried about potential income. As I learned from talking to professional tower riggers, for most of them it's barely a step up from ordinary construction work. The pay isn't that great. Only a minority do well.

I don't do tower work for money. Countless hams have tried to pay me and I have always refused. Some have literally tried to shove a handful of bills into my hand, and were surprised, and occasionally annoyed, when my fingers fail to grasp the money. In a few cases I relented in order to avoid giving offense. I have had others offer me items from their shack such as electronic parts, boxes of floppy disks (remember those?) or whatever they had on hand. I accepted in very few cases.

Not working for money also gives me the choice of whether to take on a job and where to draw boundaries. Most often that involves safety but also when the ham insists on making poor choices about which I have strong negative views. Those in the trades see all types and have learned how to manage bad situations while still taking the project and getting paid. I don't have to.

My one great weakness is food. If I'm offered a meal at the end of the job I rarely refuse. When I was young and living far from home for the first time, a home cooked meal was irresistible. Snacks and drinks while I worked were also welcome. Now that I'm older I don't worry as much about the offer of meals. I am often happy to make a quick exit and return home.

Occasionally the food being offered brings a smile to my face. Last year after doing a tower job for a repeater group, one of the members raced off to the farm he was looking after and surprised me with a carton of farm fresh eggs straight from the chicken coop.

I have been given cups and other paraphernalia customized with their club logo, my call, t-shirts and more. Then came last week when I was given an even better surprise after taking down a tower: a box of hand made cookies. But they weren't just any cookies. Here's a sampling of what was in the box:

It seems a shame to eat them. They're nicely done, and tasty. I quickly realized that it would be advisable to take a photo before the inevitable occurred.

Unfortunately I don't know who made them, other than that it's a woman starting a baking business. There is as yet no address or business name. In any case that information would be useless to a global readership. It isn't too useful to me either since her city is a 2 hour drive from here.

I'm sure there are similar artistic bakeries near to many hams. Seek one out and put a smile on the helpful ham's face with a radio themed edible gift.

Modelling Ice on Yagis

On the second day of the recent CQ WPX SSB contest we had an ice storm (freezing rain). All my antennas survived the moderate amount (up to 5 mm) of ice. The impact of the storm was far worse to the west of us, with many lost trees and antennas. The high SWR of most of my antennas made operating the contest a hassle. I don't really like WPX so it was no great loss.

This is not the first time I've dealt with ice; it's a common weather event in this region. I know from experience how ice has mechanical and electrical effects on wires and aluminum elements, and towers. The major worry is antennas and towers breaking and even falling to the ground. That is a serious hazard. The electrical effects are not often given the same attention. These may only come to the fore during important operating events like contests.

The electrical effects are unsurprising. Ice is a dielectric so it behaves as a lossy insulator. That lowers the VF (velocity factor) and therefore increases the electrical length. The same is true for antenna made from insulated wire. But the effect of wire insulation is fixed and therefore constant, and that can be quantitatively dealt with during design and construction.

The variable effects  of ice depend on the amount of ice, whether it coats all or part of the element circumference, the presence of impurities (dirt), and of course its weight changes the shape of the antenna (droop and sag). Ice, like most insulators, has loss so there is a reduction of gain, ranging from small to large.

It can easily be enough to shift an antenna far outside of the band(s) for which it is designed. Some antennas are affected more than others, depending on their shape, orientation and Q. Here is a rundown for my antennas during the recent storm:

• Beverages: They worked despite sag and icicles on the wires. Performance may have been worse but there is no easy way to determine that. Beverages can be forgiving antennas when subjected to non-destructive assaults.
• 80 meter vertical yagi: Resonance shifted down only a little so it remained usable. I suspect most of the impact was on the parasitic wire elements.
• 80 meter inverted vee: Wire sag and resonance shift were substantial but they went in opposite directions: higher and lower, respectively. The SWR was a little higher but the antenna was otherwise okay.
• 40 meter yagis: The effect was greater on the big 3-element yagi and less on the 2-element XM240. Perhaps that was due to antenna Q, although the inductively loaded elements of the smaller yagi has a similar effect.
• TH6: Moderate resonance shift. It remained usable on all 3 bands but the SWR wa a problem high in the phone segments. 
• Skyhawk: Similar effect as for the TH6 but less severe. It remained usable.
• 5-element yagis for 10, 15 and 20: The effect seemed quite acute on the 15 meter yagis and moderate to large on the other bands.

Freezing rain requires unusual atmospheric conditions so that it is uncommon even in regions with cold winters. Where I grew up in VE4 (Manitoba) they were rare; it was too cold for ice storms. In eastern Ontario they happen more often, perhaps 2 to 3 times per year, mostly in late fall and early winter, but they can occur mid-winter. It is important in this climate to design towers and antennas to survive ice loads of up to ½" (over 12 mm). Of course many do not, out of unconcern or relying on luck.

The ice coating can surround antenna elements or just sit on top. The liquid water of the cold rain has only a brief period to flow to the bottom of the element before it freezes. Wires are narrow so they tend to have a coating of roughly constant depth (left). It takes more freezing rain to fully surround yagi elements. The coating typically has a corrugated appearance and can have icicles.

The mechanical effect on wire and aluminum tube antennas is different. The weight of the ice can far exceed that of the wire and cause extreme sag, such as the inverted vee on the left. The ice lowers the resonant frequency while raising it due to the more acute interior angle. However, they are unlikely to exactly compensate.

For the same amount of ice on yagi elements the effect is less severe. Rather than sag in the centre, the elements ends droop. Moderate droop does affects antenna elements a small amount since the centre, where current is maximum, remains straight. But, as we've seen, the resonant frequency will be lower.

Even if an antenna survives an ice storm, that isn't good enough; a mechanically sound antenna can become useless when covered in ice. That can prove very inconvenient if it happens when you really need it: contests, DXpeditions, etc. As you can see in the above list, some of my antennas fared worse than others. This is especially true of the 5-element yagis.

How bad is it? It is easy to add ice to antennas in modelling software so that's what I did. The wire group editting feature in EZNEC was very helpful in this exercise. 

Fresh water ice can be modelled as an insulator. All we need to know is the dielectric constant, thickness, and loss tangent. Rainwater contains impurities that are difficult to predict so I didn't bother. This study was intended to roughly quantify the effects of ice, not to make precise predictions.

For fresh water ice, the several references I reviewed closely agree with an approximate dielectric constant of 3.15 and a loss tangent at HF of about 0.1. I first tested the model of the 5-element 15 meter yagi with ⅛" (~3 mm) of ice, with the antenna in free space. Using free space simplifies the model to focus on the variables of primary interest, rather than far field pattern and ground interaction.

The 15 meter yagi has a distinct SWR curve. The first dip (SWR of 1) at 21.1 MHz serves as a key marker. The second minimum help to further validate the integrity of the model when ice is included. The base azimuth pattern is taken at 21.2 MHz. The second pattern, with ice included, is at 20.7 MHz, in proportion to the calculated SWR shift. Notice that the patterns are identical other than loss for the ice-covered yagi.

It is assumed that the ice is constant depth for the full circumference of all tube diameters used in the elements, in this case ranging from 1" to ½". The model uses a hairpin (beta) match, since it is far easier to model than the gamma match on the real antenna. Other than the different effects of ice on the matching network (not included in the model) the SWR curves are almost identical for both matching networks.

With this modest amount of ice the antenna resonance is shifted downward by 500 kHz. That's surprisingly large for such a small amount of ice. Yet my experience agrees with the model. Ice causes the SWR to be very high across the 15 meter band, and especially on phone for the WPX contest. The noise floor when switching to the 5-element yagis (either or both in the stack) is eerily quiet.

If the scaling effect of the ice is to be trusted the full SWR curve should remain intact, as should the antenna pattern. The SWR curve scales well with only a small deviation of perhaps 25 kHz between the two minima. The patterns at 21.2 and 20.7 MHz are identical other than less gain for the latter (the plots show only the azimuth patterns). There is loss in the ice since it is not a perfect insulator.

With these favourable results in hand I collected model data for my 20, 15 and 10 meter 5-element yagis with radial ice depth from ⅛" to ½" (3.2 to 12.7 mm).

As you can see, the effects are quite large and they increase with frequency. The effect gradually tapers as the ice depth increases. Gain/loss was calibrated to 14.15, 21.2 and 28.5 MHz for each 5-element yagi, and validated by the aforementioned pattern and SWR markers.

By the point where you have ½" of ice on your antennas you probably have more important concerns than yagi performance! The SWR and gain of a broken antenna is not the concern.

Above is the roughly calibrated (450 kHz width of the the band) SWR curve of the 15 meter yagi with ½" of ice. The models became increasingly difficult to calibrate when the ice is thick. The familiar SWR markers (as noted earlier) have become vague for all 3 yagis. The expected patterns remained the same despite the changing shape of the SWR curves, so that served as a further check on the calibration. However, some uncertainty remained in these extreme cases.

This is not precision work! Even so, my method proved to be adequate since my purpose was to gain insights into the effects of ice. I only ever measured the SWR once during an ice storm and my recollection of it for the 15 meter yagi closely resembles my model.

At this point of the study, I had a question: is this a uniquely 5-element yagi problem? That is, are smaller yagis and single element antennas less severely affected by ice than large multi-element yagis? That's more modelling work than I'm willing to do! What I could do was a few models of smaller antennas, and then compare the severity of the effects.

I went through the same exercise with an ordinary 80 meter inverted vee. Its apex is at 20 meters height, has an interior angle of 90° and is resonant at 3.525 MHz (for CW). I did not attempt to compensate for sag due to ice even though it could be significant. That would lessen the change in resonant frequency and increase loss, both due to the smaller interior angle.

The effect was greater than I expected. My 80 meter inverted vee with an apex at over 30 meter did not seem to fare so badly in the ice storm -- the SWR was higher although it worked reasonably well. Yet the model shows that its performance is as bad as for the 5-element 10 meter yagi. 

I suspected that the large modelled effect was due to the larger ratio of radial ice to conductor diameter for a wire antenna. To test my hypothesis I reran the model for #10 and #18 wire, both two steps from the original #14 gauge wire. Other than a small shift in the resonant frequency, the percentage of ice induced shift barely changed. I didn't bother to substitute even thinner or thicker wire to see if that resulted in a non-negligible difference.

The loss for the ice covered 80 meter inverted vee is not as bad as for the yagis. I didn't bother with a plot since the numbers are small. With ½" of ice the loss is -1.0 db. That much better than the 5-element yagis.

I tried a few other antennas in my large model library. Since the modelled effects of ice on these antennas were unsurprising or ambiguous, I decided to stop there. I was satisfied with the insight gained, and that was my objective.

I wish I had a model of the Skyhawk tri-band yagi on hand since its SWR suffered less during the recent ice storm. However, that was only the SWR and it is certain that loss due to ice would mirror that of other yagis built with tubes. It may just be that the Skyhawk is lower Q than many other yagis. The yagis I've designed and built are higher Q than is typical since I tune them for better gain at the expense of SWR bandwidth and F/B. The SWR of my 5-element yagis for 20, 15 and 10 meters rise sharply at the high end of their respective bands.

I have noticed frequency shifts on the 5-elements yagis during heavy rainstorms, though they were of less severity. Heavy snow fall close to the freezing point when the snow is wet and sticks to the antennas has effects similar to that of ice. I was a little surprised by that when it happened the first time. However, snow is not as heavy as ice so there was no structural risk, just operating inconvenience.

I briefly considered the possibility of using these results of this study as a basis to perform reverse calculations. It should be possible to determine the average radial ice depth by the change in antenna behaviour. I don't know whether that would be useful or just a curiosity. The "average" is far from a precise characterization of the variable deposition of ice on surfaces. But it could be done.

This picture was sent to me by a contester (whose call you'd recognize) taken during a severe ice storm several years ago -- he lives hundreds of kilometers from me and there was no ice storm here. All of his antennas survived but it was touch and go. This is the bottom of the support rope for one leg of an 160 meter inverted vee going up to one of his big towers. It's astonishing that the antenna bounced back when the ice melted. Take this as a lesson to build antennas that can survive ice loads if you live in an area where they are periodic events.

The snow is gone and the weather is warming. Soon it will be recognizably spring and worries about ice storms will fade. That is, until next winter! You never know what nature will throw your way. Be prepared for whatever weather risks affect you, be it ice, flood, wind or earthquake.

Making Mistakes Along the Way

Amateur radio is practical: we use what we buy or build. A device can be ugly, deficient, cheap, less than state-of-the-art and still be immensely useful. If it helps you to put a signal on the air you will be successful. There are no asterisks on an operating award because your home brew QRP transmitter was a oddball collection of parts spread across the operating desk, held together with bits of tape and clip leads.

I use computer design tools and breadboards to test my designs. They often remain prototypes for a very long time. Witness my long running project to design and build a digital controller for my prop pitch motor rotators. The breadboard direction indicators sitting on the old 24 VDC power supply and a standalone panel meter is still there and working! Ugly and embarrassing but it gets the job done. 

The new one is getting better although not yet completed. With the weather warming it will take a while longer as my focus turns to antenna and tower jobs. That's okay. I don't need a pretty station to be competitive in contests and the DX pile ups.

Despite all the design and prototyping I continue to make many mistakes. Maybe I discover that a design is deficient and I have to research and try alternatives. Or the design is fundamentally sound but impractical when implementation is attempted. Worse, the design and implementation work, it goes up high on a tower and then fails because it isn't immune to abuse by the weather or the flow of high power. 

I have not been shy about highlighting my mistakes on the blog. And there have been many! I believe my biggest mistake was to attempt to make a hardware-based antenna switching system. At the time I believed that this would be a relatively straight-forward first step toward a software solution. I was very very wrong! It was mechanically and electrically complex, inflexible, overly large and difficult to use.

I recognized my mistake and (reluctantly) discarded it. However, my first software solution was also deficient. I am still using it but it is due for a redesign. The UI (user interface) is confusing to guest operators and even to me at times. I know because I keep pressing the wrong buttons for the wrong radio or antenna. The next version will be much better, I hope. 

There is no shame in making mistakes. That's how we learn. What is shameful is hiding mistakes or denying them and continuing on as if all is well. No one will be fooled by your evasion. Laugh at yourself, accept the lesson and try again -- to err is human.

Other examples of mistakes I've made, and the lessons learned, can be found in the blog. A few examples are gamma matches, 160 meters on the 80 meter vertical yagi, blazing the wrong trail for a Beverage antenna, breakage of scrounged aluminum, and even the choice of manually-tuned tube amps in a station intended to be used by guest operators. Searching my blog for the word "mistake" tags a lot of articles! Many of them are for operating mistakes or small things like forgetting to tighten bolts, but in this article I'll focus on design and construction.

I will also overlook ugly construction. When the ugliness is 100' up the tower, no one will notice but me. When it's on or under the operating desk, I can always cover it. No one dares ask what's underneath the cloth -- most hams have one or more ugly contraptions in their stations.

Analysis paralysis

You've certainly heard this phrase. It means to delay or avoid doing something because there are uncertainties or alternatives. Deciding what is best, let alone perfect, can be fruitless since there may not be one. This is especially fraught when the requirements are skimpy or vague, as they often are for projects hams are wont to build. That is, we really aren't certain what we are trying to achieve. That may be okay in some cases since exploration can be a good way to learn.

The key is to think through the options and research alternatives as well as you are able, then do something. You may choose a poor or non-optimum path, but so what? We aren't building rockets or nuclear reactors where failure isn't an option. When an unforeseen difficulty arises, correct your course or throw it away and try a different approach. Don't stubbornly persist with your first choice, or abandon the project because you are consumed by doubts. 

That said, I strongly advise against jumping into construction with no analysis at all. Many hams work that way, and I did as well when I was much younger. Impatience can be a curse, leading you into the wilderness without a compass. You may find yourself taking two steps backward for every step forward. 

Do at least some planning and design work before jumping into construction, which is usually the fun part. Know your destination and keep aiming at it no matter the bends in the road along the way.

Devil in the details: "minute details can have a big, often negative, impact"

Often the reason we get caught in analysis paralysis is that there can be so many options and details in a project. This is as true for an electronic project utilizing software or PCBs with dozens of components, as it is for towers and antennas where questions about static and dynamic loads, tensile strength and other challenging questions are critical for durability and safety. 

Don't be afraid to ask for help. But for many in-the-shack projects you can work it out without serious risk other than a little money and your time. However, don't simply proceed by trial and error since there are fantastic and free design tools for you to use, and online fora where you can ask questions. Those can help you to avoid many elementary mistakes. Examples of design tools include EZNEC, LTspice, KiCad, Elsie and so many more.

That said, a custom design can entail many pesky details that you will have to work through. The shape or position of one part can impact the success of the project. Let me illuminate this with a project that I am currently working on: a 40 meter reversible Moxon antenna. Let's look at the element switching system component where a relay changes each element to be either the driven element or a director. Both elements have the same component but with the roles reversed.

The linked article contains the switching system design which you may want to review before continuing. The only item from that article I'm reproducing here is the element switching schematic. The central coax switch for selecting the driven element is not included in this discussion. That's for the wire version, while the NEC5 design for the rotatable version I'm building can be found here. Full antenna detail will be published after the antenna is up and working.

On the right is the centre feed point of one of the elements, as it has been assembled in the hay field. On the left are most of the parts for the switching system that will be placed at the feed point. There are a DPDT relay, non-conductive spacers (to support the reflector coil and element connector tangs), solder lugs, custom made aluminum tangs to connect the switching system to the element halves, an ABS enclosure and a coax connector. Here are some of the details I've been considering in the design:

• Minimum enclosure size to accommodate the parts that permits good RF design and that minimizes wind load. It must be strong enough to withstand the weather.
• Minimize lead lengths to the connector, coil and element for low stray inductance.
• Fits over the element hardware and can be secured, while allowing ease of attachment and maintenance. Those tangs will be formed to fit under the hose clamps.

Consider a few of the many details that I must address:

• A high-Q coil tends to have equal diameter and length with space between the turns. For the required 1.2 μH coil those dimensions are on the order of 3 to 4 cm. The fit is tight so I've made the length a little longer than the diameter.
• To adjust the coil (mainly due to lead inductance), one tail must be solid (for support) and one flexible. That is, to compress or expand the coil.
• If the support spacers are too close the aluminum tangs must be positioned outboard of the screws, and inboard if the supports are wide apart. I need them close so that they can be bent outward to fit under the hose clamps (projecting outward from element centre). That increases the risk they could touch, but will reduce stray inductance. Also, the tangs must be appropriately sized (width and length) so that the entire enclosure can be rotated in and out of the hose clamps for insertion and removal.
• Will the enclosure stand on the tangs or be tight against the hose clamps or the large element u-bolts? For the former, the risk is that wind load will fatigue cycle the aluminum tangs and ABS enclosure. For the latter, the risk is that weather and the pressure of the clamps on the enclosure bottom could deform the plastic which might eventually crack.
• Stiff wire must be used for internal wiring so that the relay is held in place by the wires. This is the dead bug style I used in similar cases such as auxiliary coax switches and stack switches. Since #18 solid copper wire has worked in the past I'll use it in this project.

There are other details which I've omitted; there are a lot of them! Perfection is impossible but analysis is still prudent. Yet I will confidently predict that there will be one or more mistakes made in my design and construction choices. 

One mistake already discovered is that the plastic spacers (threaded for #6 screws) won't hold the screws very well when adjusting and torquing the top and bottom screws. Therefore I will place nuts on the bottom screws (for the tangs and solder lugs) to hold them securely. The coil solder lugs on top don't have to grip the spacer as solidly. The screws will have to be cut to measured lengths so that they grip well but don't get close enough to arc or short inside the spacer. 

At this point readers may wonder why I'm going into this excruciating amount of detail. But that's just it: the devil is in the details. Even a simple project entails a large number of decisions about details that will have consequences. Those choices are inter-dependent since they affect each other.

All of us must make many choices, consciously or not, when designing and building any project. When a problem is inevitably encountered, we revise and improve. Mistakes are just part of the process.

There is no escaping the details. Buying a product instead of home brewing also entails details in its installation and use: there are interconnection and system details to be worked out. A station like mine is very complex and it can be a struggle to make everything work in harmony.

Accept that mistakes will happen

Saying that mistakes happen is unsurprising and uninformative. What I tried to communicate in this article is to accept their inevitable appearance no matter how cleverly we design and build the contraptions we use in our stations. 

But it is important that we learn from our mistakes; that's one of the wonders about our technical hobby -- that there is so much to learn. Learning can be fun and fulfilling even though we may cringe when a mistake is discovered dangling from a tower or by the acrid smell of burnt electronics.

We should do our due diligence, while being careful not to obsess over preventing every possible point of failure. There lies paralysis, and that's never good. Move forward by trying something when you're stuck, or get a second opinion from someone with greater experience. 

The only time to stop is if a mistake would risk life or property. That's one time that you are advised to seek an expert. Asking for help is not a sign of weakness. Nor is admitting and correcting a mistake.