An Unwelcome Guest: Self-RFI

The modern world is full of devices which generate RFI. Few people notice because hardly anyone listens to OTA (over the air) broadcasts these days, or when they do it's local rather than fringe reception. As hams we notice. On the bright side, we can now operate a kilowatt in a densely populated area and not have neighbours pounding on our doors during a major contest. How times have changed.

Unfortunately the RFI we suffer is often self-imposed. We use the same devices as every other household and we experience their negative effects. Even though we can turn our devices off, add filters or carefully screen our purchases, problems will persist, either because of what devices our family members have and use or because amateur radio equipment also generates RFI or is not immune to its effects. We can be our own worst enemies.

It is especially annoying when the guilty equipment is found in our shacks. Transceivers, computers, audio peripherals and more can be the source or victim of RFI. It occurs more often than you might expect. When I was preparing the station for CQ WW SSB there were several issues that I spent time tracking down and resolving; some are so intermittent that they might still be lurking to show up later.

In this article I'll mention what I'm calling "self-RFI" problems that I dealt with leading up to the contest. I had allowed some to fester since I was the only one affected. I didn't want guest ops to have to be taught the how and why of working around the issues. It should be my problem, not theirs.

Audio rectification is a common form of RFI. With two kilowatt transmitters active at the same time there are many opportunities for it to manifest. It happens even though, for the most part, the audio connections to our rigs are designed to block RF or bypass it to ground. Headsets have two audio paths, mic and headphones, at risk.

I keep a bag of ferrites near at hand for dealing with these types of issues. It takes less than a minute to wrap the cord around a ferrite RFI suppressor. When it happened with the Koss SB45 headset a few years ago, the problem was resolved in less than a minute. I kept it there ever since. It's a little heavy so I try to keep it on the desk rather than dangling in midair, where its weight can damage the cord and pull down on one ear. 

Thankfully its mate on the other radio and those brought by my guest ops didn't experience the same problem. But I made sure everyone knew where to find the bag of ferrites.

USB and video cables can radiate RFI due to the high data transfer speeds they operate at. They are similarly susceptible to a high RF environment such as our shacks -- if it can transmit it can receive. Good quality cables are fully shielded and grounded at both ends, and protected from common mode with ferrite cores at both ends. Quality is such that they don't always work or work well enough in a high RF environment.

This is the back of a small Windows PC running one of the stations. There are commercial USB and video cables with a built-in cylindrical ferrite core. Between them is the USB cable to the CAT connector on the transceiver with a ferrite that I installed. All have a matching ferrite at the other end. I added the cores to the CAT cable when I previously experienced intermittent drop outs.

I did the same for the rather long USB cable to my home brew Arduino-based antenna selector. Again, there were problems before I made the addition. 

You'll notice that I selected rectangular (split) cores designed for HF and cheap import cylindrical snap-on ferrites with unknown specifications. The latter work pretty well so I keep using them. Perhaps I got lucky since their uncertain source can make for a risky online purchase. 

The USB ferrites did not fully protect the antenna selector. Therefore I added another to the 12 VDC power cord the day before the contest. There were no RFI incidents all weekend. It can be difficult to say whether that was the work of the ferrite suppressors or just luck. Regardless, they will remain. Fully encasing the antenna selector would be even better but that would be difficult due to the multitude of connectors. Perfection is nice but not always necessary.

I saved the most curious case for last. Ever since acquiring the Acom 1200S I experienced occasional problems with the control screen and buttons due to RFI. It happened on most bands and antennas but not all. There was no obvious single cause of RF ingress. 

The first symptom was that the buttons would fail to work. However the amp continued to operate normally. Band changes had to be by RF detection (I don't use external band control at present) and there was no ability to switch the amp offline (OPR/STB). Although it's annoying there is no real need to do anything, just keep operating.

The second symptom, always preceded by the first, is that the screen would freeze. Although the amp continued to operate normally, there was no way to know the power output, reflected power, or PA temperature. That's a bigger problem. You can probably continue with the same band and antenna but eventually the lack of feedback becomes unacceptable.

The third symptom, always preceded by the first two, is that the screen goes blank. Now you're flying blind even though the amp continues to operate normally. That is, for a very liberal interprettation of "normally".

The cure is simple: cycle the power, since that button at least still works. That can take up to 30 seconds, which is okay most of the time but not while running in a contest. You'll lose contacts, and your run frequency.

There are only 4 connections to the amp: RF in, RF out, PTT in and mains power. Putting a ferrite core on the PTT line had no effect. The day before the contest I added two cylindrical and one rectangular core to the power cable. I included the cylindrical cores because only a few turns of the thick power cord fit onto the rectangular core.

I added all three at once rather than test each addition since I was running out of time to troubleshoot the problem. Luckily that did the job. There were no amp glitches due to RFI for the entire weekend. 

There was one more issue that caused a headache during the contest. Whether that is considered RFI is a judgment call. Let's see what you think.

With the 15 meter stack pointed to Europe and/or Asia, those antennas and the south pointing TH6 on the north tower point at each other -- they're 60 meters apart. Of course the TH6 isn't used on 15 meters when that happens since the other radio must be on 10 or 20 meters to use the TH6, however it's a tri-band yagi that will happily accept 15 meter signals. The 1200S sees a few watts of RF at the output connector and decides it's a fault. The protection circuit takes the amp offline.

That was really irritating. We could not use the TH6 when that happened. Had the TH6 been on the other radio (band selections reversed) there would have been no problem since the Drake L7 has no protection circuits and in any case doesn't care.

Unfortunately there is no good solution to the problem other than upgrading my station to use high power BPF (between the amp and antenna) rather than the 6-band switchable low power BPF that I use. That's an expensive solution to an intermittent problem. There are alternatives that I will have to explore. This new problem and other ongoing issues may spell the end of the line for the TH6 in my station.

I fully expect that I have not seen the last of self-RFI issues in my station.

CQ WW SSB 2025: M/2

After skipping CQ WW multi-ops in 2024, we're back! It was quite a chore preparing the station to host guest ops but somehow I got it done in time. Although there were so many loose ends that I expected many technical problems to crop up during the contest, they didn't. The problems that did show up were quickly resolved. The work was responsible for the gap in my blog activity this month.

The setup 

The station is similar to what we used for M/2 in 2023, with a few important differences. The Acom 1500 tube amp is out for repair and has been replaced by the Acom 1200S. A no-tune solid state amp has advantages but is not without issues. These are discussed later in the article. The Drake L7 continues to power the second operating position. 

The IC7610 replaces the FTdx5000MP in the left position. The more recently acquired 7600 is on the right with the L7. The advantage of using similar rigs is that operators are less confused when they change chairs. They aren't exactly the same and that caused its own confusion. I am unwilling to invest in a second 7610 that will only be used several weekends each year, if that. I made adaptors so that any electret headset with stereo 3.5 mm plugs would work with both radios. I supplied headsets or the operators could bring their own.

Antenna selection software is unchanged, with the GUI on the middle PC (screen is off). There wasn't time to rewrite the software for this contest. I started but there is a long way to go. Winter may provide the quiet time I need to get it done. Despite the need to reach for the mouse and glance far to the side, there weren't many operator errors. 

The old prop pitch motor controller was replaced with my home brew Arduino-based controller. All the rotator controllers are centrally located, within reach of both ops. I have a ground issue that affects the rotators and especially the home brew device. Until that's resolved I added a software tweak to reset the 1602 LCD after each use since the display is easily corrupted by power issues (a known problem with that module). 

Software GUI for rotator control might be better (I've used those elsewhere) but are not a priority. They are more necessary for remote operations which is not currently planned for my station.

The operators sit close to each other so it is vital not to shout into the mic. As contesters we do tend to do that while chasing multipliers even though it is not only unnecessary but can cause distortion. Perhaps more importantly, it disturbs the other operator. Idle ops also needed to converse quietly or leave the room. We weren't shy about telling others to shut up or leave. I was guilty more than once!

The team 

The only returning operators from 2023 were myself and Dave VE3KG (sitting at the left radio). Shel VA3AA (right radio) was keen to join the team and willing to drive the several hours from west of Toronto. In the middle is new contest op Matt VA3UMM. I can be seen pointing out a few things to get him oriented. 

After listening in for a while (audio splitter) we set him up to run stations for a few hours. That was a gentler introduction than S & P since the latter requires the additional skills of tuning, monitoring and understanding spots and mults, selecting antennas and more. He did very well. Expect to hear more from him in the future.

The contest 

 Band     QSOs     Pts   ZN  Cty  Pt/Q
   1.8      89     176    9   18   2.0
   3.5     441     937   18   63   2.1
     7    1933    4914   34  119   2.5
    14    1479    4152   36  139   2.8
    21    1989    5682   37  137   2.9
    28    2072    5916   36  150   2.9
 Total    8003   21777  170  626   2.7
Score: 17,334,492 

As you can see we did pretty well. It helps to have a big signal since few will go out of their way to log another VE3. Only one station thanked me for the mult. That was on 160.

From the average points column (on the right) it is apparent that we worked many US stations (2 points) on the low bands. We would ordinarily work more except with little DX appearing on 80 and 160 there was no reason for them to be there. Presumably sleep was more important than working a few Canadian stations (very understandable).

On the high bands it was mostly Europe (3 points). We paid close attention the N1MM AMQ window to chase spots whenever a CQ went unanswered. Since there are no SSB skimmers (yet) it is important to spin the VFO when the rate slows since many stations, including mults, are not spotted. We were configured to spot all S & P QSOs, which is a feature everyone should use in phone contests.

We had enough operators to keep both chairs occupied for the full 48 hours. The only exception was overnight when enduring the low activity could become so bad that it was refreshing to periodically walk away for a few minutes. During the day an operator would immediately fill the chair when an op needed to take a break. We made no schedule -- my only rule was that if you've been operating for some time and you see someone hovering over your shoulder, it is polite to cede the chair to them. I have an intense dislike of rigid schedules, as do many of the contesters of my acquaintance.

My preference is to let guest operators do most of the operating. This weekend I took on both overnight shifts to let them rest and be fresh for the high band runs to Europe. I get to operate as a "big gun" every day, including most contests throughout the year. Other than Vlad VE3JM, who's station is a similar size, the others don't regularly have the opportunity. 

I manned my new espresso machine to happily act as barista. I was regularly slinging single shots, double shots and Americanos. We had so much good food on hand, healthy and otherwise, I suspect we all gained a little weight over the weekend.

The station

The day before the contest I found it necessary to make a few last minute station changes. Over the past 2 years there have been many configurations, equipment changes and software updates. There were a few unpleasant surprises while I was setting up for M/2 operation. I really ought to have started sooner and saved myself the anxiety. A few hours of intense effort had the computers properly networked and my home brew software working as they should.

One last minute headache was recording the function key messages for the new operators. So much has changed within Windows, N1MM and equipment interfaces. Luckily we found workarounds at the last minute and got everyone set up just 15 minutes before 0000Z.

One thing I did get done in advance of the contest was ridding the station of RFI. I put ferrite cores almost everywhere I had a known problem. That seemed to work since, unlike in previous contests, there were few issues this time. This is an important topic that is worth its own article so I'll defer the description of what I did and why until then. 

Outside the shack, there were numerous antenna jobs to be completed. With the return of cold weather I was able to go into the bush to repair Beverage receive antennas damaged by storms and summer growth. I did that 3 times in the week preceding the contest. 

The radials for the 160 meter shunt fed tower needed to be rolled out for the fall/winter season. I completed that Friday afternoon. A few radials were made of narrow gauge wires which got horribly tangled while being unrolled. I got but one untangled. I don't think we lost too much effectiveness on 160 by having 15 instead of 16 radials!

I ran out of time finishing the new rotatable side mount for the lower 15 meter yagi. I wanted it to have stacking gain toward east Asia and the Pacific. The QSO potential is low but there are many multipliers to be sought. For us, these paths tend to be marginal so that another 3 db of gain would be very welcome. I should have it ready for the CW weekend.

The 40 meter reversible Moxon performed admirably. This is rapidly becoming my favourite antenna. It freed the 3-element yagi for the longer path openings such as to Asia and the Pacific, and the occasional long path opening. They make a great combination. 

On 80 meters we were mostly restricted to the omni-directional mode of the 80 meter vertical array. Its yagi modes work no higher than 3650 kHz. This is a deficiency that I have on my to-do list. Those with 4-squares have gain and directivity from 3500 to 3800 kHz. I have to work a little harder to achieve that.

The 80 meter inverted vee was problematic since its SWR is high on most of the SSB part of the band (3700 kHz and up) and not really usable with the solid state amp. The L7 was fine with both the inverted vee and the 40 meter 3-element yagi in the SSB segment where its SWR rises. I used the L7 and 3-element yagi with great success on 40 to work the Pacific and east Asia after sunrise.

One antenna that did not work to expectations was the TH6. It is fixed south as a multiplier antenna to work the Caribbean, Central and South America. It may have a loose connection since the 1200S often went offline when using it, even in those parts of the high bands where the SWR is low.

Lessons learned

The Acom 1200S is not a suitable amplifier for contesting. It is generally a good performer but with many quirks that emerge during contests. The list is a long one that I won't delve into here. This is not an indictment against solid state amps or of Acom. LDMOS devices have limitations. I will have to give the matter some thought. The amp is unlikely to be replaced anytime soon.  

I'm increasingly unhappy with trapped and inductive loaded yagis. In comparison to other antennas, the TH6 and XM240 perform poorly, with narrow bandwidth and loss higher than I'd like. The XM240 is on the ground and I am unlikely to put it back up as is. Either I sell it or convert it to a Moxon. The TH6 is a greater dilemma since I have nothing handy or economical to replace it with. Having a south pointing tri-bander remains very attractive for multiplier hunting without needing to turn other yagi. Again, I'm not yet sure what to do; there are several alternatives I'm considering.

The 80 meter array really needs to become a yagi on SSB. Having just an omni-directional vertical cost us contacts and multipliers. Maybe I'll get there in time for next year's CQ WW contests. Certainly the low bands will become increasingly important in the coming years as the sunspots fade away. As noted in an earlier article, I am proceeding with improvements methodically, step by painful step. 

Accurate logging isn't easy for new contesters when the rate is high. It is too easy to send/say the call correctly and type something different. Also, many stations will not correct their call when you send it with errors -- their score doesn't suffer. Experienced contesters recognize calls and know how to use the check partial feature to confirm calls, and will check that the pre-filled exchange is correct (zone in CQ WW). It is easy to forget how complicated this can be for those new to the game. New operator training is key.

The same goes for propagation, knowing where and when to look for rate and multipliers. There is too much to remember from past experience. Top contesters make notes based on their and others' previous experience and from public logs. Being a fervent DXer is not sufficient to do well in major contests.

Inexperienced contesters run more effectively than they S & P. Just point and CQ, and watch the contacts fill the log. Hunting requires a wider range of skills and an ability to adapt to the prevailing conditions. We have lots of S & P features integrated into the major contest logging applications that make it easier to spike the score when runs subside. Learning how to use them and coming up with effective tactics can be difficult skills to acquire.

Sharing a contest experience with good friends is tremendous fun. However, bringing new contesters into the fold, whether young or old, is tremendously rewarding. Get over the anxiety and make it happen. We can all do our part. It's good for us and for the future of the hobby, and contesting in particular. I intend to do more. 

An memorable contest is the sum of many small stories, not simply the final line score. There is one small story that I'd like to share that I found amusing. On 40 meters at sunrise I was running west coast and Pacific stations, with occasional forays to work multipliers. While running I was called by a very weak station starting with an F. Another try brought a K and then, I believed, a 5. I thought it was FK8IK, except that would have been a dupe. After working a few W6/7 stations he was still in there trying to get my attention. I was finally able to determine that it was FW5K. That was a little embarrassing. I thanked him for sticking around and giving me the mult. This wasn't the first time I was called by a DXpedition from a rare entity during a contest and I'm sure it won't be the last. Running can be a great way to rack up the multipliers. 

TTS (text to speech) is rushing at us like a freight train. Within a year we should expect most of the best contesters to be using it to break records in phone contests. The rapid integration with N1MM is driving its use. Let the computer send the calls and exchanges while you copy calls and exchanges on the other radio. SO2R and 2BSIQ will become accessible to everyone, just like on CW and RTTY. No longer will you need the rare talent of being able to listen and speak at the same time.

Top tier contesting is expensive! In this blog I've shown how it is possible to design, build and raise towers and antennas on a budget. Most won't, but it can be done. But that's just the RF side. Once you enter the shack the challenge is completely different. Transceivers, amplifiers, networking and control systems are not so easily built or purchased cheaply. If you want it, be prepared to spend. 

I've reached the point of diminishing returns outdoors and now I face a dilemma: I have to decide how far I want to go, and how fast, depending on how important this side of amateur radio is to me. There is no easy answer. It can't be done with a small budget. 

Wrap up

We were all smiles at the end of the contest. You really can't ask for more than that. We'll win no plaque but we challenged ourselves and did very well in the M/2 raw scores ranking. Positions may shift after log checking but that doesn't matter too much. 

VE3 is not an attractive prefix in the major contests. Our geographic position, while better than some, will probably never host an overall winner. We do what we can with what we have and learn to appreciate what we get. For me, that's enough. 

Of course I'll use the lessons learned to make improvements but those are to do better in future and not with the dream of winning. Being the top M/2 in Canada wins no accolades, or plaques.

There will be only a few changes when CQ WW CW arrives a few weeks hence. I'll probably do another single op entry though that might change. I like to challenge myself. However, improving my score over last year will be really difficult, if only to endure for 48 hours while doing more 2BSIQ than ever before. This may my last chance since the solar cycle will decline over the next several years and I will be old when the next one, perhaps my last, arrives during the next decade.

Whatever I decide, I hope to work you in the upcoming contests. 

Photo credits: VE3JM for the team photo and the one of me on the tower.

Rebuilding the 80 Meter Vertical Yagi: First Steps

I have been planning improvements to the 80 meter vertical yagi for a long time. Too many higher priority projects have kept it in the background for several years. Also, despite its complexity it has been one of my most reliable antennas. This fall I am taking the first steps to make it better.

I am proceeding in a set of carefully chosen steps. It is critical that after each step that the antenna work; I can't afford to have my best 80 meter antenna offline for long, especially now that the fall contests are rapidly approaching along with low band DX season.

The steps are as follows:

1. Install ground anchors to support a taller tower as the driven element and parasitic element support. These are set back further from the existing augur (screw in) anchors in accord with the planned tower's greater height. Since these are separate from the existing system there is no impact on the antenna.
2. Replace the radial plates for all 5 elements. My original design is not aging well.
3. Replace the parasitic element supports with more robust structures that repel wildlife and permit easy access for maintenance.
4. Replace the tower with a taller and stronger one. The intent is to eliminate the tall stinger at the top of the tower and provide robust support for an improved 160 meter mode for this antenna.
5. Replace the parasitic sloping T-elements with bent (sort of L-shaped) wire elements. The objective is to increase efficiency by raising the radiation resistance. This is the first step that changes to the electrical design, including the matching networks.
6. Add SSB to the yagi. The wiring and control structure has been present from the start but not used. Above 3650 kHz only omni-directional mode is available. There are alternatives I am considering to conveniently switch the yagi's operating range to 3650-3800 kHz with minimal loss of efficiency
7. More radials. Combined with step 5, from 0.5 to 1.0 db improvement can be achieved by doubling the radial count. The parasitic elements in particular need help since there are at present only 16 for each of them.
8. Improve the antenna's 160 meter mode. I'd like an effective 160 meter antenna available year round. The radials of the shunt-fed tower must be removed during farming season so I miss several months each year, from approximately May through September or October. The base loaded 160 meter mode of the 80 meter antenna is only minimally effective.

When complete, the antenna will be stronger, perform better, and be more useful. Although scrapping it entirely to replace it with a 4-square might be preferred by most, I really like the yagi. It is mechanically simple, takes up less space and is ripe for experimentation. I am willing to cede 1 db so that I can have some fun with this antenna project. If you've read the blog long enough, you'll know that I am more of a builder than a fervent competitor. 

The first step is now complete. Steps 2 and 3 are planned for this fall. The 4th step completes the mechanical rebuild, which I might not be able to accomplish this fall. The weather is getting colder, and sub-freezing temperatures will make concrete work difficult. The tower work would also have to be fit in between contest weekends. That's a tall order.

In this article I'll briefly review installation of the new anchors since that may be of interest to those with a similar challenge with small guyed towers. Then I'll address the other problems with the antenna that I hope to fix. Some things you only learn after the antenna has been in place for a few years. I've taken several temporary measures but, well, they're temporary and not good enough for the long haul.

The dead man anchors for this rebuild are identical to one that I've discussed before. There are a few differences. One is that with over 1600 meters of radials it is inevitable that one or more will be encountered when you stick a shovel in the ground. The picture shows one such case. I started digging with a hand trowel, carefully lifting the vegetation layer (it's a hay field) while checking for the presence of radials. Originally laid on the surface, they have become lightly buried over the years. Only one of the 3 holes didn't strike a radial. I worked around them rather than pulling them out of the ground during the work.

The angle and alignment of each anchor is critical to best performance. They are located the usual 80% of the tower height from the tower base and inclined to an angle midway between the two guy stations. For one guy station it should be inclined to point at that guy station. 

The angle in my case is 42°. I used an inclinometer to set the angle, approximately before the pour and then exactly after the concrete has been poured and is still wet. There is the possibility of anchor movement even if you're careful pouring the concrete. Don't use a quick setting concrete mix for this job!

The anchor must also point directly at the tower. A string or thin rope is tied to the tower leg that will be guyed and pulled taut to the anchor and a marker behind the anchor. The foreground steel rod was aligned with the bottom of the anchor using a level. It's easier to align the anchor this way than trying to run the string to the bottom of the excavation.

You can see the existing screw (augur) anchor in the background. Notice that it is not quite in the correct position. I hit a rock when auguring it into the ground so I moved it a few inches. Several large rocks were found while digging the hole for the new anchor so I moved it back about 8". Small deviations from perfect anchor placement are allowed but don't use that as an excuse to be sloppy.

A block of wood and metal rods keep the anchor in its aligned position while the concrete cures. They can be removed a day or two later. That is also a good time to back fill the excavation. There isn't a lot of organic soil in this field, which is typical for the local geology. I put the soil and the lower material in separate piles so that while back filling the good soil is put back on top. The lower material displaced by the concrete and no longer needed is dumped elsewhere. Heap the soil high during back filling so that you aren't left with a depression after it settles. You might also want to tamp the material down as you refill the hole

While the concrete cures, let's review the remaining issues and steps that will come later.

There are two problems to be seen in this picture. The first is the preserved wood base. It's showing its age. It is slowly decaying and weeds are growing between the logs and at their edges. Replacement isn't urgent but must be done at some point since it supports the tower. The tower is isolated from ground and because it is not grounded there is no need for a concrete foundation. However, I will put in a 4' × 4' concrete pad when the tower is rebuilt (step 4) to keep the base clean and ease maintenance.

The radial hub is simple and easy to use but has proven not to be reliable. Electrical contact degrades over time for two reasons. One is that the copper oxidizes and the contact to the stainless hose clamp isn't strong enough to keep oxidation out of the contact area. 

The second is that the ring is "bumpy" so that a slight bend of a wire or unequal spacing of wires under the clamp causes unequal pressure on the wires. That leads us back to the first point since some of the wires are a little loose and there is no way to clamp them more securely. I've had to clean the wires and re-tighten the ring a few times. The same is true of the 4 parasitic element radial hubs.

All 5 radial hubs will be replaced with sheet aluminum and stainless studs for the radial wire connections. I will do that soon for the driven element (tower base) and later for the parasitic elements when I redo their bases. 

The bases of the parasitic elements are all protected as shown at right. It's ugly and makes maintenance difficult. But it's necessary. Deer and even coyotes love to gnaw on the antenna wires, control cables and the PVC enclosures. As currently constructed they are difficult to protect properly.

They will be replaced by wood posts in concrete and a small concrete pad. The PVC enclosures will be mounted on the posts and the antenna wires replaced by aluminum tubes up to at least head height. Those measures and a ring of galvanized mesh around the base will solve all of these problems while leaving the parasitic elements accessible for maintenance.

The tower sections have been stored for the past 3 years. When combined with the two top sections of the existing tower the total height will be over ~62' (~19 meters). It will need only a 2 meter mast to tune the tower to the same value as before, which should allow no change to the existing matching networks at the tower base. I will have two guy stations due to the greater height and a possible 160 meter stinger. The same considerations guided the design of the new guy anchors.

The stinger will no longer be needed. It has proven to be the weakest part of the 80 meter array. A short mast on the tower will be more than strong enough for the wire element support ropes. I will no longer have to be so careful when tensioning the ropes. Now when I periodically need to do that maintenance, the stinger and the insulating support pole above can bend, with the risk of excess stress leading to failure.

As mentioned, the tower will eventually have a stinger that will be used for 160 meters. There are electrical considerations to ensure the stinger doesn't strongly interact during 80 meter use and managing potential high voltages across relay contacts. I have a few alternatives to solve those issues when I am ready to proceed. But for now I intend the tower to be a plug in replacement with no changes to any of the matching networks.

Items 5, 6 and 7 will follow after the major mechanical upgrades are complete. I want to delay changes to the electrical design of the array until the mechanical rebuild is complete. It's a sensible step by step process.

Step 5 is quite simple. All that is required is to change the T-top wire parasitic elements with bent elements in the shape of an open L. Modelling suggests an improvement in efficiency due to the higher radiation resistance of the latter. It comes about because of the acute angle on the lower side of the T which causes field cancellation with the vertical section. Inverted vees with a small interior angle can suffer a similar fate.

The main benefit of the higher radiation resistance is that the series ground loss (radial system and soil) becomes smaller relative to the sum of radiation and loss resistances. Each new parasitic element will be tuned to the same director and reflector self-resonant frequencies as the existing array. It will require experimentation to determine the length of the upper segment of the element. Once the first parasite is tuned the rest should be identical. The length of the vertical segment will remain the same so that I can use the existing element anchors. In any case, geometry (tower height and antenna field perimeter) doesn't allow much variation of the wire angle.

Higher radiation resistance and a lower ground loss by doubling the radials for each parasitic element will change the feed point impedance of the array when used as a yagi. The matching networks will have to be adjusted or replaced. That will be determined by measuring the impedance after the changes are made.

Adding SSB to the yagi can be done with an additional series reactance at the base of each parasitic element. This can be a coil (shorted for the 3650-3800 kHz range) or a capacitor (shorted for the 3500-3650 kHz range). The latter can be more efficient but is perhaps more difficult in practice since suitable capacitors are expensive and unusual to run across at flea markets these days. 

I'll close by mentioning the help I receive from a couple of friends. Chris VO2AC/VE3FU transported almost a half ton of concrete mix with his pickup truck. The 30 kg bags are cheap but not easily transported in bulk with a passenger vehicle. Dave VE3KG come by one morning for the mixing and pouring of the concrete for the anchors.

The bags in the picture are what was left over after the anchors were done. It may be enough for the 4 parasitic element bases. More bags will be needed for the tower base. The rebar has tape markings at 8" (20 cm) and 30" (75 cm) for convenient measurement of anchor width, concrete depth and excavation depth. 

40 Meter Reversible Moxon - Complete

The antenna on the tower and working. Finally. I've been using it for the past few weeks and my overall assessment is that it is a very good antenna, better than I predicted. In this article I'll review the design, the repairs and what I've learned about its performance. Instead of repeating what has come before I'll sprinkle the article with ample references to earlier ones. The antenna has been a regular visitor to the blog for quite some time!

The reversible Moxon replaces the XM240 on the rotatable side mount. Despite the limited rotation, since it's reversible it covers 275 of the compass, from southeast to west (default direction) and from east to northwest (reverse feature enabled). The dead spots are between west to northwest and east to southeast, which from here are the least productive directions. The 3-element yagi at the top of the tower has full compass coverage, including over-rotation when needed.

The antenna default direction allows coverage of the Caribbean, South/Central America and almost all of the US. That leaves its big brother (the 3-element yagi) to focus on Europe and longer path DX where it shines. Even so, the lower antenna performs better in all directions before sunset and after sunrise, and during geomagnetic disturbances, when lower elevation angles encounter D-layer absorption. That's due to their respective heights, not their relative performance.

As previously noted, this antenna requires NEC5 for accurate modelling. Although it took a lot of segments to get sufficient accuracy, the built antenna tracks the model exceptionally well. It seems to be well within 0.5% of its design frequency -- about 25 kHz. That's even more impressive considering the close spacing of the capacitance hat tips (30 cm) which increases sensitivity to inaccurate modelling.

Another critical measurement is the inductance of the coil that makes an element the reflector. Careful tweaking of the design resulted in a value of 1.2 μH. Values between 1.1 and 1.3 μH would be acceptable, provided that the two coils are equal. Otherwise the frequency range will differ from the other direction, which is undesirable. The measured difference is a very close 25 kHz. That's the same as measured for the sanity check with the antenna on the ground.

The test setup at right was used to measure the inductance for both element switch boxes. To roughly account for the short leads from the analyzer, I set the inductance a little higher than 1.2 μH. The inductance of the internal leads and relays requires no compensation since it is present during use.

To test the opposite direction, a 50 Ω load is connected to the coax connector. The stray inductance of the leads and relay has an effect, however it is only a few ohms of inductive reactance. That is low enough to be considered negligible.

The DPDT relays behave oppositely in each switch box, with one connected to the element in its NO (normally open) position and the other to the coil. The relay coil is energized to test the switch box in its opposite direction. 

You can see from the alligator clips to the 13.8 VDC power supply that I use a common ground (return) path for both RF and DC. I do the same for all my switching systems. Therefore only one conductor is required to energize the relay coil. 

A stainless screw stud serves for the relay power. A 1N4007 diode in the centre box (see below) acts as the flyback (suppressor) diode for the relays in all three switch boxes. This is possible because to reverse direction all relays are energized.

One of the faults that I had to repair was the relay in one of the element switch boxes. I was surprised to find that the relay coil was essentially shorted: 5 Ω rather than the expected 340 Ω. I don't know how that happened since it worked fine on the ground. Repair of the relay isn't possible since it's a sealed unit. I've never had a problem with this line of Omron relays for RF switching. I tossed it into my NFG bin and replaced it with my last DPDT relay (I'll have to order more). 

The other fault was in the centre switch box. The stud for the relay coil loosened, allowing the internal lug to touch the enclosure wall and short. It's a tricky bit of hardware since the DC path has to pass through the metal enclosure. 

I discovered that the insulator was too high and could not be properly tightened. I replaced it with a plastic washer that has an inner rim to centre the screw within the hole. A thin layer of silicone caulk is used as a moisture seal.

Do you recognize the enclosure? I reused the one for the upper 20 meter matching network to compensate for a misbehaving gamma match. After I repaired the yagi's gamma match the network was removed. All it needed was a third coax connector and a relay stud.

The capacitance hat clamp hardware was tightened. I had forgotten to fully torque the nuts before its first raising in the spring. The clamps rotated when the tram line was drawn over them. While the yagi was on the ground I made a few other hardware adjustments to the antenna. It should now be far more robust.

The same crew from the earlier raising (and lowering) happily came over one more time -- it's good to have friends. Hopefully this is the last time we need to raise the Moxon for a long time to come! 

The only rigging change was to run the tram line under the trailing capacitance hats. As usual there were steering issues to fit the elements between the guys immediately below its mount on the tower and the TH6 above. We used tag lines and a lever on the boom to tilt up the leading element tips. It took a lot of force on the rope connected to the end of the lever to rotate the boom of such a heavy antenna. I did the same for tramming the 3-element 40 meter yagi to clear the upper guys.

It was a bit of bother to get the boom attached. The antenna is too heavy to push it into position with one hand while driving the saddle clamps home with the other and reaching around the back to thread on the lock washers and nuts. I gave my friends what were to them very confusing directions on how to use the tag lines to bring the boom flush to the plate. But it got done just the same. The little HT's VOX feature was a perfect third hand.

With the antenna secure I measured the SWR in both directions (a friend in the shack operated the reversing switch). The SWR was near perfect but about 25 kHz different between directions. That agreed closely with my sanity check on the ground. You can look at the forward SWR curve in an earlier article and imagine the reverse direction curve shifted to the left a bit.

Unfortunately I could not easily grab screenshots on the tower at the time and then the connector was weather proofed. The above plots were taken in the shack when I was back on the ground. Different but still excellent. There are two likely causes. One is the ~80 meters of transmission line and several antenna switching relays between the antenna and shack. Although I use good and tested coax, mostly LDF5 in this case, some impedance deviation is normal. The other less likely but possible cause is that there is no CMC (common mode choke).

I debated whether to put one at each feed point, at the element switch boxes, or just one at the rotation loop. Since the SWR is fine at the latter point I'll probably put the CMC there. It isn't urgent. 

Now we come to the important part: how it performs on the air. The overlaid azimuth plots at right, in 75 kHz steps from 7.0 to 7.3 MHz, set the expectation. Again, NEC5 is used in the model with the antenna at its actual height over EZNEC medium ground.

The free space gain peaks at close to 7 dbi at 7.0 MHz and gradually declines as we move up the band. F/B behaves the same. Gain, F/B and SWR are really quite good right across the band.

When I had the XM240 on the same rotatable side mount during DX contests I would often aim the 3-element yagi to Europe and use the XM240 to work the US, South America and other southerly directions. Since, like most yagis, the XM240 is uni-directional so it only covered west to southeast (~130°).

As already mentioned, the opening to Europe before sunset favours high elevation angles. With the reversing feature of the Moxon I can quickly switch directions to take advantage of its lower height in this circumstances. It also does well after sunrise towards the west while I point the 3-element yagi north for long haul contacts into east Asia.

Recent conditions turned the tables on my expectations. With so much geomagnetic activity, low elevation angles encountered high absorption. That meant the Moxon at its lower height outperformed the 3-element yagi on many DX paths, including the all important one to Europe. That shows the Moxon's value but it made comparisons difficult.

It is fair to say that the Moxon's gain is better than the XM240 with its coil-loaded element. That isn't a surprise since the latter is a high-Q antenna that also has a narrow SWR bandwidth. This is best I can do to assess gain on the air since it is very difficult to measure differences of a db or two, especially with the XM240 on the ground!

F/B is far easier to assess with the reversing feature. It is excellent on all paths within the CW segment and up into lower part of the phone segment. Once we reach 7.2 MHz the F/B becomes quite poor, possibly worse than the model predicts. I have more testing to do so this is not the final word. However, the excellent SWR means that the entire band can be worked without an ATU, even with my Acom 1200S solid state amplifier.

I have extra coax coiled up for the runs to each element that could be used to wind on a ferrite toroid to make a CMC. I used longer lengths of LMR400 for routing flexibility, not for a CMC; LMR400 is too stiff for that use. When I added the outer shield of the coax to the model it suggested that common mode may be present. A poor F/B (pattern distortion) is the most obvious consequence. However, the excellent F/B lower in the band is evidence that disfavours that explanation.

Another possible cause is the interaction due to the proximity of the capacitance hat tips to the guy wires in some directions. The distance isn't really very close except in comparison to the 40 meter wavelength. Again, that seems like another unsatisfactory explanation due the antenna's excellent F/B low in the band. The same can be said for the proximity of the TH6 above the Moxon, even after we lifted the TH6 higher earlier this year.

Clearly I have more testing to do. Even if it is the F/B is poor high in the band I am not too concerned. The reason is that as a contester I value gain and SWR more than F/B for the potential contacts to be made off the back. I can live with it. I will add a CMC later at the rotation loop (not at each element).

A picture further above shows another oddity: the rotator. That's a Hy-Gain Ham-M that I've owned it for 40 years -- I refurbished at least twice -- and it's at least 10 years older than that. It is under-powered for this large antenna. What saves the day is the symmetry of the reversible Moxon. High winds, even those that excite oscillations, place very little net torque on the mast. 

The only real problem is the 105 lb antenna's momentum. You typically have to wait several seconds for the antenna to coast to a stop before it stops turning so that the brake can be safely engaged. If it becomes a problem I'll swap in a rotator with a higher overturning moment. But not this year since I don't have one at hand.

To summarize the antenna after a years long process is, to put it simply, I like it. I could live with the reversible Moxon as my only 40 meter antenna. It's that good. I am not surprised that many contesters stack 2 or 3 W6NL Moxons on 40 meters rather than take on the far greater challenge of a 3 or 4-element yagi despite the latter's better performance.

NEC5 made this antenna possible. I still marvel at how closely the model agreed with the built antenna. It took many modelling variations and tweaking to make a symmetric and reversible yagi that performed to my satisfaction. While few will ever buy or build an antenna of this size, I believe that I've demonstrated what is possible with enough ambition. Perhaps one or more readers will be incentivized to do the same.

And the XM240? I might yet raise it again on the 20/15 tower and point it south as a multiplier antenna (similar to the how the TH6 is being used). The Moxon's reversing feature makes that job less urgent. I'll mull the idea over the winter, and whether to stack the Moxon and the 3-element yagi.

I look forward to working you with the Moxon in the upcoming CQ WW SSB and CW contests.

Elevation Pattern Nulls

Propagation prediction for most hams is a terribly inexact science. You get on the air and tune around for stations, or you click on spots. You hear the DX or you don't; sometimes they're weak and other times they're strong. But why? Well, propagation. That's as far as many hams think about the matter. 

Maybe you try another band or just turn off the radio and come back another day. And if it's a contest weekend? You understand that's the playing field for every competitor so you persevere.

The reason HF propagation is difficult is because it is a complex phenomenon, much like the weather. We understand the physics but the data to make accurate predications is sparse or not collected at all. Ionospheric propagation is rarely as simple as depicted in this diagram above. There are numerous factors to consider, some of which include:

• Polarization and Faraday rotation
• Chordal hops, within and between ionosphere layers
• Geomagnetic activity due to the sun (UV, X-rays, energetic particles and more)
• Scattering (diffuse rather than specular reflection or refraction)
• Ducting 

Our responsibility as communicators is to design and build antennas that make maximum use of the propagation paths provided by the environment. If a signal is vertically polarized, we prefer a vertically polarized antenna. If the elevation angle is large, we prefer an antenna that has a high angle lobe. If we want to work over the pole we prefer an antenna that favours the north (or south) direction. 

The important thing to remember is that the environment decides the propagation parameters, not the antenna. It can change quickly and has pronounced daily and seasonal cycles, and that's on top of solar cycles and solar activity. 

Even simple antennas can have complex patterns. This is due to their size, orientation, height, terrain, and other factors, usually with respect to signal wavelength. One antenna is never good enough for all types of propagation. We need set operating objectives and build accordingly.

The azimuth pattern of a vertically-polarized 40 meter delta loop operated on 15 meters (its third harmonic) is shown at right. You need to install it so that pattern nulls are not placed at bearing you consider important directions, and that lobes are where you want them. Compromises may be unavoidable. Rotatable antennas are less constrained but they come at a cost.

A curious situation occurred years ago when I had a much smaller station. In a now ancient article I described my initial confusion when comparing two 40 meter inverted vees. I had two so that I could cover all compass directions. Faraday rotation and polarization dependence on the azimuth angle (bearing) made for a difficult comparison since an inverted vee is vertically polarized off the ends and horizontally polarized broadside. 

They were equally effective antennas but never at the same time due to Faraday rotation. That experience highlighted how antenna comparisons can be so fraught with uncertainty. You may think you're making a fair comparison -- WSPR, A/B switching, etc. -- but in many cases all you're doing is comparing antenna patterns relative to the prevailing propagation, not performance. Assessing relative performance is never easy since propagation characteristics can make a comparison worthless or misleading.

There is a saying among many hams: you can't have too many antennas. The more antennas you have the greater the probability that one of them will best exploit the propagation to a particular station or region at any given time and band. Of course most hams don't have this option. They're left guessing how well their antennas are really performing.

My motivation to write this article came from a recent antenna selection conundrum. I was listening to the FP5KE DXpedition on 20 meters and I wanted to know which of my several 20 meter antennas was best. I was driven by curiosity, without the intention of working them, since I've worked FP countless times on 20 -- it's single-hop distance and frequently active.

I first listened on the upper 5-element yagi of the 5-over-5 stack. They were surprisingly weak, no more than S5 on the meter. This wasn't surprising since that antenna, at 40 meters height, concentrates most of its energy at low elevation angles. The lower antenna does better for high angles usually associated with shorter paths.

When I switched to the lower 5-element yagi, which is about 21 meters high, the signal was still quite weak, not much different compared to the high yagi. That seemed odd to me since it was so unexpected. I then switched the stack to BIP (both in phase) and the signal jumped to well over S9. Imagine my surprise. 

To complete the comparison, I turned the Skyhawk eas. It has 3 elements on 20 and is about the same height as the lower 5-element yagi of the stack. The signal strength was a little less than the full stack and far better than either individual yagi in the stack. 

I suspected that the differences were due to the elevation patterns of the various antennas. There are two or more lobes with nulls between lobes. The terrain is pretty flat so the lobes and nulls of antennas at similar heights but on different towers should be about the same. The patterns are likely to closely resemble those produced by modelling software that assume a homogeneous, flat ground.

This EZNEC plot contains the elevation patterns of the three selectable configurations of the 20 meter stack: upper, lower, BIP. It is taken from an earlier article, and annotated with a few lines.

The red line at 30° elevation highlights a null common to each configuration. That can occasionally pose a problem when the propagation path favours that elevation angle. If there is real difficulty due to that null I can switch to one of the tri-band yagis.

The dark red (brownish) lines highlight elevation angles where there are two traces with nulls. That is, there is one antenna selection that will work better than the others. The chart is quite busy so I hope that readers can successfully read the critical data it contains.

Despite the generally excellent performance of the stack, it is those many nulls that occasionally render it impotent. You can get better elevation angle coverage in a stack with 3 yagis, whether of similar size or smaller.

Assuming that the plot accurately resembles that of my stack can we determine what happened when I was listening to FP5KE? We are looking for an elevation angle where each individual yagi has a null and the combined stack does not. However, there isn't one!

A reasonable conclusion is that the elevation plot produced by the model does not accurately match the real stack. That isn't too surprising. Only a small deviation from perfect terrain can move those nulls. That's because a null requires an exact cancellation of fields from the space and ground reflected waves of one or both yagis -- the requirement is equal amplitude and opposite phase when the fields are summed. The major lobes of the antennas may be close to those in the model even when the nulls are significantly different since the sums of the fields can be less critically aligned

My wild guess is that the propagation path's elevation angle is close to 30° (red line) and that one of the minor lobes of the stack on opposite sides of that angle is really much larger. But it's just a guess. It also must have been the case that the signal was constrained to a narrow range of elevation angles. Otherwise pattern nulls would not have such a prominent effect. That is often not the case, with signals appearing at a range of elevation angles due to various scatter phenomena.

A ray tracing tools such as HFTA combined with an accurate terrain profile for my QTH might deliver better insight. Or it might not. As I said, null positions are very sensitive to small deviations. Those deviations can quite easily be larger than the accuracy of the terrain data fed into HFTA.

I'm not so curious about those nulls that I'm willing to go the trouble of using HFTA. I have enough antennas that I can "navigate" around pattern nulls to work what I need to work. I am not going to rearrange my towers and antennas for dubious improvements, nor will I take the trouble to add BOP (both out of phase) to my stack switches. There are many serious contesters that would never build a station without first consulting HFTA.

I'll say it again: you can't have too many antennas.

One Screw Defeats Two Antennas

When we left the 30 meter delta loop I mentioned that it had to be disconnected for the time being. The reason is that it did not have a feed line. Instead I used the one for the 160 meter antenna (shunt fed tower). To conserve Heliax my intent was to have them share the same transmission line. That required a way to switch between antennas at the tower base. It's an economical arrangement since those antennas are never used at the same time; 30 meters is not a contest band.

Switching up to a kilowatt in a low impedance (50 Ω) system is conveniently done with an 8 to 10 amp modular relay. I keep them in stock for applications such as this. I selected a 12 amp SPDT relay, the same device I use in many of my home built antenna switches. "Dead bug" style construction is ugly but it works fine in this application. No one will see it with the enclosure cover in place.

The enclosure at the tower base was previously drilled for 2 auxiliary antenna ports (UHF chassis connectors), so I removed the tape from one of them and installed the connector, relay and flyback diode (1N4007).

The default path is the 160 meter antenna. There is no external connector for it since the enclosure includes the gamma capacitor and connections to the gamma rod and ground. When the relay is energized the RF path switches to the 30 meter delta loop. 

When (if) I add another antenna port, a second SPDT relay will be added. The relays would be chained by taking the wire to the 30 meter connector to the second relay, which would switch between the two auxiliary ports. That is, to select 30 meters the first relay is energized and to select the other auxiliary port, both relays are energized. There is negligible impedance "bump" at HF from this circuit.

The configuration of the antenna selection software was changed to select the delta loop on 30 meters and manage contention in case two radios (SO2R or multi-op) attempt to select more than one of the antennas on the transmission line.

If you've read this far you may be thinking that this is routine and hardly worth writing about. What's interesting about relays switching RF? It's done all the time. Despite hinting that I would write an article about the 30 meter antenna switching I had no real intent of doing so. What changed my mind was what happened a couple of weeks later.

The antennas and switch had been working well. Then one morning they weren't. Both antennas showed high SWR and poor receive signal strength. On a warm and sunny weekday morning I grabbed a few tools and my antenna analyzer and walked across the hay stubble to discover what was going on.

There were no obvious problems. The antenna analyzer confirmed that the transmission line was okay and that the trouble lay inside the enclosure. There were no visible faults when I opened it. The DMM confirmed that the relay was operating and the RF paths were fine. This had become very puzzling indeed.

Since the enclosure is non-conducting I moved on to test the ground wiring.  Unlike in a metal enclosure those wires are necessary. 

When I measured the resistance from the ground lug (bottom centre on the right panel above) to the transmission line connector flange there was an open circuit. How was that possible? The trouble spot is pointed out by the blue arrows.

The picture above shows what it looked like after the repair. At right is the #4 screw that I replaced.

There is rust on the screw head, top and bottom, and the first couple of threads. The rust acted as an insulating layer between the screw and connector flange, both it's outer surface and the screw hole interior walls. I used plated screws since they were handy at the time. But I did not replace them with 304 stainless screws that I bought this summer for this very purpose. I forgot.

As already mention, for a plastic enclosure there is no automatic ground path from the connector flange. It must be explicitly wired. The flange screws are critical since the solder lug, which is wired to the earth (and tower) grounding lug, must have a solid electrical connection to the flange. 

I opened the box of stainless screws and replaced the rusty one. That fixed the fault. I'll have to do the rest, just in case. Although not all are used for ground connections, it's sensible to replace them all. Stainless should not be used everywhere but when you need it, you need it. The corrosion deposited on the silver plated N-connector flange was removed by light buffing with steel wool. 

Even the biggest antennas can be defeated by a tiny screw. Don't overlook their importance.