Monday, July 30, 2018

Skewed Wire Yagis

So you want to put up a wire yagi for the low bands. All you need is two sufficiently tall structures -- trees or towers -- tie a catenary rope between them, hang wire inverted vee elements from it and run coax along the catenary to the driven element. Simple and cheap high performance fixed yagis for 40 or 80 meters can be yours.

But what if those structures aren't positioned where you need them. If they are not on a line that is close to the direction (or directions if the yagi is made reversible) that is your objective it could kill the project or you'll have live with reduced performance due to the desired direction being off the centre of the main lobe or by reducing the number of elements to broaden the beam width.

For the case where the structures are towers (or just one of them is) and you plan well in advance you may be able to place the tower or towers to suit wire yagis. Otherwise you can only count on luck to have them where they're needed. I expect this is a dilemma that some readers have faced or will face since the all time most popular articles on this blog are designs for 2, 3, and 4 element 40 meter wire yagis.

The left diagram demonstrates the problem. The greater the angle between the catenary and the desired direction (two directions for an electrically reversible yagi) gain will suffer. The effect increases with the number of elements since the main lobe will be narrower. A 2-element yagi is therefore the least affected by direction error.

On the right is the hypothesized solution. The elements have been turned so that they are broadside to desired direction(s); they remain parallel to each other even. The questions to be answered include:
  1. Is the main lobe (if it survives this abuse) along the catenary (a line through the element centres), broadside to the elements or somewhere in between?
  2. How much are gain, F/B and SWR affected as the skew angle increases?
  3. If the pattern does shift direction how great an angle can be achieved while maintaining a clean pattern an similar SWR band width?
In this article I will explore the problem with wire yagi models where the elements are askew from each other, as they would be when the catenary runs at an angle to the desired compass point. This computer model study should be enlightening and may give hope to those with unfortunately located support structures.

Model parameters

Since a 2-element wire yagi has a wide main lobe and wire yagis with 4 or more elements are uncommon and take more time to deal with in the model I will restrict myself to a 3-element reversible wire yagi for 40 meters with inverted vee elements. The antenna is one I've described in detail before, and indeed is the all-time most popular article on this blog according to web site statistics.

The diagrams above are based on my EZNEC model of the antenna. Refer to that earlier article for details on antenna design and performance since I will not repeat that data in this article. Apex height of the inverted vee elements will be fixed at 25 meters since I don't expect that height will significantly alter the effects of element skewing.

The yagi is reversible by switching a coil at the centre of the two parasitic elements, such that one becomes a director and the other a reflector. The driven element is at the centre of the array so that the performance metrics are identical in both directions (symmetric). Therefore we only need to examine the model in one direction.

Skewing the elements can be done in two ways:
  • Rotate the elements in place. The greater the rotation angle the shorter the effective "boom" length. That is, the element spacing is reduced along a line broadside to the elements.
  • Constant element spacing. The distance between element centres increases with rotation angle so that the element spacing remains constant.
A 2-element reversible yagi has transmission lines between the feed point and the elements. Therefore when you skew a 2-element yagi the electrical design must be adjusted for the latter method because the physical length of the transmission line must increase. Worse, the lines will be unbalanced due to the loss of left-right symmetry. In my opinion this is more trouble than it's worth. The 3-element yagi is far simpler since it has only DC control lines running to the parasitic elements. Common mode currents due to the asymmetry can be addressed without affecting antenna performance.

For the 3-element 40 meter wire yagi I will examine both skewing methods for angles up to 30° in steps of 5°. Patterns are only computed at 7.1 MHz, which is approximately the midpoint of the yagi's design range for best performance. Based on my modelling I don't expect any surprises in performance at other frequencies over the antenna's range of 7.0 to 7.2 MHz.

For the constant element spacing method I will not compute the increased "boom" length -- distance between element centres along the catenary. This is easily accomplished with elementary trigonometry if you are interested in building an antenna of this type.

Modelling results

As I increased the rotation angle (skew) in the models I was surprised by the yagi's resilience. Performance is sustained better than I expected. However the skewing is not entirely without cost. Scan the table of the calculated key performance metrics before you continue reading. The tables have been normalized to compensate for shortcuts and orientation oddities in my rotation methodology. None of this affects the results but did reduce the time I spent modelling.

There are several conclusions we can make based on the data:
  • Gain, F/B and main lobe beam width are largely sustained for angles up to 30°. I didn't expect that I could push that far without negative consequences.
  • The centre of the main lobe is between the catenary direction and broadside to the elements, and is the same for both skewing methods. The main lobe skew is ~60% of the physical skew in all cases.
  • Although SWR bandwidth is unaffected the resonant frequency increases slightly at the largest skewing angles. Adjusting the matching network should be easy, or at least straight-forward.
Despite these promising results it was disappointing that the pattern skew fell short of the physical skew of the yagi. For a 30° physical skew the skew of the main lobe is 18° ±1°.

The azimuth patterns at a skew angle of 30° show a difference between the skew methods -- original in black and skewed antenna in blue. The rearward pattern is cleaner when the elements rotated in place. With constant element spacing the F/B and F/R are slightly worse. The F/B figures in table also clearly favour rotating the elements in place. [Note: The small skew in the left pattern plot is an artefact of my modelling shortcut discussed previously, not a true difference between the skewing methods.]

Application and items for further study

Hopefully this study can act as an incentive to a few readers with seemingly inconvenient supports for high performance wire antennas. It can pay dividends on the low bands as we survive the next 2 or 3 years the solar minimum will last.

To start, consider how far askew the line between supports is from what would be ideal for you operating objectives. Let's say the amount is 20°. Choose a skew angle of 30° and the pattern will be almost exactly where you want it. If the desired direction is 30° off the catenary you can get most of what you want by skewing 30°, since the pattern will be 18° closer to the ideal than it would otherwise be.

You can even do this with conventional yagis. The application I have in mind is side mounting a yagi flush to a tower face but have its pattern in a different direction when no tower face is suitably oriented. However this would require custom boom-to-element clamps and there may be awkward questions from visiting hams about your peculiar looking antenna. The skew of a wire yagi is less obvious to the casual observer.

If the desired skew is greater than 30° I strongly recommend developing a model rather than extrapolating from the models in this article. At some point I expect performance to rapidly degrade. I didn't attempt to find that point, but be assured that it exists. The same goes for wire yagis with more elements: do a model before jumping into construction.

Depending on my own interest I may extend the skewing concept to wire yagis with more elements, yagis with a coupled resonator and wire yagis with loop elements. It is reasonable to predict that all will benefit to a degree though it will require modelling to find how far each yagi design can be skewed before performance deteriorates.

The primary message of this study, I believe, is to use the supports you have and exploit skewing to make the best of your circumstances. Perfection is rarely necessary or even desirable. I would only caution against deployment of novel skewing arrangements without modelling beforehand to determine what you can expect and so avoid wasted effort.


Skewing yagi elements may seem an odd idea and I admit it never occurred to me until about two weeks ago. I was walking back and forth in the hay field armed with a compass, wood stakes and a 200' tape measure looking for the ideal location to plant my second big tower. Tower siting is a compromise between safety, transmission line length, yagi side mounting, interactions and wire antennas. It is rare that all objectives can be fully satisfied.

This tower's placement in my original site plan for this QTH was based on these criteria. Details matter when one moves from a rough plan to literally pounding a stake into the ground, hence my recent surveying activity.

I expect to place the tower such that there will be some skewing required for wire yagis hanging off a catenary between the towers to make the project more tractable and to minimize interference from side mounted yagis pointing directly at antennas on the other tower (towards Europe). Shifting the tower site ~20 meters solves these and several other problems.

Thus was born my motivation to study skewing. A few hours later I sat down in front of the computer to study the implications and possibilities.

That tower project is now well into the planning stages with a tentative schedule. Wire yagis for the low bands may be in my future now that I have skewing data in hand. Stakes are in the ground.

Monday, July 23, 2018

WARC Bands Without WARC Antennas

For those who have followed this blog for the past two years may recall that my primary objective in building my station is contests. HF DXing is my secondary priority despite comprising the bulk of my daily operating. Contests are episodic which leaves a lot of time in between!

As a consequence my effort on towers and antennas favours the HF contest bands: 160, 80, 40, 20, 15 and 10 meters. However I also need antennas for the WARC bands -- 30, 17 and 12 meters -- for general and DX operation. I want effective DX antennas on those bands without requiring effort better spent on antennas for the contest bands, with respect to performance, time and money.

As an interim measure last year to get back on 17 and 30 meters I once again put up my tried and true multi-band fan inverted vee. Originally this antenna was a key part of my Ottawa station soon after I got back into ham radio several years ago. I was very happy to find that it delivered fantastic results with QRP and 100 watts on 15 through 40 meters. Of course that was during the solar cycle maximum when a little antenna (and power) goes a long way.

When I moved to this QTH in late 2016 I temporarily hung it off the house so that I could operate until my first tower was up. I moved it to that tower in late 2017 when I moved the 80 meter inverted vee to the big tower so that I would have resonant antennas on 17 and 30 meters, and for short distance contest QSOs on 40 meters.

The multi-band inverted vee is about to be retired once more, perhaps for good this time. The reason is that I no longer need it. I am able to operate very effectively on 17 and 30 meters with other non-resonant antennas. As we'll see, even non-optimal antennas can do very well under favourable circumstances.

17 meters on the XM240

Unlike a full size dipole or yagi driven element the Cushcraft XM240 does not resonate on the 3rd harmonic, which would be 15 meters. Coil-loaded elements is the reason. Instead the antenna has a resonance at a lower frequency, almost but not quite on 17 meters.

This has long been known and many hams with this antenna have had some success using it on 17. The additional transmission line loss due to an SWR between 2 and 3 is modest since of the 350' (110 m) of coax ~90% is LDF5 Heliax.

When I purchased the antenna I built a model in EZNEC to learn, in part, how it might do on 17. Unfortunately an accurate model isn't possible with NEC2 due to the loaded elements, As of EZNEC 5 the stepped diameter correction (SDC) does not support loaded elements but this deficiency has been partly addressed in EZNEC 6. I don't know the reliability of the new feature.

Nevertheless it is possible to gain an insight into the antenna's pattern despite the resonant frequency being incorrect.

As you can see the azimuth pattern is bidirectional, just like a dipole. There is ~1.5 db gain due to the parasitic element (a very wide spaced yagi on 17 meters), making it slightly better than a dipole. Since the equality of the two lobes is frequency dependent the azimuth pattern above is not quite what you'll get at 18.1 MHz -- I evaluated the pattern at 19 MHz to approximately compensate for NEC2 inaccuracy. In practice I find that the pattern is slightly directional in the same orientation as it is on 40 meters, so it can help to point it at the DX rather than relying on it being bidirectional.

One big advantage this antenna has is height: 46 meters up. That makes up for its quirks. I am able to work DX very well, often breaking pile ups quickly with 200 watts. The XM240 is far superior to the multi-band inverted vee with its apex at 19 meters, and it's rotatable. With the XM240 a proper yagi for 17 meters is not currently in my plan.

30 meters on the 80 meter inverted vee

The WARC bands are not harmonically related to the others. Sometimes they come close enough to be tempting. This is the case of 80 and 30 meters, where the 3rd harmonic of 3.5 MHz is 10.5 MHz and of 3.8 MHz is 11.4 MHz. When cut for the CW segment of 80 meters the 3rd harmonic comes within 5% of the 30 meter band.

My inverted vee is cut for mid-band so the 30 meter resonance is around 10.7 MHz. Although the modelled SWR is quite high at 10.1 MHz in practice it is no higher than 4, after accounting for loss in the long run of FSJ4 and a shorter run of LMR400; that is, it is lower in the shack due to transmission line loss. I use the rig's ATU. As is usual with a harmonically fed dipole the feed point impedance is high even at resonance.

The pattern is oddly shaped but not too problematic. As is usual with an inverted vee the polarization is primarily horizontal broadside and vertical off the ends. At its apex height of 32 meters you might expect the radiation upward should be close to nil, but this is not so.

There are two current maxima on each leg of the inverted vee -- there is a single maximum at the antenna's centre when operated on its fundamental frequency. Thereforeon 30 meters the effective height is lower and there is a big lobe pointing straight up (pattern not shown)! This lowers gain in the main lobe at 30° elevation.

Despite these drawbacks the antenna works pretty well on 30 meters. Performance depends on the direction. If this antenna does poorly on a station I can always switch to another. Which brings us to the next antenna.

30 meters on the 80 meter vertical

From on high we go down low. The 80 meter yagi array construction is ongoing and is already giving an account of itself on the bands as an efficient ground mounted vertical with its (so far) 34 radials. I will have more to say on this antenna in future. In this article I'll restrict myself to its 30 meter performance.

Like the inverted vee the vertical's 3rd harmonic falls in the vicinity of 30 meters and has two current maxima along the monopole. The 34 radials, each 20 meters long, form a non-resonant ground plane that extends much further than on 80 meters, with respect to wavelength. In comparison to the inverted vee the SWR bandwidth is very broad, a characteristic that appears at its 3rd harmonic. I haven't carefully measured the SWR on 30 meters but it appears to fall between 2.4 and 3. Half the 250' (80 m) transmission line is LDF4 and the rest is RG213 and LMR400. The rig's ATU must be used.

In 30 meters it compares favourably to the multi-band inverted vee and the 80 meter inverted vee. Again, direction and elevation angle are factors. The vertical is omni-directional with a higher angle lobe common to verticals operated at its harmonics.

Sometimes the vertical does better and sometimes the inverted vee does better. In a minority of cases the multi-band inverted vee equals or slightly exceeds the 80 meter antennas.

The elevation pattern assumes 5 Ω ground loss. It is more difficult to estimate the radial system efficiency when the vertical is operated on its harmonics. I didn't try since its use on 30 meters is incidental rather than a design objective.

12 meters

I left this one last since 12 meters has two strikes against it. One is that it is rarely interesting at this point of the solar cycle. Two is that I have little interest in it. There is no rational reason for the latter.

The multi-band inverted vee doesn't work on 12 meters. In fact I've never had a resonant antenna for this band. That none of my panoply of antennas comes close to resonating is therefore no great loss. The few times I do venture onto 12 meters I pick one of tri-band yagis and use the ATU.

The high SWR is not extreme and the reduced system performance does not worry me, considering my attitude towards 12 meters. A poor antenna that is up high is usually enough to get the job done when I call a DX station.

Perhaps when the solar flux rises from the dead in a few years and most of the heavy lifting is done in my antenna farm I'll take the time to put up a resonant antenna for 12 meters, even a yagi if I get serious.

Summing up

With many antennas and a little bit of height it is often possible to recruit one of them to put you on a band for which you have no antenna. Obviously this strategy is less than optimal in most cases, yet as I've discovered not optimal can still do very well indeed. It is something to consider. However if your antennas are modest and not high it is entirely likely you will not achieve superior results. This is when a resonant antenna can be well worth pursuing.

The multi-band inverted vee is still up and occasionally put to use. In fact I'm using it more right now while summertime work on the station switching and cabling make other antennas temporarily unreachable. When that is done and the next stage of antenna and tower work begins it will be in the way. Retirement is tentatively slated for August.

The antenna won't be junked. If nothing else it make a handy portable antenna should I ever want to do that, or as a loaner to a friend. This may be optimistic since it doesn't wrap up neatly due to it peculiar construction. It ends up as a tangled mess every time I roll it up, no matter how careful I am.

Sunday, July 15, 2018


I did not make a substantial effort in the IARU contest this weekend. Not only do I not enjoy warm weather contesting -- too many opportunities to do things outside -- ongoing cabling and switching work meant I had just the TH7 at 21 meters and an inverted vee for 40 meters. The bigger antennas are not currently connected to the shack.

Despite all of this I did operate for several brief periods as CW LP. If not for this also being the WRTC contest-within-a-contest event I doubt I would have made any effort to be active. With some amusement I note that, apart from 80 meters, my station as currently configured is quite similar to what the WRTC competitors used. By accident I also used the same power -- 100 watts -- which I selected out of habit. IARU, being an ARRL sponsored contest, permits 150 watts in the low power category.

What I will do in this short article is provide my thoughts on how the competition may have played out. Of course I could be completely wrong. It is nevertheless interesting to test my understanding and make a few speculations. Eventually the full story and the facts will emerge. For background on competitor strategies and choices I recommend N3BB's excellent book, Contact Sport.

QSO totals

Compare the WRTC scoreboard and raw scores on 3830 and you'll immediately notice just how much higher the QSOs and scores are for the WRTC competitors. Their modest stations were no impediment. Like rare multipliers in any contest they attracted a lot of attention regardless of signal strength.

For this reason it is no surprise they spent the bulk of their time running. Indeed, despite spending half of my short time in the contest running I did not get called by any Y8. This continued to be true in the final 30 minutes of the contest when I'd expect their rates to be relatively low. Clearly they knew how to best utilize their time and energy.

Since there are 1,440 minutes in a 24-hour contest it is easy to calculate their overall QSO rates. The top placers were all above 4 QSOs per minute (240 per hour) averaged over the full contest period, with two stations and two operators. That's impressive! The high power multi-op stations could not match this level. I doubt the HQ stations (which are multipliers) fared any better.

Activity level and low power

In most contests those with low power and modest stations (especially QRP) usually must go high in the CW band segments to run. Big guns tend to congregate towards the low end of the band where it is difficult for smaller stations to run and be heard. Yet in this contest the low power Y8 competitor stations appeared to be evenly distributed across the band and could even be found hugging the bottom band edges.

Perhaps they could do this due to their starring role in the contest. On the other hand it is summer in the northern hemisphere when many contesters are loathe to spend time in the shack. I know I am. Despite midsummer conditions with its many attenuated propagation paths it seemed that the overall activity level was not high. That makes room available for smaller stations. Of course many of the big gun operators were congregated in Germany, not at their usual operating positions.


N3BB points out in his book that choosing the split between CW and SSB modes is a key strategic decision in WRTC. CW has the advantage of favouring low power and being picked up and spotted by the global skimmer network. SSB has the advantage of faster rate and a pool of operators who do not operate CW. In 2014 the competitive advantage appeared to favour those who emphasized CW.

From the scoreboard it appears that the CW advantage was suppressed in this year's competition. SSB totals are generally though not universally higher among the top scoring teams. If this is true why might it be so?

In 2014 the competitors were in W1 where the bulk of the valuable QSO points come from Europe. There remains a strong CW culture among Europe's contesters. From Germany looking west towards North America there may be proportionately fewer CW operators due to the now longstanding migration to SSB due to no-code licensing. Outside of the US, Canada and Europe the use of CW is even lower.

From central Europe I would expect that inter-continental QSOs favour SSB toward Central America, South America and east Asia. East Asia does not contribute many QSOs from W1 but it is a productive path from Europe. There are, for example, many non-CW operators in E2 and YB.

Could this have been a factor? I notice that North American teams are lacking among the top scorers. European operators may have better understood the need for SSB for racking up contacts on the available propagation paths beyond Europe and North America.

Low bands

Europe is a hotbed of amateur radio activity, including contesting. Pay close attention to the standings after any major contest and you'll notice that even the QRP participants in Europe work an enormous number of other European stations on the low bands, day and night. Did enough of the WRTC competitors from outside Europe understand the importance of the low bands? Intra-European contacts are worth less but there are so many more available.

In North America the low bands are mostly a wasteland during daylight hours, except in late afternoon and early morning during domestic contests, and then only in the eastern third of the continent. Europe is different.

Time will tell

Don't take my analysis too seriously. It is mostly speculation based on limited data. Time will tell whether I am right on any of my suspicions.

Being wrong does not worry me. In any competitive endeavour it is beneficial to study the competition to see what can be learned, whether to emulate or avoid the tactics they use. Answers may always remain elusive since the competitors themselves may be uncertain about what exactly they did right or wrong. Sometimes it is merely a matter of luck, be it good or bad luck.

For me contests are fun rather than a serious competition. I have no aspiration to compete in a WRTC. It is an opportunity to learn and to watch the masters in action. We should all always be learning.

Sunday, July 8, 2018

Insurance Contact Dilemma

An insurance contact is when you work a DXpedition a second time on a band-mode. This is generally frowned upon since it takes time away from the DXpedition operators to work other stations, stations that may not have worked the DXpedition at all. Sometimes it's done by mistake.

Most often it is done to be extra certain that their call is in the log. Rare DX can cause anxiety! Before the internet became prevalent in the DX sport insurance contacts were more often due to anxiety about whether they got in the log the first time.The uncertainty may be due to QRN, QSB or QRM (deliberate or accidental) that obliterates part of the QSO. The DXpedition operator can contribute to the anxiety by not clearly repeating and confirming the received call before moving on to the next QSO.

There are also less than stellar individuals who do it to show off their big antennas and big power. With computer logging braggarts risk being called out by the DXpedition operator so this behaviour is no longer as common as it once was.

With near real time log uploads to services such as Club Log there is less cause for anxiety. It used to be that you might only learn one or two years later that you weren't logged! That happened to me a few times so I am appreciative of the benefits of modern technology. Since internet access may not be available or reliable from remote locales, resulting in upload delays from hours to days or until after the DXpedition is over, anxiety still occurs.

Which brings me to the recently completed Baker Island DXpedition: KH1/KH7Z. There were two QSOs that caused me some anxiety, and ultimately a pursuit for insurance contacts.

The first was 15 meter CW. Despite the poor solar flux there is a brief propagation peak in the late evening my local time. Signals are quite weak, putting us at a disadvantage to those further west on the continent. Nevertheless I managed to eke out a QSO around 04Z with my TH6 up 43 meters and with my usual 200 watts. Having a kilowatt would have helped get through but with a lesser antenna I may not have heard them.

The QSO was not without drama. It took a few overs until he copied my call correctly. Finally I heard it, sent my report and heard the confirmation. One more band slot in the log. Or so I thought.

The QSO was not in the online log after it had been updated up to and beyond the time of our QSO. This is not necessarily a problem since it often happens that logs from all operating positions are not uploaded at the same time. So I waited another day. Still nothing.

Perhaps I had been wrong when I heard him send my call that final over. I felt (and still feel) it was correct. I could wait and hope for the best or consider trying for an insurance contact. For many this would be simple to resolve: just get in there and do it. My reticence is due to my dislike of bothering the DXpedition operator again or, worse, hearing the dreaded "QSO B4".

Nevertheless I went ahead. Propagation peaked at the same time a few nights later and I made the QSO easily enough despite the weak signals. No incorrect call to correct this time. Lo and behold the QSO appeared online the next morning! It seems I had made the correct decision.

During the intervening time I got lucky one morning and worked them on 80 CW about 10 minutes after local sunrise when signals usually peak. I heard my call clearly above the noise after one extra over to communicate my call sign suffix. There was a smile on my face as the pile up crowded onto my transmit frequency for their attempt to make it.

The smile didn't last. As you can guess the very same thing happened. Again the QSO failed to appear. There was no happy ending this time since I was not able to catch them again on 80 at the right time and mode.

How does this happen? As the DXpedition wound down I happened to be chatting with a serious DXer friend of mine and we talked it over. We could think of several possibilities.
  1. My mistake: Skill, experience and diligence can still fail us. Despite my certainty it is possible they copied my call incorrectly and I failed to hear the error and correct it. This happens quite a lot in contests, so I am accustomed to these irritating mistakes. However I tend to be especially careful when working a rare country.
  2. Accidental erasure: Endurance operating takes its toll, whether in a contest or on a DXpedition. A wrong or forgotten keystroke and a QSO can vanish without a trace. Although this certainly happens I don't consider it highly probable in the case of my missing QSOs.
  3. Typo: Operators have many different styles. Logging software can smoothly handle transmission of partial calls that are typed in and re-sending of updated call signs. Some operators are more comfortable sending the partial call by hand. Only after they believe they have the call correct do they type it in and log it. In this way I could hear my call being sent yet not be correctly logged. The likelihood this happening twice does not seem high to me. It's just a possibility.
Frankly the most probable explanation is that the mistakes were mine and mine alone, despite my confidence about what I heard. It is nevertheless a healthy exercise to explore the range of possibilities. Even if errors occurred at their end it does not absolve me: perhaps there are things I could have done to improve the outcome. That is, to make it easier for them to copy and log me. Insurance contacts are a poor substitute for accuracy.

All of this said there remains the faint possibility that my QSOs are correctly stored on the DXpedition PCs. There may be gaps in the records uploaded from the database, gaps that will be filled when they get home and process the logs. I am not very hopeful this will happen

Screen capture from Club Log

I did pretty well despite the lost 80 meter QSO. Unfortunately my 160 meter antenna is down for the summer since I did hear them on that band a couple of times. Baker Island on the low bands will have to wait for another 10 years, which is the earliest another DXpedition will be permitted on the island.

It give me one more thing to look forward to. I'd better get back to work on towers and antennas so that I'll be ready when that long awaited day arrives.