Tuesday, September 26, 2023

Beverage Lightning Protection

Beverage receive antennas are very susceptible to lightning. They are long wires that at close to the ground and grounded at both ends. While direct strikes are not rare, the greater threat is typically secondary strikes (low current lightning branch) and inducted current from a nearby strike. I can personally attest to the lightning risk, when my Beverage system has been struck not once but twice.

This summer I removed the head end electronics of my three reversible Beverage antennas, disconnected the feed lines and directly grounded the antenna wires. I rarely operate 160 meters during the warm months since noise is high, activity in the northern hemisphere is low and the radials of the primary 160 meter vertical are removed during the haying season.

One of my summer projects was to add lightning protection to the Beverage antennas. It isn't difficult and there is ample information available on how to do it. Nevertheless I moved slowly. I wanted to better understand how the protection systems work before ordering parts and making the modifications. I am now better informed though far from being an expert. 

When I was satisfied with what I learned I ordered the parts and made the modifications to the Beverage head ends. After completion I tested them to confirm they still worked as they should and then reinstalled them in the field. The Beverage system is back in operation and ready for the fall and winter season.

The design I settled on is a melding of the methods I gleaned from W0BTU and VE6WZ. Those weren't my sole sources but they were well documented with good explanations. I also read what W8JI and ON4UN (Low Band DXing) had to say on the topic and I delved deeper into the circuits and explanations from commercial lightning protection device vendors. Any mistakes or misunderstandings are my own!

I believe it will be helpful to first review a well-known circuit for coaxial lightning protection. The circuit above is used in a variety of products made by Array Solutions. I chose it in particular because they are open with their design and their products are used by many hams.

Lightning and atmospheric static discharge have both RF and DC components. A blocking capacitor on the ungrounded centre conductor is not effective on its own since the surge has nowhere to go and the potential will build until it exceed the capacitor's breakdown voltage. The capacitor holds the charge at bay, briefly, while other components ground the surge. 

There is no DC component in the RF signals hams work with so an RFC (choke) grounds the DC component of the surge or static buildup while blocking RF. If the DC charge is large enough or increases faster than the charge can be grounded via the RFC, or the RF potential is large, the GDT (gas discharge tube fires (conducts) and it has a short-term ability to conduct kiloamps of charge. The diagram text explains the other components.

It should be obvious that the choice of capacitor and RFC affects normal operation. The capacitor in particular should have a low reactance and high Q over the operating frequency range suitable to the power rating and maximum SWR. A higher power rating is recommended even with less than legal limit power unless a low SWR is certain.

A DC surge with a slow rise time and moderate current might be handled entirely by the RFC if it does not overheat and fail from grounding the surge energy. This is desirable since the GDT will fail from repeated firing and conducting high currents, so we want to reduce how often it fires. For maximum protection they should be replaced after several secondary strikes or one primary strike. Since the GDT typically fails open you cannot easily determine that it has failed.

It should be obvious that the GDT firing voltage should be higher than the maximum voltage for the transmitter power into a 50 Ω load or the higher voltage due to a mismatch. For example, 1000 watts into a 50 Ω load has an RMS voltage of about 225 and a peak voltage of 320. Increase the power and the voltage rises. SWR multiplies the maximum voltage that can be present. Array solutions selection of a 1200 volt GDT is sensible.

There is more to it than that. The capacitor should be rated to hold off the surge being grounded by the RFC and GDT. How high the voltage grows depends on how quickly and effectively the surge can be grounded. Both the surge and working voltage ratings are relevant. The longer it takes the charge to flow to ground, the longer a high voltage is applied to the capacitor.

It is not only the current capacity of the RFC and GDT, but also the ESR (equivalent series resistance) of the ground rod's connection to the ground and how quickly the earth charge within the ground rod's "reach" is depleted. An excellent ground connection is no guarantee of the protector's protection during a direct lightning strike. But that's a subject well beyond the scope of this article. 

With measurements, I estimate that the ESR of the 4' copper clad ground rods that I use for my Beverages, in my local soil, is between 100 and 150 Ω, more or less. That doesn't appear to affect Beverage performance but it is not low enough for the best lightning protection. The sooner the GDT fires the faster the charge can be grounded. That motivated my choice of a 75 rather than a 90 volt GDT.

A receiving antenna has less extreme requirements than transmit antennas because the signal level is very low. A modest amount of signal loss is acceptable and the GDT can be chosen that fires at a much lower voltage. Consider the following open-wire Beverage protection system by W0BTU.

The GDT are 90 volts, the RFC is replaced by a 33 kΩ resistor and the capacitors are ordinary ceramic bypass devices. Resistors are cheaper than RFC and can be effective, for RF and not just DC static and surges. Clearly it is far less expensive to protect receive antennas than transmit antennas. We must protect both conductors of the open-wire line, doubling the component count.

The capacitor can have a lower voltage rating commensurate with the GDT firing voltage. Since the frequencies are low -- typically down to 1.8 MHz -- the capacitor value must be high for a low reactance. At 1.8 MHz a 0.1 μF capacitor has a reactance of about 1 Ω, which is negligible in series with a 600 Ω antenna. The reactance is lower at higher frequencies so we size the capacitor for the lowest operating frequency. A smaller high voltage capacitor would also suit, except on the feed line side of the unit where the impedance is typically 50 or 75 Ω.

My implementation is modelled on W0BTU's unit. Differences include: 

  • 75 volt 5 kA Bourns GDT were specified by VE6WZ in his Beverage system and I wanted to keep the voltage as low as possible for maximum protection from even minor induction events. I am confident that Steve made a well informed choice.
  • The 33 kΩ resistors that drain charge to protect the GDT are ½ watt rather than 1 watt. It isn't a big change and I had them in stock. If a resistor fails the GDT will fire more often and fail sooner.
  • The coupling capacitors are 0.1 μF and 1000 working volts. Notice the size in comparison to the 0.1 μF capacitors on the feed line side of the unit (blue, lower right). The 630 volt devices are much smaller. I could have used the 630 volt capacitors from my stock but opted for the larger capacitors for their higher power dissipation.
  • I don't directly protect the relay as W0BTU does. The resistor and GDT on the coax centre conductor offers limited protection from lightning conduction between Beverage head ends via the remote switch; the coax shield is already grounded at the remote switch and has limited GDT protection in the head end via the secondary windings of the transformers. The current iteration of the Beverage remote switch uses SPDT reed relays to ground the coax centre conductors of all but the active Beverage. That in itself is good protection. I have being doing the same for control lines wherever feasible.
  • The capacitors are mounted on the PCB but not the GDT and resistor. By direct wiring them to ground I keep the high voltage and high current surges away from sensitive components where the narrow separation of copper pads can offer an alternative and perhaps more enticing path to ground. Stranded interconnect wires allow easy removal of the PCB and connectors from the enclosure for service. 

The above design was used for both the northeast-southwest and east-west reversible open wire Beverages.

The 470 Ω resistor between the Beverage wires was fitted temporarily to test that the added components do not affect normal operation of the unit. Above is a test of the modified north-south RG6 reversible Beverage head end. The SWR curve is not perfect since I used a 51 Ω resistor on the antenna rather than 75 Ω, the analyzer is normalized to 50 Ω and the coax between the analyzer and unit is RG6. This was merely a sanity check that I had made no serious mistakes since all the head end units were in good working order.

Here's a closer look at the reversible RG6 Beverage head end. The antenna port is on the right and the feed line port is on the left. The GDT and resistor pair protect both conductors of the antenna coax. The outer conductor protection is obviously needed. Inner conductor protection is in case of strong coupling between conductors or from the far end via the reflection transformer. I want to avoid a protection path via the fragile transformer windings.

I did not protect the reflection transformers. In previous lightning strikes the reflection transformers were unaffected so I didn't feel the urgency. I will probably go ahead and add the protection after the flurry of fall antenna projects come to an end. It isn't a priority.

The Beverages head ends are back in service -- well, after locating a cold solder joint. I am now ready for the 160 meter season, except...I am in the process of modifying my big vertical. For the time being I have only 4 radials installed, which I rolled out earlier than usual to work E51D. I'll have more to say about improvements to the vertical once the work is complete. One of my objectives this year has been to improve my 160 meter signal.

I also plan to protect the multitude of control lines and rotator cables. Although many are grounded by relays when not energized, others cannot be grounded that way. Unlike RF transmission lines, DC lines cannot be protected by series blocking capacitors. I have not yet settled on a design. 

There are many months to work on it before the arrival of the 2024 lightning season. That is also when I will find out how well the Beverage protection works. My intent is to leave them connected year round.

Wednesday, September 20, 2023

Cutting Small Tubes With Hand Tools

While working on a few antenna projects recently I cut a lot of ⅝", ½" and ⅜" aluminum tubes. This is a subject I touched on briefly in an article on cutting pipe square. Making a straight 90° cut of a round object is not as easy as it might seem. Lucky for us, extreme accuracy is unimportant in the construction of HF yagis. An ugly cut is usually hidden inside the next larger tube, where it is invisible and soon forgotten.

The 3 most common ways to cut small tubing:

  1. Pipe cutter
  2. Hacksaw (with guide)
  3. Band saw

When I first moved to this QTH with a plan to build many antennas, I considered purchasing a band saw.  They're wonderful machines that make clean and reliable cuts with little effort. New ones aren't cheap and used saws come with risks do to their age and maintenance record.

Many of us have a table or power saw that might seem suitable, but they are not. The blades are not designed to cut metal and the rotation speed is far too high. Aluminum in particular requires a slow speed saw. If you could get around these obstacles, do you really want aluminum shrapnel flying out of the back or damaging the saw? Those tools are designed to cut wood, not metal.

I chose to delay the purchase of a band saw and used hand tools for cutting metal. I never did buy that band saw and now that the construction flurry of the past several years has abated it hardly seemed worth it now. Cutting large pipe, tubes and plates with hand tools takes time and can be tiring. My time as a retiree is not a scarce resource and exercising the arm muscles has benefits. Others would choose differently.

Tool quality

To save yourself a lot of grief, do yourself a favour and buy high quality tools. Do not go by price alone since that is no guarantee of quality. Consider yourself fortunate if you have an acquaintance with metalwork expertise who can point you in the right direction. Otherwise you need to know what to look for in a tool.

It's time for a story. When I was in my early 20s I took a bicycle making course from a master frame builder. I loved bicycles and cycling and I wanted to learn more. This was back in the days when the best bicycle frames were made of high strength, thin wall steel. There were only a handful of reputable makers of high end steel frame components -- my choice was Columbus SL.

The course was held in a community college where they had classrooms equipped with enough hand tools and gas welding equipment for all of us. However the instructor urged us to buy our own high quality hand tools. Of course not everyone did, saving expense by using whatever was in the tool room.

When one student ran into difficulty cutting a (very expensive) oval tube, the instructor proceeded to demonstrate the correct technique with the shop-supplied hacksaw. It could not cut in a straight line. The hacksaw frame twisted and the blade warped no matter how much tension was applied to the blade. In an angry outburst he nearly hurled the saw across the room. 

He calmed down and turned the incident into a lesson. He carefully explained the flaws in the tool's design and how it would damage the work. He rooted through the tool room and found one hacksaw that he decided was adequate. He demonstrated the difference.

It's been many long years and I have not forgotten that lesson. I was careful to buy good tools even when I was loathe to spend the extra money. When you consider the pain you encounter with poor tools there really is no such thing as a cheap tool! Some of my best tools seem to last forever and continue to work well despite their heavy use. One example is my hacksaw, pictured above.

That is not an expensive tool. However it was pricier than many others. Although the logo is that of a popular Canadian brand, that is only the branding. I don't know the manufacturer. There are others like it to be found if you shop carefully.

When I bought it I inspected the structure and then performed a few simple tests. Firmly grip the handle with one hand and place your other hand on the back bar. First, try to rotate the back bar. The hacksaw should twist very little and instantly rebound when you release the pressure. If it moves more freely, the backbone or its mechanical bond the with handle and back bar are inadequate. Second, try to pull the back bar towards the handle. It should strongly resist and not behave like an accordion.

The final test may be difficult to perform in a store. Install a blade and bring it up to a high tension. If you can't achieve that it should be rejected -- the structure twists or accordions, or the tensioner is weak or is uncomfortable for your fingers. Next, try to twist the blade with your fingers. If you have adequate tension you will find it very difficult to twist. Check that the blade is vertical and straight.

My hacksaw passes these tests. I've owned it for close to 20 years and it continues to perform well. Longevity isn't easy to test, but it is easy to achieve with a good tool that is not misused or abused. The only maintenance I have to do is to change blades when they wear out.

The common hacksaw blade length is 12". Don't buy a smaller hacksaw because you'll find it difficult to cut plate and large pipe. Don't skimp on blade quality. The better blades are well worth the premium price. I stick with 18 teeth/inch for cutting steel and aluminum. A finer blade has little benefit for antenna work, in my experience, although it can help to start the cut without the blade skipping sideways.

The width of a hacksaw cut is no narrower than the blade's maximum width. A perfect cut isn't possible so expect it to be a little wider. I measured the cut on the tube slit at right at 0.03", using the above 18 teeth/inch blade. When you cut a 12" tube into two equal pieces and file the edges clean, each will be slightly less than 6". I've never found this to be a problem in antenna work. For finer work, position the hacksaw blade to the outside of the measured cut line.

I use a flat file to trim the small lip created by the cut. The blade primary removes material but it also pushes some to the side. A round file removes the lip and debris inside the tube. I use a small triangular file to clean the cut edges of tube slits, inside and outside.You want smooth surfaces to avoid cutting yourself when handling the cut tube and to ensure good mechanical and electrical performance when telescoped inside a larger tube.

On to the next tool. I have two pipe cutters, one small and one medium size. They are not precision tools. Care is needed to ensure clean, square cuts. This might be surprising to some. After all, you have a blade and wheel in fixed alignment rotating around the tube. What could possibly go wrong?

In a picture below you can see several tubes cut with the pipe cutter shown at right. If you look very closely you'll notice that some of the cuts are not 90°. The obliqueness is small but it's there. Again, that isn't important for antenna work. It has to do with tool design and how it's used. Even a good pipe cutter can be mishandled to cut poorly.

The tool looks simple enough.. A sturdy frame supports a roller and a round blade. The blade is thin and wedge shaped but harder than the metal to be cut. You open the gap to fit the tool over the tube and then pull down the blade until it contacts the tube, while ensuring that the roller is flush to the opposite side of the tube. You spin the tool and periodically draw the blade inward until it pushes through the tube wall.

In a perfect world this would result in perfect cuts. The world isn't perfect and neither are tools.

The blade and its body spin on a removable axle (to allow for blade replacement) between the 'C' arms of the tool body. The imperfect fit for both allows for play. The worse the tool or due to long service the greater the amount of play. My pipe cutters are of no better than moderate quality. Spending more didn't seem worthwhile since the cuts are often hidden and aluminum alloys are soft compared to other metals. The blades can last a long time when only used for cutting aluminum tubes.

Aggressive force on the blade will cause the blade to tilt at an angle. It may track in a circular pit or it may wander or spiral. I've found it very easy to trace a spiral when cutting PVC pipe since the soft plastic "grabs" the tilted blade. It's important to start the cut with light pressure no matter the material.

Despite being careful the cut might not be square. Can you tell from the picture? Expand it to full size and the tubes with a bad cut will be easier to identify. 

The other thing you should notice is the profile of the raw cut of the tube posed in the pipe cutter. It is far was than what you get with a hacksaw or band saw. While it may be easier to cut a tube with a pipe cutter, there is more filing to be done afterward.

First, the cut is not vertical. It has the same profile as the wedge shaped round blade. A flat file on the open end of the tube (for both tube halves) is needed to remove the large angled projection.

Second, the cut is not as clean as a hacksaw because the pipe cutter does not remove material. The material pushed aside by the blade piles up to form a substantial lip. It is high on the outside of the tube and shallow on the inside. Both lips should be filed flat so that the end of the tube is corrected to its original diameter. You can see the filing marks on the set of tubes above. Don't skip this step or you may have difficulty telescoping it into the next size larger tube, and even if you do the mechanical and electrical performance will suffer. 

A quick sweep of a round file is usually enough to clear the lip on the inside of the tube. For the outside lip I use a flat file while rotating the tube. Be careful not to file into the tube surface while you abrade the lip.

A pipe cutter works like a butter knife. It pushes the soft butter aside rather than removing it, or splitting it like the ways a chef's knife cuts pliant vegetables. Metal tubes are too rigid for cutting with a knife.

Does it matter?

As I alluded to earlier, pretty cuts are rarely important for antenna work. Once it's in the air no one will notice. You might think that it'll bother you but it won't. You'll soon forget, and you'll forget even sooner the more antennas you build.

What matters is safety and performance. Improperly finished cuts will draw blood when you handle the material. Even for the clean cuts of a band saw I take a few moments to remove imperfections and the small ridges at the inner and outer tube edges. For mechanical and electrical performance we want maximum surface contact where tubes overlap. The ridge formed by a pipe cutter prevents that. For those of us in cold climates, an exposed ridge at the element tip can increase ice buildup and delay sloughing off when the sun comes out.

So, yes, it matters. It's worth the small investment of time and effort to clean the imperfect cuts of a pipe cutter or hacksaw. Or you can invest in a band saw. Then you will spend the time that you saved to keep it in good working order and spend more money on those long flexible blades. In my opinion, buying and learning to use good quality hand tools is the right choice for most hams. If you're the rare exception, by all means invest in a band saw.

Thursday, September 14, 2023

New Operating Desk

Widely spaced blog articles are due to me being very busy and a lot of partially complete projects that are not ready to become blog material. I was sitting in my easy chair one evening, utterly fatigued, and looked around the shack. I realized that I had not yet mentioned the new operating desk that I put into service in late August. I'll remedy that oversight now.

The previous iteration of operating desk did not work out well. It was the product of ideas not fully worked out, my horrible carpentry skills and a few unfortunate design choices. I ripped it out of the shack and made another attempt. What I learned from my mistakes led me to an improved design. Now that it's been in use for a few weeks it is worth a look.

First, I will state the obvious: it is not complete. I continue to work on placement of the equipment and cable routing. The second (right) radio is installed but not connected. Desk setup is improving, slowly, and it will be a keeper for a year or two at least. My objectives for the design include:

  • Ergonomic improvements for SO2R contests
  • Rapid rearrangement for two operating positions in multi-op contests
  • Keeping the mess of cables out of sight, especially those for the station automation system
  • Convenient access to the back panels and cables for service and configuration changes
  • Improved appearance
  • Space to place infrequently or never touched equipment out of sight
  • Lots of legroom without the risk of knees and shins bumping into the desk structure
  • Support heavy equipment, especially amplifiers

I was unhappy with reasonably priced commercial products that met my objectives. Monitoring the local used markets found many inexpensive office desks, but all had at least one fatal flaw. Maybe one would have appeared had I waited but with contest season approaching I had to act.

A friend with a truck brought a 4' × 8' × ¾" sheet of fir plywood to his workshop where we ripped it to a 30" depth and routed the edges. The remaining 8' x 18" of the sheet was set aside for shelves that will be built later (more on this below).

He delivered the plywood to my workshop where I stained it and put on several coats of polyurethane. I kept most of the frame of the previous desk, discarding the desktop and beefing up the structure so that it is very stable and capable of supporting a heavy load. I put them together and positioned the new desk next to my (non-ham) desk to form an L. I was careful to make the desktops the same height.

My old 1980s operating desk, which I revived almost 10 years ago, has the same fir plywood top but it was not suitable. Too much space was taken by a set of drawers and the vertical side supports, and it is a foot shorter than the original 8' sheet size. Although perfectly good for SO2R, it is inadequate for two operators.

The desktop height must be identical to that of the adjacent desk. I got it right but it still wasn't good enough. At 8', the desk is so long there is a small sag even with a stiff frame. But the level match is very close, less than ¼" of sag at the interior corner. The keyboard wobbles a bit when it lies across the boundary, which is where it will be for SO2R with two keyboards. The pointy right front edge of the left desktop is exposed and can catch unwary fingers.

The horizontal lumber of the frame is rearward to take the weight of the equipment while staying beyond reach of the operator's knees. Addressing the sag with a forward beam isn't possible. However a mid-span vertical support can work. I will experiment to find one that removes the sag and doesn't limit the excellent legroom.

The operating desk was designed to be functional and not pretty. The frame is the lumber equivalent of "plumber's delight" yagi construction. Metal stiffening plates are placed where they are most effective and the projecting screws are out of the way of operator's legs and feet. Some are visible when you enter the shack but I don't worry about that. Power bars are mounted to the back of the frame's rear beam.

The lower equipment shelves support power supplies and the station automation hardware. Other "low touch" equipment will be added later. Eventually the BPF (band pass filters) will move below the desk since band switching is automatic. The station automation hardware (pix) is in the back corner close to the floor opening for the cables. The computer is at the back corner of the desktop just above it. That choice keeps the multitude of Cat5 control cables from snaking across the floor or having to be dressed along the underside of the desk. You can see excess blue plenum cables hanging from a hook. 

I can hardly wait for a full wireless control system so I can dispense with control cables entirely. The same aspiration applies to transceivers; I want all communication between rigs and computers to be wireless. We're getting closer but we're not there yet. There are so many cables that a tangled mess is difficult to avoid no matter how careful you are.

I kept a gap between the desk and wall to ease service. No more crawling under (or over) the desk to access equipment rear panels and cables. The aesthetics are not great, but at 14" (35 cm) the appearance is acceptable. That may seem too narrow unless you've met me in person; I am exceptionally slim (skinny). Keeping the space narrow may dissuade guest ops from the temptation to venture where they probably should not.

In the picture at the top of the article you can see a variety of rotator controllers. One recent change was to replace the old Ham-M controller with a newer Hy-Gain model. They frequently appear at flea markets for a reasonable price after the accompanying rotator dies. Two of them make a convenient support for the computer monitor. I may need more (there's another on the far right of the desk) when I modernize the home brew breadboard prop pitch controller sitting on top of the FTdx5000. Visitors are always startled to see that monstrosity, yet it continues to work very well.

Eventually I will use the leftover plywood to make a prettier corner shelving unit to give the monitor a proper support and to avoid direct stacking of equipment. Amplifiers are dangerous to stack because they are heavy -- the Drake L7 is an exception because the power supply is in a separate enclosure. The extra shelves will wait until I replace the FTdx5000 with a modern rig since it is quite large and won't fit the shelf design.

The walls of the shack are bare. Not only do I not apply for awards, I have never put contest plaques on the wall. For that matter, my framed university degrees have also never been hung on a wall. The shack walls are more bare than before because I removed a bookcase to make room for the longer desk. Maybe this is an opportunity to dust off the plaques and cover the empty walls. Maybe.

Since I'm pretty happy with new desk I will continue with station setup and get it ready for the fall DX and contest season. I will tinker with it this winter as time allows. Right now I have a long list of higher priority outdoor projects to complete.

Wednesday, September 6, 2023

Singing Isn't Just About the Element

Many owners of HF yagis are familiar with the "singing" phenomenon. When the wind blows the elements vibrate. Not always, since it depends on wind speed and direction and element design. It is an interesting sight (or sound) but often not healthy for the antenna. That is, we want to stop it from happening.

Aluminum can fatigue and fail at well below its yield strength when stress cycled enough times. There are many examples of yagi elements breaking and the tips falling to the ground. This is dangerous and can be expensive to repair. Rebuild the yagi with the identical replacement parts and failure is likely to recur. Some brands and antennas are notorious for singing-induced fatigue failures. Here is one example:

The tower and antennas have been derelict for several years. The ham is a silent key and before that was unable to deal with the problem. Elements broke off one by one, leaving the denuded antenna seen above. Someone gathered the broken elements and leaned them against the tower. We found another sticking in the ground like a forgotten spear. Luckily the tower is located in a rarely used field so no one was injured. The tower is tentatively on my list for removal this fall.

I know many hams dealing with these issues. Large yagis with widely spaced elements may have to be lowered or carefully manipulated on the tower to access and repair broken elements. Some hams take the trouble to redesign the elements, with varying degrees of success. Haphazard changes might change the singing but not eliminate it.

Yagis that are prone to element breakage are typically the ones that "sing" in the wind. You can often hear the thrumming from the ground. The sound is louder when you're on the tower near the antenna. When the pitch is too low to hear you can still see the vibration. It may not happen at all in a very strong wind, but in those cases you are more likely to be concerned with survival of the antenna and tower.

Why do antenna sing? When they do, how serious a problem is it and what can be done about it? These are interesting and important questions whether you buy or build yagis. I've found too many hams that simply shrug their shoulders and say that's just how it is. Others are convinced that there is one true method to defeat singing and they apply that to every yagi they own. 

Fatalism and certainty are nothing more than excuses to avoid understanding the issue. This is an opportunity to learn, and if you read this blog with any regularity you will know that, to me, learning is one of the greatest benefits of our hobby. We don't have to be structural engineers to gain an insight into the why of singing and what, if anything, we ought to do about it. I like to keep yagi elements up in the air and I believe that you want the same thing.

Singing is not an arcane branch of engineering. It is very well understood in the profession. Despite owning my first yagi in 1975, I had little understanding of the phenomenon until relatively recently. I would venture to say that my ignorance is typical of most hams. With the wonders of the internet at our fingertips there is no reason to remain ignorant. It is easy to learn the fundamentals with only a modest effort. Readers who are structural engineers may cringe a little at what follows, but please humour those of us who lack the background.

[Diagram credit unknown] When a fluid like air encounters a solid object it will flow around it. The crowding of fluid at the front and sides of the object increases the pressure, and in the lee of the object there is a low pressure region. Purposely shaped objects like wings use this effect to generate lift, while the high and low pressure regions due to hurricane force winds can lift the roof off a building.

There is a critical fluid velocity where the flow changes from laminar to vortex generation, and finally to turbulence. Vortexes form on the lee side of the object, with the low pressure regions cycling between the upper and lower boundaries of the object. This is know as vortex shedding. Objects with freedom of movement will move toward the low pressure regions. As the low pressure region moves back and forth the object moves back and forth. If that frequency is close to the structure's resonant frequency it can oscillate quite vigourously.

For cylindrical objects like antenna elements there is a formula to calculate the frequency. Of course it's never this simple, but for our purposes it is perfectly adequate. One complication among many is the incident angle of the wind. I won't discuss this and other complications in any depth.

F = 0.22 (V / D)

V is the fluid velocity and D is the cylinder's diameter in the same units. For example, if V is expressed as m/s (meters per second) then D must be expressed in meters.

Let's try an example. Please note that in the following discussion I will liberally round quantities. One significant figure is enough to understand the basics.

For a 3 m/s light breeze (about 11 kph or 7 mph) impinging on a ½" (1.3 cm) diameter element tube, F is 55 Hz. The frequency is proportional to wind velocity and inversely proportional to element diameter. Thus for a 1" diameter tube in a 6 m/s wind the frequency is identical. 

It is not a given that the tube will vibrate at this frequency since only select vortex frequency ranges will excite the element. The most common is the (mechanical) fundamental resonant frequency. At other frequencies the element will move but not oscillate; yagi elements wiggling when the wind blows is normal and is not to be confused with singing. 

If any tube or tube combination in the tapered element is resonant it can shake the entire element. The tip may vibrate most vigourously because it is smallest tube and unrestrained at one end. A stronger wind on a thin element tube may be at an audible frequency that you hear on the ground. This is what we know (and fear) as singing.

It should be evident that singing depends on wind speed and direction. Too slow or too fast and the oscillation will be mild because the element's mechanical resonance is not excited. Also, a wind direction well off from normal to the elements is less able to excite resonances.

There are several approaches to mitigate singing:

  • Mass: A heavier tube with more inertia is more difficult to excite into oscillation. Greater wall thickness or tube overlap is often all it takes to prevent singing. Very thin wall tubes used in some antennas are more prone to singing.
  • Damping: Per the ARRL Antenna Book (22nd ed., 25.2.3): "Metal antenna elements have high mechanical Q, resulting in a tendency to vibrate in the wind." Just like in an RF network, resistance lowers the Q and therefore the oscillation potential by dissipating energy. The most common method is to insert rope into the element or just the element tip. Another is to run a piping helix around the tube surface. This is uncommon on antenna elements but I've seen it used to good effect to damp oscillations on Phyllistran synthetic guys.
  • Texture: Although impractical for hams antennas, the surface of the tubes can be sculpted with patterns that reduce vortex formation and their amplitude.
  • Turn the antenna: A yagi is typically pointed in the direction of stations you want to work. But when you are not in the shack the singing can be reduced by parking the yagi normal to the prevailing or current wind direction. For example, turn the antenna north or south for prevailing westerlies.

I won't delve into the finer points of mechanical mitigation measures except to note a few items that I've learned from experience and listening to those with greater knowledge. First, don't use excessively thin wall tubes. It is used in many antennas because aluminum is expensive and manufacturers know that keeping prices low increases sales. Stay at or above the typical 0.058" wall of aerospace high strength tubes, and overlap tubes more than the 3" minimum to further increase mass and strength. Second, beware product instructions that tell you to rope the elements. This is very close to an admission that there is a deficiency in the mechanical design.

I have watched hams thread rope through under-engineered full size 40 meter yagi elements and I've seen roped element tips fall to the ground. Damping from an internal rope may be insufficient to keep singing in check. Vibrating yagis can also cause fasteners to loosen or fail, and that can prove disastrous even without tube breakage. My TH6 elements are not roped and the non-trapped 15 meter elements sing; those of the TH7 were roped yet they also sang.

Singing isn't always straight forward to resolve, whether by treating the cause (design) or the symptom  with rope. Without a comprehensive mechanical model and simulation software we are too often left guessing. When one mitigation method fails should we blindly substitute, say, a different or longer rope, or should we try something completely different?

We may be going about it completely wrong. For example, mistaking the location of the symptom as the cause of the singing. Consider the following two yagis in my station. These are the upper and lower 5-element yagis of the 20 meter stack. We are looking toward the back of both antennas.

The construction, size and placement of the elements of the stacked yagis are identical. It may surprise you to learn that one yagi sings and the other does not. I know that I was surprised. Clearly there must be a difference. However we can eliminate a difference in the directions they are pointing since the disparate behaviour remains when the direction is the same (northeast, since the lower yagi is fixed).

The lower side-mounted yagi occasionally sings at moderate wind speeds. It isn't always noticable from the ground and I haven't carefully noted the wind speed and direction when it occurs. Typical for this latitude the prevailing winds are west to southwest, and that is approximately normal to the elements when they sing.

If not the elements themselves, what is the cause of the singing. I believe the answer lies with the frequency of the singing and how the entire antenna behaves. The audio frequency is low, below what a human can hear. The singing is not a sine wave so it may be the weaker harmonics that are heard. But when the singing is bad, all the elements shake and so do the boom and boom truss. That's unusual in my experience.

It is instructive to return to the above equation for vortex induced vibration. A low frequency implies a larger cylinder than is present in the elements, the largest being the 1" × 0.120" centre tube. The boom is the only large cylinder in the antenna and it is approximately the same for both antennas: 3". The equation for the same 3 m/s breeze gives a frequency of 8 Hz. That is in the range of what I observe. I didn't note whether the wind was from the northwest or southeast, which are normal to the boom.

Why only the lower yagi? I think the answer is that the booms are constructed differently despite having the same diameter. The tubes comprising the boom of the lower yagi have a wall thickness of 1/16" other than the thicker 10' long pipe at the centre of the boom. The weight difference of the two booms is substantial. As for elements, the mass of the boom seems to make all the difference.

Compared to the elements, yagi booms oscillate at lower frequencies and boom truss cables at higher frequencies. Unlike a singing element, when the boom sings (oscillates) all the elements shake. That appears to be what is happening in this case.

The solution is simple: replace the boom. I probably have what's needed in my stock of pipes to do that. If not, I can probably find more surplus. It isn't urgent. Watching the entire yagi vibrate is worrisome but it is not an emergency. It happens only occasionally when the wind direction and speed are just right. Maybe next year.

Singing is an example of structural resonance where the oscillation frequency is usually in the audible range. The 20 meter yagi is singing but the fundamental frequency is not audible. I can't hear it from the ground. A low resonant frequency is common to large structures like houses, bridges, skyscrapers and towers. 

Perhaps the best known example of undamped large structure oscillation is the Tacoma Narrows Bridge. Large structure oscillations are more difficult to analyze and defeat because there are many structural elements whose behaviours can be both independent and synergistic. Even so, as with yagis, there are a variety of effective mitigation methods in the engineer's toolkit.

Years ago I made use of structural oscillations excited by footsteps, wind and various mechanical impulses to sense and measure activity in buildings. It was interesting work. The oscillations were damped by the structure but lasted long enough to glean useful data. Although only minimally relevant to singing yagi, or suspension bridges, the physical principles are similar.