Monday, June 20, 2011

Verticals for Contest Expeditions

Verticals for Contest Expeditions

"You're using verticals for your contest expedition - you're not serious are you!??

That was a typical response when people heard we were going to use verticals on all bands in the 1997 CQ WW CW 6Y4A M/M effort. And frankly, I think most people wrote us off as serious competitors based on this information.

The choice of verticals didn't come easy, and countless hours of research went into antenna predictions on the computer, propagation predictions (using CAPMAN and Mini Prop Plus), field testing to calibrated receivers, 6Y4A group discussions, and discussions with other well-known antenna specialists. We also had the recent experience of the 1997 ARRL DX Contest at 6Y4A to help us in our analysis. While the antenna strategy for 6Y4A was specifically tailored to that QTH, much of what we learned can be used in all vertical antenna installations.

Why Verticals?

This was a difficult decision. On the one hand, we had nine yards of "conventional wisdom" that the Caribbean was always a high angle path and to use Yagis, even if they are relatively low to the ground. On the other hand was a mounting array of computer predictions, plus our experience in the ARRL DX Contest, in which we had followed the computer predictions in the decision to use verticals. In the months following the ARRL DX Contest, we had gathered additional information on various antennas, but the selection of verticals for the CQWW CW ultimately was based on a few key pieces of information:

  • Computer predictions indicated that takeoff angles for the USA were 4-7 degrees for most of the population, and takeoff angles for DX locations were typically 1-10 degrees. Operating tests before the ARRL Contest showed the verticals to be excellent performers on DX.
  • There is no other antenna that will produce high gain at very low angles except for a vertical on or immediately adjacent to salt water, or a Yagi at high heights (in terms of wavelength).
  • Computer predictions indicated that Caribbean and W4 stations had take off angles of nearly 30 degrees.
  • When the far field is over salt water, certain vertical antennas produce a large main lobe that has consistent gain from 1-30 degrees. High horizontal antennas can produce a low angle lobe, but they will also produce multiple lobes, between which will be nulls at many useful angles (the number of lobes for a horizontal antenna is two times the antenna height in wavelengths).
  • Ground mounted verticals are easier to install than high rotatableYagis.
  • Two fixed vertical arrays per band weighed less than one tall Yagi with a rotor.

As noted above, the computer propagation models indicated that most of the signals are arriving at angles below 10 degrees. The highest we felt we could get a Yagi in the air was 30-35', which would place a 20m Yagi just under a half wavelength high, a 15m Yagi at around 3/4 wavelength high, or a 10m Yagi around 1 wavelength high. Of all those antennas, only the 10m Yagi would have usable gain at 10 degrees but had little gain below 5 degrees. Based on this information, and our results in the ARRL DX Contest (and operating before the contest), we again chose to use verticals on all bands for the CQ WW CW Contest.

Other criteria in selecting the antennas:

Selection of antennas on a per band basis was done by balancing performance, use of antennas that were already constructed (when possible), and weight. Since we had to bring all of our equipment with us, the other operators brought the equipment, and N6BT and his wife Bonnie brought the antennas. So we wouldn't have to pay for excess baggage costs, the goal was to pack all antennas into the checked baggage limit for 2 people. We used 4 golf club carriers. Each person was allowed 2 check-in bags, weighing no more than 70 lb. per bag, or 280 lb. total. (The carriers weigh 15-18 pounds each.) Getting all the antennas into these 4 carriers took a lot of pre-planning and careful selection of materials. The actual packing was performed with each carrier on an accurate scale and loaded to within one pound of the 70 pound limit. The total actual weight was 276 pounds, including just enough tools to assemble and disassemble the antennas.

Our experience with vertical antennas:

We have learned a lot about vertical antennas from these two expeditions, plus the two testing sessions performed locally. During the planning and on-site installation of the antennas, we started installing the antennas using the traditional methods for verticals. We quickly found out that these methods did not work, and that there were many unknown or incorrect facts about vertical installation and performance.

Here are some typical, and often incorrect assumptions about verticals:

  1. You can take a horizontal Yagi, turn it vertically and it will work OK
  2. You don't need radials when the antenna is over or near salt water
  3. Verticals work OK when radials are laid on the ground
  4. Verticals are easy antennas to tune up
  5. Verticals are always noisy receiving antennas
  6. Phasing verticals are the best way to achieve gain
  7. The farther you move a vertical away from the water, the weaker the signal will get

Lesson #1: You can't take a horizontal Yagi and simply turn it vertically.

One of the first ideas we came up with was to take a regular full size Yagi and turn it vertically. Besides the installation problems, there were a number of performance problems that arose. Computer predictions showed that the pattern and dimensions of the Yagi absolutely did not work when you turned the antenna vertically, with the ends of the Yagi within a few inches (or feet) from the ground.

So rather than using existing dimensions, we started off with one vertical dipole. We then added another element (reflector). This gave us an immediate 4.5 dB gain, which was predicted by AO. But the predicted dimensions were not what we expected, as most of our experience to date was with horizontal Yagis. The 2 element vertical array tuned up nicely, and had a nice broad pattern with reasonable gain. The vertical Yagi can be put into practice in two ways: either full size vertical half wave elements, or using quarter wave verticals with radials. The full size elements have slightly higher gain since the feed point is typically elevated higher than the quarter wave verticals. The increased gain compared to a 1/4 wavelength vertical comes from the lobe being compressed vertically (not as "tall"). The vertical Yagi should be elevated no more than 1 to 3 feet (for the high bands). This places the feedpoint at, or only slightly higher than a 1/4 wavelength high. Above that height, high angle lobes are created.

We then added a third element (director), and that's when all the problems started. It was extremely difficult to generate a reasonable model with 3 or more elements. The antenna gain was not what was expected, as the pattern was totally unacceptable, and it was hard to obtain a reasonable match. Since 3 elements did not work, we then tried 4, 5, and 6 elements. Apparently due to close proximity with ground, the model ran very slowly, and when the automatic optimizer was used, it skewed the elements into funny locations and with unusual dimensions. Often the model would effectively eliminate one or more elements by moving it a 100' away, or by changing its length so it was not part of the circuit. N6BT tried to manually optimize the 3 element design, and also could not achieve a proper antenna design. We spent a lot of time trying to get a multi-element (more than three) parasitic vertical antenna to work. We eventually arrived at a 6 element model after way too much effort and finally abandoned the idea. The best solution was to utilize the 2 element (driver / reflector) design in a single Yagi, or in a pair of phased Yagis (2 by 2 array). Yagis with 2 elements (either quarter wave or half wave elements) offer the "most bang for the buck", plus other benefits, as well.

As an FYI, it is our philosophy to tune all antennas close to a 50 ohm match to make installation easier. N6BT has used this technique on several commercial Yagis, going back many years (e.g. the Magnum 620 & C-3). The higher feedpoint antenna has no matching component (fewer parts), is more efficient and tends to be more broadbanded. On the expedition, we tuned the antennas for a balance of maximum forward gain and minimum front-to-back ratio. In a multiplier location like ours, we don't want to limit our signal into any geographic area. The one exception is creating a null to the USA when we were running Europe, since Europe is worth more points per QSO, and the USA stations tend to be much stronger than the Europeans.

Lesson #2: You do need radials when the antenna is mounted over or near salt water.

The salt water is great for the reflecting plane of the vertical, but does not provide a good current return. The radials are the current return for the vertical portion, and are the "other half" of the antenna. It is recommended that at least two resonant radials (elevated) be used for an effective current return. Two radials produce a balanced antenna pattern, while one radial will skew the pattern and contain a horizontal component. Moxon covers this subject very well in his writings. In our installation, we used 2 radials on all but the 160 antenna, where we installed 4 radials (we later felt they were not needed, as the 2 extra radials did not change the feedpoint impedance when they were added).

Lesson #3: Under most any circumstance, radials should be elevated.

When we first installed the 80m vertical in Jamaica during the ARRL DX Contest, we laid the resonant radials on the ground, believing (as most would) that being on the ocean, etc., that everything would tune up OK and work great. The joke was on us! While tuning the vertical, we were unable to get a proper match. We tried various matching coils, adjusting everything adjustable, all to no avail. The clue we had was that the tuning meter on the MFJ antenna analyzer showed a high value (>50 ohms) even without a matching coil. The antenna should be down in the 20 ohm range, as it is linear loaded and physically short. A full size vertical should be in the low 30 ohm range. When we added a matching coil to the already high VSWR feedpoint, the VSWR got higher! (it should have been lower) All of us glared at N6BT, who is supposed to know everything! Tom did deduce that the radials were obviously coupling into what must be extremely lossy ground (not expected). The interaction with the lossy ground was the single largest factor which impacted the tuning and performance of the vertical antennas.

What to do? Fortunately, Tom had recently read an IEEE article about adding elevated radials to AM broadcast stations that had disintegrating buried radial systems. One of the key points was that the new radials needed to be elevated, not the whole antenna. Actually, the radials can be lifted at an abrupt angle right at the antenna and then run horizontally. We left the vertical in place, but raised the radials around six feet. Immediately we were able to obtain a good match. We concluded that all radials at this location needed to be elevated due to the interaction with the very lossy ground. For the ARRL Contest, we elevated the 160m and 80m radials by around 6 feet, the 40m radials by four feet, and we left the 20m radials sloping from about a foot, down close to the ground. When we returned for the CQ WW CW Contest, we elevated all radials.

When raising the radials, the key is to get them high in as short a horizontal distance as possible, and then run them horizontally. We used two methods: using bamboo poles mounted very close to the vertical to support the radials, we also made loops in the guy ropes to suspend the radials via porcelain insulators.

K2KW has been on other expeditions where he used verticals mounted in the sand, right by the ocean. Many times he said he was unable to achieve a good match. He now knows that the key is to raise the radials.

Lesson #4: Verticals are not always easy to tune up.

While raising the radials provided an efficient current return, tuning the verticals and obtaining a proper match was not that simple. When we installed the 20m verticals for the CQ WW CW Contest, we decided to test out what happened when we raised the radials. Are you ready for this? With the radials placed on the salt-water enriched ground (coral, with salt water pot holes here and there under the radials), the frequency was 11 MHz! (as seen on the MFJ antenna analyzer) What a shock.

We then started raising the radials, and were able to watch the resonant frequency increase. At around two feet above ground, the antenna was close to resonance just under 14 MHz and the feedpoint was in the low 30 ohm range (right where it should be). We were amazed to see that the coupling to the ground could lower the resonant frequency by 3 MHz. We had used this antenna in our original empirical testing, and knew that the vertical portion and the radials were set to the correct lengths.

The basic tuning procedure was to use the radials to tune all the 1/4 wavelength verticals. The vertical portions had been calculated using the proper tapers over perfect ground. The lengths of the two radials were then adjusted together for the proper frequency, while watching the feedpoint to be sure they were high enough above ground. Tuning the verticals this way brought the antennas into resonance. We first started to tune the system by tuning only the reflector (the director must be out of the system). Once the reflector was tuned, it was shorted at the feedpoint (no matching coil required). From there we added the driven element, and repeated the procedure above. We then added matching coils (a.k.a hairpin or beta match) across the feedpoint to achieve the 1:1 VSWR. Also, we installed baluns on all driven elements.

The linear loaded verticals were tuned differently. Their radials were cut to length, and the tuning jumpers on linear loading wires were adjusted for the proper frequency, then the match coil was set. The 40 meter ZR verticals (by Force 12) are actually vertical dipoles, so they have no radials and were simply set to the right frequency. The 10 meter vertical dipoles were tuned to the proper frequencies (driver and reflector). In all the 2 element arrays, the reflector is always set first, with the driver open. The reflector is shorted to place it in the circuit and the driver is then tuned.

Lesson #5: Verticals are not always noisy receiving antennas.

During the ARRL DX Contest, we used single element, linear loaded verticals for 80 and 160m, and a 2 element 40m linear loaded vertical array, (all standard antennas by Force 12). These were selected mainly for their compactness and efficiency. If we really wanted, we could have had full size verticals for 40 and 80m. We noticed that the linear loaded antennas were very quiet on the low bands, and found them to be very good receiving antennas. We realized that the high-Q of the linear loaded antenna essentially made the antenna a filter, and only let in noise from near its resonant frequency. (This is consistent with other efficient, hi-Q antennas, both horizontal and vertical.)

We never felt the need to use separate receiving antennas. During the ARRL DX Contest, we found out by listening to our competitors, that we out-heard everyone of them. And we confirmed this by setting Multi-Two QSO records for all bands on which we used verticals (160-20m). We also set an all time QSO record on 80m - even though we were on 80m for only half the time!

Lesson #6: Parasitic vertical arrays really do work.

Even though AO said that a 2 element parasitic vertical array (with quarter wave elements) would "work" (achieve the desired gain, have a reasonable feedpoint and target pattern we wanted), not many people have used this kind of antenna. During our empirical testing, we did verify that a reflector added the predicted 4.5dB of gain over a single vertical. We also tested a 2 element parasitic array using Force 12's ZR verticals, and obtained the same results. A 2 element parasitic vertical array is easy to match, and doesn't require the complicated phasing line of a driven array. If you are are interested in obtaining gain in a single direction, then you should consider using a parasitic vertical array.

The 2 element vertical array has a very different pattern than a 2 element horizontal array. The vertical array has a large beamwidth of up to 120 degrees (depends on tuning). The beamwidth somewhat changes with the tuning of the antenna, but is still close to double the beamwidth of a typical 2 element horizontal array. This was a real plus, as we covered a wide area without using a rotator. The 2 element vertical array over salt water has around 8.5 dBi peak gain at 8 degree take off angle, with minimal difference in gain from 1-30 degrees. When over salt water, the vertical array also maintains useful low angle radiation off the sides and back of the array, which is not typical of a horizontal array.

Lesson #7: A vertical right at the water's edge may not be the optimum place.

There is no question that the salt water does wonders for the performance of a vertical antenna. If in doubt on where to locate your vertical, do it as close as possible, or over, salt water. It is only over salt water that the very low angle gain is achieved. Over dry ground, the energy below about 12 degrees (pseudo Brewster angle) is lost.

While field testing the verticals this past summer, we decided to test the effect of the land-water boundary on the pseudo Brewster angle. Since our receive site was elevated less than 1 degree across the bay, we could see any change in the low angle energy. To our knowledge, there has not been any published tests of this kind. The goal was to see how far from the water the vertical would loose the benefit of the salt water on the pseudo Brewster angle. The tests were done with a 20m ZR vertical, and we moved the antenna away from the water in 5' steps. The water's edge was considered the reference point. As the vertical was moved back from the water, there was little change until we came close to 1/4 wavelength from the water. At that point there was a 3 dB increase in signal level! Moving farther, the received signal level dropped, indicating a loss of low angle energy. This was most significant at 1/2 wavelength from the boundary, being down about 3dB from the waters edge. Moving farther back to 3/4 wavelength, the signal picked up again, to more than 2dB enhancement from the water's edge. We could not move the antenna farther due to obstructions. During the tests, we did not believe the data, and reran the test. We also observed the same results on the second test. At the time we only had 20m antennas, so we could not confirm that enhancement was truly frequency dependent. But based on these results, more testing is warranted.

During the CQ WW CW test, we attempted to position the antennas not more than a 1/4 wavelength from the ocean. This would place them in the region of best enhancement, unless they were right in the surf. Of course the tide would come in and out, so the actual impact on our signal was unknown. We did assume the enhancement would fluctuate at certain tide levels. Maybe next time we will find out the tide schedule and set up a test vertical so it sees the enhancement at specified times and target locations.

Other key lessons:

During the two contests, we had problems with the salt water spray coating all the antennas, guy ropes, and radials. We had anticipated this problem, but we were not expecting that things would be as bad as they were. We had assumed that the nylon rope used to guy the verticals and radials was a fairly good insulator. It is when it's dry, but when it becomes coated with salt water, the rope becomes conductive! During the ARRL Contest, it was a fight to keep the radials in the air, as the nylon mason line would melt at high voltage points along its length.

We got a little smarter in the CQ WW Contest. We insulated all the verticals and radials from the nylon rope with 4 inch long insulators on the vertical elements, and small egg insulators (or tie wraps when we ran out) on the radials. This worked very well, but we did melt one guy rope on the 80 meter driver. After seeing what had happened, we knew we had seen this before, on the same antenna last February! We had (again) attached the guys above the linear loading, which is a higher voltage point. So, when the insulator became coated with salt water, it too became conductive: energy zipped across to the salt-saturated rope and the rope melted. The next time we will make sure we are truly awake and that the guy ropes are attached at low voltage points. The large tie wraps work very well for a lightweight, easy to use insulator. They can, however, also become so saturated with salt that they will melt at the high voltage end of a radial. This happened on the 15 meter driver during a very high tide, with heavy surf and strong winds, when the entire antenna (including the radials) was being drenched in a lot of salt spray.

Even with the occasional problems of keeping the antennas in the air, the 6Y4A experience with these antennas is best captured by one of the 6Y4A operators, Mas, JE3MAS: "I never thought such small and simple antennas could work so well".

Here is a summary of the 6Y4A antennas used in the CQ WW CW Contest:

  • 160m: A single element 55' linear loaded vertical. Vertical was approximately 30' from the sea, and had 4 elevated radials. The antenna weighed 35 pounds, but could be made lighter.
  • 80m: 2 element parasitic linear loaded array, fixed due north, and optimized for minimal front-back ratio. Each element had 2 elevated radials, and the reflector was approximately a quarter wavelength from the sea. The elements were made with 18' of 2" pop-together expedition mast, and 18' of lightweight small diameter tubing on the tip. Each quarter wave element weighed 15 pounds.
  • 40m: We used two antennas on this band, one fixed on Europe, the other fixed on the USA/JA. Each array was a 2 element parasitic ZR vertical by Force 12 Antennas (see front cover). The ZR is an electrical and physical half wave length long, center-fed shortened vertical dipole, and on 40m is only 15' tall. The rings on each end of the ZR form a single turn open inductor (canceling out radiation from each other), providing end loaded inductance. Thus the vertical section is the only radiating point. The vertical section of the antenna was made of light weight 1.25" diameter tubing, and each element weighed 15 pounds. The big advantage of the ZR is they are very quiet antennas, and close to the technical description of a magnetic antenna.
  • 20m: Two antennas were used on this band: 1) We called the first antenna the "2 by 2" array. The 2 by 2 array is comprised of two, 2 element parasitic arrays which are spaced ~ 5/8 wavelength apart, and are fed in phase. Each of the 4 quarter wave elements had 2 elevated radials. The antenna was fixed on Europe, had a beamwidth of around 55 degrees, provided at least 3 dB gain over a single 2 element array, and had a very large null off the side. The null was strategically oriented towards the USA so it would be easier to run Europe when the band was open. There was at least a 20db difference (and often larger) between the two antennas. 2) The second antenna was a 2 element parasitic array made with quarter wave elements, and was fixed to the USA/JA. The beamwidth of this antenna was approximately 100 degrees, and was located a little over a quarter wavelength from the sea. A WX0B Stackmatch was used to drive both antennas together, or separately.
  • 15m: We used 3 antennas on this band: The main antenna was a 2 element parasitic vertical array made with quarter wave elements with raised radials and was just a few feet from the ocean (often the driver was awash with breaking waves). The beamwidth was 100 degrees, and the antenna was fixed at 10 degrees to cover Europe, USA, and Japan. The second antenna was a 2 element horizontal Yagi at 28', and was turned by hand. The horizontal Yagi was supported by 24' of pop-together expedition mast that was taken down with the team. The mast was mounted on a small concrete equipment house which was 4' tall. The third antenna was an 8 element inverted-V Yagi fixed on Europe. The reflector was hung from a tree and was 40' high. The front director was at 20' high, and was around 50' from the ocean. This antenna was installed a few hours before the contest in an attempt to add more gain to this band due to the report of a solar flare. At the height of the opening this antenna was 3-5 dB stronger into Europe than the verticals, but the verticals opened and closed the band. A WX0B Stackmatch was used to drive all antennas together, or in any combination.
  • 10m: We used two antennas on this band: 1) The main antenna was a two element full size vertical dipole parasitic array fixed at 10 degrees to cover Europe, USA, and Japan. The antenna was a little over a quarter wavelength from the sea. 2) The second antenna was a 2 element horizontal Yagi at 33'. The mast was made of sea-drift bamboo poles that were lashed together. A WX0B Stackmatch was used to drive both antennas together, or separately.

No comments: