Due to some recent happenings, there needs to be some clarification about antenna shootouts. One of those is reconfirmation of their validity, or lack of it. The second issue is how the results are used. Some amateurs believe the results are all-encompassing, as do some manufacturers. This said, unless you know all of the parameters used in any specific shootout event, any conclusions drawn from the results are suspect at best.
One of the most popular mystery novels of all time, gave away the whodunit in the first few sentences. It was the story behind the perpetrator which made the novel what it is. I'll do the same thing here by stating; "Antenna shootouts prove (almost) nothing!" Why they don't is very difficult to put into simple, easily understood terms.
Not too long ago, I made reference to two cars with two different antennas, where in switching the antennas would improve the performance of both mobiles. I even predicted at least a 3 to 5 dB increase in both. Why this is so, it part of the mystery we're trying to solve here.
Part of the problem is understanding the factors which determine an antenna's efficiency, and this article, Antenna Efficiency, covers the basics. But it's more than just efficiency per sé, it is antenna performance which counts. Unfortunately, very little thought goes into antenna performance at purchase time. Nominally, the choices are made based on ease of installation, including size and length, and to a lessor degree on purchase cost.
The best quality, highest efficiency, antenna money can buy, isn't any better than the ground plane it is mounted atop. As pointed out in the aforementioned article, and others on this web site, mounting methodology (where and how) is the prime factor in reducing ground losses.
One very important point needs to be inserted here. A vehicle is not a ground plane, but rather a capacitor between the antenna and the surface under the vehicle which acts as the ground plane. Since the surface in question is a poor conductor of RF, ground losses occur. The term ground plane in the following text is therefore a bit of a misnomer, but is used to differentiate it from DC and RF grounds.
Mobile HF antennas come with a variety of coil sizes, overall length, diameter, and where within the antenna the coil is located. For any given antenna, coils mounted higher up require more inductance to obtain resonance, than ones mounted closer to the bottom. What's more, the optimum position of the coil is directly related to how much ground loss there is, and to a lessor degree the coil's Q.
Ground losses aren't easy to calculate or measure, and neither is coil Q, especially when the coil is mounted within the antenna. Making matters worse, we really don't know, at least with certainty, the radiation resistance (Rr). One thing we do know, Rr is a function of length. Double the length of the antenna, and the Rr increases by four times. This means, that a 9 foot antenna will have twice the Rr as a 6 foot one.
Another issue which comes to light quite often is using NEC and EZNEC to predict a given installation's efficiency. While these programs are good for what they are, they do not predict ground losses accurately. When folks enter the data manually, they almost always under estimate the losses (and over estimate coil Q). The Antenna Efficiency article does cover a methodology which compares calculated to empirical measurements to estimate an antenna's efficiency, but that's exactly what it is, an estimate. One unfortunate aspect of calculating efficiency, folks tend to take the data out of context, and apply it where it really doesn't fit. Apples to oranges as they say.
There's another important item with respect to ground losses, and that is consistency in ground conductivity. While the mean deviation over a large statistical area may be fairly narrow, over a small statistical area the mean deviation can be rather drastic. What's more, the mean deviation in soil conductivity changes as the moisture content, and surface temperature changes. In fact, the changes are often great enough, that you can measure the difference in input impedance between morning, and evening. I have never witnessed a shootout where this factor was considered, or even mentioned!
Rudy Severns', N6LF, wrote a series of white papers on the effects of ground loss. His empirical testing confirmed an important fact about vertical antennas. To wit; Any practical ground system will not affect the radiation angle or far-field pattern! The ground system around the antenna does nothing for the far-field pattern except to increase the power radiated for a given input power.
With the last two paragraphs in mind, there is one more factor to consider. Every single object in the near field has an effect on the antenna—and its measured parameters—within that field. Those objects may be trees, bushes, vehicles, other antennas, signs, billboards, people, you name it, all will have an effect.
The term abstraction refers to the process of considering something independently of its (other) associations. In a recent post on eham.net, Todd Moore, K1TM, said it best: "Antenna shoot outs crack me up. We live in a world where it is the complete system that counts and the shoot out is trying (to) extrapolate the performance of the antenna from a measure of the system and none of the systems have been normalized. They are indeed a very unscientific comparison."
The key word here is, normalized. In fact, most shootout organizers go out of their way to make sure they're not, albeit unintentionally. One very popular way to minimize the changes in ground loss between vehicles, is to use just one vehicle as a test bed. All of the antennas under test are then mounted on this vehicle. Aside from the mounting logistics, this doesn't address any issues with respect to ground losses versus coil position.
The next question should be, where on the test vehicle are the antennas mounted? If the mounting position is low, trailer hitch mounting for example, then the antenna with the least mast diameter is going to have an advantage. In simple terms, the larger the mast diameter, the more shunt (capacitive) coupling to the ground and vehicle body becomes a factor. You can't measure this coupling with any degree of accuracy, so it's typically ignored. It should be noted that you can easily measure the increase in input impedance caused by shunt coupling, but not shunt coupling (capacitance) directly.
I should add here, that the shunt coupling increases with increasing frequency. As a result, antennas mounted close to the body become less efficient as the frequency increases, which is the exact opposite that one would otherwise conclude. If one wanted to address this issue, then multiple measurements (multiple bands) would be necessary. I've not seen this factor considered either.
One obvious need is a calibrated RF generator. Ideally it will be battery powered, and exhibit a long-term stability of at least one-tenth dB. There isn't an amateur transceiver on the planet that's this stable, yet most shootouts employ one. Irrespective of the use of a wattmeter, the reliance on an amateur transceiver as a repeatable, calibrated output source is pure folly.
You also need a calibrated receiver. Here too, you cannot rely on an amateur transceiver, no matter how you go about it. This includes the use of switched attenuators, and padders of any kind. Perhaps not so obvious, most transceiver readouts vary with ambient temperature changes with no correlation between them.
Most modern transceivers are capable of receiving very weak signals. For example, the Icom IC-7000 will produce a S+N/N ratio of 10 dB (SSB, 1.8 to 30 MHz), with as little as .15 uV. However, the vast majority of the time, band conditions require signals much stronger than this to produce the same S+N/N ratio. How strong is the issue, but even more of an issue is the fact band conditions change very rapidly, irrespective of the sampling period (minutes, hours, or days). These changes effect both near and far field signals, and cannot be normalized.
And what about the antenna? Not the antenna under test mind you, but the one you're using to transmit with (or receive on)? The first question is, why do most shootouts use a loop receiving antenna? If you know anything about antenna patterns, you sure wouldn't be using a loop! Especially one with an unknown impedance. Keep in mind that amateur transceivers are designed to transmit into a 50 ohm load. That fact does not mean the receiver is also 50 ohms. In fact, most aren't. As a result, the change in signal strength caused by a variation in VSWR can easily exceed the measured difference between the various antennas under test. If you doubt this premise, you can empirically test it with a stepped attenuator and calibrated signal generator.
At the onset of this article, I mentioned antenna performance, rather than gain (or lack of it). The reason is simply this; While the reciprocity gain rule applies to all antennas, transmit versus receive performance does not! This is especially true of HF mobile antennas where the body of the vehicle distorts the pattern. Albeit considerably less than most folks believe, the difference is enough to negate any single point of measurement scheme, no matter the style of the opposing antenna!
There are additional problems hidden in the last two paragraphs. Considering the receiver's input impedance, the antenna's capacitive coupling and input impedance, and the distortion in the antenna's pattern caused by the body, there will be a difference between transmit and receive signal strengths, all else being equal. This fact could favor one antenna over another, irrespective of their quality, perceived or otherwise.
As mentioned in the Antenna Efficiency article, Dr. Belrose, VE2CV, authored an article for the September 1953 issue of QST (page 30). An updated version appears in the ARRL Antenna Compendium #4, starting on page 83. The established ground plane losses based on the original article are still used today (2 through 10 ohms, 10 meters through 80 meters, respectively). However, poor mounting techniques can easily double this, and remember, ground and shunt losses can't be measured directly.
Taking all of the variables into account, it is evident that normalizing an antenna shootout is almost an insurmountable task. That said, there is some data one can take from those shootouts albeit general in nature. First, the antennas which win are almost always mounted in the center of the roof. The main reason is, ground losses are at a minimum. What's more, they're usually longer than local conditions would otherwise allow. In other words, they're there to win shootouts, not to drive around with.
Almost as good are those well mounted atop quarter panels or bed rails. Then there is usually a big measurement gap (no middle ground seemingly). These toward-the-bottom-of-the-list antennas tend to be spirally wound and/or short and stubby, and mounted on lip type mounts.
Probably the most glaring data is this fact; antennas mounted on trailer hitches and frame mounts (low to the ground), are typically at the bottom of the list, regardless of their overall length. Adding a little insult, most are heavy antennas with large diameter masts. All of these factors increase apparent ground losses, and are reflective in the (unmatched) input impedance. As John Belrose, VE2CV, stated in the aforementioned Antenna Compendium #4, "There is more to consider than appearance and convenience when installing your HF mobile antenna, since the frame and body of the vehicle are a part of the radiating system."
There are some variances of course, especially when cap hats are used. However, far too many folks are led astray by the change in input impedance when they install a cap hat. Thinking, of course, that the Rr increased, when in reality the Rc losses increased (they're both part of the total input impedance). You can generally say they do increase efficiency, but that largely depends on implementation.
For example, too low (close proximity of the coil as shown left), and they're a detriment, not an asset. Small ones, no matter how they are configured, aren't worth the effort. Putting all of this another way, cap hats are (almost) always too small in diameter, their spokes are usually not electrically connected, and they're typically mounted too close to the coil.
If you want a cap hat to work correctly, it must be at the top of the whip, at least as large as its height above the coil, and the ends must be electrically connected.
Digressing. Since the original publication of this article was written, I have added one specifically about cap hats. However, there are a few points which need to be emphasized with respect to shootouts.
First, any cap hat will exhibit the same capacitance effect qualitatively, no matter where within the antenna's superstructure it is mounted. And, the changes seen in the input impedance will be about the same; it will increase! Whether that increase is caused by an increase in radiation resistance, or by a reduction in coil Q, depends on the implementation of said cap hat. Let's look at this from a slightly different angle.
Assuming we have two, absolutely identical antennas, replete with cap hat installed. It really doesn't make much difference if the cap hat is correctly, or incorrectly implemented, although the effect is greater in the latter case as will become apparent. We mount said antennas on different vehicles, but with identical ground losses (a rare case indeed). Could we therefore assume the measured results would be the same? The answer is no! If you read the cap hat article, you'd already know the answer, but here's a hint; it's the difference in capacitive loading between the antenna structure, and the body of the vehicle it is mounted atop. This fact adds a level of complexity which can't be normalized, because you can't measure the coupling directly. It is also why, multiple, and large tubular cap hats are a lousy idea! It should also become apparent why large cap hats should be mounted as far away from the vehicles superstructure as mechanically possible.
As alluded to above, the mean angle of radiation (point of maximum power) doesn't vary with the amount of ground loss. However, the power at any given take off angle does vary as can be clearly seen in the chart at left. Since the receiving antenna is at a fixed height, a case could be made wherein a winning antenna would actually have less low angle radiation, than a losing one. The question then would be, which one is a better DX antenna? The answer is moot of course, but it does add one more variable to the equation.
It also exemplifies (one reason) why antennas have different levels of receive verses transmit performance. Please note, I didn't say gain (or lack of it). This concept is difficult for some to understand, even when they model their antennas in EZNEC. Again, this harkens back to ground loss figures, which are almost always under estimated, even by the well informed.
Less obvious is the need to measure antenna performance on more than just one receiving height. While I realize the added cost, and record keeping these multiple measurements would entail, the results from such tests would bring us closer to the real winner!
It is difficult for some people to comprehend the notion, that a lowly hamstick could outscore a well designed, sturdily made, remote-controlled, HF mobile antenna. Yet, it does happen, and for good reason. The Antenna Efficiency article lists the parameters required to achieve an efficient design; at least as efficient as the inherent design constraints allow; overall length for example.
One major design faux pas, is to use large, metallic end caps, and place them within the field of the coil. Doing so greatly reduces the coil's inherent Q. In some designs, there is also a metal shorting plunger. Besides the additional losses incurred by short tapping the coil (shorting out unused turns), these metal masses cause the coil to operate very close to its self resonant point which also increases coil losses (reduces Q). At some point, these Q reducing factors add up, and the Q effectively becomes zero. At that point, the coil starts acting more like a rather lossy capacitor (operating above self resonance) than an inductor. At which point this occurs in any given antenna design, depends on a lot of factors, and ones which aren't easy to measure or calculate. They aren't easy to comprehend either, unless you have a fair understanding of how loading coils behave.
Cap hats have become ubiquitous fair for those wanting the best out of their antennas. However, when we incorrectly install a cap hat (as shown in the top left photo), by mounting it close to the coil instead of at the top of the whip, we add additional metal within the coil's field, further exacerbating the aforementioned problem (and don't forget; capacitive coupling is also an issue). Therefore, it is easy to comprehend why a lowly hamstick which has been optimized with a correctly installed cap hat, could outscore an apparently, well-designed antenna with an incorrectly installed cap hat. This can occur regardless of the measuring methodology, or its Art of Abstraction.
All of this points out yet another esoteric truth, and that is, mobile HF antennas are a lot more complex than a synergistic assembly consisting of a mast, a coil, and a whip. Design one correctly, and you have a winner. Design one incorrectly, and you have the proverbial dummy load on a stick!
As mentioned above, you must view every component, especially the unknown ones (variables), as part of the overall system. When you don't, the differences between individual antennas, measured in a few micro volts or tenths of a dB, becomes very suspect.
If shootouts rated antennas in a tiered system (poor, good, best for example), their data might be more meaningful.
If there is but one thought to take away from this article, it is this; Reducing ground loss (and to a lessor degree, shunt capacitance losses) is the key to improving efficiency. While you can't measure it directly, you can measure a change in ground loss by using an antenna analyzer, but you still have to be careful. That is to say, assume nothing you can't measure and/or compute directly.