You can't win.
You can't break even.
You can't even get out of the game.

And playing the game is not optional.

-with apologies to C. P. Snow


Breaking the laws of physics is a popular pastime. Who would not want to get a thousand miles per gallon, lose weight by taking a pill, or receive a fortune in cash from a Nigerian prince by email? But, most of us understand it just ain't going to happen. And so it is with antennas. Previously, I wrote about the tradeoffs associated with antennas. The most contentious corner of our tradeoff triangle seems to be size; it's the easiest to measure, and the one in most visible conflict with other aspects of wireless product design.

You want your antenna small, admit it.

Small antennas have been around a long time. Remember pagers? They most often operated in the VHF band, where wavelengths were on the order of two meters. Yet, they were tiny (although the volume of cell phones is approaching what later model pagers were) and the antennas they contained were even tinier. The facts that enabled their good performance were: narrow frequency range and receive-only operation. You can have small size if you give up bandwidth and efficiency. No problem. For many years, AM radios operated from 510-1800 KHz (with wavelengths in the hundreds of meters) using a ferrite loopstick at the core of their antenna; narrow bandwidth, low efficiency, and completely successful.

But, these days who wants to give up bandwidth? If size is the easiest thing to measure, then impedance bandwidth is a close second. An RF network analyzer can measure impedance bandwidth in a few milliseconds, and generate a curve which can be put in a data sheet, or used to compare to an existing data sheet. So, let's make a small antenna with wide bandwidth!! If we're lucky, nobody will ask about the efficiency (which is pretty hard to measure, but easy to experience).

And so, some companies manufacture Antennuators. These are a cross between antennas and attenuators. Between good and evil. Between protein and carbohydrates. Between Apple and Microsoft. Let's radiate some energy and burn some energy as heat. That way, it's kind of usable, and the measured impedance bandwidth is glorious!

There are several ways to build an Antennuator. The easiest method is to build it with lossy materials in the parts that carry RF current. One example of such a material is stainless steel. Stainless steel is a rather poor conductor, but its nice to look at and resists corrosion. There are many whip antenna made from stainless steel, and often the environmental considerations outweigh the loss from the material selection. And in certain antenna types, which are inherently high impedance (such as the Kraus Helical) the difference in material losses will be insubstantial. But, when antennas are made physically small the currents can get rather high, and this is where lossy materials will rear their ugly head.

A few years ago I was working for a company that manufactured various specialized receivers and transmitters operating in the VHF, UHF and low-microwave range. I was the in-house antenna engineer as well as a product designer. A good friend of mine, John D., came to my office with a question. He ran the test department which had the responsibility of tuning and testing all the equipment before shipment. A group of "tactical repeaters" were being tested, and one was on his bench transmitting a few watts of RF at VHF into a six- or eight-inch "rubber duck" antenna (a helically wound short monopole).

"Spence, should the antennas get warm?"

I turned and looked at him incredulously. After careful consideration I answered with a incisive question of my own, "HUH?!!?"

"After about five minutes the antenna is warm, and I don't remember this happening before. Is it normal?" John asked.

"Hell, no!" was my measured response which began a careful investigation into this Antennuator.

I verified that, indeed, the antenna was warming up substantially, and that it was not some other portion of the system making heat and warming up the antenna. After some digging, John and I determined that there was a recent vendor change from vendor "C" to vendor "A". I grabbed some examples of the parts from each vendor and started cutting them open (this is a recurring theme, as you will see, Dear Reader). The helical conductor from vendor "C" was copper-colored, and that from vendor "A" was stainless-steel colored. Hmmmm.

Then I took scraps of the plastic that encased each antenna and carefully tested the dielectric loss using an UHF RF Thermal Conversion Exposure Cavity. Yep. . . a microwave oven. The difference in thermal dissapation between the two plastics was very significant: after about 15-20 seconds of exposure (with a cup of water at the opposite corner of the microwave for loading) the plastic from vendor "A" got hot whereas that from vendor "C" was not noticably warmer -- lossy dielectric confirmed. (While this test was conducted at about 2450 MHz, it is still indicative of losses at VHF.)

The Antennuator from vendor "A" looked great on the RF network analyzer, and even better in the purchasing department's scorecard. But, it was burning precious (battery powered) RF in the process. I immediately wrote an ECO (Engineering Change Order) eliminating vendor "A" as a supplier, and specifically naming vendor "C" as the supplier for this particular part. Problem solved. Purchasing had been credited for saving money on cheaper antennas, and Engineering looks evil throwing out inventory. The Earth continued turning on its axis.

More recently, one of my clients brought me a product from a Serious Defense Contractor, which was labelled "Antenna, Broadband, 50-2000 MHz". It was about twelve inches long, about 5/8-inch in diameter, with a BNC male connector on one end. It was flexible, black, and no doubt expensive. Since my client was taking a training class, I immediately used it as an example of an Antennuator. I started a discussion with the class as to how a twelve-inch long antenna can be rated for operation from 50-2000 MHz. After a healthy amount of discussion, one of the students put the antenna on an RF network analyzer and swept it from 50-2000 MHz, measuring the input impedance.

"It looks good to me!" he said, showing a respectable "knot" in the middle of the Smith Chart, indicating a VSWR of less than about 2:1 (ref. 50-ohms) over the range.

I said, "You'd think so, but remember the engineering rule: if you can't fix it, at least you can break it!". I extended the low end of the measurement down to 2 MHz. And do you know what we saw? It still had a VSWR of about 2:1. . . . but, it shouldn't have. It SHOULD have looked lousy at 2 MHz. Now think, what has a VSWR of 2:1 at 2-, 50- and 2000-MHz? That's right -- a RESISTOR!

I made a bold (and risky) statement to the class: "There's a 100-ohm resistor in parallel with the connector.", I confidently proclaimed, hoping like hell nobody would call me on it. Well, I was with a group that did not see any impediment to cutting open the antenna and finding out Right Now. And so, they did.

I was wrong. There were two, 200-ohm resistors (each rated 3-watts) in parallel across the connector. As I see it, when RF current flows in a conductor the only thing it can make is radiation, but when RF current flows in a resistor the only thing it can make is heat. So, the ONLY possible reasons those resistors were there were to make the transmitter happy over that frequecy range (no spurious oscillation, no VSWR alarms), to make the datasheet look magnificent (and meet procurement requirements), or to prevent the buildup of ice. Unfortunately, my clients are professional communicators and in need of their equipment actually... you know... communicating, and they were already acutely aware of the inefficiency of this product. Putting "50-2000 MHz" on the body of an antenna does not make it so. Notice the efficiency was not stated on the same placard. Pity.

Once I designed an antenna for a small surveillance product that had a resistor in it. The product was battery operated, and had a very high efficiency transmitter. Sometimes, however, if the antenna came too close to a conductor, it presented the transmitter with a very low impedance that caused a spurious oscillation. This was Very Bad in the intended application. I ended up using a small series resistor in the feed loop of the electrically small loop antenna. The cost of this resistor was a fraction of a dB of transmitted signal in normal operation, but it prevented the spurious oscillation completely and the associated loss of signal. This was a carefully weighed decision, it solved the problem, and the efficiency cost was calculated and acceptable. The problem could have been solved further upstream in the power amplifier, and there would have been an efficiency cost there, too.

Resistors turn current into heat. That's what they're supposed to do, and thank goodness they do it so well. But, when you find 'em in antennas, it's worth asking why they're there.

Sometimes, it's useful to burn undesired energy of the wrong polarization as in a Terminated Bifilar Antenna or a Log Spiral Antenna, for example. In that case, polarization purity (axial ratio) is more valuable than efficiency. That is an engineering decision, and a good one.

Sometimes, it's critical to keep the transmitter happy and invest in a bit of heat to do so.

Sometimes, it's to make the datasheet look miraculous and score a big order from the government.

Perhaps someone should ask the professional communicators if the solution to their very real problems is an Antennuator. I think not.


Spencer Webb