Designing the Giant Antennas

Synopsis of a presentation by Boynton Hagaman
At the April 2000 AMRAD Meeting

Thanks to AMRAD for granting permission to post this document.

The career of Boynton Hagaman, AA4QY, spans many years and has had many interesting turns. He is an engineer, designer and builder of some remarkable antennas. These include the Trideco antenna used by NSS, Annapolis, Maryland, the antennas used by NAA, Culter, Maine and many others. His talk encompassed a period of three decades, spanning the time period from the end of the 1950s through the Cold War and until today. This is a story of the giant antennas unique in their own right, and rare. There are a decreasing number of opportunities to see giant antennas. International broadcast stations used them, for example, the Voice of America transmitting stations. HF antennas are conventional in many ways because they operate at reasonable wavelengths. Not so conventional are the giants used to send signals at just a few kilohertz in the very low frequency band that ranges from 3 to 30 kilohertz. Wavelengths here are measured in the thousands of meters, making it difficult to design an efficient antenna. These antennas are extremely sharp and their bandwidth at resonance is very narrow. It is so narrow, in fact, that one experimenter actually used this sharpness as an SSB filter to eliminate an unwanted sideband from a double-sideband AM signal.

The antennas we heard about at the AMRAD meeting all operate in the band between 10 and 30 kilohertz and are used by the US Navy for one-way transmission to submerged submarines.

In the early 1960s, Lester Carr, Hagaman and a few associates formed a company called DECO, short for Development Engineering Co. At about the same time, a partner, Jim Weldon, started Continental Electronics, a company in Dallas Texas that manufactures high power medium, short-wave and VLF transmitters. At that time the Navy wanted to upgrade their VLF transmitting capability, which included stations located in different geographical locations to achieve worldwide coverage. Each station sent continuous coded traffic using a narrow form of frequency-shift keying of a few Hertz wide. This allowed submerged submarines to receive slow speed commands while remaining submerged.

Earlier, RCA built a station in this chain at Jim Creek in Washington state. This was another giant station generating a million watts into massive antennas that consisted of multiple cables stretched between adjacent mountaintops across a large valley. These antennas are known as valley span antennas. Although the performance of this station was not as efficient as planned, primary power was available at low rates and the antenna is still in use.

The Navy required a more powerful and more efficient antenna. Hagaman and his associates felt they knew of such a design and prepared a proposal. After proposing what they thought was a unique design they now worried that the Navy would ask them to actually build one. And they did.

Anyone who has done practical engineering knows this problem. Most times, the electrical problems are well understood and itís the practical engineering problems that occupy most of the time. This project was no different. The antenna design for NAA was called the Trideco and the chosen site was at Cutler, Maine. The antenna was required to have an efficiency of at least 50%, which, at that time, had never been achieved.

A German antenna was erected during the Second World War called Goliath. It was quite efficient but was limited both in power and efficiency by the wartime lack of copper and other suitable materials. It was used for submarine communications. After the war, the then Soviet Union seized the Goliath antenna, packed up its massive cables and towers, and assembled it inside the Soviet Union near Gorky. (Iím not sure where we learned about the design of the Goliath. I wonder if it was from captured German plans?)

NAA is located on a peninsula at Cutler, Maine. A diagram of this station shows two giant sub-antennas, each as large as several football fields and requiring twenty six 800-to-900 foot towers fed by two large transmitters. The antenna is split into a North and a South array. Each array is fed from a separate helix building and a transmitter building housing two transmitters is also located midway between the arrays. The transmitters can each supply one megawatt. Each antenna array resembles a six-sided compass rosette. The antenna feed point is located at the center of the rosette where the special building called the helix house is located.

Looking over a schematic of the helix house equipment you see what looks like a simple matching network consisting of a few coils called variometers and something called a reactor. A variometer consists of two coils, a stator and a rotor. The rotor is located inside of the stator and is connected in series with the stator. When the rotor is rotated the mutual coupling either aids or reduces the effective inductance of the variometer. Since the antennas are operated below their natural resonance their impedance is capacitive and must be tuned by a series variometer. However, when looking at the equipment inside the helix house you see nothing familiar. In reality, this is a link-coupled antenna matching circuit, but that is where the similarity to familiar low powered equipment stops. The helix house is jammed full of giant coils and openly wound transformers, most larger than a large truck. The inductor wires are about 4 inches in diameter consisting of multiple strands of Litz wire. The antenna itself is tuned with a series variometer while a shunt variometer couples the signal into the antenna tuner from a coaxial feed line leading from the transmitter through a long underground tunnel.

A part of this matching network called a "saturable reactor transformer" or a "stretchable" reactor is used to slightly change the tuning of the antenna. Direct current from the modulator is fed into a separate winding on the reactor transformerís core and saturates the transformer magnetic core slightly in response to the requirements of the frequency-shift signal. This causes a slight variation in the transformerís inductance thereby slightly shifting the point of resonance to keep the antenna in tune. Otherwise the frequency shift may cause the antenna band-pass to be out of tune. This change accommodates the antenna frequency shift even though the shift may be only a few Hertz away.

For example, NAA is known to have a very narrow bandwidth, depending on the operating frequency. The reactor increases the antenna bandwidth by slightly shifting the center frequency back and forth, along with the keystream. The reactor transformer is designed to allow up to thirty-five words per minute FSK keying, which is well above the rate that this system usually operates. In other designs, a resistor, called a bandwidth resistor, is used to reduce the antenna Q instead of a stretchable reactor transformer.

The last stage of the matching network is the helix that is simply a multi-turn link-coupling transformer. Like all the other components, it too is massive. A picture of the helix house interior shows the helix coil standing two stories in height and with the same diameter. A technician standing atop it is dwarfed by its size.

The selection of the metal alloy from which to strand the antenna wires was critical. At Cutler the wires (over one inch in diameter) were stranded of Calsum Bronze. This material has a higher resistance than copper and is very strong. The higher resistance was necessary so that the antenna conductors could be heated during icing weather conditions. It requires a megawatt of 60 Hz power to de-ice either antenna array.

Aluminum is relatively cheap and lightweight however it does not have the tensile strength required in long spans. Stranded Calsum Bronze cable is capable of providing low resistance to RF current due to the skin effect at radio frequencies yet has an appreciable resistance at 60 Hertz.

Sending current into the strand causes it to heat up and this heat is used to melt off antenna ice accumulations during the winter. The power required to melt the ice is supplied by a large Diesel generating plant located on the NAA property. At NAA during the winter months, one of the arrays is sometimes shut down to melt ice accumulations while the other continues to operate.

After installation of the antenna, engineers encountered something unexpected called Aeolian vibration. This was an effect known to builders of high voltage electric transmission lines. The Aeolian vibrations result from an interaction of the wind with the suspended cables. It starts being noticeable when a steady side wind reaches about five miles an hour and continuous for an extended period of time. Aeolian vibration caused unexplained breakage of some wires in the cable strand and stress on the antenna. Attaching "Stockbridge dampers" on the conductor strands at strategic locations can prevent it. These dampers absorb the vibration energy and dissipate it in the form of heat. Aeolian vibration may also be controlled to some extent by the very heavy insulators weighing up to 800 pounds each, that are used in the cables and tower guys.

Static electricity is also a problem in the large antennas. The accumulation of worldwide lightning storms, rain static and other similar weather phenomena result in the generation of a large potential difference between the ionosphere and the earth. Electrical charge is constantly leaking down to the earth and is not ordinarily detectable without special equipment. However, the charge accumulates around highly conductive structures such as tall towers thereby generating dangerous voltages at their top. A recorder placed at the top of the 1200-foot NSS tower at Annapolis recorded a three to five kV/meter potential difference in the atmosphere during an approaching thunderstorm. These high voltages cause flashover to occur across the insulators that can disrupt the transmitter and take the station off the air. Similar voltages have also been observed during snowstorms. Since flashovers cannot be completely eliminated, a special device was designed to deal with it. It uses an ultraviolet or current sensor to detect when current flow begins at the start of a flashover and briefly cuts the RF excitation power to the final amplifier for the duration of the flashover. This solves the problem in most cases but not under all conditions. The device is called the CCO (carrier cut-off) unit and is now used in some form at all of the Navyís VLF stations.

As previously mentioned, the antennas for the Navy VLF stations are giants. They are constructed to have a lot of capacitance on top of the towers. In fact, they are nothing so much as a large capacitor with a one plate formed by the tower- supported wire panels suspended 1500 feet or so in the air and the other plate being the earth. A million watts or so are fed into this circuit and, as you would expect, preventing corona discharge from the very high voltage is very important design issue.

Determining in what part of the antenna structure corona may be a problem is interesting. A scaled model of the antenna is built and as much as 50,000 volts at 60 Hertz is pumped into the antenna. During darkness, and when the adjustable test voltage is at its highest, the antenna lights up like a Christmas tree with hundreds of small individual corona discharges occurring at critical points on the suspended mesh. Obviously, some coronas are in the higher voltage locations and may require the antenna configuration to be modified. Boynton gave us an interesting picture showing a model antenna under test, and the individual coronas were visible. It looked like the model was decorated with many small Christmas lights that illuminated its outline against the night sky. These locations were modified until the corona did not appear.

The Trideco antenna, used until recently by NSS in Annapolis, is similar to the antennas used by other VLF stations in that they have several large supporting towers between 700 and 1500 feet tall. For example, in Annapolis several towers were used to suspend the top wire panels; the tallest being 1200 feet and its base was insulated from ground. The VLF station at Lualualei, Hawaii has similar towers (two base-insulated towers 1500 foot in height) although the antenna design is quite different. The towers are made of steel and remind me of those used in a large suspension bridge. The amazing thing about these towers is that they rest on a single porcelain insulator.

This design is something like an amateur vertical monopole using a Coke bottle for a base insulator. In this case, however, the insulator holds up an antenna that weighs several million pounds. All that weight concentrated upon this one insulator is truly amazing to see and it is a wonder how such a thing could hold all that weight. In fact, as we learned, there have been problems with them. Cracks have been found in base insulators that required them to be changed out! How does one hoist up an antenna tower supporting a mile square of steel cable over 1500 feet in the sky? Well, amazingly, a system of hydraulic jacks are available to do just that. Modern towers are fabricated with "jacking pads" provided on the corners near the tapered lower section. The antenna tower can be lifted, the insulator changed and the tower lowered back into place. Although there are no pictures of the equipment is use, you can image what it must be like when you consider that these insulators are nearly two stories tall.

So far, Iíve given you a pretty good idea of what Hagaman talked about, describing his days developing and building these giant antennas for the Navyís VLF stations. He told us that his company was located out in Leesburg, Virginia and thatís were they did testing on small-scale mock-ups of the giant antennas they built. He said that newer designs use a top-loaded monopole and this kind of antenna is a good radiator and economical to build. The now defunct Ground Wave Emergency Network (GWEN) stations, also designed by DECO, used this type of antenna thought they operated at LF and not VLF. Some Omega navigation stations also use the top-loaded monopole.

The US and Russian Navies use VLF for ship and submarine communications and both countries use giant antennas. The firm of Kershner, Wright & Hagaman recently designed a VLF antenna for a Southwest Asian country. It is similar to the Trideco array located in the NW part of Australia. Several other nations are planning VLF antennas also. With the end of the Cold War you might expected no further orders for giant VLF antennas but its not so. Additional nations now have submarines that must remain hidden and require one-way communications too. In that, they are practicing a Cold War art in place for the past fifty years.

AMRAD