Home Featured Articles There’s No End in Sight for Waveguide

There’s No End in Sight for Waveguide


by Steven Pong, Product Manager, Fairview Microwave

The RF and microwave industry may be awash in chipsets, but less flashy components still make up a sizeable portion of the market, and there’s no better example than waveguide. Considered archaic “plumbing” even by many microwave engineers, waveguide’s advantages still make it the best and sometimes only choice in many applications. Even as many microwave applications move to low RF power and millimeter wavelengths, waveguide has a lot to offer.

Over the years, waveguide has been used in a broad array of applications, from delivering the power from radar, electronic warfare, and other defense systems, to satellite terminals, broadcast transmitters, microwave ovens, medical systems, linear accelerators, guided wave testing for pipeline inspection, and dozens more. Perhaps the most ambitious application of waveguide ever undertaken was AT&T’s WT4 Long-Distance Buried Waveguide System that was intended to provide a nation-spanning voice, data, and video communications network at millimeter wavelengths using waveguide installed underground.

Waveguides transfer RF energy with greater efficiency than any medium, as their dielectric is air rather than a plastic like polyethylene or PTFE used in coaxial cables. Their large surface area also reduces copper losses below those incurred by two-conductor cable because although the latter’s outer conductor may be large compared to wavelength, the surface area of the inner conductor is small. When combined with the skin effect that causes current to flow on outer surfaces, the conductor’s small available area for carrying current makes the cable less efficient. Dielectric loss is also greater in coaxial cable owing to the heating of insulation and breakdown resulting from standing-wave induced voltage spikes. Breakdown can also occur in waveguide but at much higher voltages.

Figure 1: Large rectangular waveguide delivers the massive amounts of RF power required to excite the Argonne National Laboratory’s Advanced Photon Source

Waveguide also has disadvantages, the most obvious of which is that  its physical size increases as frequency decreases (Figure 1), making it unusable at frequencies below UHF: a rectangular waveguide for 300 MHz is 23 x 11.5 in. and 500 ft. at 1 MHz. When compared with coax, installation of waveguide is a major project, especially on towers for UHF television broadcast, where it has long been a standard feature. Special couplings are required between waveguide lengths, and in rotating radar, a coupling called a rotary joint is required to connect stationary waveguide to the rotating antenna. Radar has been a major application for circular waveguide as turning a rectangular waveguide would distort the magnetic field. A cost disadvantage of waveguide is that the metals often plated on the interior to reduce losses are silver or gold. These disadvantages are often mitigated by the fact that for some applications, no other transmission media can handle the required power or has the very low loss of waveguide.

Dueling Inventors

Waveguide and all the applications it made possible owe much to George Clark Southworth (Figure 2). His experiments in the 1930s arguably made it possible for the development and use of radar in World War II (for which he was awarded the IEEE Medal of Honor). His work and that of others, played a major role in enabling the first radar systems, which were instrumental in winning the Battle of Britain and on other fields of battle. It is not an overstatement to say that without waveguide and waveguide components such as the rotary joint, there would have been no way to transfer RF energy from magnetron-based transmitters to the rotating antennas used in the radar systems during World War II, the Vietnam and Korean wars, and to a large extent in every other conflict since then.

In what must hvae been one of the most frustrating announcements in the history of electromagnetic energy, Southworth at Bell Laboratories and former colleague Wilmer Barrow, who moved to MIT, learned by accident that they were doing the same waveguide work. Both had planned to announce their latest achievements at a meeting of the International Scientific Radio Union, a potentially embarrassing situation. Southworth found this out when he looked at the conference program and noticed that Barrow was to present his paper on the second day of the conference while Southworth was to deliver his the day before.

Figure 2: George Southworth in 1943 standing in front of an experimental waveguide line he used in his original research during the 1930s

This situation resulted in a flurry of letters between the two, each trying to keep from stepping on the toes of the other. Barrow finally discussed the subject with Professor Vannevar Bush, then an MIT professor and later head of the Office of Scientific Research and Development, which was responsible for development of the atomic bomb. Bush suggested Barrow cooperate with Southworth, which more or less settled the issue for the moment.

Power up the Tower

UHF television broadcast applications are another classic waveguide application, as insertion loss is a critical specification and long runs of transmission line are required along with the need to handle hundreds of kilowatts of RF power. The FCC’s recent “incentive” auction is a good example of the continuing importance of waveguide in this application.

The auction, concluded early in 2017, was the first to provide an incentive rather than a mandate for over-the-air broadcasters to change channels, in this case to free up spectrum for use by mobile broadband services. The incentive was straightforward: if a broadcaster agreed to abandon its current channel and move to another one, it would get a share of the auction proceeds in exchange. As the auction was a success, more than 1,000 stations in the U.S. will be participating.

A change of this magnitude presents an opportunity for broadcasters to modify some of their equipment, including the type of transmission line from the transmitter up the tower. Broadcasters have three types of waveguide to choose from: rectangular, circular, and elliptical. Rectangular waveguide has long been the choice at UHF, but it has the highest wind loading of all waveguide types, and while circular waveguide provides even greater efficiency, lower installation cost, and resistance to twisting, its dimensions must be precisely maintained along its length to avoid changes in polarization characteristics.

Elliptical waveguide, which has long been used in semi-flexible (Figure 3) systems for various microwave applications, has not been widely used in the broadcast industry. It can cover broad bandwidths, is comparatively easy to install, very efficient, and has far less wind loading than rectangular waveguide. However, it can be large at UHF broadcast frequencies and has been difficult to manufacture. At least one manufacturer is now providing an appealing solution, so broadcasters can now decide to retain their rectangular waveguide or evaluate circular and elliptical waveguide as well.

Figure 3: Fairview’s SMW112TF005-12 WR-112 twistable, flexible waveguide operates from 7.05 GHz to 10 GHz with 0.12 dB insertion loss for a 12 in. section

Millimeter Wavelengths and 5G

The wireless industry, in which waveguide has been used for macro base stations, may present a new opportunity in the coming years, as 5G systems will eventually use millimeter-wave frequencies at 28, 60, and 72 GHz. Millimeter-wave frequencies have never been used for cellular communications other than for microwave backhaul, as propagation distance is limited by high signal attenuation caused by the atmosphere, the weather, and almost any type of obstruction.

Communication at millimeter-wave frequencies requires a line-of-sight path from end to end, which mandates the use of highly directional, high-gain antennas with very narrow beamwidths. But as there are few services competing for spectrum at millimeter-wave frequencies, an enormous amount of interference-free bandwidth is available, which is a necessity for accommodating the extremely high data rates of current and future cellular applications. It should be possible to use millimeter-wave spectrum not just for connecting base stations to mobile devices, but also to link base stations to other base stations or back to the switch.

The inherently low loss of waveguide makes it well suited for use at very high frequencies, where the cost and insertion loss of coaxial cable are very high, making it viable only in very short lengths. As the dimensions of waveguide (and coaxial cable) decrease with frequency, waveguide becomes quite small and of little consequence in many systems. For example, while a WR-15 waveguide operating between 50 and 75 GHz has an inside dimension of only 3.7 x 1.9 in.

Waveguide is also inherently more rugged than flexible coaxial cables at very high frequencies, can handle more power than semi-rigid cables, and has far less insertion loss. As a result, waveguide and waveguide components such as filters are a staple of millimeter-wave measurement systems, radio links, satellite earth stations, and residential receive terminals.

Two examples of millimeter-wave waveguide components from Fairview Microwave illustrate the size differences and other characteristics of millimeter-wave waveguide. The SANT-2010 (Figure 4a) is a WR-15 vertically-polarized, omnidirectional waveguide horn antenna that measures 4 in. x 1.8 in. and operates from 58 to 63 GHz. The SMF-12S001-06 WR-12 (Figure 4b) is an instrumentation-grade, 6 in. WR-12 waveguide section made from oxygen-free copper with a gold finish to optimize RF performance and built to tolerances of +/- 0.001 in. that operates from 60 to 90 GHz. VSWR is only 1.06:1 and insertion loss is 0.44 dB/ft.

Figure 3: Fairview’s SMW112TF005-12 WR-112 twistable, flexible waveguide operates from 7.05 GHz to 10 GHz with 0.12 dB insertion loss for a 12 in. section

So, while highly integrated system-on-chip devices will dramatically reduce the need for discrete transmission line at the highest frequencies, networks operating at lower millimeter-wave frequencies can make use of waveguide’s low loss for transmission lines, as well as components such as filters, diplexers, duplexers, and antennas.


Since it achieved the feat of helping to bring radar systems to fruition during World War II, waveguide has proven to be one of the most valuable components ever made by the microwave industry. Thousands of ground, air, sea, and space platforms rely on its ability to transfer energy with the lowest loss of any transmission line and its unparalleled power handling ability.

And although most future radar and electronic warfare systems will use the Active Electronically-steered Array (AESA) architecture and thus will have no need for waveguide, there are many other commercial, industrial, and scientific applications that will. One of the most surprising is the fifth generation of cellular technology, which will have large numbers of very small, highly-integrated devices. However, 5G marks the first time the industry has ventured into the millimeter-wave region, where like all components including waveguide become very small. In short, the day when microwave plumbing is no longer useful is a very long way off.


When Waveguide Buried the Competition

by Steven Pong, Product Manager, Fairview Microwave

As waveguide has the lowest insertion loss of any type of RF and microwave transmission medium, it would seem to be a perfect fit for long-haul communications. Or so it seemed in the 1960s and 1970s when AT&T and Bell Laboratories developed the details for building, installing, and operating a 4,000 mi. long network of buried circular waveguide for voice, data, and video communications. The massive initiative that evolved over more than a decade was ultimately abandoned after optical fiber made lightwave communications possible, but even by today’s standards the waveguide system’s performance would have been remarkable.

The WT4 Long-Distance Buried Waveguide System would have accommodated 238,000 two-way voice circuits or various combinations of voice, data, and video using two-level, differentially coded, phase-shift-keyed modulation to achieve a continuous data rate of 274 Mb/s. The system could be upgraded to four-level modulation to provide capacity of 476,000 voice channels by modifying the electronics, without adding repeaters.

There would be 124 broadband channels, of which 59 would be used for protection in each direction, and are designed to have higher reliability than the current transmission media, such as buried coaxial cable and microwave repeaters, with predicted outage of 0.025 per two-way, 4000 mi. connection. A fully-loaded system was expected to have a cost per circuit mile far lower than any existing long-haul system and was to be installed on rights-of-way using standard construction techniques.

Figure 1: AT&T had great expectations for WT4, which is was none too shy in promoting in this ad in Time magazine

One of the factors driving development of WT4 was the world’s first videoconference system, called Picturephone (Figure 1), on which AT&T would ultimately spend more than $500 million (nearly $2.3 billion today), that was introduced at the 1964 New York World’s Fair. Although a Picturephone system was commercially deployed in Pittsburgh in 1970 and in Chicago the following year, it was taken off the market a few years later.

Its failure was the result of several factors, not the least of which was the cost to the user, which initially was $160 per month for the equipment and 30 min. of calls. Additional calls cost 25 cents per minute. This was later reduced to $75 per month for 45 min. of calls to increase the number of customers, but at its peak the Picturephone service generated only 453 calls. As it was usable only if parties at both ends had the equipment and service, there was obviously little demand. And it turned out that most people really didn’t like being seen while they were talking.

A Massive Challenge

Bell Laboratories’ scientists were old hands at making the seemingly impossible a reality, and as waveguide had the bandwidth and exceptionally low loss required in such a network when excited in the TE01 mode, the idea of creating such a huge expanse of waveguide seemed achievable. Like all emerging applications, there were enormous technical hurdles to be overcome as well as those encountered in building a 4,000 mile long, waveguide-filled pipeline in widely varying terrain and environmental conditions.

The first exploratory development work was conducted in 1959 using 2 in. waveguide and traveling-wave-tube (TWT) repeaters, but was abandoned in 1962 because of TWT cost and reliability problems. There also wasn’t enough Bell System traffic at the time to warrant continued work. But in 1968, after six years of continuous long-haul growth and the emergence of solid-state devices such as IMPATT diodes, an all-solid-state system was undertaken as a joint project of Bell Laboratories, Western Electric Company, and the Long Lines Department of AT&T. The new system, the WT4A, would use all-solid-state regenerative repeaters, replacing the problematic TWTs.

Figure 2: Illinois Bell Telephone President John de Butts in Chicago talks to First Lady Lady Bird Johnson in Washington in 1964.
(Source: AT&T Archives and History Center)

The dielectric-lined circular waveguide consisted of a steel tube copper-plated on the inside and lined with a thin layer of polyethylene that could reduce mode coupling (and thus loss) in bends. This type of waveguide would be used in 98% of the transmission path, with sections of helix waveguide inserted periodically along the length to limit transmission deviations. Both types had an inside diameter of 60 mm, which was determined to provide the least loss across the frequency band from 38 GHz to 104.5 GHz. Loss was calculated to be an incredibly low 1 to 2 dB per mile, which would allow repeater spacing to be every 37 mi. in gentle terrain and 31 mi. in rugged terrain.

The final shot across the bow of WT4 was the first commercial fiber-optic communication system developed in 1975 that, along with the GaAs semiconductor laser, made long-distance communication possible and effectively eliminated the need for other transmission media. When compared with the enormous capacity and speed of optical fiber, waveguide seems archaic, but more than 40 years ago there was simply nothing that could compete with its extraordinary low loss and high throughput.