After more than a quarter century of development, it seems logical that the traveling-wave tube has become all that it can be. If that’s so then it’s equally logical that as solid-state technology marches forward, TWTs will slowly fade away, to be found only in museums, textbooks, and a page on Wikipedia. Neither assumption is likely. What’s more likely is that everyone reading this will have expired before the last TWT rolls off the production line.
This rather contrarian projection flies in the face of the experience of other technologies that have tried and failed to survive the onslaught of semiconductors, but it’s happening today with TWTs and other Vacuum Electron Devices (VEDs). How will they survive? The answers range from the mundane (hundreds of thousands of TWTs installed in an enormous number of air, sea, land, and space EW and radar systems require spares), to technical (no other technology can generate as much power from a single device, especially at millimeter wavelengths).
The technology is also still vibrant after all these years, with active research being conducted throughout the world. Some of the results are already in production, as techniques developed have increased operating life by 50%, reduced “light up” time from minutes to a second, and increased RF output power levels hundreds of watts–CW–can be achieved at 80 to 90 GHz and up to 40 W at 230 GHz. Combined with the already impressive bandwidths achievable by helix-type TWTs, reasonable efficiency, reliability high enough to power transmitters in space for more than 15 years; the result is a technology that is, if not in ascension, then certainly formidable for many years.
Of course, like all technologies, TWTs have disadvantages. Unlike gallium arsenide (GaAs) and gallium nitride (GaN) MMICs they aren’t well suited for powering the tiny T/R modules required by AESA-based systems like the Next-Generation Jammer or the Air and Missile Defense Radar (AMDR). They still require kilovolts of DC power to create their electron beams and (currently) have a maximum operating lifetime of 100,000 hours, while GaN power amplifiers can (potentially) achieve 1 million hours of operation.
There is no question that GaN discrete transistors and MMICs will power most of the radar systems and EW systems in the future, thanks to the exceptional characteristics of this compound semiconductor technology. GaN can produce more RF power than any other competing solid-state technology (at least four times higher than GaAs and 10 times higher than silicon) and it can do so on a die the size of a pinhead.
GaN is also more efficient than GaAs and silicon, can operate up to 225° C, has a lower noise figure than GaAs, and can rearch well into the millimeter-wave region. GaN won’t realize its full potential for many years and possibly decades. But even then, there will still be a need for VEDs, especially at millimeter wavelengths when high RF output power is required.
To understand the future of the TWT requires some knowledge of its components and basic operation. A TWT, or any VED for that matter, bears no technological resemblance to its solid-state counterparts, electrically, mechanically, or physically (Figure 1). Using a helix type as an example, the device consists of an electron gun, interaction circuit, a collector (housed in a vacuum inside a tube) and magnets that surround it on the outside. The walls of the tube are made of heat- and corrosion-resistant metals like tungsten and molybdenum, or materials such as iron, high-purity copper, and high-temperature ceramics.
A high voltage is applied to the cathode, which heats up to about 1000° C and produces a stream of electrons that are then highly focused using the magnets surrounding the tube to form a very thin beam that travels axially away from the cathode. Along the way to the collector at the other end of the tube, this pencil beam of electrons passes inside of a helical coil (the helix) at the beginning of which an RF input signal is injected via a directional coupler.
The result is that the beam current, which was unmodulated as it entered the helix, now has an RF component at the input frequency. The modulation induces electromagnetic fields on the helix, which then act on the electrons, producing a massive amount of gain, dramatically increasing RF power. This RF energy, now amplified but still at the input frequency, is removed from the tube with another directional coupler.
However, only about a third or less of the beam power is converted to RF energy, and the rest is sent to the collector. So, if the collector can extract more power, the TWT’s efficiency will increase. To obtain this benefit, modern TWTs used a depressed collector with up to five stages (a multi-stage depressed collector or MSDC), which is essentially a series of electrodes. “Collectively,” they can recover more than 80% of beam energy that would otherwise be lost. Overall efficiency of 65% or greater can be achieved this way.
Helix-type TWTs are inherently broadband, with the ability to cover up to one octave and a maximum operating frequency of about 70 GHz. Their disadvantage versus the coupled-cavity type is that, being a wire, the helix is limited by its thickness to a comparatively low RF output power. The coupled-cavity type solves this problem by replacing the helix with a series of coupled cavities placed along the beam, producing a helical waveguide structure that can handle much higher power levels. However, it can only operate over a 10% bandwidth and is considerably heavier than the helix type.
As noted earlier, advances have been taking place in VED technology over the years, resulting in achievements that have reduced some of their limitations and expanded their capabilities. For example, TWTs generate their electron beams through thermionic emissions (i.e., producing electrons from a heated source), and it has traditionally taken 3 minutes or more for the cathode to heat up to a temperature sufficient to produce the desired emissions level. In many applications and especially missile seekers, rapid warm-up times are critical, so techniques have been developed to reduce warm-up time to between 1 and 3 seconds.
Although TWTs have demonstrated exceptional reliability up to about 100,000 hours, their life is ultimately limited by the heated cathode that uses a barium film as the means of “liberating” the electrons from the cathode structure. Advances have also been made in Microwave Power Modules (MPMs) that are essentially small “mini-TWTs” and a solid-state driver amplifier combined with an electronic power conditioning circuit (Figure 2). They operate at lower supply voltages than traditional TWT amplifiers (TWTAs) and compensate for the gain reduction caused by the shortened helix with the driver amplifier.
The Department of Defense and other agencies have long considered the millimeter-wave or Extremely High Frequency (EHF) region a potentially lucrative resource waiting to be exploited if only technology could allow it to be done. EHF officially begins at 30 GHz it extends to 300 GHz where the low infrared optical region begins. Nothing is simple at EHF, where wavelengths are measured in millimeters or microns, but there are a few significant benefits to be obtained there. Which is why DARPA and other agencies are increasingly interested in it. The TWTA (Figure 3) is one of the critical technologies that will allow systems at these frequencies to be built.
Strangely perhaps, the normally hostile propagation characteristics of millimeter wavelengths can be beneficial. For example, in the Earth’s atmosphere, RF energy at these frequencies is attenuated by almost anything and propagates only over very short distances, which makes communication difficult at best. The only way to do it is by using very high gain, very directional antennas that develop large amounts of gain by concentrating the beam over a very small area, effectively increasing the output of the RF power amplifier by a factor of 10 or more. As the beam is narrow the signal is also very difficult to detect or jam. To deal with this problem, EW systems must be able to deliver even higher power, for which solid-state devices are wholly inadequate. Advantage: tubes.
In space, the hindrances of the atmosphere don’t exist, which makes remote sensing at EHF a very interesting challenge with huge rewards. It’s possible to capture images with exceptionally high resolution, communicate over huge distances, and send data at the highest rate a system can generate, as bandwidth is basically unlimited. But once again, generating RF power is a key factor, and semiconductor technologies, if they work in this region at all, produce miniscule amounts of power. TWTs, on the other hand, can produce “real power” at frequencies well into the hundreds of GHz, and climbing.
An excellent example of what is currently being achieved in this area is an MPM for high resolution airborne radar designed and built by L3 Electron Devices and IPG Photonics along with researchers at the Standard Linear Accelerator Center. It has an instantaneous bandwidth of more than 3 GHz from 231.5 to 235 GHz (G-band) and produces a peak RF output power of 32 W with drive of just 10 mW and efficiency of about 9%.
The MPM is based on a serpentine waveguide TWT that uses high-energy-density, temperature-compensated samarium cobalt magnets to focus the beam and a very small, 20 kV electronic power conditioner. Beam power through the slow-wave circuit is about 5 MW/cm2. The input and output ports are chemical vapor deposition diamond and a four-stage MSDC is employed for beam energy recovery. The MPM operates from a 270 V power source. The radar sensor has been flown on an airborne test bed with high resolution real time video imagery obtained under cloud obscured operating conditions.
The serpentine structure, a variant of the coupled-cavity architecture, is a meandering reduced-height TE10 rectangular waveguide that “kind of shines at higher frequencies but because it can handle high power levels and can also be useful at lower frequencies at even higher power levels,” says David Whaley, the now retired chief scientist at L3 Technologies.
“We use an all-copper structure because at these frequencies the device becomes very small. You have high heat loads, so copper construction allows you to keep things cool. In addition to the 230 GHz devices, we have others working at 84 and 94 GHz using this structure.” It also does not have the abrupt rectangular bends of a folded waveguide circuit that it otherwise resembles, which makes wideband coupling to the structure easier with less gain variation from internal reflections and fewer problems with stability.
The beam collector was designed for the 50% duty cycle needed for the radar application. A four-stage collector is used to recover up to 90% of beam energy. The serpentine-waveguide TWT design can be scaled both down and up in frequency. So, for example, power levels of 600 W at 50 GHz and 20 W at 300 GHz are achievable. At frequencies higher than 300 GHz, “tens of watts” can be realized with a modified version of the serpentine waveguide.
If any tube concept could be considered the Holy Grail, it’s the cold-cathode TWT. Although early work on the cold cathode was done in various places, the seminal work is generally attributed to Charles “Capp” Spindt and Kenneth Shoulders of SRI International, who published a paper in 1966. Their approach remains the most likely to succeed today.
For more than five decades, researchers have been trying to surmount the formidable challenges posed by the cold-cathode TWT and have managed to overcome some of them while others remain to be solved. When this day comes, it will be a momentous day for the TWT. First, it operates at ambient temperature, so a cathode heater isn’t required, and as the cathode isn’t heated, the traditional factor limiting tube life ceases to exist. According to Whaley, “there is no inherent degradation mechanism, so it could presumably last forever.”
Without a cathode to heat, warm-up time would be eliminated from the list of TWT disadvantages, as operation would be virtually instantaneous. Current density (the amount of current emitted per unit area) could be much higher as emission would no longer be limited by operating temperature, so focusing the electron beams would be much easier. The beam’s current could be modulated directly at the cathode as well.
The biggest challenge to the cold-cathode TWT has always been reliability, as the cathode consists of tens of thousands of micrometer-size molybdenum cones deposited on a circular silicon substrate with an area of about one square millimeter. The high fields within the structure and the thin-film gate electrode make it possible for an electrical short to occur between the gate and one of these cones. When that happens, the entire array of emitters burns up and the device fails catastrophically. In a traditional thermionic TWT, degradation is “graceful,” allowing its end of life to be predicted by the amount of barium remaining.
In 2015, Carter Armstrong, vice president of engineering at Stellant Systems (formerly L-3 Electron Devices), reported that their colleagues at SRI had developed a way to reduce the damage caused by such a short. They were able to interrupt the breakdown path between the base of the cones and the gate by adding a dielectric layer between them. L-3 tested the SRI cathodes in a TWT that generated up to 10 W at 18 GHz and its results showed that the cathode better resisted individual emitter failures.
So, when will the breakthrough occur that will pave the way for the cold-cathode TWT to move toward a commercial product? Armstrong said in a 2015 IEEE Spectrum article that he believed it will be sometime in the early 2020s. However, “I have some cold cathodes sitting here from SRI,” says Whaley, “whose configuration has been shown to improve device life, and we need to test them. We have a thousand hours of life over four prototypes.”
Searching for Talent
The biggest single threat to the TWT industry and its suppliers isn’t a competing technology, it’s the fact that few graduates of engineering schools, even those with advanced degrees, rarely have much or even any knowledge of vacuum tube technology. For them, as well as people in general, tubes are what people used before there were transistors. This situation is nothing new; it’s been an increasing problem for years. In the U.S. the problem is exacerbated by the fact that the major tube manufacturers are in California where housing, if available, is notoriously expensive.
To solve the first problem, manufacturers have become tube talent scouts, scouring colleges and universities for likely candidates, bringing them into the fold as interns to show them that tubes are here to stay and that there are lots of terrific electrical and mechanical design challenges remaining to be solved. One of the most important takeaways for potential candidates from these activities is (or should be) that expertise in VEDs makes a design engineer a rare breed. That is, with comparatively few new engineers coming into the industry, short of committing some heinous crime, they’ll never want for a job, and one that pays very well.
“We look at the technical colleges and there a lot of really good schools for engineers in California, so it’s easy for us to recruit there,” says Amanda Mogin, director of investor relations at CPI. “One of the great things about this company is that when we bring in an engineer, he or she gets hands-on work pretty quickly in many areas, which is typically not something you would not get at many other Silicon Valley companies.”
To address the second issue, the company recently moved its satellite tube division to South San Jose, where housing is available and reasonably affordable, and commuting isn’t as great a problem. The facility is also the company’s newest design center, where many of the design challenges are addressed. The older people mentor the younger ones. If a new person wants to advance into management the tube industry is a great place because you have a lot of people who are ready to retire and will teach you everything they know and promote you quickly.
Hopefully, this discussion has demonstrated that not only are TWTs still viable, they have a significant roadmap for the future. For example, many of the advances being developed today haven’t reached fruition and some are still years away. As adversaries move to higher and higher frequencies, EW systems powered by TWTs will be the only solution able to deliver the required RF output power to counter them. And finally, the EHF region is still to be explored for radar and remote-sensing in space, offering TWTs immense potential.