Novel Design and Manufacturing Techniques Revitalize mmWave TWTs

The mmWave spectrum offers many compelling advantages for communication applications. Compared to microwave systems, mmWave offers larger blocks of less contested and less regulated bandwidth for high data rates. The shorter wavelength allows reduced antenna size for a given antenna gain for compact systems. Compared to optical, mmWave losses in the atmosphere are modest and it is possible to “burn through” inclement weather to maintain a link. Additionally, with enough power the spot size at the receiver can be relatively large, providing tolerance for imperfect antenna pointing accuracy.

Wireless communications systems are finding increasing applications because of their reduced capital costs, ease of deployment and reduced environmental impact over physical carriers such as fiber. Wireless transmitters have long played an essential role in satellite communications (satcom) and are used terrestrially as point-to-point relays to carry backbone traffic where the deployment of physical lines is difficult. Recently mmWave systems have seen rapid adoption for point-to-point terrestrial links up to W-Band and in satcom for gateway uplinks up to V-Band. Access to bandwidth at these frequencies enables competitive data rates with those available over fiber optic cables. Figure 1 shows the large swaths of frequency blocks available with the 71 to 76 GHz and 81 to 86 GHz bands each offering 5 GHz of continuous bandwidth. At W-Band, 92 to 114 GHz and at D-Band, 130 to 174.5 GHz, even larger bandwidths are being considered for near-term network growth.¹ Future systems will operate at G-Band frequencies spanning 200 to 300 GHz ranges.²

Figure 1 mmWave spectrum bands.

Figure 2 Summary of available SSPAs.³ Elve power amplifiers have been added.

Figure 3 Range enhancement with an E-Band TWT.

The power amplifier (PA) is usually one of the last components in the RF chain before the antenna, playing a key role in system performance. Making use of mmWave for communications and imaging requires acquiring the signal at the receiver with a suitable signal-to-noise ratio. Practical systems that function over appreciable distances in various weather conditions often need tens of watts of power to meet requirements.

Achieving this at mmWave frequencies is a challenge. The advantages of mmWave systems have been acknowledged for decades, but the lack of availability of mmWave PAs has impacted mmWave deployment. Local heat dissipation limits achievable power in a single MMIC, so reaching watts of output power requires power combining that reduces efficiency. Low efficiency, low system-level power density, high thermal load and design complexity of mmWave solid-state power-combined systems are some of the challenges of deploying mmWave systems. While several technologies have been explored to deploy mmWave PAs,³ GaN and GaAs are the solid-state solutions that offer the most potential for high power levels as shown in Figure 2. New technologies offering increased power efficiently in compact forms are needed.

Traveling wave tube (TWT) amplifiers (TWTAs), consisting of a TWT and its power supply, or electronic power conditioner are a well-established highly reliable technology^4,5 that has demonstrated high power efficiency in a compact form at mmWave frequencies. We believe that the technology outperforms solid-state power amplifiers (SSPAs), but is often overlooked for deployment in high data rate communications systems. As shown in Figure 3, a 100 W TWTA allows data to be transmitted at the same rate 10x as far as a 100 mW amplifier.

LINEAR BEAM AMPLIFIER OVERVIEW

Many vacuum devices are used to generate or amplify mmWave power. Linear beam devices, such as klystrons, TWTs and backward wave oscillators (BWOs) provide power that is unachievable in solid-state devices. Klystrons are narrowband amplifiers with resonant circuits, producing high peak power. They are used in radar and some communication systems. TWTs employ a non-resonant circuit that allows significantly wider bandwidth, typically at lower power levels than klystrons. BWOs have a circuit that is designed for an unstable interaction with a backward traveling wave so the devices generate RF without an input signal, effectively amplifying noise.

THE TWT

TWTs are the vacuum amplifiers used most commonly in communication systems. Dr. Rudolf Kompfner is credited with the invention of the TWT,⁶ but Dr. John R. Pierce quickly realized the potential of the device to enable the type of communications he was working on at Bell Laboratories. He developed much of the engineering needed to design and build practical devices.⁷

TWTs amplify a signal using the kinetic energy carried by electrons traveling in a vacuum environment. The operation starts with an electron gun that creates an electron beam that is electrostatically focused into a narrow stream. Most TWTs employ a thermionic cathode, where a low work function material is heated to emit electrons into the vacuum. The hot cathode evaporates the emissive material, leading to a finite lifetime of electron emission. Electron energies are given by a Maxwellian distribution in the cathode and only those with energies above the cathode work function can travel into the vacuum. Higher electron emission requirements mean a hotter cathode for a given cathode material work function. A voltage applied to the anode accelerates the electrons and lenses electrostatically focus them into a compact beam. If the cathode surface is large and the focused beam small as in a high frequency TWT, this focusing may reduce the beam’s cross-sectional area by a factor of a hundred, requiring extreme precision in the lenses.

Next, the electron beam, carrying kinetic energy established by the electrostatic acceleration, enters a magnetic field that counteracts the electrostatic repulsion of the electrons, maintaining a constant cross-section as the electrons travel through the interaction circuit. The circuit starts with an RF input port where power is injected. The RF is carried on a transmission line that wraps around the beam so that the electric field from the RF input power is aligned with the electron beam’s axial motion. The alternating electric field speeds up some electrons and slows down others, forming electron bunches. As the modulated electron beam travels with the RF wave, the electron beam induces current on the circuit, causing the amplitude of the circuit wave to grow at the expense of the electron kinetic energy.

The electron beam and the electromagnetic wave must travel at similar speeds to form the electron bunches. Otherwise, the electron sees a sinusoidally-varying electric field with velocity increasing and decreasing, but on average retaining its initial energy. Electrons moving at a similar speed to the wave are continuously accelerated in the accelerating phase and continuously decelerated in the decelerating phase.

The amplified RF is coupled out at the end of the circuit and the spent electron beam goes to a collector. In most TWTs, the collector contains multiple electrodes, each depressed below ground to a different electric potential level. The electrons give up kinetic energy as they climb the potential hill created by these biased electrodes, allowing the power supply to recover energy, significantly improving the overall operating efficiency of the device. This energy recovery is one of the reasons that TWTs can achieve significantly higher efficiency than solid-state amplifiers.

The technique of bunching the beam by adjusting electron velocity with an electromagnetic signal is commonly used in vacuum amplifiers. First demonstrated in klystrons in the 1930s, velocity modulation can produce high gain in the interaction circuit since small changes in velocity result in patterns of high and low current density downstream. Since all this happens in a collisionless vacuum environment, the approach allows these devices to scale to very high frequencies.

The most used TWT circuit is a helix, or more specifically a coaxial transmission line with the center conductor twisted into a helical path. The quasi-TEM mode on the transmission line follows the helical path, causing the axial velocity to slow. The electron beam travels through the center of the helical line where the electric field of the electromagnetic wave acts on the beam along the direction of the beam propagation.

On a coaxial line with a vacuum dielectric, the electromagnetic wave propagates at the speed of light. In a TWT, the helix pitch, the distance between each turn, reduces the electromagnetic wave’s net velocity in the beam direction. Figure 4 shows this with a “dispersion curve,” the relationship between frequency ω=2πf and wavelength. Frequency is plotted on the y-axis and inverse wavelength, called β on the x-axis. Since the velocity of a wave is given by frequency multiplied by wavelength, the velocity at any point on the curve is ω/β .

Figure 4 Dispersion curves for coaxial line and helical delay line.

Figure 5 Waveguide and folded-waveguide dispersion diagram.

Figure 6 Folded-waveguide periodic dispersion curve.

TWT circuits based on two-conductor transmission lines can be extremely wideband since they use the TEM mode, which has no cutoff frequency. Unfortunately, the center conductor must be electrically isolated from the outer conductor. This requires ceramics to support the helix, resulting in non-ideal heat paths for electrons that intercept the circuit and ohmic losses generated in the helix. For applications requiring smaller bandwidths, single-conductor transmission line circuits are preferable. These circuits can be all metal, improving thermal power handling. Many traditional TWTs are made from coupled cavities that use a series of resonant cavities connected with irises or slots to create a winding RF path.

A folded-waveguide circuit employs a waveguide bent back on itself many times, reducing the effective speed of the RF along the beam propagation.⁸ A beam tunnel hole is punched through the circuit. Starting with the dispersion curve of the waveguide, the net velocity of the RF following the waveguide path is reduced as shown in Figure 5.

The periodic structure results in a periodic dispersion curve shown in Figure 6. The direction of the electric fields reverses each time the waveguide folds back on itself. As the electron beam passes through the folded-waveguide structure, it sees an additional 180-degree phase shift every half period as shown in Figure 7.

Figure 7 Folded waveguide field as seen by electron beam.

Figure 8 Dispersion curve for folded-waveguide TWT interacting with an electron beam.

Figure 8 shows the resulting dispersion curve. Appropriate values of waveguide cross-section and path can be chosen to achieve a phase velocity that matches the beam velocity. The circuit can be optimized for a relatively constant phase velocity over the band, resulting in flat gain over frequency for a wideband amplifier.

MMWAVE TWTS

Many TWTs using helix circuit designs to generate hundreds of watts of power at Ka-Band for communications systems are available today. Suppliers include Stellant, CPI, Thales, Photonis, Teledyne and NEC. These TWTs can have efficiencies over 50 percent with output power densities around 100 mW/cm³.

For applications where size or weight are at a premium, mini-TWTs are often used. These devices have shorter circuits and reduced gain that is offset by higher-power solid-state drivers. Lower voltages allow for a very compact high-voltage power supply to be packaged with the TWT. At Ka-Band, up to 100 W is available with power densities of hundreds of mW/cm³. At E-Band frequencies there are fewer commercially available options, as shown in Figure 9.

Figure 9 Comparison of E-Band amplifiers. 

ELVE TWTAS

The construction of vacuum electronic devices, such as TWTs, is often an artisan process; it requires extremely high-precision machining and assembly. The tolerances become more exacting as the frequency increases. Each mmWave circuit is constructed and assembled individually and can take months to complete. Fabrication techniques for the circuit include micromachining (milling or EDM) as well as electroplating around LIGA molded photoresist, etched silicon or 3D-printed polymer structures.^9,10,11,12 These processes do not easily accommodate design changes to individual circuits with minimal process adjustments. The processes used to date have significant limitations in the rate of production.

Figure 10 Elve E-Band TWT.

Elve has developed TWT design and fabrication techniques suitable for making mmWave TWTs in volume. The TWTs employ nanocomposite scandate tungsten emitters, which have a significantly lower work function than traditional TWT emitter materials. These special materials allow the emitted electron current density to be higher for the same temperature. As a result, a smaller emitter can be employed enabling the devices to be robust to minor dimensional errors in the beam-focusing structures while maintaining a long lifetime.

Elve TWTs use a “sheet” beam with an elliptical, rather than round, cross-section of the electron beam perpendicular to the direction of travel. The elliptical geometry reduces space charge density and power density in the beam, reducing the magnetic field requirements to confine the beam. Maintaining one of the ellipse dimensions small relative to wavelength enables good circuit efficiency, the ratio at which electron beam kinetic energy is converted into RF energy. The planar sheet beam configuration is well-suited for modern manufacturing techniques.

Figure 11 EPC powering the TWT heater, cathode and collectors.

Figure 12 Elve E-Band PA including TWT and EPC.

Elve has developed an additive manufacturing technique to fabricate the circuits. Using this approach, circuits of different frequencies can easily be fabricated using the same process. Other devices that interact with electron beams, like klystrons or gyrotrons, can be made with this approach. The circuit technology is critical to Elve’s ability to rapidly iterate TWT designs. In production, it allows circuits and TWTs to be made quickly and consistently at volume. The compact planar design of an Elve TWT is shown in Figure 10.

Traditional microwave TWTs have demonstrated decades of reliable operation in space applications. Elve is designing and testing amplifiers to meet the same rigorous standards. The cathodes are the most sensitive portion of the TWT, so samples from each batch of powder are tested to verify the work function and emitted current. Elve is putting complete units through environmental testing including cathode heater cycling, operational on/off cycling, vibration testing and operation at temperature extremes to identify and resolve any potential reliability issues.

A complete TWT-based amplifier contains an electronic power conditioner (EPC) shown in Figure 11, which produces the operating voltages for the TWT. A compact TWT requires a negative cathode voltage of several kilovolts, typically in the range of -3 to -20 kV. The cathode voltage must be tightly regulated with extremely low ripple to enable ideal RF performance from the TWT. The cathode heater, floating at cathode potential, requires a few watts of power. The multi-stage depressed collector is biased with voltages between cathode potential and ground to enable efficient recovery of spent electron beam energy. In addition to generating the TWT electrode voltages, an EPC also provides the control logic and user interface to allow system integration.

The Elve Vermillion E-Band amplifier shown in Figure 12 covers 81 to 86 GHz. The amplifier has a small signal gain of 20 dB, with other parameters shown in Table 1. Transfer curves are shown in Figure 13 with simulated linearity performance shown in Figure 14.

Figure 13 Simulated RF transfer characteristics.

Figure 14 Simulated linearity characteristics.

The behavior under multi-tone input waveforms is shown in Figure 15. When driven by 30 dBm input power per carrier to produce more than 100 W of total output power, split between two carrier frequencies, third-order intermodulation distortion is around 16 dB.

Figure 15 Two-tone intermodulation distortion at band center.

There have been impressive mmWave and THz TWT results reported by Stellant^13,14 and Northrop,¹⁵ demonstrating a technological path to mmWave vacuum power devices. The volume has been low, but Elve has spent the past 18 months developing a high volume TWT fabrication process. Elve’s process is evolving with early prototype amplifiers providing feedback to improve subsequent units. Prototype gain performance is shown in Figure 16 with temperature performance shown in Figure 17.

Figure 16 Simulated and measured data for an early Elve prototype TWT.

Figure 17 TWT temperature performance.

CONCLUSION

TWTs have been a workhorse in communications, radar and imaging applications. They have an opportunity to return to the spotlight and showcase unprecedented performance at mmWave frequencies. Decades of demonstrated reliability, high power and efficiency are some of the advantages TWTs offer. Elve’s focus on large-quantity manufacturability ensures access to TWT advantages to enable the next generation of connectivity.

References:

1. Ericsson Microwave Outlook Report, October 2022, Web: https://www.ericsson.com/en/reports-and-papers/microwave-outlook.
2. “DARPA Asks Industry to Develop G-Band RF and Microwave Enabling Technologies for Communications and Sensing,” Web: https://www.militaryaerospace.com/rf-analog/article/14211440/rf-and-microwave-gband-communications-and-sensing.
3. H. Wang, T. Y. Huang, N. Sasikanth Mannem, J. Lee, E. Garay, D. Munzer, E. Liu, Y. Liu, B. Lin, M. Eleraky, H. Jalili, J. Park, S. Li, F. Wang, A. S. Ahmed, C. Snyder, S. Lee, H. Thong Nguyen and M. E. Duffy Smith, “Power Amplifiers Performance Survey 2000-Present,” Web: https://gems.ece.gatech.edu/PA_survey.html.
4. NASA Spinoff, Traveling-Wave Tubes Travel Far, Web: https://spinoff.nasa.gov/Traveling-Wave-Tubes-Travel-Far.
5. R. Ludwig and J. Taylor, “Descanso Design and Performance Summary Series Article 4: Voyager Telecommunications,” Washington: NASA (2002): 1–6.
6. R. Kompfner, “The Invention of the Traveling-wave Tube,” San Francisco Press, 1964.
7. J. R. Pierce, “Traveling-wave tubes,” The Bell System Technical Journal 29, No. 2, 1950, pp. 189–250.
8. R. GE, Hutter and S. W. Harrison, “Beam and Wave Electronics in Microwave Tubes,” van Nostrand, 1960.
9. R. L. Jaynes, A. M. Cook, C. D. Joye, J. C. Rodgers, A. N. Vlasov, I. A. Chernyavskiy, J.P. Calame et al, “Microfabrication and Micromachining for Millimeter-Wave Traveling Wave Tubes,” 2020 IEEE International Conference on Plasma Science (ICOPS), pp. 316–316, 2020.
10. A. Baig, D. Gamzina, M. Johnson, C. W. Domier, A. Spear, L. R. Barnett, N. C. Luhmann and Y.-M. Shin, “Experimental Characterization of LIGA Fabricated 0.22 THz TWT Circuits,” 2011 IEEE International Vacuum Electronics Conference (IVEC), 2011, pp. 275–276.
11. J. C. Tucek, M. A. Basten, D. A. Gallagher, K. E. Kreischer, R. Lai, V. Radisic, K. Leong and R. Mihailovich, “A 100 mW, 0.670 THz Power Module,” IVEC, 2012, pp. 31–32.
12. A. M. Cook, C. D. Joye and J. P. Calame. “W-band and D-band Traveling-wave Tube Circuits Fabricated by 3D Printing,” IEEE Access 7, 2019, pp. 72561–72566.
13. R. Kowalczyk, A. Zubyk, C. Meadows, T. Schoemehl, R. True, M. Martin, M. Kirshner and C. Armstrong, “High Efficiency E-band MPM for Communications Applications,” IVEC, 2016, pp. 1–2.
14. N. Robbins, D. Eze, H. Cohen, X. Zhai, W. McGeary, W. Menninger, M. Chen and E. Rodgers, “Space Qualified 200-Watt Q-band Linearized Traveling-wave Tube Amplifier,” IVEC, 2018, pp. 13–14.
15. Jack C. Tucek, M. A. Basten, D. A. Gallagher and K. E. Kreischer, “Operation of a Compact 1.03 THz Power Amplifier,” IVEC, 2016, pp. 1–2.
16. W. L. Menninger, T. K. Phelps and J. Lingenfelter, “4.2: Performance and Reliability of Recent Production Space Linearized Traveling-wave Tube Amplifiers,” IVEC, 2010, pp. 49–50.