Major Advances in Phased Arrays: Part II
The entry of MMIC arrays into space-age telecommunications, and ongoing research and development programs looking to use phased arrays effectively
Major Advances in Phased Arrays: Part II
This article is the second in a two-part series describing the major accomplishments in phased arrays and new-age MMIC ground-based and airborne arrays. Part one covered significant advances made in phased arrays over the last 35 years and the entry into the new era of active MMIC phase-phase steered arrays for ground and airborne systems. Part two covers the entry of MMIC arrays into space-age telecommunications — IRIDIUM,“ Globalstar and, potentially, the Teledesic system. Also discussed are research and development in the areas of the revolutionary new electronically steerable laser and optical beams using row/column steering, the microwave row/column ferroelectric scanned array, the electronically steerable plasma mirror antenna, an aggressive wafer-integration MMIC 94 GHz missile seeker phase-phase scanned array, the row/column scanned RADANT array antenna, the C- to Ku-band multi-user Advanced Shared Aperture Program (ASAP) MMIC phase-phase steered array and digital beamforming (DBF), including space-time adaptive processing (STAP).
The revolutionary commercial global satellite personal communications systems that use low orbiting satellite constellations also employ active phased arrays populated with MMIC modules, in this case in space. These commercial communication satellite systems, listed in Table 1 , are now in production or just entering production and involve large numbers of phased arrays and MMIC modules.
Table 1 -- Satellite Personal Communications MMIC Phased-Array Systems in Production
# of Satellites per Constellation
# of Modules per Antenna
# of Modules per Constellation
The IRIDIUM and Globalstar Systems
The IRIDIUM constellation consists of 66 satellites in six circular 700 km altitude near-polar orbits (11 satellites per orbit), and provides coverage for global communications. Each satellite has three antennas pointed toward the earth for horizon-to-horizon coverage.1 The antennas are active L-band (1.6 GHz) phased arrays using approximately 100 MMIC transmit/receive (T/R) modules and patch radiating elements. As a result, there are approximately 20,000 MMIC T/R modules and radiating elements per constellation. A subscriber holds a telephone, similar to those used for conventional cellular telephone systems, which communicates directly with one of the satellites. In turn, the signal is cross linked to other satellites for final passage down to the earth to another subscriber with a hand-held IRIDIUM telephone or to a gateway, which directs the call to a conventional telephone user via land lines. Alternatively, the signal could be transmitted directly from the satellite to a user in view of the same satellite as the first user. Figure 1 shows the three L-band active phased arrays mounted on the IRIDIUM spacecraft bus for testing in an anechoic chamber.
There has been much talk about the transfer of military technology to the commercial market, especially with the end of the Cold War. The IRIDIUM program is a good example of this technology transfer. The technology for the IRIDIUM space-based L-band antennas is derived from the Air Force and Navy’s space-based radar (SBR) program. The SBR, shown in Figure 2 , was developed originally for surveillance of continental US and US fleets to warn of possible attack by USSR bombers.2 The lightweight (1 oz) L-band module, shown in Figure 3 , was developed by the Air Force for the SBR program and forms the basis of the IRIDIUM L-band module. The SBR was never deployed by the Air Force and Navy because the Cold War ended and also because of the expense of building the system. The IRIDIUM system is taking the place of the SBR for a low cost of $3.5 B, which includes the price of satellite launch services, ground stations and software. The competing Globalstar system provides global coverage up to latitudes of approximately 65° north and south and will use a total of approximately 7300 T/R modules in its constellation.
The Teledesic System
Following on the heels of the IRIDIUM and Globalstar voice telephone communication systems are high data rate digital communication systems (1 Mbps) for computer-to-computer and video conferencing communications. One such system is the Teledesic low orbiting satellite system, which plans to use approximately 840 satellites. One possible implementation of this system will require 60 20 GHz transmit arrays and 60 30 GHz receive arrays per satellite, each requiring 400 MMIC T/R modules and radiating elements for a total of 40 million MMIC T/R modules and radiating elements per constellation — a phased-array engineer’s and MMIC engineer’s dream. The seed money for the Teledesic system is being provided by William Gates of Microsoft and Craig McCaw, formerly of McCaw Cellular Communications.3 The only other system that requires a larger number of MMIC modules is one proposed by an engineer at the Jet Propulsion Lab for voice communications. For this system, each user would wear a phased-array antenna on top of his or her head, as shown in Figure 4 . If each of these antennas required 100 radiating elements and MMIC T/R modules and one billion people in the world each wore one, the total number of MMIC modules and radiating elements that would be manufactured is one hundred billion — wishful thinking.
Arrays under Development
Electronically Steered Phased Arrays for Lasers and Optical Beams
Engineers and scientists have been talking about achieving electronic scanning of lasers since the 1960s. Some thought this was a pipe dream, but these doubters have since been proven wrong. An electronically steered phased array for laser and optical beams has been demonstrated.4 This array, which is carried around in a briefcase, represents a major breakthrough in the scanning of laser and optical beams.
The technique uses two liquid-crystal sheets. One sheet consists of N columns of liquid-crystal phase shifters spaced l/2 apart and used to scan the beam in azimuth while the other consists of N rows of liquid-crystal phase shifters also spaced l/2 apart and used to steer the beam in elevation. The phase gradient is produced by stepping the voltages across the liquid-crystal columns and phase shifter rows. By using row/column steering instead of element-by-element steering, the number of phase shifters and controls is reduced from N2 to 2N. Further reduction in the number of phase shifters and control lines is achieved by using coarse/fine steering. For example, in 1991, a 4.3 ¥ 4.1 cm array was built by Raytheon, which required only 43,000 phase shifters. Raytheon has developed techniques for scanning large angles electronically. In production, the cost per phase shifter for an optical phased array will be pennies.4
The phase shifters are formed by photoetching strips of transparent conducting material on one side of the liquid crystal. A transparent conducting ground plane is placed on the other side of the crystal. This assembly of transparent conducting strip, liquid crystal and transparent ground plane forms a phase shifter. Applying a voltage across that transparent conducting strip and the conducting ground plane changes the dielectric constant of the liquid-crystal strip and, in turn, changes the velocity of propagation of the laser or optical signal passing through it. The result is that the signal passing through the strip experiences a phase shift.
A Row/Column Ferroelectronic Scanned Array
A technique performed at the Naval Research Laboratory (NRL)5 has been described for achieving a low cost phase-phase steered array at microwave frequencies by using row/column phase shifters just as was performed at optical frequencies previously.3 Instead of rows and columns of liquid-crystal phase shifters, rows and columns of ferroelectric phase shifters are used, as shown in Figure 5 . The dielectric constants of the rows and columns of the ferroelectric phase shifters are determined by the voltages placed across these rows and columns. Consequently, by applying an appropriate stepping of this voltage, a phase gradient is generated to steer the beam in elevation and azimuth as was done for the optical phased-array scanner. The ferroelectric lens, consisting of columns of phase shifters, steers the beam in azimuth. A second such lens rotated 90° steers the beam in elevation.
For a ferroelectric lens, it is necessary that the electric field is polarized linearly and perpendicular to the conducting plates. Hence, a 90° polarization rotator is needed in between the lens that scans the beam in the azimuth direction, requiring a horizontally polarized signal. The lens that scans the beam in the elevation direction requires a linearly vertically polarized signal. Considerable work is necessary before a practical ferroelectric phased array is produced; this work is being performed at present.
An Electronically Steerable Plasma Mirror
NRL is also pursuing the development of an electronically steerable plasma mirror to provide electronic beamsteering, as shown in Figure 6 . Here, a plasma sheet is rotated to steer the beam in azimuth and is tilted electronically to steer the beam in elevation. The plasma mirror is rotated by switching to different initiation points in the cathode. The plasma mirror is tilted by tilting the magnetic field around the plasma. This tilting is performed using coils placed around the plasma. These coils are placed so as not to block the microwave signal. A 50 ¥ 60 cm plasma mirror has been generated with measured antenna side lobes of approximately –20 dB.6
A 94 GHz Reflect Array Using Four-inch Wafers
An aggressive effort has been described wherein MMIC technology was taken to the point of wafer integration using 4" wafers.7 Specifically, Thomson-CSF is developing a missile seeker antenna, which uses two 4" wafers.7 One wafer has the dipole elements and one-bit PIN diode phase shifters printed on it. The second wafer contains the driving circuits, which are linked to the first through bumps. The antenna has 3000 elements. The beam width is 2° and can be steered ±45°. Reportedly, low side lobes have been obtained.7
The RADANT Antenna
Thomson-CSF has developed a RADANT antenna for the Dassault Aviation RAFALE multiroll combat aircraft.7 The RADANT technique is similar to the ferroelectric scanning technique described previously. Diodes are inserted in place of the ferroelectric material between the vertically oriented conducting plates. Many diodes are placed one behind the other in the direction of propagation. The propagation velocity of the signal through the plates is varied by controlling the number of diodes that are on in the direction of propagation. A phase gradient in the vertical direction is generated by turning on different numbers of diodes at different parts of the lens. Consequently, the beam is scanned in elevation. A gradient in the azimuth direction is obtained using a RADANT lens employing horizontally oriented plates with diodes between them. Hence, the beam is scanned in azimuth. No polarization rotation plate is needed between the two RADANT lenses as is the case between the ferroelectric lenses. Many diodes are needed in this type of antenna to achieve the phase shift but these diodes are reportedly inexpensive.
A C- to Ku-band Multi-user ASAP MMIC Array
The Naval Air Weapons Center (NAWC) and Texas Instruments are developing a phased array for a strike/fighter aircraft that could be shared by many users.8 Specifically, this broadband array would have continuous coverage from C- through Ku-band that would share the functions of radar, passive electronic support measures (ESM), active electronic countermeasures (ECM) and communications. A flared notch radiating element is used to achieve this wide bandwidth. Cross notches allow horizontal, vertical or circular polarization to be obtained. A solid-state T/R module has been built that provides coverage from C- to Ku-band continuously. The module has a power output of 2 to 4 W per element, a noise figure between 6.5 and 9 dB, and a power efficiency between 5.5 and 10 percent over the full bandwidth. A 10 ¥ 10 element scanned array incorporating eight active T/R modules was built for test purposes. A full-up array would measure approximately 29" wide ¥ 13" high. With this type of array it would be possible ultimately to simultaneously use individual parts of the array as a radar for ESM, ECM and communications. The parts used for each function would change dynamically and could be nonoverlapping or overlapping, depending on the need.
The first radars to use DBF are the over-the-horizon (OTH) radars; specifically, the GE OTH-B and the Raytheon relocatable OTH radar (ROTHR). The ROTHR receive antenna is approximately 8500' long. More recently, the Dutch company Signaal has employed DBF for its three-dimensional stacked-beam SMART-L and SMART-S shipboard radars. The DBF is performed only on receive. In the SMART-L system, the antenna consists of 24 rows. The signals from each row are downconverted and pulse compressed with surface acoustic wave lines and then converted from analog to digital with a 12-bit 20 MHz Analog Devices analog-to-digital converter. The signal is then modulated onto an optical signal and passed through a fiber-optic rotary joint to a digital beamformer where 14 beams are formed.
Rome Laboratory (Hanscom Air Force Base, MA) has built an experimental DBF system that consists of a 32-column linear array at C-band, which can form 32 independent beams and uses a novel self-calibration system.9 Also, Rome Lab has developed a fast digital beamformer that uses a systolic processor architecture10 based on the quadratic residue number system.9 MICOM (US Army) built a 64-element feed that uses DBF for a space-fed lens.11
Roke Manor Research Ltd. has built an experimental 13-element array using DBF on transmit as well as on receive.12 This experimental system uses the Plessey SP2002 chip running at a 400 MHz rate as a digital waveform generator at every element. Performing DBF on transmit allows nulls to be placed in the direction of an antiradiation missile threat or where there is high clutter.
The experimental British MESAR S-band system performs DBF at the subarray level.13 This system incorporates 16 subarrays and a total of 918 waveguide radiating elements and T/R solid-state modules.
The National Defense Research Establishment of Sweden has built an experimental S-band antenna operating between 2.8 and 3.3 GHz, which performs DBF using a 25.8 MHz sampling rate on a 19.35 MHz IF signal.14 The advantage of using IF sampling rather than baseband sampling is that the imbalance between the two channels is of no concern, that is, the I and Q channels, or the DC offset. It was demonstrated that DBF could compensate for amplitude and phase variations that occur from element to element across angle and across the frequency band. Element-to-element gain variation over angle due to mutual coupling was reduced from ±1 dB to approximately ±0.1 dB. Using equalization, a ±0.5 dB variation in the gain over the 5 MHz bandwidth was reduced to less than ±0.05 dB variation. With this calibration and equalization, side lobes of 47 dB (peak) over a 5 MHz bandwidth were demonstrated. A 50 dB Chebyshev weighting was used. It was demonstrated that the calibration was maintained fairly well over a period of two weeks. This work demonstrates the potential advantages offered by DBF.
The power of adaptive DBF was demonstrated in recent flight tests for an L-band experimental airborne radar system that used STAP.15 This system has a phased-array antenna along the side of the aircraft with 11 columns with two degrees of freedom per column for a total of 22 degrees of freedom. This system provided 52 to 69 dB of STAP clutter cancellation. Typically, the clutter was at levels of –30 and –45 dBi before and after STAP, respectively.
Based on these reported accomplishments, ongoing developments, research and large numbers of programs that are looking to use phased arrays effectively, it is apparent that the future for phased arrays is promising and should lead to exciting further developments. Phased arrays have come a long way and can be expected to make major strides in the future. For further reading on recent developments in phased arrays around the world the reader is referred to the proceedings of the 1996 IEEE International Symposium on Phased-array Systems held October 15–18, 1996 in Boston. More than 500 individuals attended and 92 papers were presented by authors from 16 countries.
1. J.J. Schuss, J. Upton, B. Myers, T. Sekina, A. Rohwer, P. Makridakas, R. Francois, L. Wardle, W. Kreutel and R. Smith, “The IRIDIUM Main Mission Antenna Concept,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, pp. 411–415.
2. E. Brookner and T.S. Mahoney, “Derivation of a Satellite Radar Architecture for Air Surveillance,” Microwave Journal , Vol. 29, No. 2, February 1986, pp. 173–191; see also M.I. Skolnick (ed.), Radar Applications , IEEE Press, New York, NY, 1988.
3. A. Kupfer, “Craig McCaw Seas and Internet in the Sky,” Fortune , May 27, 1996, pp. 61–72.
4. T.A. Dorschner, L. Friedman, M. Holz, D.P. Resler, R.C. Sharp and I.W. Smith, “An Optical Phased Array for Lasers,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, pp. 5–10.
5. J.B.L. Rao, G.V. Trunk and D.P. Patel, “Two Low Cost Phased Arrays,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, pp. 119–124.
6. J.R. Mathew, R.A. Meger, J.A. Gregor, D.P. Murphy, R.E. Pechacek, R.F. Fernsler and W.M. Manheimer, “Electronically Steerable Plasma Mirror,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, p. 58–62.
7. Jean-Marie Colin, “Phased-array Radars in France: Present and Future,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, pp. 458–462.
8. C. Hemmi, R.T. Dover A. Vespa and M. Fenton, “Advanced Shared Aperture Program (ASAP) Array Design,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, pp. 278–282.
9. H. Steyskal, “Digital Beamforming at Rome Laboratory,” The Rome Laboratory Technical Journal , Vol. 1, No. 1, June 1995, pp. 7–22.
10. E. Brookner, Aspects of Modern Radar , Chapter 2, Artech House, 1988, Norwood, MA.
11. J.F. Rose, B.A. Worley and M.M. Lee, “Antenna Patterns for Prototype Two-dimensional Digital Beamforming Array,” Proceedings of the 1993 IEEE Antennas and Propagation International Symposium , June 28 - July 2, 1993, University of Michigan, Ann Arbor, MI, pp. 1544–1547.
12. A. Garrod, “Digital Modules for Phased-array Radar,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, pp. 81–86.
13. E.R. Billam and D.H. Harvey, “Mesar — An Advanced Experimental Phased-array Radar,” Proceedings of the IEEE International Conference on Radar , October 19–21, 1987, pp. 37–40.
14. L. Pettersson, M. Danestig and U. Sjostrom, “An Experimental S-band Digital Beamforming Antenna,” Proceedings of the 1996 IEEE International Symposium on Phased-array Systems and Technology , October 15–18, 1996, Boston, MA, pp. 93–98.
15. D.K. Fenner and W.F. Hoover, Jr., “Test Results of a Space-time Adaptive Processing System for Airborne Early Warning Radar,” Proceedings of the IEEE 1996 National Radar Conference , May 13–16, 1996, Ann Arbor, MI.
Eli Brookner received his BSEE degree from the City College of New York in 1953, and his MS and DSc degrees in electrical engineering from Columbia University in 1955 and 1962, respectively. Since 1962, he has been at Raytheon Co. where he is presently a consulting scientist in Sudbury, MA. Brookner conceived and helped design the Wake Measurement Radar, the first TWT radar put into space. He has been technical director for a number of space-based radar programs. Recently, he completed work on future improvements to the worldwide satellite-cellular IRIDIUM communications system and future commercial high data rate communications systems. Brookner is a member of Eta Kappa Nu and Tau Beta Pi, a fellow of the American Institute of Aeronautics and Astronautics, an IEEE fellow and a member of the International Union of Radar Science (URSI; commissions B and C).