Comparable U.S. Technology

The North American program that most resembles EMS is the joint AMSC/Telesat Mobile Incorporated (TMI) effort whose MSAT/M1 and MSAT/M2 spacecraft are scheduled for a 1994 launch. The AMSC MSAT will provide six L-band fixed transmit and receive beams, corresponding roughly to the U.S. time zones, Alaska and Hawaii, and Mexico and Puerto Rico. These will be generated by two array-fed elliptical unfurlable offset-fed mesh reflectors (one for transmitting and one for receiving) whose dimensions are approximately 6 m by 5 m. The feed array is 23 cup dipoles. The L-band beams are for communicating with mobiles; feeder links will operate at Ku-band. The Ku-band system will provide one beam for all of the coverage area except Hawaii; a separate spot beam will cover Hawaii. A single 0.76 m diameter shaped reflector with two feed horns will generate beams for both transmitting and receiving at Ku-band (Whalen and Churan 1992).

Vehicular Antennas

In many respects the vehicular antennas currently available for satellite mobile applications constitute the weak links of the system. If the vehicular antenna has a high gain, it has to track the satellite, following both vehicular and orbital motions. This is difficult and expensive. If the vehicular antenna has low gain, the capacity of the communications link is limited.

Representative specifications for perhaps a minimum practical L-band GEO system appeared in a recent Inmarsat RFP (Inmarsat 1992) for an antenna that will mount on the vehicle roof, trunk lid, or window with a simple clamp or magnetic mount. It will not track but should be omnidirectional in azimuth with fixed elevation coverage. The transmit band is 1626.5 to 1660.5 MHz; the receive band is 1530.0 to 1559.0 MHz. The minimum receive gain is 10 dBi. Polarization is RHCP receive and transmit with an axial ratio of less than 2 dB on axis. Power capability is 50 W.

Except for a few straightforward prototype Inmarsat-like antennas (Figure 2.18), none of the European companies visited reported activity in vehicular antennas. Considerable work is in progress in the United States, Canada, and Japan, where bifilar helices and phased arrays are both under consideration for L-band. The lack of indications of similar work in Europe implies that vehicular antennas for mobile satellite systems are receiving less emphasis there than elsewhere in the world.

The bifilar helix provides a beam which is omnidirectional in azimuth and which can be steered in elevation by mechanically changing the pitch angle (using stranded wire), changing the length while keeping the pitch fixed (the antenna telescopes) or by using a dual feedpoint and mechanically rotating the bottom helix with respect to the top one. Typical gains are 7 to 8 dBi (McCarrick 1991).

A mechanically steered phased array typically has a fixed-pointing beam that is scanned in azimuth by rotating the array in response to motions of the vehicle or satellite. The Jet Propulsion Laboratory (JPL) of Cal Tech has developed a mechanically steered L-band planar microstrip array of four 4-element yagis. The array has a 50 degrees elevation beamwidth and provides 15 dBi peak gain and a minimum gain of 10 dBi over the operational range of elevation angles (Huang 1991). A Japanese design using four crossed-dipole printed elements with a reflector provides a gain of about 14 dB essentially flat from 1.5 to 1.68 GHz (Kazama et al. 1992).

Figure 2.18. A Mobile Antenna for the PRODAT-2 System (Courtesy ESA)

An electronically steered phased array developed in Canada uses 18 circularly polarized microstrip patches, arranged around an inclined cylinder. It achieves 10 dBi gain over elevation angles from 10(superscript ) to 50(superscript ) for a bandwidth of 1530 to 1660.5 MHz (Petosa, Strickland, and Wight 1992).

With the ETS-V/EMSS (Experimental Mobile Satellite System) as motivation, Japanese investigators have studied vehicular antennas for satellite applications since the mid 1980s. Most of the initial work focused on flat panel phased arrays that are mounted about 45 degrees from the horizontal and mechanically steered in azimuth. Japanese researchers feel that elevation steering is not required at L-band but will be needed at Ka-band. Current NEC opinion is that a flat or conformal phased array cannot achieve the required gain at low elevation angles.

A sophisticated antenna developed by NHK for direct broadcast services (DBS) reception on tour busses (Takano 1992) offers a combination of mechanical and electronic tracking with a 64-element microstrip adaptive self-phasing array on a turntable. The array consists of six 8-element flat panel subarrays. The turntable can point the array in any azimuth direction; electronic tracking provides plus/minus 2.5 degrees beam steering in azimuth and plus/minus 6 degrees in elevation.

The Kashima Space Research Center (KSRC) of the CRL is looking at two vehicular antenna technologies: a mechanically steered array and a very low profile electronically scanned phased array. Both apparently use an open loop tracking system that combines a geomagnetic sensor and an optical fiber gyroscope. The geomagnetic sensor serves to reduce the cumulative error in the gyro. Experimental results were presented comparing power received by the mechanically steered array and power received by a reference omnidirectional antenna.

The low-profile array offers 14 dBi gain at 90 degrees elevation angle and 11 dBi at 45 degrees. Its half-power beamwidth is plus/minus 14 degrees and the size is 60 cm diameter by 3 cm thick. The array compensates for Doppler shift but the method used was not described.

Prototype antennas have been built in Japan for hand-held, aircraft, and maritime applications. The aircraft unit can track electronically. Our hosts were asked to estimate the cost of a commercial version, but they were unable to do so. The maritime antenna includes a diversity reception system (Iwai et al. 1992) that takes advantage of differential fading on the cross-polarized and co-polarized components of the incoming signal to compensate for multipath effects.

Japanese companies and laboratories have been active in the development of hand-held satellite terminals. NEC delivered six "briefcase terminal" prototypes for the Inmarsat system using mechanically steered 5-element arrays that achieved 13 dBi gain for elevation angles between 5 degrees and 90 degrees. A similar hand-held "Message Communicator" for the L-band ETS-V/EMSS used two patch antennas (Figure 2.19) and developed 6 dBi gain.

Antennas for ISLs

Herein phased array and reflector antenna technology developed explicitly for ISLs are discussed. More general aspects of phased arrays are treated in the sections which follow.

The first European microwave ISL system is the S-band Multiple Access Data Relay Payload being developed by Alenia Spazio for ARTEMIS. It will provide S-band and Ka- band crosslinks to one LEO satellite at a time. The antenna design has been scaled back from an earlier version capable of developing four simultaneous beams (two for transmitting and two for receiving) and working several LEO satellites simultaneously. The present single-satellite system uses a receive array of 64 microstrip patch elements. Alenia already has made the MMICs and BFN for it. The company is ready to go to the final design and production, pending a decision by ESA on whether the S-band crosslink package will actually be launched.

Figure 2.19. A Hand-Held L-Band "Briefcase Terminal" with Two-Element Array Antenna (Courtesy KSRC)

The proposed ARTEMIS data relay payload includes separate arrays for transmitting and receiving because of interference and PIM problems, and the general complexity of trying to build a single array for receiving and transmitting. The individual MMICs must have intermodulation distortion (IMD) levels 150 to 160 dB below the carrier if the PIM products are to be kept at or below the thermal noise level. The relatively close spacing of the transmit and receive frequencies also makes the design difficult.

The European Data Relay Satellite (DRS), to be launched in 1996, will carry S-band and Ka-band ISLs according to current plans. DRS's Ka-band (27.5 GHz) antennas must be able to scan across the earth (approximately 20 degrees) in 120 seconds. If this is done mechanically it can create unacceptable changes in spacecraft momentum.

Alenia is developing an antenna positioning module (APM) for DRS that involves a mount (one for each antenna) with two degrees of freedom and containing two rotary joints. It rotates the antenna about the feed, minimizing in some way the momentum transferred to the spacecraft. The details of how this is done were not revealed. Alenia rejected scanning the antennas by moving the feeds because that would degrade the antenna patterns unacceptably.

Alcatel is working on the same problem and has designed a Ka-band beam waveguide antenna with an ellipsoidal reflector and mechanical scanning. It will scan up to plus/minus 155 degrees and employs six reflectors in the beam waveguide system.

An earlier DRS design by Alcatel (Lenormand et al. 1988a, 1988b) featured an array feed and a deployable umbrella mesh reflector. The 140-element receiving array would apparently be "self-phasing," meaning that it would adjust itself to optimize the received signal. The antenna system would provide four receive channels and two transmit channels.

A later Alcatel proposal calls for using a corrective lens between the main reflector and the subreflector in an array-fed offset paraboloid Gregorian configuration. This would allow scanning over an angle of 40 beamwidths instead of the 15 beamwidths that could be achieved without the lens (Lenormand et al. 1992).

Comparable U.S. designs include a Cassegrain antenna with a movable subreflector that could scan over 20 beamwidths (Lapean and Stutzman 1992) and a Gregorian with a rotatable tertiary reflector and an array feed that compensates for the resulting phase errors (Werntz et al. 1992).

The Japanese counterpart to ARTEMIS is the ETS-VI S-band Intersatellite Link Experiment (SIC), designed to establish the Japanese technology of S-band multiple data relay channels. One "forward link" beam (transmission from ETS-VI to other spacecraft) and two independent electronically steerable "return link" beams (transmissions to ETS-VI) are generated by an active direct radiating phased array manufactured by MELCO. The antenna has 19 elements; all are used on receive and 16 are used on transmit. Each element is itself an array of seven microstrip circular patches and has its own low-noise amplifier (LNA) with a 1.5 dB noise figure (NF); those used on transmit also include 1.26 W solid state power amplifiers (SSPAs) and diplexers. The array transmits at 2.1 GHz and receives at 2.3 GHz. The beams are fully steerable over plus/minus 10 degrees, covering the surface of the earth plus orbital heights up to 1000 km. The minimum gain is 27.1 dBi over the full field of view, and the total EIRP is 35.8 dBW. The gain of the individual elements is 14.3 dBi.

The ETS-VI SIC is comparable to NASA's TDRSS, and the electrical performances of the TDRSS and ETS-VI ISL antennas seem to be similar (Okubo et al. 1989). The Japanese system is more advanced than ARTEMIS's since it can handle two satellites simultaneously and uses a single array for both transmitting and receiving. Also, ETS-VI is ready for launch now, while, when this report was written, the final configuration for ARTEMIS had not been selected.

For the ETS-VI K-Band Single Access Antenna, Toshiba has developed a pointing mechanism (Figure 2.20) that provides 0.0029 degrees resolution over a plus/minus 15 degrees scan range (both axes). A unit currently under development (Figure 2.21) offers excellent resolution with rapid scanning over wide angles. The degree to which these mechanisms compensate for the problem of momentum transfer to the spacecraft was not reported.

COMETS will carry a multiple use reflector antenna that serves both for conventional millimeter wave and terrestrial links. It is described below. COMETS will also carry an independent single access antenna for S- and Ka-band ISLs. The feeder link for the ISL shares the same dish with the millimeter wave system.

Other Space and Ground Antenna Technologies

European C-Band and Ku-Band Reflector Antennas for Space. Alenia has developed C- band and Ku-band reconfigurable feed systems for Intelsat applications that employ variable power dividers and variable phase shifters. These allow a spacecraft to adjust its footprint to compensate for changes in orbital position. The version shown in Figure 2.4 achieves 100 MHz of bandwidth and can operate anywhere in Ku-band.

Alcatel has done extensive work on developing software and design techniques that optimize horn cluster reflector feeds (Courtonne, Dusseux, and Brunet 1992). With the availability of better software for antenna design (and their own accumulating experience) they have been able to trade off reflector shaping against feed complexity. Thus a EUTELSAT 2 antenna system that required 22 feed horns with mid 1980s design and construction techniques can now be made with three horns and a shaped reflector.

ANT designed a sophisticated 4.7 m Ku-band offset Cassegrain antenna in a technology development project. It had 15 feed horns covering the former West Germany and one horn for Berlin. The antenna was of a petal design that would unfold after launch. In space, the reflector would be adjusted by an open loop system based on the predicted position of the surface.

ANT demonstrated a breadboard model of an 8 GHz antenna for COLUMBUS which had an unusual shaped pattern with two peaks -- one on each side of the antenna's central axis. This allows the antenna pattern to compensate for the increased path loss that occurs when the polar orbiting spacecraft is not overhead. (COLUMBUS will be in a polar orbit.) The pattern was achieved by rippling the reflector surface. While this is not a communications satellite antenna, it demonstrates ANT's high level of competence in antenna design and in shaping reflectors to achieve particular patterns.

Figure 2.20. Antenna Pointing Mechanism for the ETS-VI K-Band Single Access Antenna (Courtesy Toshiba)

Figure 2.21. A Development Antenna Pointing Mechanism (Courtesy Toshiba)

Japanese C-Band and Ku-Band Reflector Antennas for Space. The widespread adoption of satellite broadcasting in Japan and the need for shaped antenna beams that cover the country have driven Japanese techniques for designing and manufacturing shaped reflector antennas. NHK has a sophisticated design procedure whereby the main reflector (and subreflector, when used) are shaped to achieve complex patterns using a single, circular corrugated horn (Shogen et al. 1992). A 2.3 m 12 GHz test model provided excellent agreement between computed and measured patterns, delivering at least 40 dBi gain over the intended coverage area with WARC-mandated sidelobe levels and polarization purity. Toshiba makes a shaped Gregorian reflector whose single beam covers the Japanese islands; NEC makes a shaped Gregorian and shaped offset parabolic reflector the single beam of which likewise covers the Japanese islands. Measurements at 22 GHz show excellent performance. NEC uses an NHK-proprietary CAD system for designing shaped reflectors.

Combined S-band, C-band, and Ka-band Antennas. The Japanese ETS-VI spacecraft described above offers advanced sharing of reflectors for multi-band operation. The 3.5 m antenna (used for 20 GHz transmission and 2.5/2.6 GHz S-band links) and the 2.5 m antenna (used for 30 GHz reception and 6/4 GHz C-band links) were built by Toshiba. The 3.5 m antenna has a mass of 43.6 kg and an rms surface accuracy of 0.17 mm under ambient temperature conditions and 0.23 mm worst case. It is made of honeycomb sandwich with CFR face sheets and is folded by segments for launch and then deployed by a spring. The pointing accuracy for both the 3.5 m and 2.5 m antennas is plus/minus 0.015 degrees (the corresponding figure for ACTS is plus/minus 0.025 degrees). Pointing information is derived from a 30 GHz monopulse receiver that tracks a beacon station located in Hokkaido. The monopulse receiver drives an XY table which holds the subreflector of each antenna. The pointing accuracy offered by the ETS-VI antennas and their sophisticated multi-frequency reflector sharing are in advance of anything available in the U.S. or Europe.

European Space Antennas for Ka-Band and Above. Alenia Spazio was the principal company the panel visited who reported significant antenna research and development above Ku-band. Among their products is an impressive Ka-band feed system for multibeam antennas. See Figure 2.22.

For Italian military satellite communications applications, Alenia has developed a 20/44 GHz space antenna using a flat dichroic reflector. (A dichroic material is transparent at one frequency and reflective at another.) It is one of the first dichroic antennas made for 44 GHz. The antenna uses a cluster of rectangular feed horns and provides a "fully controlled" shaped beam with Italian domestic coverage.

At the component level, ANT displayed some impressive hardware that included Ku- and Ka-band ortho-mode transducers (OMTs) with bandwidths of 56% and a set of 30 and 60 GHz low pass filters. ANT has developed a dielectric backfire radiator consisting of a precisely-machined plastic structure that can be inserted in the end of a circular waveguide to make a feed. It offers a cross-pol level of 40 dB and a return loss of 30 dB with a bandwidth in the 13% to 15% range.

Figure 2.22. Feed System for a Ka-band Multibeam Antenna (Courtesy Alenia Spazio)

Alcatel Espace has done work in the past on antennas for Ka-band and mm wave frequencies. They do not see opportunities to sell antennas for this frequency range to commercial customers.

Japanese Space Antennas for Ka-Band and Above. The ETS-VI mm wave experiment (OCE) will fly a 0.4 m 43/38 GHz reflector which is mechanically steerable (by moving the entire OCE platform) over a range of plus/minus 9.8 degrees.

The COMETS 2 m Ka/mmw antenna is an offset Gregorian design with a complicated feed network (Figure 2.23). It develops co-located mm wave and Ka-band spot beams using one wideband single-mode feed horn that covers 20 to 46 GHz. This Ka-band spot beam is also used for ISLs. Two more horns generate additional Ka-band beams, one directed at Tokyo and one pointed at Nagoya. The antenna will be space deployable without active surface control. The reflector material has not yet been selected.

Figure 2.23. Feeder Link Antenna Block Diagram (Courtesy NEC)

European Multibeam Antennas for Space. Alenia has developed a multibeam Ku-band antenna whose main reflector is 3.7 m in diameter and has hinged tips. The feed uses a 1 m diameter dichroic subreflector. The subreflector is Kevlar and the main reflector is made of carbon fibers. The 20 GHz BFN is was very complex and is manufactured by electro-erosion. The antenna system is fully qualified and gives "very good performance."

Alcatel showed the panel a variety of BFNs for frequencies from L- and to Ku-band. One 1:9 model using waveguide techniques had measured losses of only 0.25 dB and a very low standing wave ratio (1.065:1). It was designed and constructed in a highly automated CAD/CAM process in which instructions went directly from a computer to the machinery without any manual intervention.

Japanese Multibeam Antennas for Space. The ETS-VI S-band antenna generates five beams each 2 degrees wide, and is a precursor of N-STAR which will have similar antenna beams.

The COMETS broadcasting experiment is part of a plan to develop a regional satellite broadcasting system for Japan with six beams, one for each of the country's cultural or ethnic regions. Each beam will use one of two downlink frequencies and the system will depend on physical isolation of the beams to prevent interference. Isolation between two neighboring beams which share the same frequency must be greater than 35 dB. The COMETS advanced broadcasting experiment will test this concept with a single-frequency two-beam prototype, made by Toshiba (Figure 2.24) that will provide service to the Kanto and Kyushu regions. Each beam will have 44 dBi edge of coverage gain. The feed system is an array of circular horns; Figure 2.25), illustrates a test model.

U.S. technology in multibeam space antennas is best represented by ACTS which develops three fixed spot beams and two independent hopping beams at 30 GHz (uplink) and 20 GHz (downlink). Both the 20 GHz and the 30 GHz antennas are offset Cassegrain systems with dual gridded subreflectors. The main reflector diameters are 3.3 m for 20 GHz and 2.2 m for 30 GHz. The spacecraft also carries a separate, mechanically steered 1.1 m diameter reflector that is capable of tracking a LEO satellite.

Superconducting Antennas. The development of high temperature superconductors has prompted research on the properties of superconducting antennas and a few superconducting antennas have been built. (For an excellent review of the potential benefits and the practical challenges of this work, see Williams and Long 1990.)

Figure 2.24. Test Model of the COMETS Broadcasting Antenna (Courtesy Toshiba)

Figure 2.25. Test Model of the Feed System for the COMETS Broadcasting Antenna (Courtesy KSRC)

While superconducting antennas are not without loss, superconductors offer an opportunity to reduce the ohmic losses in feed lines and BFNs. But a critical problem in making useful superconducting antennas is the substrate on which the thin-film high- temperature superconductors will be grown. Substrate losses can dominate conductor loss. Other practical difficulties are that there exists a maximum frequency at which a superconductor will remain superconducting and a critical magnetic field which cannot be exceeded in the superconducting state.

None of the organization visited reported work on superconducting antennas, although several knew of university work. No European work on practical antennas was reported at the 1992 IEEE Antennas and Propagation International Symposium. U.S. developments reported there included a 16 element (4 x 4) 12 GHz microstrip array with a superconducting feed network (Herd et al. 1992) and a Ka-band (30 GHz) superconducting microstrip antenna with about 2 dB more gain than a gold antenna of the same design (Richard et al. 1992). These small samples indicate a U.S. lead in superconducting antennas, but this has no obvious significance for commercial satellite communications.

Photonic Antennas. Photonic antennas (antennas with opto-electronic components) are not an active subject for R&D at European organizations, being primarily of interest to academic researchers. U.S. and Japanese work, reported at the 1992 IEEE Antennas and Propagation Society International Symposium included ten papers, principally described components rather than complete antennas, except for one (Daryoush et al. 1992) which describes a C-band active array design using optical fibers to distribute carrier signals. Photonic antennas is of relatively little European interest. American work has not yet lead to practical applications to commercial satellite communications.

No Japanese work on photonic antennas was directly reported, but an undated document "Investment and Loan Service Projects Promoted by Japan Key Technology Center" provided by ATR Optical and Radio Communications Laboratories (ATR-O)reports "Successful experiments on microwave band radiation beam formation and phase measurement in an optically-controlled array antenna."

Phased Array Antennas. Active and passive phased array antennas with microstrip radiating elements (Figure 2.26) constitute one of the busiest areas of antenna research. The practical problems of making these antennas work are well described by James (1990):

"The simplicity of printed microstrip elements is notably absent when an array is constructed, and this is true for quite small arrays. Array architecture dominates the design, which invariably necessitates complex, multilayer assemblies. Mutual coupling continues to be a difficult issue, which is made more acute if a scanning facility is required. An outstanding problem with microstrip antenna arrays is the realization of side lobe and polarization levels consistent with the knowledge of unwanted radiation effects due to surface-wave radiation, feeder radiation, etc. Herein lies the big challenge ahead for microstrip arrays -- namely that of creative, reliable, CAD packages, to totally dictate the optimal array architecture, for specifications in a wide range of applications.

"... the challenge is to engineer known circuit and antenna concepts into an integrated assembly which not only achieves the scanning specifications, but is robust enough and affordable. An example of the complexity encountered is illustrated by a thin, conformal, GaAs monolithic array, having only 16 elements, yet 768 wire bonds. The wafer yield here was quoted as, typically, 42 functional RF chips out of a batch of 87."

Figure 2.26. A Phased Array Antenna with Microstrip Radiating Elements
(Courtesy D.M. Pozar and D.R. Jackson)

Concerning other issues in realizing practical phased arrays, James says:

"Digital beamforming, combining RF and digital techniques, is an exciting prospect, in view of the continued advances in computer processing and their hardware realization. Many problems have to be solved, including dynamic range limitations, the sheer volume of data to be processed, self-calibration procedures, and presently-unrealistic costs."

European companies are actively working on phased array for space applications with ESA support. Except for the previously noted antennas for ISLs, the work is either generic or aimed at scientific or remote sensing applications.

Remote sensing applications are particularly significant because of the sophistication of the arrays involved. For example, Alenia is building a distributed phased array antenna for the Spaceborne Imaging Radar-C (SIR-C) which can scan 23 degrees. It uses adjacent L-band and C-band apertures, each radiating orthogonally polarized signals. The L-band section has 972 elements and the C-band has 5,184 (Huneycutt 1992).

Other impressive array work is in support of the European Remote Sensing Satellite, ERS-1. Ericsson and Dornier have developed a waveguide slot array with 576 elements, in metallized carbon-fiber reinforced plastic. Ericsson is concentrating on microstrip antennas with low losses, and much of this work "has been directed toward material process developments, in order to meet the demands for low weight and low losses, which are important parameters in space applications" (Dahlsjo 1992).

Alcatel representatives believe phased arrays will be an important technology in future commercial space programs and their company has invested heavily in related research programs. For radiating elements they favor cavity-backed metallic slots; in some designs used individually; in others in 4-element subarrays. The feed points are brought out to coaxial connectors at the back of the array and the elements or subarrays fed at the connectors by active or passive circuitry. All parts are extremely light. Alcatel's sentiments are shared by at least one representative of ESTEC who felt strongly that direct radiating phased arrays would ultimately be the antennas of choice for mobile satellite systems.

DLR has developed design and analysis techniques which demonstrate the capability to develop microstrip antenna designs for new applications. Work has focused on the 9 to 10 GHz band, but most results could be extrapolated to higher frequencies such as Ka- band. Electromagnetically coupled square and circular patch microstrip antennas have been designed, built, and analyzed. Units include 8 x 8 arrays and 3 x 8 arrays.

MBB has also identified active arrays as an important area for R&D and has concentrated particularly on the components and subsystems required, including high power amplifiers (HPAs), LNAs, and BPNs from L-band to Ku-band. MBB has ESA funding to work with Plessy on the realization of BFNs as multi-function MMICs. The intent is to combine all amplitude and phase steering and control functions in a single ASIC (application specific integrated circuit).

MBB's "PAMELA" linearized power amplifier modules for antenna arrays give typical L-band performance of 47 dB gain at 1,545 MHz plus/minus 25 MHz delivering more than 40 W output power at an efficiency better than 40%. The techniques involved can be extended to C- or Ku-band without difficulty, and the company is now working on a Ku- band model, skipping C-band. Parasitic problems in the diodes presently preclude Ka- band.

Recent U.S. work in microstrip phased arrays is described in (Huang 1991). Much of the current emphasis in U.S. phased array developed is on slot-coupled and aperture-coupled patches (Figure 2.27). These are reportedly simple to fabricate because they lack coaxial feedthroughs and are easier to integrate with active devices (Chen and Hamada 1992). Other advantages include wider bandwidth and possibly reduced PIM products (Hall and Sanford 1992). This did not seem to be an area of immediate interest to the European companies visited.

U.S. developments in active arrays include a TRW design for a Q-band (44 GHz) integrated active phased array transmitting antenna using InGaAs/GaAs pseudomorphic high electron mobility (HEMT) technology. Results presented were for a 2 x 2 monolithic 4-element subarray. Components developed included a high gain, high efficiency monolithic amplifier and a 3-bit switched line, monolithic phase shifter (Yen et al. 1992).

A major Japanese phased array project is Toshiba's 128-element 13.8 GHz planar slotted waveguide array for the Tropical Rainfall Measurement Mission (TRMM) precipitation radar (Figure 2.28). TRMM is a joint NASA/NASDA mission to be launched in 1997. The antenna provides 47.7 dBi peak gain with a 17 degrees scan angle and a sidelobe level better than -30 dB. Toshiba is also developing multi-frequency and multi-polarization planar arrays for future precipitation radars and arrays for 23/26 GHz ISLs.

ATR-O is developing active arrays that are small and cheap. These include planar arrays of 16 elements and spherical section arrays. Research includes work to optimize the elevation tracking range and beam formation with minimized sidelobes. ATR-O is investigating slot-controlled microstrip antennas.

NEC is not currently making phased arrays for space because of a lack of customer interest.

Figure 2.27. An Aperture-Coupled Patch Antenna (Courtesy D.M. Pozar and D.R. Jackson)

Figure 2.28. TRMM Radar Antenna (Courtesy of Toshiba)

Other recent Japanese activity in phased arrays include a planar 248-element receive array for 12 GHz satellite broadcasting with a gain of 33.1 dB and an efficiency of 78.6%. The design used rectangular patches directly connected to stripline with slots above the patches (Ota et al. 1992).

Ground Antennas for C- and Ku-Band. Of the European sites visited, only Deutsche Telecom Laboratories (DTL) reported significant R&D activities in ground antennas for C- and Ku-band. This situation reflects the maturity of that market and parallels that in the U.S. DTL has developed an innovative design for large earth station antennas that employs curved struts (rather than linear ones) to support the feed structure. This eliminates the side lobes that are commonly caused by linear struts. DTL also appears to be doing good work in developing corrugated feed horns. One example is a Ku-band feed horn for KOPERNIKUS that had 89 slots with tapered depth and spacing.

Also noted was an unexpected lack of European interest in antennas for DBS applications. This probably reflects the maturity and "consumer electronics" nature of that market, since the antenna price is already low and U.S., European, and Far East designs all seem to compete world-wide. (See Baker 1992 and Bulloch 1992 for a discussion of the European DBS equipment market.)

The success of direct satellite broadcasting in Japan has promoted commercial interest there in DBS antennas. Roof space is scarce, and needed are flat, small aperture antennas that are capable of pointing a beam at two or more satellites. Thus KSRC would like to develop an intelligent array antenna for the COMETS broadcast subscribers that would allow them to receive both COMETS and commercial DBS signals. The antenna would point a beam at one and discriminate against the other. Ideally it would be a flat plate antenna. Nippon Telephone and Telegraph (NTT) has developed a dual beam antenna with one primary reflector and two subreflectors that can simultaneously track two satellites. MELCO makes a flat 12 GHz linearly polarized TV receive only (TVRO) antenna which is 1 m square and which offers 47 dBi gain and a G/T greater than 16 dB. The company has also made a TVRO antenna using three separate feeds to receive from three different satellites without repointing.

Antenna Design, Manufacturing, and Testing

Design Software. All companies visited had sophisticated software for antenna design. The GRASP package from TICRA in Copenhagen, Denmark, is widely used in Europe and North America. Companies tended to have their own enhancements and supplemental programs. For example, DTL has a package that can accurately account for the effect of reflector paint on antenna performance.

ANT uses a sophisticated finite element antenna design software package called MAFIA, developed by the University of Darmstadt. It solves Maxwell's equations by overlaying the antenna with a system of finite elements that is not tied to a particular coordinate system.

Manufacturing. MBB demonstrated a clever technique for making dichroic (frequency selective) reflector surfaces. First they make a Kevlar substrate for the reflector and coat it with a conductive material. Then a laser burns off successive stripes from the coating, leaving an array of parallel conducting lines that reflect electromagnetic waves with one linear polarization while being almost transparent to waves with the orthogonal linear polarization. With this technique, the company can make very lightweight dual shaped reflectors. The technique is easy to understand; MBB's advantage is that it has perfected the manufacturing process.

Japanese experience with truss-honeycomb reflectors instead of conventional honeycomb construction indicates that a two-thirds reduction in weight can be achieved.

All of the manufacturing companies visited exhibited considerable skill and experience in making antennas and related components.

Testing. MBB has impressive antenna test facilities, particularly a spherical near field range of their own design and construction (Figure 2.29). They have sold similar ranges throughout Europe and North America.

The ANT antenna laboratory includes a standard outdoor range for far-field measurements, a semi-closed range for far-field measurements in a controlled environment over a limited angular range (30 degrees from boresight), and a spherical near field range. The ANT semi-closed range was particularly interesting. Its antenna positioner uses a true spherical coordinate system. All three axes intersect at a fixed point that is independent of the antenna position. The accuracy is better than 1/1,000 of a degree in any direction with the positioner fully loaded. The limiting cross-pol level of the range is 45 dB. When the range is used for far-field measurements a panel is removed to open a window and allow the antenna under test to "see" the outdoors.

The ANT semi-closed range has a sophisticated scanning mechanism that allows the antenna to be scanned along a contour of constant power. The range can produce a plot of the difference between the predicted or specified contour and the measured one. The system is also capable of doing raster scans, but the contour scan procedure is a way of developing a set of contour plots much faster. Obtaining the necessary raster scans to get 10 to 20 contours takes 30 to 40 hours of range time; a single contour scan takes five minutes, and the system plots only the contour that is wanted. This method does run a risk of missing any unwanted lobes or "islands" that may exist in the pattern outside the contour being measured.

Figure 2.29. The MBB Compensated Compact Range for Antenna Measurements (Courtesy MBB Deutsche Aerospace)


The companies visited in Europe and Japan with major antenna capabilities are all doing state-of-the-art work and show an impressive commitment to the careful manufacture of quality products. So far as space and ground antennas for today's market (C- and Ku- band fixed, broadcasting, and maritime satellite services) are concerned, there is no clear world leader. For the expected future satellite mobile market, Japan is well on the way to perfecting extremely large, space deployable antennas, particularly those that will be assembled by robots. Japan also leads in developing multiband space antennas that cover from C-band or L-band to mm waves; the degree to which these become commercially important will depend on the future popularity of frequencies at and above Ka-band for personal and mobile communications.

Published: July 1993; WTEC Hyper-Librarian