A connectorless module for an EHF phased-array antenna.
Mobile SATCOM Terminals
Future military requirements call for a quick reaction global presence, using fewer resources in all categories. MILSATCOM terminals are necessary to steer these resources to instantaneous worldwide communications and facilitate rapid deployment and effectiveness of mobile military elements. Highly reliable, low maintenance mobile terminals that operate unattended are necessary to achieve enhanced performance while using fewer personnel. This paper describes a high-performance EHF phased-array antenna packaging architecture that brings autonomously operated, mobile terminals to the smallest military element for communicating situation awareness, targeting updates and threat information, over-the-horizon, and worldwide. Table 1 lists the potential benefits of this architecture. Such antennas, supplied economically, will find widespread application in mobile SATCOM terminals used in aircraft, ground vehicles, ships and spacecraft.
Packaging is the principal technology challenge facing industry in providing affordable, high-performance EHF phased arrays. The problem is largely imposed by the small feature sizes available. At 44 GHz, for example, separation between elements in a wide angle scanning array must be less than 0.15". The region behind this area must house at least an amplifier, a phase shifter, a means for RF, DC and logic interconnections, and an antenna element. The various involved technologies, including antennas, semiconductors, microwave circuits, hermetic sealing, thermal management, assembly, test and inspection, often dictate conflicting requirements within the same, extremely-limited physical space. Analysis of cost factors suggests that major packaging improvements are required. Also, module and module qualification costs must be reduced to achieve affordable phased-array antenna subsystems. Objective
In 1987, a research project was initiated to develop active EHF phased-array antennas for the airborne terminal of a SATCOM system. Separate aperture 44 GHz transmit and 20 GHz receive arrays were to be provided, each containing hundreds of elements. As shown in Figure 1, 140 |degrees~ of angular coverage, to within 20 |degrees~ of the horizon, using a single pair of apertures was a subsystem objective. Overall array thickness, including heat exchanger, was limited to less than 6". Operating bandwidth is five percent. Since the SATCOM is a frequency hopping system, instantaneous bandwidth need only be a small fraction of the operating bandwidth. In both arrays, a single, agile, sum beam was to be provided. Tracking for the satellite terminal is provided by electronically nutating the receiving antenna beam.
The array apertures were required to operate for exceptionally long periods between maintenance intervals, preferably longer than the structural life of the platform itself. Perhaps the most critical requirements were that the arrays be available in the mid-1990s and be affordable, that is, with complete antenna subsystems costing less than $200 per radiating element in quantity production. Trade Studies
It was determined that hybrid technology was essential in order to achieve affordable antennas in the mid-1990s. After evaluating several approaches, it became clear that availability of fully monolithic phased-array antennas at 20 GHz and 44 GHz was unlikely. Potential yields of subarrays of MMIC circuits containing hermetically sealed antenna elements, amplifiers, phase shifters, logic circuits and passive components at these frequencies with quarter micron feature size were viewed as stretching credibility on both affordability and availability issues. In addition, an all-monolithic package would not deliver satisfactory performance in power, noise figure, reliability and an antenna coverage. Consequently, attention was focused on hybrid approaches, that is, approaches incorporating discrete MMIC circuits for amplifiers and phase shifters, that are bonded into hermetically-sealed, conductor-shielded, hybrid packages. A key feature of the hybrid approach is the ability to use the best available MMICs, independent of technology and/or supplier.
The single-channel, hybrid module per radiating element approach was selected for its overall cost and performance advantages. The principal hybrid module candidates considered were single electronic module (single-channel module) per radiating element, and an arrangement of four, eight or 16 electronic channels side-by-side in a tray-type architecture, servicing four, eight or 16 radiating elements. While combining multiple channels into a single package has the advantage of minimizing the number of difficult interconnections, the selected approach was to develop improved interconnect technology to take advantage of the lower cost and superior performance of the single-channel module approach. An important aspect of cost control is minimizing labor in assembly, testing, qualification, transport and installation of modules. The simplicity of the single module is suitable for hands-free, automatic machine assembly, testing and qualification.
The single-channel, hybrid module per radiating element architecture has advantages over tray architecture in terms of isolation, single-point failure, antenna element, polarizer, semiconductor protection, radar cross section and conformality. Loss of isolation causes degradation of amplitude and phase control. The independent single-channel module architecture provides superior isolation between channels, and between input and output on each channel.
Although all phased arrays are credited with graceful failure, a single-point failure in a tray is potentially more damaging. The damage is caused by multiple radiating elements being affected.
The selected connectorless approach to module embedding enables the use of open-ended, dielectric-loaded waveguides as radiating elements that have broad E- and H-plane patterns over wide scan angles. These elements differ from dipole radiators that are often used with tray packaging.
Waveguide elements enable a choice of polarizers, not only the meanderline polarizer. A waveguide-mounted polarizer provides lower loss and better axial ration over wide scan angles and is protected from physical and electrical damage.
The single module per radiating element semiconductor is mounted behind waveguide radiators. Therefore, it is protected from damage caused by lightning-induced surface currents or high powered radar.
An EHF antenna of open-ended waveguide radiators provides reduced radar cross section at frequencies below guide cutoff. Conformal antennas with complex shapes are easier to design using single modules than using trays.
A simple and robust antenna packaging architecture is described wherein antenna assembly and all interconnections are quickly accomplished using mechanical fasteners without the need for wire nor solder connections. Figure 2 shows the hybrid module-honeycomb architecture, common to both transmit and receive arrays. The architecture contains three main subassemblies, an antenna honeycomb, a module honeycomb and a feed honeycomb, as well as two multilayer wiring boards. The honeycomb-like structures are fabricated using numerically controlled electric discharge machining (EDM).
The antenna subassembly is a triangular lattice of dielectrically-loaded, circular waveguide antenna elements. To prevent onset of grating lobes, spacing between radiating elements was selected to be |lambda~/|square root of 3~, where |lambda~ is free space wavelength. This close spacing dictates dielectric loading to lower the waveguide cutoff frequency. Internal circular polarizers serve as these loads. The external antenna cover is a wide angle, impedance-matching (WAIM) surface. These multilayer, radome-quality, low-loss dielectric WAIM designs have been described previously.|1~ Less than 1 dB of reflection loss can be achieved over a five percent bandwidth for scan angles out to 70 |degrees~ off broadside for these arrays.
The module-honeycomb subassembly consists of individual hermetically sealed modules positioned within a honeycomb-like support structure. The modules are located on a triangular grid matching the antenna lattice pattern. Compared to a square-element lattice, this approach maximizes the size of each module to provide circuit space, and minimizes the number of modules for a given aperture size to minimize array cost. An amplifier, a phase shifter and a logic chip are located within each module. The module substrate extends beyond the two faces of the honeycomb into the lattice of waveguides making up the antenna on one side, and into the waveguides forming the RF divider/combiner on the other side. The extensions provide RF radiation coupling into and out of the module. There are no metal-to-metal RF connectors on the modules. The RF divider/combiner subassembly contains an array of waveguides geometrically identical to the antenna assembly except that the dielectric loads do not embody circular polarizers. Bonded to this honeycomb is the waveguide distribution network (WDN), which is a slotted waveguide array. Slot parameters were adjusted to provide a uniform amplitude distribution across the array. Each slot in the WDN is aligned with and excites one of the waveguide holes in the feed subassembly. By using waveguide throughout the WDN design, RF losses are minimized. Waste thermal energy is conducted from the modules via the feed honeycomb through the WDN to the mechanical and thermal interface on the back. The WDN design and performance has been described previously.(2,3) In the representative hermetic hybrid MMIC module, shown in Figure 2, the ends of the module contain transitions that couple RF energy into and out of the module without mechanical contact. On the ends of the module carrier are metal contacts that conduct DC power, ground and logic signals into the module. During assembly, modules are inserted into the module honeycomb holes and are locked in place using a spring retainer. Sandwiched between the honeycomb subassemblies are two copper/gold multilayer wiring boards. Power and logic signals are conducted to each of the modules through these two boards. Electrical contact is achieved at assembly by compressing elastomeric connectors between the modules and the wiring boards, as shown in Figure 3. Thermal vias are provided through the logic board to ensure efficient thermal transport to the heat exchanger. The entire assembly is held together using simple mechanical fasteners. The assembly process is easily reversed, thereby providing ready access to the modules for repair or replacement.
A test-bed antenna was fabricated to validate and reduce the key features of the design to practice. The test bed consisted of a 91-element array, where the central 19 elements are active. The remaining 72 radiating elements were left empty. They could have been resistively loaded to simulate a larger array. A multimode feed horn is used to excite the centermost active elements for economy. Figure 4 shows the test-bed antenna during assembly, including the module honeycomb, and the central portion of the antenna before and after modules are inserted and wiring boards with pressure plates are pulled into place with four bolts. Protruding RF transitions of the 19 active module substrates in the assembled module honeycomb are shown. These noncontact RF transitions couple RF energy into and out of this waveguide assembly on each side of the module honeycomb.
Assembly of the entire test-bed antenna is completed in minutes, and is reversible to the module level. It is assembled and held together with 14 bolts, four associated with the module honeycomb center assembly, and 10 to attach the feed assembly and radiating elements. During assembly, the RF signal into and out of each module is accomplished by 38 radiation-coupled connectorless interconnections, one on each end of each module. The DC, logic and ground connection to each module consists of 114 interconnections, three on each end of each module.
The receive test-bed antenna was mounted on a two-dimensional motorized platform that simulated motions of flight, and demonstrated by tracking and receiving an FM modulated television program from either of two ceiling-mounted antennas simulating relay satellites. Additional demonstrated features include global search and satellite acquisition, sector scan satellite tracking, built in test and angular coverage over 140 |degrees~.
The beam steering controller (BSC) positions the antenna beam by adjusting the settings of the phase shifters in the individual modules. Phase shifter settings are calculated in a central processing data and distributes them to the logic multilayer wiring board. The phase shifters for an entire column of elements are updated in parallel, selected by the active clock. The BSC is capable of supporting up to 64 x 64 element phased arrays, updating all the elements within 40 ||micro~seconds~, with less than a 0.6 ||micro~seconds~ interruption in the beam (settling time). The controller accepts beam position of azimuth and elevation angles, frequency and synchronization information from a terminal, and computes the desired four-bit, formatted phase shift settings. The BSC accommodates adjustments in phase shift settings to compensate for WDN variations, frequency dependent performance of the modules or component temperature dependencies at up to 64 frequencies, or at 16 frequencies and four temperatures. These adjustments are achived on less than 40 square inches of board area.
For low cost, the phased-array modules are designed to be simple electronically and physically. Figure 5 shows a schematic diagram of a receive module, which consists of a MMIC low noise amplifier (LNA), a MMIC four-bit phase shifter and a silicon BiCMOS logic integrated circuit. A transmit module schematic diagram is identical except the LNA is replaced by a power amplifier and the signal direction is reversed.
Each module housing consists of a ceramic substrate, a carrier, a hermetic seal ring and a lid. Figure 6 shows the bottom view of the module, which displays the carrier with DC terminals mounted on the end (logic terminals are identically located on the opposite end). These DC terminals are connected to the multilayered wiring boards through compressible elastomeric connectors. The terminals are fed through the module substrate to the interior DC and logic traces on the top side through metalized, hermetic via holes. The carrier provides an efficient thermal path from the MMICs to the module-honeycomb, and from there to the feed honeycomb and heat exchanger through metal-filled via holes in the multilayer wiring board. This carrier is made of copper-tungsten with end contacts provided on small ceramic substrates. In later versions, the contacts are printed directly on the ends of a BeO carrier, which provides reduced parts count, significant weight-savings and improved thermal performance.
The RF transitions (slot line fed dipoles) on each end of the module substrates provide radiation coupling between the modules and the circular waveguides of the antenna and feed subassemblies. The thickness of the ceramic substrate is chosen to load the circular waveguide to avoid cutoff and to provide the impedance match with the adjacent dielectrically filled circular waveguide section.
During the past six years, all the necessary components have been designed, fabricated, tested and assembled into working antenna models. Results of this work at 20 GHz and preliminary results at 44 GHz are described.
Figure 7 shows the circuit board layout and the measured losses and gains for a typical 20 GHz module. The entire module has less than 1.5 dB of back-to-back insertion loss from 19.6 to 21 GHz, when microstrip lines are used to replace both MMICs. Thus, the module transition loss is about 0.6 dB to a point on the substrate where the MMIC amplifier is placed on the antenna end of the module. The back-to-back return loss is less than -20 dB from 19.7 to 21.2 GHz without the MMIC chips. Separate measurements and theory show the microstrip-to-slotline transition has over 20 percent bandwidth. The reduced bandwidth associated with the entire module transition, |approximately~7 percent, is limited by the slotline/dipole radiating element.
MMICs and Modules
One-half micron MESFET MMIC devices were used for the demonstration array because they were readily available, passivated, reliable and of moderate cost. The average gain for the amplifier chips was 17.5 dB. The phase shifters provided four bits of phase shift and had an average insertion loss of 11 dB. The module gain is 6 dB from 19.5 to 21.5 GHz using this chip set. The input return loss is -10 dB over this range and is primarily set by the input impedance of the MMIC amplifier chip. The module's output return loss is -20 dB and is set by the MMIC phase shifter chip's output impedance.
The subarrays fabricated and tested to date have performed close to prediction. This experience bolsters confidence that this packaging architecture represents the leading edge for future phased-array designs. The 20 GHz receive test-bed antenna with 19 active modules and beam steering controller, proved to be a cost-effective mechanism for verifying array design and scan performance out to 70 |degrees~ off broadside.|1~ A 91-element, 20 GHz receive array and a 91-element, 44 GHz transmit array subsequently have been fabricated, tested and accepted.|4~ Figure 8 is a photograph of both of these antennas with the WAIM removed for clarity.
The research has been focused on airborne SATCOM terminals, including tactical aircraft. However, the phased-array architecture has potential for application on a wide variety of military platforms, including spacecraft, ships and ground vehicles. Research is also on-going to extend the technology to EHF T/R modules.|5~ Other potential applications include deep space communications and direct broadcast satellite (DBS) terminals for receiving television programs onboard commercial aircraft. The described phased-array antenna architecture can be used at frequencies through at least 60 GHz.
The objective for this research, to develop an EHF phased-array antenna packaging architecture that is affordable and available in the mid-1990s, has been accomplished. This achievement, in conjunction with subsequent contract results,|4~ has prepared the architecture for the next step, full scale development. A module manufacturing cost target approaching $100 each for relatively large quantities of |is greater than~ 50,000 modules per year was established near the beginning of the program in 1988. Such costs are within reach today with additional manufacturing engineering.
We are grateful to John P. Turtle, Rome Lab., and Don McMeen, Boeing for the photograph of the two antennas shown in Figure 8. MMIC on-wafer measurements are courtesy of Glenn H. Martin. Micro-electronic assembly was performed by Marjorie Benezra. John J. Rankin supplied the graphic art.
1. B.J. Lamberty, W.P. Geren, S.H. Goodman, G.E. Miller and K.A. Dallabetta, "Wide-Angle Impedance Matching Surfaces for Circular Waveguide Phased-Array Antennas with 70 |degrees~ Scan Capability," Proceedings of the 1992 Antenna Applications Symposium, Monticello, IL, 23-25 Sept. 1992.
2. B.J. Lamberty, G.E. Miller, D.N. Rasmussen, G.W. Fitzsimmons and E.J. Vertatschitsch, "Waveguide Distribution Network for EHF Phased-Array Antennas," Proceedings of the 1992 Antenna Applications Symposium, Monticello, IL, 23-25 Sept. 1992.
3. B.J. Lamberty, E.J. Vertatschitsch and G.W. Fitzsimmons, "Distribution Network for Phased Arrays", US Patent No. 4,939,527, July 3, 1990.
4. "Integrated Circuit Active Phased-Array Antenna (ICAPA)," Contract No. F19628-90-C-0168, Rome Lab., John P. Turtle, Technical Monitor/EEAA, Don McMeen, Boeing Program Manager.
5. "MMW T/R Module," Contract No. F-33615-92-C-1116, Wright Lab., Keith Stamper, Technical Monitor/ELM, Raynor Shimoda, Boeing Program Manager.
TABLE I POTENTIAL BENEFITS OF EHF MOBILE SATCOM ACTIVE PHASED-ARRAY ANTENNAS
* Autonomous operation (no crew set-up)
* Long MTBF (|is greater than or equal to~ platform life)
* Graceful failure
* Lower maintenance & training costs
* Mobile aperture for tactical forces
* Force multiplier
* Small & lightweightfor
* Low observable
* Low drag
* Long MTBF
* Agile beams
* Multiple beams
* Adaptive nulling
George W. Fitzsimmons received his BSEE and MSEE from the University of Washington in 1959 and 1963, respectively. He joined the Boeing Company in 1960 and is now an associate technical fellow in the Research & Engineering Department, Boeing Defense and Space Group. He is also a task leader in the RF and Photonics Laboratory with a focus on mm-wavelength communication phased-array antennas and MMIC insertion. At Boeing, Fitzsimmons has conducted applied research on a wide variety of passive and active microwave circuits and subsystems. He is member of the IEEE. Fitzsimmons has received 10 patents. Bernard J. Lamberty received his BSEE from the Illinois Institute of Technology in 1950 and his MSEE degree from Stanford University in 1961. He has 42 years experience in applied R&D in antennas and EM, the last 16 years at Boeing, where in 1990 he was made an associate technical fellow. Earlier engineering experience was at ESL Inc., GTE Electronic Defense Laboratory and ITT Gilfillan. He was an instructor at New Mexico State University and California State Polytechnic College. Lamberty holds 15 patents (five others pending). He is a member of the IEEE, Eta Kappa Nu and Tau Beta Pi.
Donn T. Harvey received his BSEE (honors) from the Milwaukee School of Engineering in 1985. At that time, he joined the Boeing Company, where he worked at the High Technology Center designing microwave and mm-wave MMIC and hybrid circuits for radar applications. Since 1989, Harvey has been working on the development of EHF active phased-array antenna packaging technology. Harvey holds one patent. He is a member of the IEEE and past chairman of the Seattle MTT Section for two terms.
Dietrich E. Riemer received a Master's degree in electrical engineering in 1955 and a PhD in mechanical engineering in 1987 from the Technical University Berlin. He joined Boeing in 1971. He is responsible for electronic packaging in the Boeing RF/Photonics laboratory. He has an extensive background in microelectronics manufacturing and in packaging design analysis. Riemer has been awarded six patents. He is a fellow of ISHM and a recipient of the ISHM Technical Achievement Award and the ISHM Daniel C. Hughes Award. Edward J. Vertatschitsch received his BSc in applied physics in 1980 and his PhD in electrical engineering in 1987 from McMaster University in Hamilton, Ontario, Canada. He joined the Boeing Company in 1987 and is now a senior principal engineer in the Research & Engineering Division. He also is a project lead engineer in the RF/Photonic Systems section. Vertatschitsch has received three patents. He serves on the AIAA technical committee on sensor systems and is an IEEE member.
Jack E. Wallace received his BSEE and MSEE degrees from the University of Washington in 1972 and 1978, respectively. In his 16 years of experience at Boeing, he has designed a variety of RF and microwave systems, subsystems and components, including GaAs monolithic circuits.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||extreme high frequency|
|Author:||Fitzsimmons, George W.; Lamberty, Bernie J.; Harvey, Donn T.; Riemer, Dietrich E.; Vertatschitsch, E|
|Date:||Jan 1, 1994|
|Previous Article:||The smallest personal communication systems voltage-controlled oscillator.|
|Next Article:||A high precision outdoor cyclindrical near-field test facility.|