|
Dual Band Multifunction Radar Antenna Final Report
FAA FEASIBILITY STUDY DECEMBER 1993 DUAL BAND MULTIFUNCTION RADAR ANTENNA CONTRACT #: DTFA01-93-C-00023 FINAL REPORT Westinghouse Electric Corporation Dr. Gary E. Evans Peter D. Hrycak Linda S. Parrish Timothy G. Waterman James B. Yon Federal Aviation Administration ARD-90 800 Independence Avenue, SW Washington, DC 20591 DUAL BAND MULTIFUNCTION RADAR ANTENNA FEASIBILITY STUDY TABLE OF CONTENTS Page LIST OF ILLUSTRATIONS i LIST OF TABLES iii 1 OVERVIEW 1 2 REQUIREMENTS 6 3 ARRAY DESIGN 8 3.1 Array Pattern Summary 8 3.2 Beacon Sidelobe Suppression Review 10 3.3 Aperture Design Formulation 11 3.4 Surface Wave Considerations 12 4 DUAL BAND APERTURE ANALYSIS 28 4.1 Band-to-Band Isolation Requirements 29 4.2 Analytical Modeling Background 30 4.3 Integrated Radiator Development 31 5 HARDWARE DEVELOPMENT 40 5.1 Transition Into Hardware 40 5.2 Manufacturability Review 41 6 EVALUATION OF HARDWARE DEMONSTRATION 44 6.1 Scan Loss Measurements 45 6.2 Band-to-Band Isolation Measurements 47 7 CONCLUSIONS AND RECOMMENDATIONS 64 LIST OF ILLUSTRATIONS Page 1-1 Integrated Sensor Functions 4 1-2 Basic Components Of Integrated Aperture 4 1-3 Multimode Wave Guide Simulator 5 3-1 Integrated Array Geometry 14 3-2 L-Band Azimuth Patterns At Upper Frequency And Scan requirements 15 3-3 S-Band Azimuth Patterns At Upper Frequency And Scan Requirements 16 3-4 L-Band Elevation Coverage Demonstrating Sharp Horizon Cut Off And Coverage Beyond 30 17 3-5 Type Of Frequency Scanning Dividers 18 3-6 Phaser Pairs On Each Element 19 3-7 Recommended Mounting For S-Band Coverage 20 3-8 S-Band Elevation Patterns 21 3-9 Beacon Operation 22 3-10 Problems Due To Ground Lobing 22 3-11 Control Pattern With The Array Plus A Backfill 23 3-12 Preferred Control Pattern 23 3-13 Notched Patterns At 0 And 45 Scan 24 3-14 Surface Wave Concern 25 3-15 Resonance Shown In Element Pattern Due To Surface Wave 26 3-16 Element Pattern Of Similar Patch 27 4-1 Two Types Of Coupling In The Array 34 4-2 S-Band Rejection Built Into L-Band Feed 34 4-3 S-Band Balanced Feed Design 35 4-4 L-Band Feed Design 35 4-5 Integrated Dual Band Aperture 36 4-6 Vertical S-Band Computed VSWR 37 i 4-7 Horizontal S-Band Computed VSWR 38 4-8 L-Band Computed VSWR 39 5-1 Multimode Wave Guide Simulator 43 i LIST OF ILLUSTRATIONS Page 6-1 Wave Guide Simulator Wall Configuration 51 6-2 Multimode Wave Guide Simulator Operation 52 6-3 S-Parameter Set Required For Impedance Match 53 6-4 L-Band Measured Impedance Match 55 6-5 Vertical Polarized S-Band Measured Impedance Match 56 6-6 Horizontal Polarized S-Band Measured Impedance Match 57 6-7 Summary Of Isolation Measurements 58 6-8 Measured Isolation From L-Band To V-Pol S-Band 59 6-9 Measured Isolation From L-Band To H-Pol S-Band 59 6-10 Measured Isolation From V-Pol S-Band To L-Band 60 6-11 Measured Isolation From H-Pol S-Band To L-Band 60 6-12 Filter/Stub Interaction 62 6-13 Filter/Stub Design Approaches 63 ii LIST OF TABLES Page 2-1 Integrated L/S Antenna Requirements 7 6-1 Simulated Scan Angles 54 6-2 Isolation Estimates 61 iii 1 OVERVIEW As part of the "Dual Band Multifunction Radar Antenna Program", contract # DTFA01-93-C-00023, Westinghouse has successfully designed and developed an aperture for use in a multiple frequency band, shared-aperture array configuration as an option for the FAA's integrated sensor functions (Figure 1-1). It operates at three frequencies and two polarizations, scanning ±45 in azimuth and 0 to 30 in elevation. Beacon operation is provided vertically polarized at 1.03 to 1.09 GHz. L-Band radar operation is provided vertically polarized at 1.25 to 1.35 GHz. S-Band weather radar is provided both vertically and horizontally polarized at 2.7 to 2.9 GHz from which any sense of polarization can be generated. L- and S-Band frequencies share the same aperture without interference and with sufficient isolation to allow independent timing. Such a combination would allow a single antenna at an airport to provide beacon interrogation and communications, intruder monitoring, aircraft surveillance, and weather surveillance. This program was aimed at developing and demonstrating the radiators by building and testing a segment of a shared aperture. Such a dual band aperture would provide beacon, communications, aircraft and surface radar and weather radar in the area normally required for any one of them. The size requirements are comparable (approximately 15'X 25') so the overall area could otherwise become prohibitive for some installations. A rotating array without a radome would be such an example. To guarantee operation without pattern interaction and associated grating lobes, the ground rules required aperture Dual Band Multifunction Radar Antenna Final Report -1- periodicities no greater than the S-Band element spacing. As a result, the L-Band radiator is positioned behind the S-Band aperture where the gaps between multiple S-Band rows comprise the aperture area for a single L-Band element. Furthermore, the structure behind the S-Band aperture appears with the same periodicity as the S-Band element spacing. Figure 1-2 illustrates these basic components of the integrated aperture configuration. The shared aperture demonstration proved successful through a combination of analytical development followed by experimental testing and evaluation. The baseline for the shared aperture was developed with the aid of computer modeling for solving Maxwell's equations using a finite difference approach. During this stage, the integrated configuration was sufficiently developed to assure successful operation. This entails minimizing scan impedance variation across the band while maintaining sufficient band-to-band isolation needed for independent operation. The second stage provided empirical verification by conducting a hardware demonstration containing two L-Band and eight S-Band elements in a multimode waveguide simulator, as shown in Figure 1-3. The results demonstrate that all bands have VSWR of approximately 2:1 across the ranges of frequency and scan. Furthermore, the demonstration verifies that there are no blind angles or resonances due to coupling from one band into the other radiator. Isolation from L- to S-Band is always below 34 dB and from S- to L-Band it is below 29 dB. These numbers are low enough to allow independent simultaneous operation. While the intended goals were achieved, there were two main challenges to overcome. The first results from the L-Band aperture being limited by the gap size between S-Band rows. The underlying difficulty is to achieve the 30% bandwidth radiating Dual Band Multifunction Radar Antenna Final Report -2- from the limited available aperture. This is compounded by the requirement for dual polarization at S-Band, which causes difficulties in maintaining sufficient S-Band co-polarization rejection without compromising the L-Band performance. It was found that the placement of S-Band chokes in the gaps used for L-Band radiation used too much of the limited aperture, potentially limiting L-Band performance. As a result, the S-Band rejection was designed into the L-Band feed. Such a feed design requires reasonable precision to obtain both S-Band rejection and L-Band match. The S-Band element used for this program resulted from an IR&D developed dual polarized S-Band patch radiator. The patch consists of a three layer construction driven by a balanced feed. This design provides good isolation between S-Band polarizations and works well in the integrated aperture configuration. However, the complexity of the balancing network would make it expensive for final production. Therefore, we recommend modifying the feed design as part of future efforts. Attention has also been directed at the overall antenna configuration for combined beacon, aircraft and weather surveillance at airport locations. The objective was to investigate a low cost antenna architecture. The original L-Band approach using a shaped elevation beam and modules at the base of each column appears to be valid. S-Band, on the other hand, was to be frequency steered in elevation with base modules for azimuth. At least 400 MHz would be necessary to limit live length to a reasonable size, and each site would require the full frequency allocation. This extensive allocation, coupled with the potential for interference between sites, has turned out to Dual Band Multifunction Radar Antenna Final Report -3- be impractical. With frequency scan thus eliminated, individual phasers would be required on every S-band radiator if electronic dual axis scan is to be implemented. The overall conclusion is that the integration of the bands is difficult but practical. For applications in which projected area is critical, such as a mechanically rotated array, the shared aperture is the correct solution. Where surface area is not critical, as in a four-faced fixed array, separate apertures are recommended to simplify the assembly. Dual Band Multifunction Radar Antenna Final Report -4- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -5- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -6- 2 REQUIREMENTS A dual band antenna is required for a combination of beacon interrogation, communication, airport surveillance and weather surveillance, all with 360 azimuth coverage and 0 to 30 elevation coverage. Beacon and airport surveillance are performed at L-Band whereas weather surveillance is performed at S-Band. L-Band elevation coverage is achieved by shaped beam. Electronic scan is required in azimuth and elevation at S-Band and in azimuth only at L-Band. This would be true even if the antenna is mounted on a rotating platform. Furthermore, difference beams are required for monopulse capability in azimuth at both bands in addition to elevation at S-Band. To guard against false alarms, a control pattern is required to provide sidelobe suppression. Based on the general knowledge at hand, a multiface planar or cylindrical array are the likely candidates. In the absence of a detailed system design, the tentative array specifications are as shown in Table 2-1. They assume either a four-sided array or a cylinder with one quadrant active. The arrays are large in order to get the 1 beamwidth necessary for the S-Band definition. Despite this, low cost is a key parameter in the overall system design. Dual Band Multifunction Radar Antenna Final Report -7- Click HERE for graphic. Additional Beacon Requirements SLS Notch -15 to -22 dB at the sum pattern crossover SLS Gain 4 dB above the sum sidelobe and backlobe levels SLS Notch Level: <4 dB wrt crossover in sum pattern region * Suggested Parameters for reference only ** Not including azimuth beam forming *** Shaped Beam. -3 dB beamwidth. Need roll off at- 1.8 dB/deg. at 0 elevation. Table 2-1 Integrated L/S Antenna Requirements Dual Band Multifunction Radar Antenna Final Report -8- 3 ARRAY DESIGN While the main focus of this study concentrates on the development of the integrated aperture, the overall system performance should be addressed from a top level approach. This section summarizes theoretical array pattern performances and establishes basic guidelines for integrating the two independent apertures. Beamwidth and scan requirements govern the overall array dimensions and element spacings, respectively. Patterns are computed for the beacon, L-Band radar, and S-Band radar using the intended array dimensions. These patterns demonstrate grating lobe free performance at maximum scan of the highest frequency in both bands. Furthermore, the synthesized patterns demonstrate the desired sidelobe characteristics. In summary, the collection of these patterns illustrates the L-Band azimuth monopulse and elevation shaped beam, and S-Band monopulse in both azimuth and elevation. A review of the sidelobe suppression (SLS) is also summarized in this section. The beacon produces azimuth sum, difference, and control (or SLS). The control can be made with a notch by driving a central element with a small fraction of sum power subtracted. Finally, this section establishes the fundamentals for integrating the S-Band and L-Band apertures. 3.1 Array Pattern Summary Dual Band Multifunction Radar Antenna Final Report -9- Figure 3-1 shows the array geometry. For the L-Band azimuth pattern, a -32 dB Taylor weighting with n bar of 4 was assumed. This allows 7 dB of tolerance below the -25 dB sidelobe specification. For a difference beam the gain is 2 dB lower, so a -30 dB weighting was used. When the beams are scanned to 45 we must check for grating lobes at the highest (1.35 GHz) frequency. These are shown in Figure 3-2 and are seen to be lobe-free. At S-Band the corresponding azimuth patterns are shown in Figure 3-3. These are also good, showing that the element spacing is satisfactory. The L-Band elevation pattern must be shaped for sharp horizon cutoff and coverage to 30 (with a goal of 40 ) as shown in Figure 3-4. Grating lobes are not a problem at L-Band since the element spacing in wavelengths is smaller than at S-Band. Pattern shaping is accomplished with a stripline column beam former. At the base of each column is phasing, and probably amplification, provided for azimuth beam steering and beam forming. The beacon and radar would be separated and multiple sum, difference, and control patterns formed in azimuth. At S-Band, the elevation beam must be steered in both planes. Conventionally, this requires 64 times as many control elements as would be needed for azimuth alone. Consequently, frequency scanning in elevation was seriously considered at the outset. This approach is shown in Figure 3-5. Unfortunately, with the band restricted to 2.7 to 2.9 GHz, and possibly assigned airport by airport, the line lengths required for 30 steering are unrealistically large, and this technique was abandoned during the study. Dual Band Multifunction Radar Antenna Final Report -10- This leaves the baseline approach of Figure 3-6, with phaser pairs on every element. Note that the dual polarization doubles the RF device count. The recommended mounting and coverage for a planar array are shown in Figure 3-7. The scanned coverage extends from -6 to +30 in elevation which need only require ±18 of steering. However, +30 scan capability is required if the array is mounted vertically such as for a cylindrical configuration. In either case, the S-Band vertical element spacing is suitable for the maximum scan of ±30 . The scanned S-Band patterns are shown in Figure 3-8 for the highest frequency (2.9 GHz). 3.2 Beacon Sidelobe Suppression Review Conventional beacon operation is illustrated in Figure 3-9. Pulse P1 is transmitted on the sum beam and P2 on the control beam (nominally omnidirectional). If the aircraft receives P1 sufficiently stronger than P2, it is in the main beam and prepares to responds. P3 is transmitted on the sum beam and the received signals are picked up on the sum and difference beams. In the presence of ground reflections, the main and omni signals can have lobing differences that allow false responses if the antenna heights are different, as shown in Figure 3-10. Consequently, sharing the main aperture for the control pattern is presently preferred. Unfortunately, a column of the array only radiates in the forward hemisphere, requiring a backfill for 360 coverage. Figure 3-11 shows that this results in a crossover region which Dual Band Multifunction Radar Antenna Final Report -11- may have interference lobes. Fortunately, this occurs where the normal sidelobes are very low. The ideal pattern has a notch at the main beam to give a crisp and definite separation between response and no-response regions as shown in Figure 3-12. One way to achieve this is to radiate from a central column, with just enough main beam subtracted to make a null. This makes a nice notch and the notch scans with the main beam. Figure 3-13 shows the notched pattern at 0 and 45 scan superimposed on the main beam. These show that the crossover is proper and the sidelobe coverage is excellent. 3.3 Aperture Design Formulation The radiating elements can be located on either a triangular or rectangular gird. It is considerably easier to make an array of columns using a rectangular grid, especially with interlaced radiators at two frequencies. For such a grid, the maximum spacings allowed without grating lobes can be readily calculated from Click HERE for graphic.4.gif with impedance matching favoring even smaller spacings. When this is done for the two bands, it must then be compromised to allow an integral number of S-Band radiators per L-Band radiator. The resulting multiplier is 2 in both azimuth and elevation, so that the L-Band elements are slightly smaller than usual. Dual Band Multifunction Radar Antenna Final Report -12- Three factors are critical for an aperture that is shared by several bands. 1. Low frequency structures (e.g.; dipoles) must not block the high frequency radiation. 2. Available low frequency apertures must not have a periodicity larger than the high frequency spacing. 3. Each band needs a fraction of the aperture approaching its bandwidth. The first factor leads towards a slot-type of low frequency radiator that has a low profile or is flush. The second factor leads us to use a continuous slot in the H-plane driven below the surface, and multiple gaps in the E-plane. In this manner, every S-Band radiator sees the same environment. Coincidentally, the slot pair has good scanning properties. It also helps to meet the third factor, by using all of the available aperture area. The dual polarized radiator is comparably difficult to achieve. A patch seems the obvious choice, as the bandwidth is modest and dual polarized designs are available. 3.4 Surface Wave Considerations In the process of designing the radiators to scan, the possibility of resonances due to surface wave build up must be considered, especially with patches supported by dielectric. Figure 3-14 illustrates the problem. The conventional formulas for element spacing assume that propagation is at the speed of light in free space. This requires that the phase increment applied to steer left does not Dual Band Multifunction Radar Antenna Final Report -13- create a phase decrement equal to the element spacing to the right. Part a) of the figure shows that under these conditions the individual contributors to the grating lobe do not add up. If dielectric is added as in b), the desired lobe is little affected, but the surface wave can be slowed enough to build up indefinitely. Even the loading effect of radiators being passed over can contribute to the delay. The experimental simulator work includes all of these effects but works at discrete scan angles and frequencies, and could miss a narrow resonance. Some computer techniques may ignore the dielectric or the interaction of other apertures. Fortunately, the "MAXX" program does not. It also allows arbitrary scan angles and need not miss the resonance. A secondary check on surface wave possibilities could be made with an element pattern taken with the radiator "embedded" in a reasonably sized array of terminated radiators. This pattern represents the envelope of the array gain, as the array is scanned. It would therefore show a resonance as a dip at the offending scan angle as in Figure 3-15. The dip is broadened by the limited size of the test array. Such patterns were not possible on the present program, but were taken on the patch that was the basis for the current design. That patch had similar dielectric layers and had the "embedded" pattern of Figure 3-16. Clearly no sign of resonance is present at the three frequencies measured. Dual Band Multifunction Radar Antenna Final Report -14- [GRPHIC] \dual-bnd5.gif Dual Band Multifunction Radar Antenna Final Report -15- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -16- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -17- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -18- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -19- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -20- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -21- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -22- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -23- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -24- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -25- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -26- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -27- 4 DUAL BAND APERTURE ANALYSIS This section discusses an integral L- and S-Band aperture which was designed and analyzed via an electromagnetic (E-M) field analysis program. The dual band antenna is capable of vertical polarization in the 1.0 to 1.4 GHz band simultaneously with dual polarization in the 2.7 to 2.9 GHz band. The analysis show VSWR less than 2.2:1 across the L-Band with a scan volume of ±45 in azimuth and ±30 in elevation. The S-Band element was originally developed under IR&D and later incorporated into the integral aperture model for this study. In the integral configuration, this band showed VSWR less than 2.3:1 with a scan volume of ±45 in azimuth and ±30 in elevation. To ensure independent operation of the integrated dual band aperture, requirements for band-to-band isolation were initially established. It was concluded that a minimum rejection of 20 dB right at the aperture provides near independent operation. However, higher numbers contribute directly to the isolation of the full operating system. Several obstacles were to be overcome during analytical development. The first results from limitations on the L-Band aperture imposed by the S-Band structure. The challenge is to achieve the scan volume over the 30% L-Band bandwidth from its limited available aperture. Furthermore, this is compounded by the requirement for dual polarization at S-Band, which implies maintaining sufficient S-Band co-polarized rejection without compromising L-Band performance. Conventional S-Band chokes could not be used without subtracting significantly from the already small L-Band aperture and impacting bandwidth Dual Band Multifunction Radar Antenna Final Report -28- performance. This has resulted in the resonant elements built into the L-Band feed network. Overall, the analytical design demonstrates the required performance using a compact dual band configuration which was designed using the following guidelines: 1) Establish requirements for sufficient band-to-band isolation for independent operation. 2) Utilize an IR&D developed dual polarized S-Band patch as part of the model. 3) Design a distributed L-Band feed which works integrally with the S-Band aperture. 4.1 Band-to-Band Isolation Requirements A critical factor in the success of collocated apertures for independent radars is isolation. It is quite unlikely that the L-Band radar will operate at the same prf as the S-Band radar. Consequently, transmissions in one band will occur during reception in the other. With the stretched pulses commonly used, a major amount of blanking will be involved unless the isolation is large. Leakage will occur by one of three paths. The simplest to control is the low frequency interference due to shared equipment and cabling in common enclosures. This must be eliminated regardless of whether the aperture is shared. Coupling that involves the aperture is our more immediate concern. Referring to Figure 4-1, aperture coupling can take place in two ways, depending on where the frequency translation takes place. The receivers will reject out-of-band signals, but cannot Dual Band Multifunction Radar Antenna Final Report -29- distinguish signals that have been translated into the band. If transmitter A generates harmonics in band B, they will couple into the B elements and pass freely through receiver B. Let us call this "out-of-band" coupling. If transmitter A leaks into the B radiators at a sufficient strength to cause non-linearities in the LNA or mixer, we will generate intermodulation products with any other signals present, such as clutter returns. We will call this "in-band" coupling. The radars are not yet designed, but we can estimate these isolation requirements roughly. The A transmitter total average power is approximately 170 dB higher than the noise level in receiver B. The out-of-band signals 1/3 octave away can probably be held to -80 dB, so we need 90 dB rejection in the array. Likewise, the in-band peak power transmitted per large element is typically +40 dBm and the onset of LNA non-linearity is typically -10 dBm, coincidentally requiring about 70 dB rejection in the array. The bulk of the rejection can be with filters, but part must be in the element itself as in Figure 4-2. Coupling from band A into element B would cause relatively unpredictable VSWR and pattern effects. To avoid this, we feel that at least 20 dB rejection is required right at the aperture. These rejections are simple to obtain in the S-Band patch. For the other direction, we must add specifically designed chokes. 4.2 Analytical Modeling Background The analytical design for the integral aperture was developed using a field analysis program. This program solves Dual Band Multifunction Radar Antenna Final Report -30- Maxwell's equations using a time domain difference approach. The format allows one to construct a sufficiently detailed 3-D model by selecting sufficient grid resolution. The model consists of metal and dielectric structures, source excitations, loads, and boundary conditions. Through proper development, it is capable of accurately modeling element performance in large arrays. The output data provides the element's match as a function of frequency and scan. Furthermore, the output data also provides a means of monitoring band-to-band rejection which is critical to the integrated dual band operation. The 3-D model uses boundary conditions similar to those in a wave guide simulator. The main difference is that theoretical boundary conditions can be modeled for the smallest periodic aperture section which normally cannot be realized in practice. In our case this consists of one L-Band element and two S-Band elements. By changing the boundary condition properties, the element performance can be simulated for scanned array conditions. Even with the simplifications made by theoretical boundary conditions, the analytical development is very time consuming as a result of the model intricacy. Sufficient grid resolution is required to accurately model fields, but in turn greatly increases run time. A grid resolution of 1/16" provides sufficient accuracy at our frequencies. This requires a run time of roughly 15 minutes for one frequency and one scan angle. To evaluate the design, as many as three to six frequencies must be analyzed for each band and at various scan angles. Therefore, while computer modeling provides an efficient means of isolating a functioning aperture configuration, it proves less effective Dual Band Multifunction Radar Antenna Final Report -31- for fine tuning. This task is better suited during breadboard development. 4.3 Integrated Radiator Development The analytical task focuses on the development of a distributed L-Band feed that radiates in an integrated aperture environment. To save on the overall effort, without sacrifice to the objectives, an IR&D developed S-Band patch comprised the high frequency aperture. The S-Band design that occurred under IR&D proved satisfactory for use in our study. The S-Band radiates from patches having a parasitic for increased bandwidth. The main reason for using this radiator hinges on the balanced feed design for driving the patch dual polarized as shown in Figure 4-3. The symmetrical feed, similar to a crossed dipole, drives both sides with a balanced 2-wire line. The conventional patch, which is driven unbalanced, has potential problems under scanned conditions in the integrated configuration. This was avoided by utilizing a balanced drive, which leads to a more complex combination of dual polarized baluns. Although this design works, considerable simplification would be required before a full antenna is built. For wide angle scanning, the L-Band radiator should have an isolated pattern that is nearly a cosine shape in power. This matches the unavoidable variation in projected array area. A slot can be made to approximate this in the H-plane, but two parallel slots are needed in the E-plane. The shared aperture has natural scanning capability, but has a limited aperture. The limited aperture forces us to use multiple tuning components to Dual Band Multifunction Radar Antenna Final Report -32- achieve the required 30% bandwidth. These components comprise the L-Band feed, its surrounding cavity, and transition to free space. A necessary precaution during L-Band development is that line lengths joining slot pairs must be minimized. Otherwise, when the array is steered the odd symmetry coupling from neighbors may be terminated in a reactance that resonates to produce a blind angle. An additional complication to the L-Band design results from the dual polarization requirement at S-Band. Because the band separation is only 1/3 octave, it is impractical to insert S-Band chokes at the aperture. That was found to significantly complicate L-Band matching. Instead, special tuning stubs were adapted to the L-Band feed to serve as filter elements without impeding L-Band match. The L-Band feed design is shown in Figure 4-4 along with the labeling of the individual components. Figure 4-5 illustrates the complete integrated dual band aperture. it has been optimized to minimize the overall impedance variation with frequency, azimuth, and elevation scan, while maintaining isolation. The computed impedance match for S-Band operation is summarized in Figure 4-6 and 4-7 for vertical and horizontal polarization, respectively. VSWR performance are summarized at four frequencies for four separate scan conditions. In general, the vertical polarization exhibits slightly larger VSWR resulting from the slots for L-Band radiation. Overall, the vertical polarized VSWR was always better than 2.3 and the horizontal polarized VSWR always better than 1.7:1. Dual Band Multifunction Radar Antenna Final Report -33- The computed impedance match for vertical polarized L-Band operation is summarized in Figure 4-8. Again, VSWR performance are summarized at four frequencies for four separate scan conditions. Overall, its VSWR was better less than 2.2:1. Dual Band Multifunction Radar Antenna Final Report -34- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -35- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -36- [GRAPHIC] dual-bnd20.gif>HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -37- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -38- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -39- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -40- 5 HARDWARE DEVELOPMENT Final verification of the analytical design is shown via hardware demonstration. This section discusses the development of a portion of the integrated aperture along with the wave guide simulator used to test its electrical performance. The developed hardware consists of an L-Band aperture one element high by two wide integrated with an S-Band aperture two elements high by four wide as shown in Figure 5-1. The simulator walls are constructed for dual use so that both apertures can be tested using the same hardware. With such intricate structures as the integrated aperture, the realization into hardware can consume a significant portion of the overall effort. Initially, the dimensions for the integrated aperture are transitioned into manufacturable hardware components. This process generally forces minor alterations for ease of manufacturability. Furthermore, some interpolation is mandated due to finite grid resolution in the analytical model. Other important facets during this phase are the selection of appropriate materials and assembly sequences. These greatly impact ease of tuning in order to achieve the electrical specification. An important component to any theoretical design is its suitability to low cost in production. While this hardware demonstration only deals with prototype hardware, the main concerns when going to production are briefly summarized. 5.1 Transition Into Hardware Dual Band Multifunction Radar Antenna Final Report -41- The process of transitioning into hardware begins with the dimensions derived from computer analysis. These dimensions are subject to interpolation due to the computer model grid resolution. This becomes more profound during the L- and S-Band feed construction where line widths tend to be small and precise. This results from the dielectrically loaded striplines with close groundplane spacings. Because the model contains many facets, the stripline widths were simulated with a minimum amount of grid points. This requires some experience to successfully transition the design into hardware and maintain similar effective impedance. Another aspect of transitioning into hardware, is its ease of manufacturability. The computer model is free of the burden imposed by such things as intermediate connections, structural support, or assembly sequence. The computer model depicts complicated structures that must be formed by many individual pieces. Some simplifications allow substantial savings during manufacturing and assembly with little impact to electrical performance. For instance, the computer model assumes that we have sharp corners where surfaces meet. This is not possible in locations where the edges are formed since a minimum bend radius is required to prevent the material from cracking and deforming. Furthermore, other simplifications allow us to make intermediate connections of the individual piece parts for ease of assembly. Because this study involves prototype development, many parts were made of brass for solderability and thin metal for forming. Some of the closely spaced surfaces presented intricate forming challenges. In such cases, it is usually desirable to use fewer pieces and avoid soldering. This results in better Dual Band Multifunction Radar Antenna Final Report -42- electrical performance as well as a more manufacturable product. Where the intricacy becomes excessive, a composition of brass pieces are connected by soldering. 5.2 Manufacturability Review Several valuable lessons were learned during the manufacture of the dual band simulator. Previously our lower thickness limit for punching on the Weidomatic automatic punch press was about 0.030 inches thick. However, forming the intricate parts mandated thinner material. Through experimental testing, we were able to achieve high quality punching on 0.010 inch thick brass, which proved suitable for our application. The forming of the L-Band components was less complicated since there are fewer elements at a larger element spacing. The L-Band cavities were formed by constructing two thin walled brass boxes. The larger of the two boxes was a 1.5" square and the smaller box was a .825" square. We discovered during the manufacture of these boxes that even though the material was very thin the bend reduction was important. The L-Band feeds and their ground planes utilize a simple construction that can easily be transitioned into production. The S-Band design however is significantly more complicated. As previously stated the S-Band was designed under an IR&D task and would need several modifications before production. The S-Band feed mechanism proved to be the most difficult, and redesign is recommended. Dual Band Multifunction Radar Antenna Final Report -43- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -44- 6 EVALUATION OF HARDWARE DEMONSTRATION The hardware demonstration provides empirical evaluation of two important performance criteria: impedance match .vs. scan angle and band-to-band isolation .vs. scan angle. The former, referred to as scan loss, can be inferred from element tests in several ways. The most accurate method is with wave guide simulators. Furthermore, a wave guide simulator can be used to accurately estimate the band-to-band isolation. In this section, scan loss and isolation are evaluated for the dual band aperture through the use of a multimode wave guide simulator. Because both bands are being evaluated, the wave guide simulator walls must be capable of providing two configurations as shown in Figure 6-1. During L-Band testing, the top and bottom walls are positioned half way between their feeds to produce the proper boundary conditions for simulating an infinite array environment. Likewise, the top and bottom walls are positioned half way between S-Band elements during S-Band testing. In both cases the side walls remain the same. In this fashion, the L-Band simulator consists of 1 element high by 2 wide, whereas, the S-Band simulator consists of 2 elements high by 4 wide. The multimode wave guide simulator allows for the determination of scan loss and isolation at multiple scan angles with just a minimal set of recorded measurements. Wave guide theory is well known for the rectangular boundary conditions imposed on the test aperture. These conditions support TE and TM propagation at a finite set of modes determined by the operating frequency. Dual Band Multifunction Radar Antenna Final Report -45- Each mode correlates to plane wave propagating conditions of the test aperture at a known scan angle which is a function of frequency. Figure 6-2 illustrates this phenomenon. The test aperture plane wave propagating conditions are constructed through superposition of a set of S-parameter measurements made during each test. There were a total of six test performed: three for scan loss measurements and four for isolation measurements. For each set of S-parameter measurements, all propagating modes can be constructed and its response for the given test determined. 6.1 Scan Loss Measurements A primary objective of the radiator design is to minimize reflection losses as a function of scan angle. The gain must fall off as the projected area or cos(SCAN) and it is common practice to assume cos1.5(SCAN) falloff to allow for reflections. That would allow a 2.33:1 VSWR at 45 scan. Our goal has been to hold this to 2.0:1 to keep scan losses less than 0.5 dB max. Because of minor differences between the theoretical and physical simulator the two sets of results approximately correlate but not exactly. Therefore, the match for each band was tuned only for sufficient representation of the theoretical results. As a result, some frequencies exceed a 2:1 VSWR at some scan angles. Given more time all frequencies could be reduced to approximately a 2:1 VSWR max. Figure 6-3 summarizes the S-parameter sets needed to determine impedance match for L-Band and dual polarized S-Band. Only those modes are constructed that fall within our scan limits. At L-Band, only one mode is useful and it simulates a Dual Band Multifunction Radar Antenna Final Report -46- beam at 0 elevation and 41 at 1.0 GHz to 29 at 1.35 GHz in azimuth. At S-Band, there are enough elements to simulate a number of beams for each polarization. The beam directions are summarized for both bands in Table 6-1. These angles provide sufficient coverage of the scan region to validate the overall predicted performance analyzed in the theoretical model. The final L-Band impedance match results are shown in Figure 6-4 for the single usable wave guide mode. The match is better than a 2:1 VSWR except at the high end of the band with the worst VSWR being 2.5:1. This is comparable to predicted performance of 2.2:1 maximum. A 2% frequency shift would correct it. The final L-Band impedance match was achieved with three modifications to the initial hardware: 1) Removal of the horizontal septums due to an unforeseen sharp resonance around 1.1 GHz, 2) Changes to the L-Band feed short circuit, and 3) Changes to the L-Band feed open circuit stub. The short circuit stub and open circuit stub together form a high Q matching circuit for L-Band while providing the needed rejection at S-Band. For this reason, the match is fairly sensitive. Figure 6-5A summarizes the S-Band vertical polarization impedance match for the 0 elevation scan modes and Figure 6-5B for the 25 (average) elevation scan modes. For all of the 0 and most of the 25 coverage, the VSWR is generally better than 2.2:1. This is comparable to predicted performance of 2.3:1 maximum. At 25 elevation, 3 particular points at the band edge exceed our goal, and warrant further matching. Dual Band Multifunction Radar Antenna Final Report -47- The final S-Band vertical polarization impedance match was achieved with two modifications to the initial hardware: 1) Increased the parasitic patch vertically to compensate for air gaps between patch layers resulting during assembly, and 2) Slightly widen the vertical dipole arms to help center match as well as reduce impedance match sensitivity. Figure 6-6 shows the S-Band horizontal polarization impedance match for the 25 (average) elevation scan modes. A maximum VSWR of 2:1 was achieved across the band and scan modes. This is comparable to predicted performance of 1.7:1 maximum. The final S-Band horizontal polarization impedance match was achieved with two modifications to the initial hardware: 1) Increased the parasitic patch horizontally to compensate for air gaps between patch layers resulting during assembly, and 2) Slightly shorten the quarterwave open on the horizontal feed circuit. The overall scan loss tests are in good agreement with analytical predictions. Furthermore, both indicate that the two bands can be operated integrally without interaction, resonances, or blind angles. 6.2 Band-to-Band Isolation Measurements As discussed in Section 4.1, the total isolation between systems includes several contributors. While filters provide the bulk of the rejection, some isolation is mandated at the aperture to ensure independent operation of each band. This section summarizes testing and results of the aperture isolation. It was Dual Band Multifunction Radar Antenna Final Report -48- found that, as a minimum, the aperture isolation needs to be better than 20 dB but higher numbers contribute significantly to the isolation of the full operating system. RF coupling between bands must involve a frequency translation to be accepted by the narrow band IF receivers. This either occurs in the transmitter (modulation envelope), or in the receiver (intermodulation). The former requires aperture coupling out-of-band, while the latter results from in-band coupling. Both cases have been determined from S-parameter measurements made on the simulator. Isolation between bands was determined from S-parameter measurements from each element of one band to a single element of the other band as shown in Figure 6-7. By using vector addition of the coupling values, with a weighting corresponding to the various scan modes, the band-to-band isolation can be found for various scan angles. The isolation between L-Band and vertical polarized S-Band, with an L-Band signal, is shown in Figure 6-8 for the single usable wave guide mode. The simulator wall configuration for this measurement allows isolation to be determined at four of the eight S-Band elements. Due to symmetry, only the two unique isolation measurements were recorded. The isolation was found to be better than 34 dB across the band for the single scan mode. This data represents the L-Band level at the S-Band LNA, and consequently is used to determine the intermodulation level. Figure 6-9 shows the isolation between L-Band and horizontal polarized S-Band for the single usable wave guide mode. The isolation is nominally better than 45 dB with a worst case of 40 Dual Band Multifunction Radar Antenna Final Report -49- dB. This data is used in the determination of the overall isolation for the other S-Band receiver. As a result of reciprocity, the data in Figures 6-8 and 6-9 can also be used to determine the out-of-band isolation when the S-Band transmitter is on and inadvertently radiating a small L-Band component. The primary difference is that in one case we use peak power affecting receiver linearity, and in the other it is average power contributing noise. The isolation between vertical polarized S-Band and L-Band slots with an S-Band signal is shown in Figure 6-10. The simulator wall configuration for this test permits isolation measurements at two of the four L-Band elements. Due to symmetry only one of the isolation measurements was recorded. At vertical polarized S-Band, there are three scan modes within our scan volume corresponding to 14 , 28 , and 45 average azimuth and 0 elevation scan. The largest scan mode showed the worst isolation. However, for this mode, frequencies below 2.78 GHz correspond to azimuth scan angles greater than 45 . With the exclusion of those frequencies the isolation is better than 29 dB. Figure 6-11 shows the isolation between horizontal polarized S-Band patches and L-Band slots with an S-Band signal. For this measurement the average scan angles are 0 , 14 , and 45 azimuth and 25 elevation. The isolation for the 28 azimuth scan mode can not be determined because it produces a null at the measured L-Band element. As before, the worst isolation occurs at the largest scan mode where the lower end of the band is excluded because it exceeds our scan limits. Dual Band Multifunction Radar Antenna Final Report -50- Therefore, the isolation is better than 40 dB for all frequencies and scan angles within our interest. As before, reciprocity allows us to use this data to determine out-of-band isolation when the L-Band transmitter is on. From this empirical evaluation, we can estimate the filter requirements as shown in Table 6-2. In estimating these requirements the interaction between the filter and the stubs in the radiator must be considered, although the filter design is not part of the present program. Figure 6-12 illustrates the concern. Both the stub and the conventional low pass filter are highly reactive and can cause interacting reflections. The result is typically a degradation of the rejection band at certain frequencies, with 6 dB changes not uncommon. When the time comes to design the filter, several approaches can be taken as shown in Figure 6-13. Ideally, the reactive elements can be spaced as shown in Figure 6- 13a). The odd number of quarter wavelengths makes the rejections add instead of cancel over a particular bandwidth. If this bandwidth were insufficient, the filter could be integrated with the stub as shown in Figure 6-13b). We prefer to avoid this, as it complicates the lines joining the halves of the L-Band radiator. This is not recommended. If the spacing approach were not acceptable a third alternative (Figure 6-13c)) is to make the filter a diplexer with Dual Band Multifunction Radar Antenna Final Report -51- the stop band terminated. It is very unlikely that this will be necessary. Dual Band Multifunction Radar Antenna Final Report -52- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -53- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -54- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -55- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -56- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -57- Dual Band Multifunction Radar Antenna Final Report -58- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -59- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -60- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -61- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -62- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -63- CONTRIBUTORS TO ISOLATION (L-S) TRANSMIT Transmit spectral purity 90 dB Filter >21 dB Coupling of radiator to collocated subarray 29 dB Incoherent combination in combiner 30 dB __________ 170 dB db RECEIVE Fraction of peak power in vicinity 30 dB Coupling to one radiator 40 dB Receive filter >31 dB ___________ 100 dB db Table 6-2 Isolation Estimates Dual Band Multifunction Radar Antenna Final Report -64- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -65- Click HERE for graphic. Dual Band Multifunction Radar Antenna Final Report -66- 7 CONCLUSIONS AND RECOMMENDATIONS The objective of the Dual Band Radar Antenna Program was to determine the technical feasibility of a single Electronically Scannable Integrated Aperture which can serve the Vertically Polarized Beacon Function (1.03-1.09 GHz), the L-Band Radar Intruder Detection Function (1.25-1.35 GHz), and the Polarization Selectable Hazardous Weather Detection Function (2.7-2.9 GHz). The integrated system of radiators was designed to accommodate 0 to 30 coverage in elevation and ± 45 coverage in azimuth. Both the theoretical calculations and the waveguide simulator measurements have demonstrated feasibility. VSWR's less than 2:1 across all ranges of frequency and scan are practical for this integrated system of radiator. The calculations and measurements also verify that blind angles or resonances due to coupling from one band to another do not exist with this design. In order to operate both bands simultaneously, we calculated that the band-to-band isolation must be at least 20 dB at the radiators (Section 4). The measured band-to-band isolation was measured to be better than 29 dB with this design. The measured results indicate that margin exists for an integrated design making it even more feasible. The system design originally employed a shaped elevation beam at L-Band with T/R modules and phase shifters placed at the column level. This would provide elevation coverage and azimuth scan capability for the MSSR and intruder functions. Including the active array control elements at the column level would allow for an easily maintainable system. The T/R modules (or Æ-shifters) would be able to be replaced while the system was operational. We also considered a low cost approach for the S-Band function. One means of reducing the cost of the S-Band subsystem Dual Band Multifunction Radar Antenna Final Report -67- would be to use frequency scan in elevation and phase scan at the column level. This would allow us to have the array active control elements only at the column level. This configuration would result in a low-cost (few T/R modules), easily maintainable system which offers full azimuth and elevation beam agility. We learned during the program that the entire frequency band of 2.7-2.9 GHz would not be available. This coupled with the fact that the line lengths required for 30 elevation steering would be large forced us to abandon this type of system architecture. An alternative, low-cost architecture at S-Band needs to be studied in subsequent efforts. Several recommendations are made. First, if an integrated dual band system with dual polarized S-band becomes a requirement, the S-Band element should be simplified. The one used for the simulator measurements was derived from an IR&D Program which investigated and built a dual polarized element with good axial ratio. Its feed mechanism needs to be simplified. However, the results of this study are still valid. No change would be apparent from the array face, but the construction behind the array would be simplified and the cost reduced. It is further recommended that an alternative architecture for the S-Band functions be explored. The weather mapping function requires electronic san and polarization diversity. Existing array architectures which provide these functions are prohibitively expensive. Trade- off analysis and performance requirements need to be done to drive out prioritized system parameters and postulate potential antenna architectures. Experiments should then be done on novel architectures to both verify technical claims and drive out potential technical problems. Dual Band Multifunction Radar Antenna Final Report -68-
