210 19 24MB
English Pages 282 Year 2023
Practical Aspects of Active Phased Array Antenna Development
7065_Agrawal_V6.indd 1
4/6/23 1:42 PM
For a complete listing of titles in the Artech House Radar Library, please turn to the back of this book.
7065_Agrawal_V6.indd 2
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development Ashok K. Agrawal
artechhouse.com
7065_Agrawal_V6.indd 3
4/6/23 1:42 PM
Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the U.S. Library of Congress. ISBN-13: 978-1-63081-989-7 Cover design by Mark Bergeron © 2023 Artech House 685 Canton Street Norwood, MA 02062 All rights reserved. Printed and bound in the United States of America. No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the publisher. All terms mentioned in this book that are known to be trademarks or service marks have been appropriately capitalized. Artech House cannot attest to the accuracy of this information. Use of a term in this book should not be regarded as affecting the validity of any trademark or service mark. 10 9 8 7 6 5 4 3 2 1
7065_Agrawal_V6.indd 4
4/6/23 1:42 PM
To the memory of my parents
7065_Agrawal_V6.indd 5
4/6/23 1:42 PM
7065_Agrawal_V6.indd 6
4/6/23 1:42 PM
Contents Preface xv
1 Practical Aspects of Active Phased Array Antenna Development 1 1.1 Introduction 1 1.2 Active Phased Array Antenna System 4 1.3 Passive Phased Array Antenna 5 1.4 Passive Phased Array Antenna Limitations 8 1.5 Active Phased Array Antenna 8 1.6 Key Radar System-Level Advantages of Active Phased Arrays over Passive Phased Arrays 10 1.6.1 Increased Sensitivity 10 1.6.2 Improved Target Detection in Clutter 10 1.6.3 Improved Waveform and Pattern Flexibility 10 1.6.4 Improved Wideband Operation 10 1.6.5 Increased Reliability 11 1.6.6 Reduced Prime Power Requirement 11 1.6.7 Reduced Cost 11 1.6.8 Lower Noise Temperature/Figure 11 1.6.9 Adaptive and Digital Beamforming 11 1.7 Tracking Radar Performance Metric 12 1.8 Introductions to the Chapters 13 vii
7065_Agrawal_V6.indd 7
4/6/23 1:42 PM
viii
Practical Aspects of Active Phased Array Antenna Development
1.9 Concluding Remarks 17 References 18
2 Analysis and Design of Linear and Planar Phased Arrays 19 2.1 Introduction 19 2.2 Analysis of Linear Arrays 19 2.3 Low Sidelobes for Linear Arrays 23 2.4 Low Sidelobe Aperture Distributions 23 2.4.1 Dolph-Chebyshev Aperture Distribution 23 2.4.2 Taylor Distribution for Linear Arrays 25 2.4.3 Bayliss Distribution for Difference Patterns 26 2.4.4 Implementation of Monopulse Beams for an Active Planar Phased Array Antenna 26 2.5 Analysis and Synthesis of Planar Arrays 27 2.5.1 Rectangular Grid 28 2.5.2 Triangular Array Element Grid 31 2.6 Comparison of Rectangular and Triangular Grids 33 2.7 Minimize the Number of Elements for a Grating Lobe-Free Pattern Using a Tilted Array 34 2.8 Directivity and Gain of Active Arrays 38 2.9 Effect of Amplitude and Phase Errors on the Phased Array Antenna Performance 40 2.9.1 Quantization Errors 40 2.9.2 RMS Sidelobe Level Due to Amplitude and Phase Errors 42 2.10 Beam Pointing Error Due to Phase Quantization 44 2.11 Bandwidth Criteria for Active Phased Array Antennas 44 2.11.1 Instantaneous Bandwidth 45 2.11.2 Phased Array Operating Bandwidth 46 2.12 Moderate Instantaneous Wide Bandwidth Array by Applying Amplitude Taper in the Receiver 47 2.13 Concluding Remarks 50 References 50
7065_Agrawal_V6.indd 8
4/6/23 1:42 PM
Contents
ix
3 Transmit/Receive Modules 53 3.1 Introduction 53 3.2 T/R Module Architecture 54 3.2.1 Control Module 56 3.2.2 Integration of T/R Module with DC-to-DC Converter 57 3.2.3 Common Leg T/R Module Architecture 58 3.3 Active Phased Array Performance Improvement 59 3.3.1 GaN Wide Bandgap Power Amplifiers 60 3.4 T/R Module Key Performance Parameters 61 3.4.1 Power-Added Efficiency 61 3.4.2 T/R Module Noise Figure 62 3.4.3 Noise Figure of a Cascaded Network 62 3.4.4 T/R Module Noise Temperature 63 3.4.5 1-dB Compression Point 64 3.4.6 Third-Order Intercept Point 65 3.5 T/R Module Architecture Trade-Offs 67 3.6 T/R Module Architectures for Circular Polarization 69 3.7 T/R Module Construction 70 3.8 Thermal Stack-Up of the T/R Module 71 3.9 Integration of MMIC, Control Module, and DC-to-DC Converters 73 3.10 T/R Module Stability 74 3.11 T/R Module Reliability 75 3.12 T/R Module Cost 76 3.13 Performance Requirements of T/R Modules 78 3.14 Application of Silicon Germanium (SiGe) BiCMOS Technology in T/R Modules 79 3.15 Concluding Remarks 81 References 81
4 Beamformer Architectures for Active Phased Array Antennas 83 4.1 Introduction 83 4.2 Beamformer Networks for Passive Phased Array Antennas 86 4.3 Beamformer Networks for Active Phased Array Antennas 90
7065_Agrawal_V6.indd 9
4/6/23 1:42 PM
x
Practical Aspects of Active Phased Array Antenna Development
4.3.1 Multiple Independent Receive Beams 91 4.4 Impact of Beamformer Architecture on System Noise Figure 94 4.5 Beamformer Architectures for High Reliability 97 4.6 Beamformer Networks for Wideband Active Phased Array Antennas 97 4.7 Concluding Remarks 101 References 102
5 Radiating Elements 103 5.1 Introduction 103 5.2 Printed Circuit Radiating Elements 104 5.2.1 Printed Circuit Wideband Radiating Elements 106 5.3 Waveguide Radiating Elements 107 5.3.1 A Wideband Tapered Double-Ridged Waveguide Element Fed by a Coaxial Probe 109 5.4 Radome Heating for Ice Inhibition 110 5.5 Wideband Parallel Waveguide Phased Array Radiator 112 5.6 Mutual Coupling Between Radiating Elements 114 5.7 Selection of the Radiating Element Type 116 5.8 Radiating Element Design Process 117 5.9 Phased Array Radiation Pattern Calculation by Using the Mutual Coupling Between Elements in a Small Array 121 5.10 Concluding Remarks 122 References 122
6 Beam Steering and DC Power Distribution 125 6.1 Active Phased Array Antenna Beam Steering Controller 125 6.1.1 Active Phased Array Distributed Beam Steering Controller 126 6.1.2 Active Phased Array Centralized Beam Steering Controller 128 6.2 Active Phased Array Power Distribution 129 6.2.1 DC-to-DC Converter Key Requirements 130 6.2.2 Distributed Power System 130 6.2.3 Centralized Power System 132
7065_Agrawal_V6.indd 10
4/6/23 1:42 PM
Contents
xi
6.2.4 Average versus Peak DC-to-DC Converters 133 6.2.5 Comparison of Distributed and Centralized Power Systems 134
6.3 Concluding Remarks 134 References 135
7 Active Phased Array Antenna Packaging 137 7.1 Introduction 137 7.2 Array Packaging Concepts 140 7.2.1 Tile Array Construction and Cooling Methods 141 7.2.2 Brick Array Packaging 141 7.2.3 Components of an LRU 142 7.2.4 Thermal Management 143 7.3 Active Array Antenna Brick Packaging Schemes 144 7.3.1 Sliding Vertical Cold Plate Active Array Packaging 145 7.3.2 Edge-Cooled, Horizontal Cold Plate Array Packaging 146 7.3.3 Vertical Fixed Cold Plate Packaging Concept 148 7.4 LRU to the Radiating Element RF Connections 149 7.5 Structural Design 150 7.6 Active Array Antenna Radome Design 150 7.7 Concluding Remarks 153 References 153
8 Active Phased Array Antenna Design for High Reliability 155 8.1 Introduction 155 8.2 Antenna MTBF 156 8.3 Active Phased Array Antenna Architecture Description for High Reliability 158 8.4 Maximizing the Array MTBCF 160 8.5 Antenna MTBF for Different Cluster Sizes 162 8.6 Increasing Array MTBCF with Redundant Power Supplies 166 8.7 Driver Amplifier Boosters in the Active Phased Array Beamformers 169
7065_Agrawal_V6.indd 11
4/6/23 1:42 PM
xii
Practical Aspects of Active Phased Array Antenna Development
8.8 Lifecycle Maintenance Cost Estimation of an Active Phased Array Antenna 170 8.9 Active Phased Array Antenna Availability and Sparing 173 8.10 Concluding Remarks 174 References 174
9 Active Phased Array Design for High Clutter Improvement Factor 175 9.1 Introduction 175 9.2 Centralized Phased Array Architecture 176 9.3 Distributed Array Architecture 179 9.4 Concluding Remarks 182 References 183
10 Active Phased Array Antenna Calibration 185 10.1 Introduction 185 10.2 Active Array Calibration Using Mutual Coupling Between Array and External Elements 186 10.3 Active Array Calibration Technique Using Mutual Coupling Between Array Elements 188 10.4 Active Array Calibration Technique Using Mutual Coupling Between One Calibration Element and All Array Elements 190 10.5 Active Array Calibration Technique Using Mutual Coupling Between a Few Dedicated Internal Elements and the Array Elements 191 10.5.1 Calibration Procedure 193 10.5.2 Required Number of Calibration Elements 195 10.5.3 Calibration Accuracy 196 10.5.4 Effect on Array Packaging 197 10.6 Concluding Remarks 198 References 198
7065_Agrawal_V6.indd 12
4/6/23 1:42 PM
Contents
xiii
11 Digital Beamforming for Active Phased Array Antennas 201 11.1 Introduction 201 11.2 Dynamic Range Improvement 203 11.3 Digital Beamforming at the Subarray Level 204 11.4 Digital Beamforming of Multiple Simultaneously Independent Receiver Beams 206 11.5 Angle Tracking Accuracy 207 11.6 Adaptive Digital Beamforming 208 11.6.1 Adapting Nulling in Analog Arrays 209 11.7 Exciter Noise and Clutter Attenuation 210 11.8 Concluding Remarks 211 References 211
12 Cost Reduction Strategies for Active Phased Array Antennas 213 12.1 Introduction 213 12.2 High Cost of Current Active Phased Array Antennas 215 12.3 SPY-1 Array Antenna Cost Reduction 216 12.4 Improvements in Technology and Manufacturing Processes 217 12.5 Paradigms 218 12.5.1 Legacy Systems 218 12.5.2 Commercial Parts and Processes Are Not Adequate for Military Applications 218 12.5.3 Cost-Plus Contracts 219 12.5.4 Lack of Incentives 219 12.5.5 Schedule Limitations Do Not Permit Any Design Changes 220 12.5.6 The Benefits of Competition to the Buyer: An Automobile Industry Example 220 12.5.7 Use the Best Available Technology 220 12.5.8 Changes Will Increase Program Costs and Schedule Delays 221 12.6 Design Philosophy 221 12.6.1 Bottom-Up 221 12.6.2 Top-Down 222
7065_Agrawal_V6.indd 13
4/6/23 1:42 PM
xiv
Practical Aspects of Active Phased Array Antenna Development
12.7 Cost Reduction Strategies 223 12.7.1 Optimizing T/R Module RF Output Power Levels for Phased Array Antenna Cost, Size, Prime Power, and Dissipated Heat 223 12.7.2 Trading the Number of Array Faces for a Hemispherical Field of View 227 12.7.3 Band-Aid Solutions 228 12.7.4 Antenna Architecture 229 12.7.5 Minimize the Number of Interfaces 230 12.7.6 LRU Size versus Cost 231 12.7.7 Radiating Element 232 12.7.8 T/R Modules 232 12.7.9 Module Packaging 234 12.7.10 DC Power Distribution 235 12.7.11 Beamformers, Cables, and Connectors 235 12.7.12 Power-Added-Efficiency and Cost 236 12.7.13 Active Phased Array Antennas for Wide Bandwidth Operation 236 12.7.14 Antenna Assembly and Test 237
12.8 Concluding Remarks 238 References 239
Appendix 241 T/R Module Requirements and Flow Down to the Components 241 T/R Module Requirements Flow Down (Transmit Channel) 241 T/R Module Requirements Flow Down (Receive Channel) 243
List of Acronyms 247 About the Author 251 Index 253
7065_Agrawal_V6.indd 14
4/6/23 1:42 PM
Preface After working for more than 35 years designing and developing active phased array antennas, I was motivated to write this book that provides the practical aspects of an active phased array antenna system’s design, analysis, and development. An active phased array antenna system consists of a radiating aperture, T/R modules, beamformers, beam steering controller (BSC), DC-to-DC converters, antenna structure that contains all components, thermal management to keep the device junction temperatures below a prescribed value, active phased array antenna calibration in the field, reliability, maintainability, and digital beamforming. Phased array antennas have been addressed in many books, but the authors tend to focus on the aperture design and address some of the other issues but do not provide a guide for the design and development of active phased arrays. This book provides a comprehensive design and development of an active phased array antenna system. An active phased array radar can provide orders of magnitude performance improvement over its predecessor—the passive phased array antenna. In addition, it can provide improved reliability and target detection in the presence of large clutter, wideband operation, reduced prime power requirements, and reduced total ownership costs. As a result, virtually all ground-, air-, and shipbased high-performance radars recently fielded or under development today employ an active array antenna. In addition to military applications, active array technologies are now used in space-based
xv
7065_Agrawal_V6.indd 15
4/6/23 1:42 PM
xvi
Practical Aspects of Active Phased Array Antenna Development
satellite active arrays for high-speed internet, FAA aircraft aviation radars, and 5G wireless systems. Active phased array system design involves inputs from multiple disciplines, such as aperture design, T/R module design, hybrid circuits, BSC, mechanical engineering, and manufacturing. In addition, the focus is on the lifetime costs and not just the manufacturing cost. Therefore, an active phased array system engineer must know the system’s multiple aspects to coordinate the antenna system design. Failure to identify interrelated issues early may result in complications that may require the redesign of some of the components of the antenna, resulting in increased costs and schedule delays. The T/R module is the heart of the antenna system and contributes approximately 50% of the cost and most dissipated heat. Detailed design and development of the T/R modules and cost reduction strategies are provided in Chapter 3. Although the cost of active arrays has come down significantly, their lifetime costs are still considered high, and at the same time, the requirements are increasing to meet the new threats. The high lifetime cost can result in cancelling the program or reducing the total number of systems significantly, as is the case with DDG-1000 ships and F-35 aircraft. Many paradigms impede the design for a lower cost. It takes a mindset to design active arrays for a lower cost. The cost of the SPY-1 antenna (passive) was reduced significantly without changing the architecture of the antenna, even after the antenna was in production and a large number of antennas had already been produced. The lifetime cost of active arrays can easily be reduced by 20% to 30% following the guidelines presented in Chapter 12. Digital beamforming (DBF) at the element level leads to active array capabilities far beyond conventional, analog beamformed active arrays. For example, DBF can generate many simultaneous independent receive beams, increase receiver dynamic range, provide adaptive beamforming by placing nulls in the array pattern in the directions of several jammers, provide true time delay digitally, and generate ultra-low sidelobes to reduce clutter from land and sea. I designed several active phased array antennas during my tenure at Lockheed Martin (RCA, GE, Martin Marietta). Later working at Johns Hopkins University’s Applied Physics Laboratory and Alion Science and Technology, I provided oversight of several U.S
7065_Agrawal_V6.indd 16
4/6/23 1:42 PM
Preface
xvii
government active phased array programs developed by the three major active array antenna providers. It gave me a unique understanding of how to improve communications with the government to avoid issues later on, improve the development process, meet customer requirements, and reduce lifetime costs. It would be beneficial if we could show the detailed design of a real active phased array antenna system. But unfortunately, active phased array designs are proprietary to the corporations that own them. However, this book presents all aspects of the active phased array antenna system development. This book aims to provide a practical guide for designing and developing an active phased array antenna system. Although this book concentrates on active array antennas for land-based and seaborne platforms, the underlying principles would also apply to all other platforms. I am very grateful to Dr. Eric Holzman for untold hours reviewing the whole manuscript and providing significant editorial and technical comments that have improved this book’s content. In addition, he contributed a section on the use of SiGe technology in active phased array antennas. In the past, we had done joint research, and his contributions are included in this book. I would also like to thank Allan Jablon for reviewing the chapters. Finally, I must express my gratitude to my wife, Sudha, for her patience and understanding during the completion of this endeavour.
7065_Agrawal_V6.indd 17
4/6/23 1:42 PM
7065_Agrawal_V6.indd 18
4/6/23 1:42 PM
1 Practical Aspects of Active Phased Array Antenna Development 1.1 Introduction An active phased array radar can provide orders of magnitude performance improvement over its predecessor passive phased array radar while at the same time improving reliability and reducing total ownership costs. As a result, virtually all high-performance radars under development today employ an active array antenna. Navy radar development programs for which active phased arrays are a critical enabling technology include the AN/ SPY-3 multifunction radar, AN/SPY-4 the Volume Search Radar for long-range surveillance, AN/SPY-6 Air and Missile Defense Radar, and advanced radars for Navy-Theater-Wide (NTW) Tactical Ballistic Missile Defense. Other active array radars include the Theater High Altitude Area Defense (THAAD), Space Fence Radar for tracking small objects in space, Long-Range Missile Defense Radar (LRDR), and the AN/TPS-80 Ground/Air Task Oriented Radar (G/ATOR). In addition, F-15, F-16, and F-18 gimbaled fixed beam fire control radars have been upgraded with active phased array radars. Besides military applications, active array technologies are being used in commercial communications applications, including Iridium and Globalstar systems; however, these systems have not been currently proven economically viable. Active phased arrays 1
7065_Agrawal_V6.indd 1
4/6/23 1:42 PM
2
Practical Aspects of Active Phased Array Antenna Development
are now being developed for high-speed satellite communication systems and high-speed internet service using many satellites with active phased array antennas. Current and future military radar requirements are driven by rapidly evolving threats, including cruise missiles and tactical ballistic missiles. To address these threats, array antennas must operate over wider bandwidths and with enhanced sensitivity, higher radiated power levels, improved stability, and improved electronic protection to address reduced target radar cross-sections. In addition, there is a growing need for reduced array signatures and a practical need to control costs, including acquisition, operational, and support costs. As a result, active phased array antennas have emerged as a fundamental technology for addressing these evolving military radar systems’ needs. In addition to enhanced sensitivity, improved system stability will be required to detect low-flying cruise missiles in the sea or land clutter. Furthermore, wider bandwidths will be needed to perform discrimination and target identification functions. At the same time that radar demands are increasing, there is a practical need to reduce acquisition as well as operation and support costs, improve reliability, and reduce manning requirements. Ground and shipboard radar systems are typically required for surveillance of thousands of angular locations, tracking hundreds of targets, and guide missiles within relatively short reaction times. These requirements can only be met with phased array antennas that provide for the electronic repositioning of radar beams to widely diverse angular locations within microseconds. Active array technology offers the capability of achieving the required performance improvements while at the same time offering improvements in reliability, maintainability, availability, and lifecycle costs. Both transmit and receive functions in an active array are moved to the aperture by placing a transmit/receive (T/R) module at each radiating element. The T/R modules provide power amplification during transmit, low noise amplification during receive, and amplitude and phase control for beam steering and sidelobe reduction. Because this configuration places the power amplifiers and low noise amplifiers at the aperture, transmit and receive losses of passive arrays are significantly reduced, resulting in increased radar sensitivity for a given amount of generated microwave power. The transmit and receive losses in passive arrays
7065_Agrawal_V6.indd 2
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
3
can be 8 to 10 dB, increasing the radar sensitivity by that amount. Further improvement in radar sensitivity for active phased array antennas will be discussed later. In the 1990s, gallium arsenide (GaAs) devices were used for power amplifiers, and the available output power was limited. The ground and shipboard radars require much higher amplifier power. With the advent of the wide bandgap gallium nitride (GaN) around 2010, the amplifier power has been improved by an order of magnitude. In the 1990s, the cost of active phased arrays was considered unaffordable, except for military applications. Although the cost of active phased arrays has been reduced significantly, they are still considered expensive. It is desired to reduce the cost further. In addition, future systems will require higher powers as new threats are identified. It would be beneficial if we could show the detailed design of a real example of an active phased array antenna system. But unfortunately, active phased array designs are proprietary to the corporations that own them. Therefore, this book presents a comprehensive design process for active phased array antenna systems using the published material and commonsense processes. A passive phased array antenna uses a single high-power transmitter based on microwave tubes or power-combined solid-state amplifiers, which is a single point of failure. Furthermore, the mean time between failures (MTBF) of the tube-based transmitter is very low compared to the T/R modules. The power efficiency of an active phased array antenna is much higher than the tube-based transmitter since the efficiency of the T/R modules is much higher than the tubes. In addition, the transmit and receive losses are eliminated, resulting in much lower prime power requirements. This is important because phased array radars require their own power generation and conversion systems. The high MTBF of the T/R modules provides much higher operation availability and allows for the graceful degradation of performance. Active phased array antennas offer high reliability and fewer maintenance requirements resulting in lower lifetime costs. Several books have presented phased array antennas [1–8]. However, they do not address the design of all active phased array antenna system components and tend to focus on the radiating aperture. This book presents a comprehensive design of an active phased array system, addressing all the key subsystems.
7065_Agrawal_V6.indd 3
4/6/23 1:42 PM
4
Practical Aspects of Active Phased Array Antenna Development
The development of an active phased array antenna system requires inputs from many disciplines such as radio frequency (RF) integrated circuit technology, radiating element design, beamformers steering controller, DC-to-DC power conversion, structure design, packaging of components for ease of assembly and maintenance, reliability and maintainability, thermal management, and lifecycle cost. The antenna systems engineer is responsible for coordinating all these array design aspects to meet the system requirements. Some requirements may compromise the system requirements and components/subsystem performance. Therefore, the systems engineer must continuously evaluate the system’s performance based on the actual performance of components and subsystems. Sometimes, all aspects of the system design, such as a shortfall in components performance, are not addressed at the beginning of the program, and issues are discovered later. Many times, the schedule and cost do not allow any major changes in the design. This book aims to provide a practical guide for designing and developing a complete active phased array antenna system, including transmit and receive RF electronics, DC-to-DC converters, beamformers (analog and digital), beam steering controllers, antenna packaging, thermal management, and antenna calibration in the field. In addition, antenna design for high reliability and detection of targets in clutter and strategies for cost reduction are also discussed. In addition, we emphasize on addressing all issues with multiple disciplines at the beginning of the program rather than addressing them after the designs have been completed, and the schedule may not permit any changes.
1.2 Active Phased Array Antenna System An active phased array radar system consists of a radome for environmental protection, radiating aperture, T/R modules, DC-to-DC converters, beamformers, a beam-steering controller (BSC), a receiver/exciter (REX), and a data and signal processing computer, as shown in Figure 1.1. The active phased array antenna system is a major part of the radar system and consists of a radiating aperture, T/R modules, DC-to-DC converters, beamformers, a beam steering
7065_Agrawal_V6.indd 4
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
5
Figure 1.1 Block diagram of an active phased array radar system.
controller, and a circulator/duplexer is used to separate the transmit and receive fluid cooling system. The shaded components are part of the radar system and are not addressed in this book. To appreciate the benefits of active phased array antennas, let us look at the performance characteristics and limitations of passive phased array antennas.
1.3 Passive Phased Array Antenna Figure 1.2 shows an RF block diagram of a passive linear phased array antenna. Passive phased array antennas are powered by highpower tubes, use high-power phase shifters, and require separate transmit and receive beamformers. Transmit beamformers use waveguide beamformers to handle the high power, and receive beamformers use printed circuit board-based stripline/microstrip line beamformers. A duplexer is used for separating the transmit and receive paths. A low noise amplifier (LNA) is placed at the end
7065_Agrawal_V6.indd 5
4/6/23 1:42 PM
6
Practical Aspects of Active Phased Array Antenna Development
of the receive beamformer. Figure 1.3 shows the RF block diagram of a passive linear array antenna for producing sum and difference beams. To produce the sum and difference beams, 180º hybrid couplers are used. An RF block diagram of a planar passive phased array antenna is shown in Figure 1.4. Row and column beamformers form the sum, azimuth delta, elevation delta, and delta-delta beams. Details of this beamformer architecture will be discussed in Chapter 4. Signals from elements in each column equidistant from the array centerline are fed to the column beamformers of sum and difference hybrids, where the two vector signals are added and subtracted. The sum and difference outputs are separately weighted and combined to form one elevation sum and one elevation difference signal per column of radiators. The column sum and difference signals are combined symmetrically about the array centerline in the horizontal beamformers to form the three monopulse signals. A Taylor weighting is applied in the sum beamformers to obtain low sidelobes in the receive mode. To obtain Taylor weighting in two dimensions, the weighting in the column and horizontal beamformers is such that their multiplication results in the desired two-dimensional weighting function. Similarly, a Bayliss weighting is applied in the beamformers to provide low sidelobes in the difference beams.
Figure 1.2 An RF block diagram of a linear passive phased array antenna.
7065_Agrawal_V6.indd 6
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
7
Figure 1.3 An RF block diagram of a linear passive phased array antenna for producing receive sum and difference beams.
Figure 1.4 Block diagram of a receive beamformer for a planar passive phased array antenna.
7065_Agrawal_V6.indd 7
4/6/23 1:42 PM
8
Practical Aspects of Active Phased Array Antenna Development
Radar systems often use monopulse techniques to derive angle tracking information from a single-echo pulse, requiring a separate receive beamformer for each of the three monopulse beams, that is, sum, elevation difference (EL DELTA), and azimuth difference (AZ DELTA) [8]. The target can be anywhere within the 3-dB beamwidth of the sum beam, while the deep null in the difference pattern provides the accurate angle of the target. In passive phased arrays, the transmit power from a single transmitter is distributed to the array elements through a passive beamformer network. Separate transmit and receive beamformers are used for the transmit and receive beams. The receive beamforming network provides the amplitude taper to provide low receive sidelobes.
1.4 Passive Phased Array Antenna Limitations Passive array systems have several inherent performance limitations and inefficiencies. For example, the transmit beamformer typically has significant losses, and the transmitter must generate a large amount of power to overcome these losses. In essence, a substantial portion of the RF power generated by the transmitter is dissipated as heat in the beamformer before being radiated. In addition, high-power centralized transmitters generally employ microwave tube-based technologies that operate at lower duty factors and have limited waveform flexibility. Also, high receive beamformer losses significantly degrade the receive sensitivity, particularly when low sidelobes are required. In addition, transmitting noise from a centralized source is often limited in clutter-driven radar applications. Finally, high-power tube-based transmitters and their attendant high-voltage power supplies have lower reliability and higher maintenance and replacement costs than solid-state technology. The latter issue is significant for shipboard applications involving relatively long missions and a strong desire to avoid maintenance at sea.
1.5 Active Phased Array Antenna An RF block diagram of an active linear array antenna is shown in Figure 1.5. As mentioned earlier, both transmit and receive functions in an active array are moved to the aperture by placing transmit and
7065_Agrawal_V6.indd 8
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
9
Figure 1.5 RF block diagram of a linear active phased array antenna.
receive functions at each radiating element. An RF block diagram of a two-dimensional active phased array antenna is shown in Figure 1.6. A T/R module is placed at each element to provide both transmit and receive functions. The column and horizontal combiners produce the sum, delta azimuth, and delta elevation beams. Notice that a single beamformer can be used for both the transmit
Figure 1.6 Block diagram of a two-dimensional active phased array antenna.
7065_Agrawal_V6.indd 9
4/6/23 1:42 PM
10
Practical Aspects of Active Phased Array Antenna Development
and receive functions, eliminating the need for separate transmit and receive beamformers in the passive phased arrays.
1.6 Key Radar System-Level Advantages of Active Phased Arrays over Passive Phased Arrays 1.6.1 Increased Sensitivity Lower transmit and receive beamformer losses coupled with the ability of solid-state T/R modules to operate at higher duty cycles than conventional tube-based transmitters generally enable order-of-magnitude improvements in radar sensitivity. 1.6.2 Improved Target Detection in Clutter In an active array, key sources of transmit noise and instabilities (e.g., T/R modules and power supplies) are distributed at the aperture. Consequently, their noise contributions do not add coherently in the same fashion as the transmitted signal. Instead, their contributions to pulse-pulse variations undergo an averaging effect. The result is a significant improvement in the ability of an active array radar to detect small moving targets in the sea or land clutter. 1.6.3 Improved Waveform and Pattern Flexibility The solid-state technology facilitates the multiple functions of detection, tracking, target identification, illumination, kill assessment, and missile communications can be better optimized. Also, since the T/R modules provide both amplitude and phase control at the element level, radiation patterns are more readily optimized for the radar mode of operation, including null synthesis techniques. The waveform, amplitude, phase, frequency, and timing can be dynamically controlled as a function of the threat and environment. 1.6.4 Improved Wideband Operation The solid-state technology employed by active arrays can support inherently wideband microwave frequency operation. In addition, active array architectures are conducive to implementing true practical time delay devices, which support wide bandwidth, highrange resolution waveforms, and target imaging capability.
7065_Agrawal_V6.indd 10
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
11
1.6.5 Increased Reliability Because of the independent nature of the T/R modules, the radar remains operational if a few percent of the modules fail. In contrast, a failure in a conventional radar transmitter can take the system entirely out of service, possibly at critical times. Active phased arrays can be serviced at a periodic interval when the failed T/R modules can be replaced. Performance degradation is mainly assessed by the increase in sidelobe level, while the decrease in gain is minimal. 1.6.6 Reduced Prime Power Requirement The prime power requirements of an active phased array antenna are significantly lower than the tube-based systems, for the power efficiency of the T/R modules is much higher than microwave tubes. Furthermore, the transmit and receive losses are reduced by 8 to 10 dB, and the cooling requirements for removing the dissipated heat are much reduced, resulting in further reduced prime power. 1.6.7 Reduced Cost The lifecycle cost of active phased arrays, including manufacturing, can be lower than that of passive phased arrays. In addition, since the MTBF of T/R modules is much higher than the tube transmitters, the increased reliability results in lower maintenance costs over the lifetime of the antenna. 1.6.8 Lower Noise Temperature/Figure Since the low noise amplifier receiver is distributed at the aperture, the losses of the receive manifold are eliminated, resulting in a lower noise figure (NF). The improvement in the NF can be 3 to 8 dB, depending on the size of the array. 1.6.9 Adaptive and Digital Beamforming By placing receivers and A/D converters behind each element, the signal processor can process the outputs from each A/D converter, and a large number of simultaneous beams can be formed digitally;
7065_Agrawal_V6.indd 11
4/6/23 1:42 PM
12
Practical Aspects of Active Phased Array Antenna Development
and provide adaptive beamforming by placing nulls in the direction of several jammers.
1.7 Tracking Radar Performance Metric The radar range equation for a tracking radar can be written as [9] 4 Rmax =
( 4p )
2
PtGt Ae s
(
kTo Bn Fn So /N o
)
min
where Pt = total transmitted power Gt = transmit antenna gain Ae = receive antenna effective aperture
σ = radar cross-section kToBn = input noise power Bn = noise bandwidth k = Boltzmann’s constant = 1.38 × 10–23J/deg To = standard temperature of 290°K (approximately room temperature) Fn = noise figure (So/No)min = minimum signal-to-noise ratio The range at which tracking and discrimination can be performed depends on the figure of merit power aperture gain (PAG): PAG ∝ power • receive aperture • transmit gain PAG = (PelementNelements)(AelementNelements)(PelementNelements) PAG = PelementA2elementN3elements This metric can be used to compare the performance of radars. It is easy to compare the radars’ performance without each radar’s details. For example, the performance of a new radar can be expressed in terms of the PAG improvement in decibels over its predecessor.
7065_Agrawal_V6.indd 12
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
13
Assuming a T/R module is used at every element, doubling the number of elements increases the PAG by a factor of 8 (9 dB). This extra 9 dB provides a 70% increase in the range. PAG can be increased by increasing the number of modules or by increasing the T/R module output power. However, increasing the antenna size may be limited by real estate availability. This factor will be discussed in Section 12.7. Increasing the antenna size with lower power modules reduces prime power requirements.
1.8 Introductions to the Chapters Chapter 2 provides the analysis and design of the phased array aperture. Since this topic has been covered in many texts and articles, only the pertinent material that would help select the aperture parameters for the design of active arrays is presented. These design parameters include element spacing to provide a grating lobes-free pattern in the scan volume and optimize element spacing for the minimum number of elements for a given aperture. In addition, phased arrays require low sidelobe patterns to avoid detection in the sidelobes. Also discussed is the impact of periodic and random errors on the antenna performance parameters, such as sidelobes and gain. Finally, the array architectures for wide and moderate instantaneous bandwidths are presented. Chapter 3 presents the design of T/R modules. T/R modules are considered the heart of an active array, for they determine all the performance parameters of an active array antenna, such as power radiated by the radiating elements, beam switching, pulse width, sidelobe levels, amplitude, and phase settings for beam scanning, intermodulation products, third-order intercept point, and array efficiency. T/R modules use the majority of the DC power. T/R modules dissipate a large amount of heat that must be removed to maintain the T/R module temperature to a specified value. Although active array antenna costs were very high in the early 1990s and power was low to meet the requirements for military radars, a significant development in GaN technologies has increased the output power by almost an order of magnitude. The array costs have also come down to levels that justify using solid-state technologies. However, there is a continuous desire to reduce the active phased arrays cost as it is still considered expensive.
7065_Agrawal_V6.indd 13
4/6/23 1:42 PM
14
Practical Aspects of Active Phased Array Antenna Development
Chapter 4 presents beamformer architectures for active phased array antennas. Beamformers provide amplitude distribution at the array aperture radiating elements to obtain the low sidelobe patterns. Radar systems often use monopulse techniques to derive angle tracking from a single-echo pulse, requiring a separate receive beamformer for each of the three monopulse beams, that is, sum, elevation difference (EL DELTA), and azimuth difference (AZ DELTA) [9]. Corporate beamformers are generally preferred in phased arrays that need wide instantaneous bandwidth and good impedance match. High-power passive phased array antennas use waveguide beamformers in the transmit mode and stripline/microstrip line in the receive mode. On the other hand, in active phased arrays, both transmit and receive functions are placed at the aperture. This function is achieved using a T/R module for each radiating element. The amplitude distribution in the active phased arrays can be provided either in the T/R modules or the beamformers. The two beamformer architectures are compared regarding their impact on the antenna noise figure (NF). Beamformer architectures for the wideband active phased arrays are also presented in this chapter. Chapter 5 presents the various choices for the radiating elements for different platforms. In addition, the design procedures for the radiating elements in an array environment are discussed. Since the input impedance of the radiating element is different from the isolated element, the element should be designed to provide the best impedance match with the feeding networks and space in the array environment. Therefore, the procedure for the radiating element design in the array environment is described in this chapter. Finally, the radiating element designs are verified using waveguide simulators and pattern measurement in a small array before commencing the antenna production. Chapter 6 presents the beam steering controller (BSC) and DC power distributions. The BSC performs the following functions: steer the antenna beam, switch the antenna to transmit or receive mode, beam switching time, fast switching within a dwell, transmit waveform from radar to antenna, provide built-in test (BIT) from individual elements, switching of all drain and gate voltages, provide fault detection and fault isolation function, mutual coupling measurements between elements for calibration, and report all data to central radar computer. Finally, the distributed and centralized BSC distributions are discussed.
7065_Agrawal_V6.indd 14
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
15
Chapter 7 presents phased array antenna packaging concepts. The predominant packaging considerations associated with the mechanical design of active phased array antennas include a design for ease of maintenance, thermal management, packaging of the DC power distribution system, RF beamformers, radiating aperture design/interface, and structural design. The antenna operating frequency is the overriding driver in what packaging options are available to the designer. As the frequency increases, element spacing decreases, requiring tighter spacing of the supporting electronics. Fortunately for the designer, higher-density arrays tend to have lower T/R module output power requirements; hence, the worst-case thermal design problems typically do not correlate with the worst-case packaging densities. Thermal design is critical for maintaining junction temperatures of the electronic devices at desired levels to support reliability requirements and maintain control of temperature-induced module-to-module phase errors. The T/R modules and DC power supplies account for 70% to 80% of the heat generated within an array. Liquid cooling is typically required because of ever-increasing power density and heat dissipation in modern ground-based and shipborne active phased arrays. The predominant cooling techniques in use today employ conduction away from the T/R module into liquid-filled cold plates or direct liquid flow-through cooling on individual line-replaceable unit (LRUs). In addition, maintenance requirements usually dictate that LRUs be easily removable in the field. Chapter 8 discusses the active phased array antenna for high reliability. Active phased arrays provide graceful degradation of the antenna performance. Since active phased arrays consist of thousands of solid-state components such as T/R modules, failure of a small number of elements results in negligible antenna performance degradation. Since the loss of a small number of elements results in negligible degradation in antenna gain, performance is generally defined in terms of the rise in peak and average receive sidelobes from the nominal level, the level without any component failures. Furthermore, the failed components can be replaced in the field. Therefore, active phased arrays can have high reliability with the judicial choice of the beamformer architecture for components. The antenna MTBF is defined as the period when the antenna performance falls below a certain specified level, and the failed components must be repaired or replaced. Therefore, the antenna MTBF
7065_Agrawal_V6.indd 15
4/6/23 1:42 PM
16
Practical Aspects of Active Phased Array Antenna Development
must be increased to increase the array reliability. For a phased array deployed in the field for a long period, the life cycle cost can be reduced to a minimum if the antenna can be designed so that its MTBF exceeds the deployment time. Furthermore, maximizing the MTBF minimizes the frequency of repairs, resulting in lower lifecycle costs. Chapter 9 presents the array architecture selection to improve radar performance in a high-clutter environment. One way to improve the radar’s ability to detect moving targets against high clutter is to use an active phased array, which has the potential for a higher high clutter improvement factor (CIF) because of its distributed architecture. The pulse-to-pulse amplitude and phase instabilities of the T/R modules tend to be randomized between modules and, thus, combine in a noiselike fashion, while target returns add coherency. We discuss in detail how an active phased array antenna can be designed to achieve high CIF by examining the benefits of the distributed architecture, the placement of components in the beamformer, and the layout of the power distribution network. Chapter 10 summarizes active phased array antenna calibration techniques. A phased array antenna may be deployed for an extended period of time. As a result, the performance of an active phased array antenna may deteriorate over time due to changes in the solid-state components. In addition, failed T/R modules must be repaired or replaced in the field. As a result, the amplitude and phase settings need to be adjusted to compensate for active component drift and replacement or repair of components. This measurement and readjustment of the element amplitude and phase settings during array deployment is known as field calibration or simply calibration. Initial radiating aperture calibration is generally accomplished by setting the amplitudes and phases of all elements equal to reference points in a near-field range. This calibration is usually referred to as array alignment. Next, mutual coupling between all elements is measured in the near-field range, and this data, referred to as the factory data, is stored for comparison for field calibration. Field calibration is accomplished by comparing mutual coupling measurements taken during deployment with the factory standard data. Several techniques are compared, and a new calibration technique is presented in this chapter. This calibration technique uses mutual coupling between a few dedicated internal passive
7065_Agrawal_V6.indd 16
4/6/23 1:42 PM
Practical Aspects of Active Phased Array Antenna Development
17
elements and all other active array elements. The dedicated calibration elements are passive in that they do not have T/R modules behind them. Chapter 11 discusses the digital beamforming arrays. In an all-analog phased array described so far, the received signals from each radiating element are combined using an analog beamformer and digitized using a receiver and A/D converter. If receivers and A/D converters are placed behind elements, the outputs from each A/D converter can be processed by the signal processor. As a result, many simultaneous beams can be formed digitally. In addition, a digital array significantly increases the receiver dynamic range, it can provide adaptive beamforming by placing nulls in the array patterns in the directions of several jammers, and it can provide ultra-low sidelobes and reduce clutter from land and sea. Chapter 12 discusses cost reduction strategies for active phased array antennas. As discussed earlier, the active phased array antenna costs are still too high. The reduction in the acquisition and production costs would require paradigm shifts in design and production methods. The industry has successfully applied six sigma and other techniques to improve quality and reduce cost. It is necessary to go beyond the six-sigma process and change how systems are developed by learning from the commercial industry to achieve our cost goals. In the early 1990s, the Navy demanded that the cost of the SPY-1 radar be reduced by 20% to 30% to continue with a single source. At that time, the SPY-1 antenna had been in full production for many years. It was a tough challenge to meet. We cannot underestimate engineers’ ingenuity when faced with challenging requirements; they can develop innovative methods to meet the challenge. The SPY-1 antenna production cost was reduced significantly close to the requirement. Similarly, the cost of the SPY-3 antenna was reduced considerably. The development, production, and acquisition costs can be reduced by applying the techniques described in this chapter.
1.9 Concluding Remarks Almost all modern military radar systems use active phased array antennas, for they can provide an order of magnitude performance
7065_Agrawal_V6.indd 17
4/6/23 1:42 PM
18
Practical Aspects of Active Phased Array Antenna Development
improvement over their predecessors’ passive phased array antennas. In addition, active phased arrays provide reduced prime power and lifecycle cost, higher reliability, calibration of the antenna in the field, and improved performance in a cluttered environment. This book presents a comprehensive design of an active phased array system, addressing each system’s components design.
References [1] Elliott, R., Antenna Theory and Design, John Wiley and Sons, 1983. [2] Mailloux, R., Phased Array Antenna Handbook, 3rd ed., Artech House, 2018. [3] Stutzman, W., and G. Thiele, Antenna Theory and Design, John Wiley and Sons, 1981. [4] Hansen, R., Phased Array Antennas, John Wiley and Sons, 1998. [5] Fourikis, N., Phased Array-Based Systems and Applications, John Wiley and Sons, 1997. [6] Fourikis, N., Advanced Array Systems, Applications and RF Technologies, Academic Press, 2000. [7] Brown, A., Active Electronically Scanned Arrays: Fundamentals and Applications, John Wiley and Sons, 2022. [8] Sturdivant, R., et al., System Engineering of Phased Arrays, Norwood, MA: Artech House, 2019. [9] Skolnik, M., Introduction to Radar Systems, McGraw-Hill, 1980.
7065_Agrawal_V6.indd 18
4/6/23 1:42 PM
2 Analysis and Design of Linear and Planar Phased Arrays 2.1 Introduction The subject of analysis and design of linear and planar phased array antennas is very broad and has been covered by several textbooks [1–4]. We address the key aspects important for the design of active phased arrays. These aspects include (1) the design of the radiating element aperture, (2) amplitude and phase distribution across the aperture for the shaping of the sum and difference patterns, (3) maximizing element area to achieve a minimum number of elements for a given aperture size while maintaining a grating-free pattern for a specified scan volume, and (4) minimizing the impact of correlated and random errors on the array sidelobes.
2.2 Analysis of Linear Arrays Figure 2.1 shows a linear array of N isotropic radiators equally spaced at a distance of d apart. An incoming plane wave arrives at an angle θ. The isotropic radiators radiate equally in all directions. When the outputs of all radiators are added together, each weighted by its current In, a directional radiation pattern at angle θ is formed. The incoming wave at element 1 arrives before element 0 since the distance is shorter by an amount dsinθ. 19
7065_Agrawal_V6.indd 19
4/6/23 1:42 PM
20
Practical Aspects of Active Phased Array Antenna Development
Figure 2.1 Equally spaced linear array of isotropic point sources.
The phase lead of waves at element 1 relative to those at element 0 is kdsinθ, where k = 2π/λ . Place a phase and amplitude control behind each element, as shown in Figure 2.2, so that the nth element leads the (n − 1)th element phase by α . Where α is a phase shift between two adjacent elements
α = kdsinθ (2.1) When the radiated fields from each element are combined, the far-field pattern is given by E(θ) = Ea(θ) ⋅ AF(θ) (2.2)
Figure 2.2 Linear array with amplitude and phase control at each element.
7065_Agrawal_V6.indd 20
4/6/23 1:42 PM
21
Analysis and Design of Linear and Planar Phased Arrays
where Ea(θ) = element pattern (assumed same for all elements); and AF(θ) = array factor Then the array factor is given by the following equation [2] AF ( q ) =
N −1
∑ ane jn( kd sin q+a) = 1 + a1e jy + …an−1e j(n−1)y
n=0
y = kd sin q + a
(2.3)
Where an are the amplitude coefficients. For a uniformly illuminated array, the coefficients an are set equal to unity, and the array factor is given by AF ( q ) =
∑
e jn( kd sin q+a) =
n=0
(e (e
jNy
AF =
n= N −1
jy
) = sin ( N/2 ) y sin ( 1/2 ) y − 1) −1
n= N −1
∑
n=0
=
e jny (2.4)
sin ( N/2 ) ( kd sin q + a ) (2.5) sin ( 1/2 ) ( kd sin q + a )
The radiation pattern of a 30-element uniformly excited linear array is shown in Figure 2.3. The first sidelobe of the uniformly excited array is at −13dB. For array steering, let θ 0 be the angle for which the array factor is maximum. Then kdsinθ 0 + α = 0
(2.6)
α = −kdsinθ 0 α is the element-to-element phase shift required to produce an array factor maximum in a direction θ 0 AF(θ 0) = N
7065_Agrawal_V6.indd 21
4/6/23 1:42 PM
22
Practical Aspects of Active Phased Array Antenna Development
Therefore, the array factor normalized to its maximum is given by AF ( q ) =
( (
) )
sin ⎡⎣ Np ( d/l ) sin q − sin(q0 ) ⎤⎦ (2.7) N sin ⎡⎣ p ( d/l ) sin q − sin(q0 ) ⎤⎦
The array maximum occurs at
π (d/λ)(sinθ − sin(θ 0)) = π p; p = 0, ±1, ±2, . . .
(2.8)
p = 0 corresponds to the main lobe Sinθ 0 − λ/d and sinθ 0 + λ/d determine the angular locations of extra main beam peaks or grating lobes, as shown in Figure 2.4. Grating lobes appear in visible space when array elements are too far apart, effectively undersampling the aperture, much like aliasing in analog-to-digital signal conversion. To avoid the appearance of grating lobes in the visible region, we must choose element spacing d such that
d/l ≤
1 1 + sin qm (2.9)
Figure 2.3 Radiation pattern of a 30-element uniformly excited linear array.
7065_Agrawal_V6.indd 22
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
23
Figure 2.4 Grating lobe locations when the array is scanned to an angle θ 0.
2.3 Low Sidelobes for Linear Arrays Phased array antennas require low receive sidelobe patterns to avoid unintentional detection of signals in the sidelobes. For example, if sidelobes far away from the main beam peak are reduced by 30 dB relative to uniform, a transmitting jammer would need 30 dB of additional output power to have the same effect, and transmit power is a costly resource. In addition, it is harder to determine whether the signals arriving in the sidelobe region are coming from the direction of the main beam when sidelobes are high. The array receive sidelobes can be reduced by modifying the aperture amplitude distribution for a small increase in cost. In addition, the array excitation can be closely controlled to produce extremely low sidelobes. The penalty for lowering the sidelobe levels is that it does broaden the main beam and reduces its directivity.
2.4 Low Sidelobe Aperture Distributions 2.4.1 Dolph-Chebyshev Aperture Distribution It is desirable to have a narrow main beam and low receive sidelobes in most radar applications. Therefore, it would be helpful to find an aperture distribution that would provide an optimum compromise between the beamwidth and the sidelobe level. For a given sidelobe level, the Dolph-Chebyshev synthesis method yields
7065_Agrawal_V6.indd 23
4/6/23 1:42 PM
24
Practical Aspects of Active Phased Array Antenna Development
arrays with the highest directivity and narrowest beamwidth, or vice versa; for a specified beamwidth, the sidelobes would be as low as possible [5]. The optimum beamwidth-sidelobe performance occurs when there are as many sidelobes in the visible region as possible and at the same level. Chebyshev polynomials have this property; Dolph [5] applied them to determine an array aperture distribution. Figure 2.5 shows a 20-element Dolph-Chebyshev array pattern with elements spaced at 0.5λ . The highest sidelobe for the uniform distribution is about 13 dB below the main beam peak, and the far-out sidelobe amplitudes vary as sinNx/Nx. The uniform distribution provides the highest directivity, narrowest beam, high near-in sidelobes, and low far-out sidelobes. On the other hand, the Dolph-Chebyshev distribution provides all low sidelobes equal in amplitude and the narrowest main beam for a given sidelobe level. Compared to the uniform distribution, far-out sidelobes are higher and pick up undesirable interfering signals and clutter reflections. In addition, the Dolph-Chebyshev amplitude distributions are not monotonic and are challenging to realize. Therefore, Dolph-Chebyshev distributions are seldom used by phased array antennas.
Figure 2.5 Twenty-element Dolph-Chebyshev array pattern d = 0.5λ (after Hansen [3]).
7065_Agrawal_V6.indd 24
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
25
2.4.2 Taylor Distribution for Linear Arrays As an improvement on Dolph-Chebyshev, Taylor [6] proposed a distribution that provides a better compromise between directivity and low sidelobes. The Taylor distribution is a compromise between the uniform and Dolph-Chebyshev distributions. Taylor’s amplitude distribution provides an array pattern with lower near-in sidelobes with equal height (Dolph-Chebyshev) and lower far-out sidelobes similar to the uniform distribution. Figure 2.6 shows a radiation pattern for the Taylor distribution for a continuous line source, with (nbar = 6) and SLL −20 dB. Taylor’s line source can easily be discretized for element spacing ≤ λ/2 with a very close approximation to the line source distribution. The number nbar of near-in equal sidelobes is specified for the Taylor distribution. The near-in sidelobes are not precisely at the design level of −20 dB. The closest pair is slightly below it; the pair out is a bit lower, and so on. So, there is a slight droop to the envelope of the near-in sidelobes. This droop is less if the nbar is selected as a larger number.
Figure 2.6 Radiation characteristics of Taylor sum pattern for a line source (nbar = 6, SLL, and −20 dB) with the main beam at the broadside. The length of the line source is 2a (after Elliott [4]).
7065_Agrawal_V6.indd 25
4/6/23 1:42 PM
26
Practical Aspects of Active Phased Array Antenna Development
The Taylor distribution is widely used in phased array radar antennas. The Taylor distribution can easily be extended to planar and circular apertures. 2.4.3 Bayliss Distribution for Difference Patterns Radar systems often use monopulse techniques to derive angle-tracking information from a single-echo pulse returning from a target [7]. A difference pattern has a null in the direction of the peak of the main sum beam and can be used to accurately determine the arrival angle. Bayliss [8] modified the Taylor synthesis method for a difference distribution. The difference aperture distributions have one-half of the aperture 180° out of phase with the other half. Figure 2.7 shows a Bayliss difference radiation pattern with a −30-dB sidelobe level. 2.4.4 Implementation of Monopulse Beams for an Active Planar Phased Array Antenna Active phased array antennas can use a single beamformer for both transmit and sum and difference receive patterns. Sum and
Figure 2.7 Bayliss difference pattern for a −30-dB sidelobe level main beam at broadside. The length of the line source is 2a (after Elliott [4]).
7065_Agrawal_V6.indd 26
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
27
difference beams are formed simultaneously during monopulse operation. The implementation of the Bayliss distribution for the difference pattern would require a separate beamformer for the elevation and azimuth difference beams, as discussed in Chapter 4. However, in most cases, a single beamformer is used for both sum and difference patterns. The difference patterns are formed by the sum and difference of the four quadrants of the array by providing a 180° phase shift between quadrants by the phase shifters, as shown in Figure 2.8. Using the sum distribution for difference patterns with a single beamformer slightly increases the difference pattern sidelobe levels. The accuracy of the angle of arrival is a function of the null depth in the difference pattern. Asymmetries between the two halves of the array and the amplitude and phase errors across the aperture would affect the null depth. The beamformer networks for generating independent sum and difference patterns using multiple beamformers are presented in Chapter 4.
2.5 Analysis and Synthesis of Planar Arrays The planar array elements can be arranged in a rectangular or triangular lattice. Our goal is to design an array lattice that achieves
Figure 2.8 Sum and difference beams of a planar array.
7065_Agrawal_V6.indd 27
4/6/23 1:42 PM
28
Practical Aspects of Active Phased Array Antenna Development
the minimum number of elements for a given array area and scan requirements. 2.5.1 Rectangular Grid A rectangular array having M elements along the x-axis and N elements along the y-axis is shown in Figure 2.9. Extending the linear array analysis to planar arrays, the array factor for rectangular lattice can be written as [9]
AF ( q, f ) =
∑ ∑ Amn exp ⎡⎣⎢ jk sin q ( mdx cos f + ndy sin f )⎤⎦⎥ (2.10) M N
ky = sinθ sinϕ kx = sinθ cosϕ Δ x = dx/λ; Δ y = dy/λ
(
)
AF k x , k y =
∑ ∑ Amn exp ⎡⎣⎢ j2p ( mΔ x k x + nΔ y k y )⎤⎦⎥ (2.11) M N
Figure 2.9 A planar array with a rectangular element lattice. Each black dot represents one array element.
7065_Agrawal_V6.indd 28
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
29
To scan the array beam, we set Amn different in phase from A00 by the phase factor, so the array factor is given by
(
)
AF k x , k y =
∑ ∑ Amn exp ⎡⎣⎢ j2p ( mΔ x ( k x − k x 0 ) + nΔ y ( k y − k y 0 ))⎤⎦⎥ (2.12) M N
Where kx0 = sinθ 0 cosϕ 0, and ky0 = sinθ 0 sinϕ 0, θ 0, ϕ 0 is the beam scan direction. We can show that AF (kx + 1/Δ x,ky) = AF (kx, ky) (2.13) and AF (kx, ky + 1/Δ y) = AF(kx, ky) (2.14) The array factor is periodic with period 1/Δ x in kx and 1/Δ y in ky. Therefore, the maximum values of AF occur at kx − kx0 = p/Δx; p = 0, ±1, ±2, ±3, . . . ky − kx0 = q/Δx; q = 0, ±1, ±2, ±3, . . .
(2.15)
p = 0 and q = 0 represent the main lobe and p = ±1, ±2, ±3, . . . and q = ±1, ±2, ±3, represent grating lobes. kx = sinθ cosϕ; ky = sinθ sinϕ kx2 + ky2 = sin2θ represents a unit circle (kx + p/Δ x)2 + (ky + q/Δ y)2 = 1
(2.16)
Equation (2.16) defines the boundaries of the grating lobe circles centered at points kx = p/Δ x and ky = q/Δ y (2.17) Scanning past the boundaries causes grating lobes to appear in visible space. Grating lobe circles for a rectangular array element lattice for dx = 0.5λ , dy =0.5λ are shown in Figure 2.10. Notice that the visible region is the entire hemisphere. Grating lobe circles for a rectangular lattice dx = 0.6λ , dy = 0.6λ are shown in Figure 2.11. In this case, the grating lobes enter the visible region, and the grating lobe-free scan area has been reduced, as shown by the darker area. For ±60° grating lobe-free scan, in both azimuth and elevation, the maximum element spacings can be determined from (2.9) as dx = dy = λ/(1 + sin(60°)) = 0.536λ (2.18)
7065_Agrawal_V6.indd 29
4/6/23 1:42 PM
30
Practical Aspects of Active Phased Array Antenna Development
Figure 2.10 Grating lobe boundaries for a rectangular lattice with dx = 0.5λ , dy = 0.5λ .
Figure 2.11 Grating lobe boundaries for a rectangular lattice dx = 0.6λ , dy = 0.6λ .
7065_Agrawal_V6.indd 30
4/6/23 1:42 PM
31
Analysis and Design of Linear and Planar Phased Arrays
The single-element area is given by Ae = dx dy = 0.536λ × 0.536λ = 0.287λ 2 (2.19) The number of elements NR in a given aperture is A having a rectangular lattice that is given by NR =
A (2.20) 0.536l × 0.536l
In the next section, the number of elements will be compared for two arrays of equal area, one with a rectangular array lattice and the other with a triangular array lattice, with the same grating lobe-free scan volume. 2.5.2 Triangular Array Element Grid The array factor for a triangular lattice, shown in Figure 2.12, is obtained by combining two rectangular arrays and is given by [10] ⎧ AF ( q, f ) = ⎨1 + ⎩
(
))
(
⎫ ⎡ j2p Δ k − k + Δ y k y − k y 0 ⎤⎬ x x x0 ⎢⎣ ⎥⎦⎭
(
(
(
)
)
(
))
∑ ∑ Amn exp ⎡⎢⎣ j2p 2mΔ x k x − k x 0 + 2nΔ y k y − k y 0 ⎤⎥⎦ M N
(2.21)
where θ 0, ϕ 0 is the array beam scan direction ky = sinθ sinϕ kx = sinθ cosϕ ky0 = sinθ 0 sinϕ 0 kx0 = sinθ 0 cosϕ 0 The maxima occur at kx − kx0 = p/2Δx, p = 0, ±1, ±2, ±3, . . .
(2.22)
and ky − ky0 = q/2Δ y, q = 0, ±1, ±2, ±3, . . .
(2.23)
p = 0 and q = 0 represent the main lobe.
7065_Agrawal_V6.indd 31
4/6/23 1:42 PM
32
Practical Aspects of Active Phased Array Antenna Development
Figure 2.12 A planar array with a triangular lattice.
(kx0 + p/2Δ x)2 + (ky0 + q/2Δ y)2 = 1
(2.24)
The value associated with p = 0, q = 0, is the main beam direction and all other values of p and q represent the location of grating lobes. Equation (2.24) defines the grating lobe boundaries for a set of unit circles centered at the nominal grating lobe points, as shown in Figure 2.13. The darker area in the main lobe shows the area over which the main beam can be scanned without forming real grating lobes. Grating lobes form when the beam is scanned to the point that is inside the intersection of the primary unit circle and the circle centered at (1/2Δ x, 1/2Δ y), the grating lobe originally at (−1/2Ax, −1/2Δ y) will be translated inside the primary unit circle and become visible. For an equilateral triangular lattice providing a grating lobefree scan over a 60° cone (at all phi angles), the element spacings are given by [11]: dx = 0.536λ , dy = dx/(31/2) = 0.536/(31/2) = 0.309λ (2.25)
7065_Agrawal_V6.indd 32
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
33
Figure 2.13 Grating lobe circles for a triangular lattice.
and the single-element area is given by Ae = 2 dx dy = 2 × 0.536λ × 0.309λ = 0.332λ 2 (2.26) The number of elements for an equilateral triangular lattice NT in a given aperture area A is given by NT = A/Ae = A/0.332λ 2 (2.27) NR = A/Ae = A/0.309λ 2 (2.28) From the above equations, the ratio of the number of elements for rectangular and triangular lattices can be written as NR/NT = 1.16
(2.29)
2.6 Comparison of Rectangular and Triangular Grids The antenna designer’s goal is to choose an array grid that gives the minimum number of elements for a given aperture size since the
7065_Agrawal_V6.indd 33
4/6/23 1:42 PM
34
Practical Aspects of Active Phased Array Antenna Development
array cost is proportional to the number of elements. Equation (2.29) shows that an array aperture with a triangular lattice with equilateral elements has 16% fewer elements than the same array aperture with a square lattice for the same grating lobe-free scan coverage. In addition, a square aperture with a width of D contains 4/π or 27% more elements than a circular aperture with a diameter of D, even though the square and circular apertures have the same beamwidths in azimuth and elevation. However, the square aperture radiates 27% more power than the circular aperture. Since the cost of the array is proportional to the number of elements, it is more economical to increase the circular array’s transmit power per element by 27%, if possible. Therefore, most active phased array antennas use circular apertures. However, an equivalent elliptical aperture would be best suited for the rectangular aperture. We shall now consider optimizing the selection of an array element grid for ground-based and shipborne antennas that are required to scan from 0° to 90° in elevation relative to the surface of the Earth (known as earth coordinates).
2.7 Minimize the Number of Elements for a Grating Lobe-Free Pattern Using a Tilted Array As discussed in the previous section, the triangular element grid significantly reduces the number of elements over the square grid. Ground-based and shipborne antennas are required to scan from 0° to 90° in elevation in earth coordinates. To avoid having to scan 90° in elevation and to optimize the antenna performance, the antenna is tilted backward. Figure 2.14 shows the tilted array geometry. Therefore, the array tilt angle is also a significant factor in determining the optimum grid dimensions. The graphical technique [10] is an iterative method to determine the optimum grid dimensions and the tilt angle so that the grating lobes do not form in the scan volume when the array is scanned to the maximum azimuth and elevation angles. Let us consider an example where the antenna is required to scan ±60° in azimuth and 0° to 90° in elevation. We use the technique described by Corey [10] to determine the antenna grid and tilt angle. In the grating lobe technique [12], the array cannot form grating lobes in the area defined by a circle formed by the three extreme
7065_Agrawal_V6.indd 34
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
35
Figure 2.14 A backward-tilted antenna geometry.
scan angles. In this example, the tilt angle is 27.5°, and the grid dimensions are dx =1/(1 + sinθ m) and dy = dy/(√3), θ m is the extreme beam scan angle. From Figure 2.15, sinθ m = 0.887 is the radius of the grating lobe circle. As shown in Figure 2.15, the grating lobes are tangential to the circle formed by the three extreme angles. Therefore, the single element area is Ae = 2dxdy = 0.325λ 2. The optimum element spacings are obtained when the grating lobe boundaries (GLB) nearly touch the required scan boundary (RSB). Therefore, we can improve the design by changing the antenna tilt angle and grid dimensions by bringing the grating lobes closer to the RSB. For example, in Figure 2.15, the region between the grating lobe circles and the azimuth scan angles where the antenna is not allowed to scan. This unused area results in an inefficient design. In the second example, shown in Figure 2.16, the array geometry is set for the same tilt angle of 27.5° with dx = 0.535, dy =0.323λ . The element area is Ae = 2dxdy = 0.346λ 2. This design results in 6% fewer elements in the array and reduces the cost of the antenna by the same amount.
7065_Agrawal_V6.indd 35
4/6/23 1:42 PM
36
Practical Aspects of Active Phased Array Antenna Development
Figure 2.15 A triangular grid design for a tilt angle of 27.5°, dx = 0.535λ , dy = 0.306λ (Az = ±60°, El 0° to 90°).
Figure 2.16 Triangular grid design for a tilt angle 27.5° Az ±60°, El 0° to 90° dx = 0.535λ , dy = 0.323λ .
7065_Agrawal_V6.indd 36
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
37
To further optimize the design, we need to bring the grating lobes closer to the scan boundary as closely as possible. In the third example, shown in Figure 2.17, the array geometry of tilt angle = 37.5°, and dx = 0.535λ , dy = 0.342λ is chosen. This design results in the element area Ae = 0.366λ 2. This design uses 11% fewer elements than the grating lobe technique design. The optimum design uses 11% fewer elements, maximizing element area for the required scan volume. Corey’s graphical technique makes it possible to use the space inside the GLB efficiently by matching the GLB and RSB as closely as possible. It should be noted that the number of elements within a given aperture does not change the antenna’s beamwidth; however, the total radiated power is proportional to the number of elements. Our goal is to use the minimum number of elements for the given required scan volume and the area of the array aperture. For certain situations, the tilt angle may be fixed; for example, the array tilt angle may be fixed by the structure design for some platforms, such as ships. In that case, the array dimensions are the only variables determining the optimum grid dimensions. For
Figure 2.17 Optimized grid design with tilt angle 37.5° Az = ±60°, El 0° to 90°, dx = 0.535λ , dy = 0.342λ .
7065_Agrawal_V6.indd 37
4/6/23 1:42 PM
38
Practical Aspects of Active Phased Array Antenna Development
shipborne phased arrays, additional elevation coverage is required to compensate for the ship’s roll and pitch. Figure 2.18 shows an optimum grid design for a tilt angle of 15° for −30° to +90° elevation coverage to compensate for the ship’s roll of ±30°. For the ground-based radars that scan the beam near the horizon for targets at large distances, the scan loss can be significant if a large tilt angle is selected. So, there is a compromise between the tilt angle and the beam scan loss. Figure 2.19 shows the ground-based antenna design tilted at 15°. The array grid spacing in Figure 2.19 is slightly larger than the shipboard array.
2.8 Directivity and Gain of Active Arrays Directivity is a measure of an antenna’s ability to concentrate power in a preferred direction. Directivity = radiated power density in direction (θ, ϕ) divided by the power density averaged over all regions D ( q, f ) =
1/4pr 2 ∫
p 0
2p
∫0
P ( q, f ) P ( q ′ , f ′ ) r 2 sin q ′ d q ′ d f ′
(2.30)
Figure 2.18 Array grid design for a shipboard array tilt angle 15°, Az ±50°, El −30° to 90°; dx = 0.5337λ , dy = 0.2935λ .
7065_Agrawal_V6.indd 38
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
39
Figure 2.19 Array grid design for a ground-based array tilt angle 15°, Az ±50°, El 0° to 90°; dx = 0.563λ , dy = 0.313λ .
The maximum directivity of a planar aperture having area A is achieved for uniform amplitude and phase illumination at a wavelength λ as given by D=
4pA cos q0 (2.31) l2
The factor cosθ 0 accounts for the directivity decrease due to the aperture’s projected area in the beam scan direction. For a large array with separable amplitude distributions in x (azimuth) and y (elevation), the directivity is approximately given by [3], D = πDxDy cosθ 0 Max directivity in the main beam in terms of the azimuth and elevation beamwidths is given by
7065_Agrawal_V6.indd 39
Dmax =
(
)
PT / BWAZ ⋅ BWEL 4p = (2.32) BWAZ ⋅ BWEL PT /4p
4/6/23 1:42 PM
40
Practical Aspects of Active Phased Array Antenna Development
PT = Total power radiated in watts Beamwidths are in radians Antenna Peak Gain = Peak Directivity − Antenna Losses The antenna losses for both passive and active arrays include aperture mismatch loss, radome loss, and amplitude taper loss. However, for passive arrays, the beamformer losses in transmit and receive reduce the transmit power and increase the receive NF, respectively. However, since the transmit and receive functions are distributed at the aperture, the beamformer losses do not impact the active array performance.
2.9 Effect of Amplitude and Phase Errors on the Phased Array Antenna Performance Due to the imperfection of the components and signal beamforming networks, random and correlated errors are introduced across the array. Active phased arrays use digitally controlled attenuators and phase shifters. Correlated errors due to the quantization of phase shifters and attenuators can degrade the peak sidelobe performance of the array [13]. In addition, the random errors due to imperfection in the desired amplitude and phase versus the actual impact RMS sidelobes. Correlated errors in the distribution network have a more significant impact on the sidelobe performance if not corrected. Some of the correlated errors in the distribution networks can be corrected by adjusting the settings of the attenuators and phase shifters. Examples of correlated errors are caused by these discrete components’ periodic amplitude and phase errors across the array and frequency-dependent phase errors for wideband operation. The phase errors also slightly degrade the beam-pointing accuracy of the array. Random errors are not deterministic and can be minimized by controlling the manufacturing process. In the following section, we evaluate the impact of these errors on the gain, sidelobe levels, and main beam pointing accuracy. 2.9.1 Quantization Errors Active phased arrays use digital attenuators to set the aperture amplitude distribution and digital phase shifters to steer the beam
7065_Agrawal_V6.indd 40
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
41
in any direction within the antenna’s field of view. The attenuation and phase ranges are divided into the number of bits. The array is required to produce a smooth phase taper; however, an N-bit phase shifter has phase states separated by the least significant bit [13]. f0 =
2p 2N
Peak Phase Error = ±
p 2N
RMS Phase Error s f =
(2.33) 1 p 3 2N
The actual phase shift for an N-bit digital phase shifter is implemented as the staircase approximation of the desired phase shift, as shown in Figure 2.20. The net phase error is shown as a sawtooth along the horizontal axis. Equation (2.33) says that the peak phase error from the desired phase shifter setting is equal to half the size of the least significant phase bit. The first peak lobe due to the phase quantization errors is given by [14]
Figure 2.20 Phase error due to phase quantization. (From [14].)
7065_Agrawal_V6.indd 41
4/6/23 1:42 PM
42
Practical Aspects of Active Phased Array Antenna Development
1 2N (2.34) PGL ( dB ) = −10Log2 2 N = −20NLog2 = ∼ −6N PGL ( dB ) =
For six-, seven-, and eight-bit phase shifters, the first grating lobes would be −36 dB, −42 dB, and 48 dB, respectively. For most phased array antennas, six phase bits are sufficient; a seven- or eight-bit phase can provide ultra-low sidelobes. An eightbit phase shifter was implemented by Komiak and Agrawal [15] to reduce the quantization errors. Small phase shifter bits do not add much real estate to the variable gain amplifier (attenuator) MMIC. The quantization lobes can also be reduced by adding a random phase offset at each element and correcting the added phase in the phase shifter [14]. The insertion phase variation between T/R modules in large arrays can also decorrelate quantized errors in a phased array since the T/R module insertion phase varies due to variations in the manufacturing process. Miller [14] calculated the average sidelobe level due to the triangular phase error shown in Figure 2.20. The mean squared phase error σ ∅ is given by
sf =
1 p (2.35) 3 2N
2.9.2 RMS Sidelobe Level Due to Amplitude and Phase Errors The mean squared sidelobe level (MSSL) from random amplitude and phase errors is given by MSSL =
s2 (2.36) hN e 1 − s 2
(
)
Where σ2 = total RMS error = sum of the RMS amplitude and phase errors In (2.36), the RMS sidelobes are relative to the peak of the main beam. The RMS sidelobes are a function of the number of elements in the array; the RMS sidelobe level relative to the peak of the main beam goes down when the number of elements increases.
7065_Agrawal_V6.indd 42
4/6/23 1:42 PM
Analysis and Design of Linear and Planar Phased Arrays
43
σA = RMS Amplitude Error Volts/Volt; σ ϕ = RMS Phase Error in Radians; σ2 = σA2 + σ ϕ2; σ = Total RMS Error; η = Aperture Efficiency (107 hours. Furthermore, the GaN MMIC power amplifiers used in the T/R modules can provide approximately 100 to 200W output power versus approximately 10 to 30W for GaAs in the L through C microwave radar bands. Active phased arrays are significantly improved over passive arrays as the failure of a few T/R module channels does not degrade the antenna performance significantly. In addition, the higher efficiency of the T/R modules reduces the prime power requirement, which is important for fielded systems that have to generate their own power.
3.4 T/R Module Key Performance Parameters This section discusses the definitions of the key T/R module design parameters and their impact on the antenna and radar performance. Often, a compromise must be made between these parameters to choose their optimum values. 3.4.1 Power-Added Efficiency The power-added efficiency of the T/R module is defined as the ratio of the RF output power minus RF input power to the total DC power input to the T/R module Power added efficiency (PAE) =
Pout − Pin PDC
The total input power to the module is equal to the total power required by all its components PDC = PLNA + PDA + PHPA + PDriver + PPhaser + PAttenuator + PControl + PSwitches + PLosses A goal of T/R module design is to provide the required RF output power with the highest efficiency since the power-added efficiency determines the required input DC power. The difference
7065_Agrawal_V6.indd 61
4/6/23 1:42 PM
62
Practical Aspects of Active Phased Array Antenna Development
between DC input power and RF output power is the amount of heat generated that has to be removed by the cooling subsystem. Besides, the input DC power greatly impacts the size and weight of the power conversion subsystem. Therefore, as DC power increases, so do the weight and cost of the phased array system. Furthermore, since many phased array radars are deployed in remote locations and on ships, the DC power must be generated locally, which is very expensive. In the past, GaAs device technology has been used in most phased array antennas. However, GaN wide bandgap power amplifiers are more efficient than GaAs amplifiers and are used in most of the active arrays. 3.4.2 T/R Module Noise Figure The NF of a linear network is defined as the noise output of a practical receiver/NF of an ideal receiver at standard temperature T0 [5] Fn =
N out Sin /N in ΔN = = 1+ kT0 BnG Sout /N out kT0 BnG
Where, Sin = available input signal power in watts Nin = available input noise power (equal to kT0Bn) in watts Sout = available output signal power in watts Nout = available output noise power in watts k = Boltzmann’s constant = 1.38 × 10–23 J/deg T0 = 290°K (approximately room temperature) Bn = noise bandwidth in Hz G = receiver gain (linear) ΔN = the additional noise in watts introduced by the network In dB, NF = 10LogFn 3.4.3 Noise Figure of a Cascaded Network The T/R module minimizes the active array’s receive NF by being near the radiating element. The basis for this benefit lies in understanding the NF of a cascaded network (see Figure 3.6).
7065_Agrawal_V6.indd 62
4/6/23 1:42 PM
Transmit/Receive Modules
63
Figure 3.6 A cascaded receiving network.
The NF of a two-port cascaded network is given by [5] F = F1 +
F2 − 1 F3 − 1 F −1 + + 4 +… G1 G1G2 G1G2G3
where G1, G2, G3, . . ., Gn and F1, F2, F3, . . ., Fn are the gain and NFs of the stages of the network. The previous equation shows that the gain of the first stage reduces the noise contribution from the second stage, and the gain of the second stage further reduces the noise contribution from the third stage. Therefore, if the LNA gain in the receive chain is high enough, it can significantly reduce the noise contribution from the following stages to the receiver. 3.4.4 T/R Module Noise Temperature The noise introduced by a network may also be expressed as an effective noise temperature, Te, defined as that temperature at the network’s input, which would account for the noise ΔN at the output. Fn = 1 + Te/T0 Te = (Fn − 1)T0 The system noise temperature is defined as the receiver’s effective noise temperature, including the effects of the antenna temperature Ta. For an antenna at temperature Ta, the system noise temperature is given by [5] TS = Ta + T0 = T0 Fs where Fs is the system NF, including the antenna temperature’s effect.
7065_Agrawal_V6.indd 63
4/6/23 1:42 PM
64
Practical Aspects of Active Phased Array Antenna Development
The noise temperature of a cascaded two-port network is given by [5] Ts = T1 +
T T2 + 3 +… G1 G1G2
For radar receivers, the use of a NF is more common. 3.4.5 1-dB Compression Point The amplifier goes into saturation when no further output power increases occur for an input power increase. The output power flattens, meaning that the amplifier becomes saturated at high input signal levels. Its response becomes nonlinear, producing signal distortion, harmonics, and intermodulation products. The 1-dB compression point P1 dB is the input power that causes the gain to decrease by 1 dB from the linear gain. It is the point where the amplifier goes into compression and becomes nonlinear. It is essential to know when the receive compression begins to occur so input levels can be restricted to prevent distortion. For example, the 1 dB decrease in gain may be specified at the input or the output, as shown in Figure 3.7.
Figure 3.7 The output power versus input power of an amplifier.
7065_Agrawal_V6.indd 64
4/6/23 1:42 PM
.
Transmit/Receive Modules
65
3.4.6 Third-Order Intercept Point There are two critical measurements in determining receive amplifier quality, efficiency, and linearity: the third-order intercept (abbreviated TOI or IP3) point and the 1-dB compression (P1 dB) point. These quantities determine the amplifier specifications and performance. No amplifier is perfectly linear. TOI allows one to describe just how linear an amplifier is. The TOI point is measured using two signal inputs into the amplifier at frequencies f1 and f2 within the amplifier’s operational bandwidth. At the output of the amplifier, besides the fundamental signals at f1 and f2, intermodulation products are generated by any nonlinearities. Most intermodulation products can usually be filtered out. However, some fall in the amplifier passband. The most troublesome of these are the third- order products, 2f1 − f2 and 2f2 − f1, as shown in Figure 3.8. The TOI point is defined as the point where the third-order products intercept the linear gain line, as illustrated in Figure 3.9. Power series analysis shows that TOI = P1 dB + 10 dB. An amplifier’s spurious free dynamic range is defined as the input power range at which the input power and TOI rise above the noise floor, as shown in Figure 3.9. The compression dynamic range is the difference between the output power noise floor and the 1-dB compression point. The following equations give the relationship between the TOI point and the third-order intermodulation product [6]
Figure 3.8 Third-order intermodulation products.
7065_Agrawal_V6.indd 65
4/6/23 1:42 PM
66
Practical Aspects of Active Phased Array Antenna Development
Figure 3.9 Third-order intercept point and spurious-free dynamic range.
IM3 = 3(Pi + G) − 2Q3 DR = 2/3(Q3 − G − Ni) Ni = FT − 114 + 10LogBR IM3 = Ni + G Pl = Pi = 1/3(Ni − 2G + 2Q3) where, Pi = the input signal power level in dBm Pl = the input power that will generate a third-order modulation product equal to the noise floor in dBm DR = the two tones spur free dynamic range in dB Ni = the noise level of the receiver at its input in dBm G = the gain in dB BR = receiver bandwidth Q3 = the TOI at the output in dBm IM3 = a third-order intermodulation product in dBc referenced to the output signal
7065_Agrawal_V6.indd 66
4/6/23 1:42 PM
Transmit/Receive Modules
67
3.5 T/R Module Architecture Trade-Offs T/R modules must provide a high TOI and low NF to meet dynamic range requirements and maintain radar performance in the presence of high levels of electromagnetic interference (MIL Standard 461). The selected module architecture must simultaneously satisfy these requirements, some of which may conflict. In particular, the trade-off between receive gain, NF, and TOI will drive most T/R module circuit architecture choices. For example, the high firststage gain desired for NF minimization directly conflicts with the need for a high TOI. Since the T/R module cost is almost 50% of the antenna cost, selecting an architecture that meets all requirements with a minimum number of MMIC chips is imperative. We consider two module architectures [3]. The first architecture uses a common phase shifter and attenuator for both the transmit and receive chains (Figure 3.10(a)). The second architecture uses the common-leg approach and shares major functional groups between transmit and receive (Figure 3.10(b)). For this discussion, the module’s total receive gain is fixed at 30 dB, and the same receiver protection and T/R duplexing are used in each case. Spreadsheet performance budgets are shown in Table 3.1 for the common phase shifter architecture. Table 3.2 shows the performance budgets for the common leg architecture. Finally, the receive performance comparison of the two architectures is shown in Table 3.3. There are apparent differences in NF, TOI performance, T/R module, and antenna complexity for the constant receive gain cases studied here. Each of the architectures studied has advantages and disadvantages, which will be weighted differently for different applications [3]. The common phase shifter module has a NF of 0.3 dB better and an input TOI of 3 dB worse than the common leg module. The common leg module has a 10 dB lower internal module gain than the common phase shifter module. Although the module gain is 30 dB for both architectures, the gain in the common phase shifter architecture at a certain point inside the module has a 10 dB higher gain that could result in higher coupling between transmit and receive chains. This is an important feature, as the high internal gain of the common phase shifter module can cause oscillations (instability). Therefore, this module will need more care to provide the necessary internal isolation. For the common leg module, all functions of the
7065_Agrawal_V6.indd 67
4/6/23 1:42 PM
68
Practical Aspects of Active Phased Array Antenna Development
Figure 3.10 T/R module RF architecture: (a) shared phase shifter; (b) common leg circuit. Table 3.1 Common Phaser/Attenuator T/R Module Receive Performance
Component
Signal Gain (dB)
Output TOI (dBm)
Noise Figure (dB)
Cumulative Signal Gain (dB)
Cumulative NF (dB)
Cumulative Input TOI (dBm)
Two Tone IMD (dBc)
Circulator
−0.50
100.00
0.50
−0.50
0.50
100.50
−271.00
Limiter
−0.50
100.00
0.50
−1.00
1.00
97.73
−265.47
LNA
25.00
25.00
1.50
24.00
2.50
1.00
−72.00
8.00
35.00
4.50
32.00
2.52
−1.12
−67.75
Gain block
9.50
40.00
8.00
41.50
2.53
−4.33
−61.35
Switch
−1.50
100.00
−1.50
40.0
2.53
−4.33
−61.35
Phaser
−10.00
100.00
10.00
30.00
2.53
−4.33
−61.35
VGA
7065_Agrawal_V6.indd 68
4/6/23 1:42 PM
69
Transmit/Receive Modules Table 3.2 Common Leg T/R Module Receive Performance
Signal Gain Component (dB)
Cumulative Signal Gain (dB)
Noise Output TOI Figure (dBm) (dB)
Cumulative Noise Cumulative Two Tone Figure Input TOI IMD (dB) (dBm) (dBc)
Circulator
−0.50
100.00
0.50
−0.50
0.50
100.50
−271.00
Limiter
−0.50
100.00
0.50
−1.00
1.00
97.73
−265.47
LNA
20.00
20.00
1.50
19.00
2.50
1.00
−72.00
Switch
−1.50
100.00
1.50
17.50
2.51
1.00
−72.00
Gain block Phaser Gain block VGA Switch
8.00
30.00
4.50
25.50
2.59
−0.60
−68.79
−10.00
100.00
10.00
15.50
2.65
−0.60
−68.79
8.00
35.00
4.50
23.50
2.77
−0.86
−68.27
8.00
40.00
9.00
31.50
2.84
−1.34
−67.32
−1.50
100.00
1.50
30.00
2.84
−1.34
−67.32
Table 3.3 Receive Performance Comparison of Two Architectures Common Phase Shifter Gain (dB) Noise figure (dB)
Common Leg
30.00
30.00
2.53
2.84
Maximum internal gain (dB)
41.50
31.50
Input TOI (dBm)
−4.33
−1.34
−61.35
−67.32
IMD (dBc)
common leg circuit can be integrated into a single chip, resulting in a lower module cost.
3.6 T/R Module Architectures for Circular Polarization Many radar systems use linear polarized radiators, typically vertical. However, returning signals from complex targets may contain horizontal and vertical polarized electromagnetic field components. Therefore, radar systems sometimes use circular polarization to increase target signatures and discrimination [5]. Weather radars
7065_Agrawal_V6.indd 69
4/6/23 1:42 PM
70
Practical Aspects of Active Phased Array Antenna Development
also use circular polarization to improve performance from back scatterers and clutter from precipitation. Circular polarization has a 90° phase shift between the vertical and horizontal electric-field vectors, and their magnitudes are equal. An RF block diagram of a T/R module capable of transmitting left-hand circular polarization and receiving right-hand circular polarization is shown in Figure 3.11. The outputs of the two power amplifiers are connected to the horizontal and vertical inputs of a circularly polarized element. In both transmit and receive, a quadrature hybrid creates the 90° phase shift between the horizontal and vertical radiating elements required for circular polarization.
3.7 T/R Module Construction Figure 3.5 shows a photograph of the MMIC part of an S/C-band T/R module. This T/R module was chosen to describe the construction of a typical T/R module. The basic multichip module (MCM) construction techniques are common to most T/R modules; however, there will be some differences in the process used by different companies. The T/R module package in Figure 3.12 is a low RF parasitic, hermetic structure composed of a metal base, ceramic mother substrate, ceramic wall, and lid [7]. The base/ground is molybdenum, providing an excellent thermal interface and a
Figure 3.11 T/R module architecture for circular polarization.
7065_Agrawal_V6.indd 70
4/6/23 1:42 PM
Transmit/Receive Modules
71
thermal expansion coefficient compatible with the alumina substrates and GaAs MMIC chips. The substrate incorporates all RF circuitry on the top surface with a thin film, printed TaN resistors for voltage conditioning/50 Ohm terminations, thick film multilayer DC, and control distribution on the backside. An entire thick film plane has been devoted to HPA voltage distribution. Backside thick film ground planes minimize coupling and provide the reference ground plane for the frontside microstrip circuitry. Laser-drilled holes filled with thick film gold connect to the top surface, and cutouts are provided for the MMIC chips. Spacers or shims underneath the MMICs are utilized to maintain planarity between their upper surfaces and the substrate. Before attaching the MMICs, the substrate/frame combination is soldered to the metal base.
3.8 Thermal Stack-Up of the T/R Module Figure 3.13 shows an example of the thermal stack-up to the MMIC device junction from the base plate for the T/R module in Figure 3.5. The materials used in the stack-up selection depend on the MMIC’s allowable maximum junction temperature. Since the
Figure 3.12 T/R module packaging cross-section [7].
7065_Agrawal_V6.indd 71
4/6/23 1:42 PM
72
Practical Aspects of Active Phased Array Antenna Development
power amplifier generates the most heat, we consider the power amplifier for this illustration. The power amplifier’s output power and efficiency determine the cooling required. The power amplifier is mounted on a molybdenum (Moly) spacer using a Gold/Tin (Au/Sn). Most of the temperature change often occurs in the MMIC substrate because of its poor thermal conductivity. Power amplifier chips require an eutectic attachment with a minimum of air gaps (voids) because the temperature drop in an air gap can be substantial. The Moly spacer is attached to the base plate using thermal epoxy, and the base plate is attached to the cold plate with a thermal interface material such as GRAFOIL®. For repair and replacement, the T/R module must be easy to remove from the cold plate. The fluid running through the cold plate cools the base plate. The fluid temperature and the flow rate are determined based on the amount of heat that must be removed from the power amplifier so as not to exceed its maximum allowable junction temperature during normal operation. A liquid-to-liquid heat exchanger removes the heat away from the antenna cooling system. Stack-up material coefficient of thermal expansion (CTE) should be as similar as possible to avoid stress on the interface. The conductivity of the HPA’s Moly spacer should be as high as possible. Other spacer candidates include copper moly synthetic diamond, and silicon carbide (SiC). Silicon carbide has excellent thermal properties and is used for GaN power amplifiers. GaN devices are grown on SiC substrates to address their high heat dissipation by depositing the GaN epitaxial layer directly onto a SiC base substrate.
Figure 3.13 Thermal stack-up for the high-power amplifier MMIC in a T/R module.
7065_Agrawal_V6.indd 72
4/6/23 1:42 PM
Transmit/Receive Modules
73
Unlike MMIC amplifiers, externally matched power amplifiers can be made with thinner substrates to reduce the temperature rise through the chip. However, the TCE of the materials should be the same or as close as possible to avoid damage by stress. High or low-temperature cofired substrates are an alternative for alumina because of their lower cost and multilayer construction. Low-temperature cofired ceramic (LTCC) substrates have lower RF losses, but high-temperature co-fired ceramic (HTCC) costs less. In addition, their multilayer construction accommodates higher- density circuitry, and components such as resistors and capacitors can be embedded in the substrate.
3.9 Integration of MMIC, Control Module, and DC-to-DC Converters It is very expensive to provide a T/R module for each radiating element in the array partly because the T/R module package, as discussed in Chapter 5, is very complex and challenging to build. Therefore, the package size can be reduced by moving the control electronics to a separate control module that drives multiple T/R MMIC modules. A cost saving is achieved by having fewer multichannel control modules. Similarly, a DC-to-DC converter can also feed multiple T/R modules. Figures 3.14, 3.15, and 3.16 show LRUs for one control module for one, two, and four T/R modules. The optimal size of a line replacement unit (LRU) is one containing one control module and four T/R modules. The optimization of the number of T/R modules per control module will be discussed in the design of active phased arrays for high reliability in Chapter 8.
Figure 3.14 An LRU containing one control module for each T/R module.
7065_Agrawal_V6.indd 73
4/6/23 1:42 PM
74
Practical Aspects of Active Phased Array Antenna Development
Figure 3.15 An LRU containing one control module for two T/R modules.
Figure 3.16 An LRU containing one control module for four T/R modules.
3.10 T/R Module Stability T/R modules are required to be stable under all operating conditions. Stability is defined as module operation without oscillations in the operating and out-of-frequency bands. However, the coupling between T/R circuits often leads to oscillations, which can be exacerbated by resonant modes in the enclosed module cavity [8, 9]. Therefore, a metal wall separates the T/R module’s low- and high-power sections to avoid coupling between the T/R circuits and reduce the size of internal cavities to push resonant modes as high in frequency as possible, as shown in Figure 3.5. In
7065_Agrawal_V6.indd 74
4/6/23 1:42 PM
Transmit/Receive Modules
75
addition, the external coupling between the T/R module’s output and input ports can also result in oscillations as the small-signal gain of the transmit channel is much higher than the high-power gain. Therefore, the array packaging is designed to provide high isolation (greater than 50 dB) between the input and output ports of the T/R modules.
3.11 T/R Module Reliability T/R module reliability is defined in terms of its MTBF. Designing a T/R module for the highest MTBF possible is highly desirable, for the antenna system’s availability and lifecycle cost are inversely proportional to the MTBF [10, 11]. The reliability of active phased arrays is discussed in detail in Chapter 8. The MTBF of a high-power T/R module often depends on the MTBF of the HPA, which depends on its maximum operating junction temperature. The junction temperature also depends on the efficiency of the thermal stack-up and cooling system. The power amplifier generates most of the T/R module’s heat, concentrated in a very small area of the semiconductor device, resulting in very high heat density. We generally want to connect T/R modules directly to the radiating elements with as little interconnecting loss as possible. So, for brick-shape T/R modules, which insert perpendicular to the array radiating face, the array element lattice spacings usually determine the T/R module cross-sectional dimensions. As frequency increases, the available volume for the T/R module decreases; however, the power generation capability of semiconductor amplifiers also decreases with frequency. The array thermal management subsystem must remove the dissipated heat, often with liquid cooling. The module’s control section is generally less reliable than the MMIC section, as it contains high-power HEXFET switches to turn amplifiers on and off, regulators, and memory storage devices. We show in Chapter 8 that the active array MTBF often is dominated by the least reliable components, such as the DC-to-DC converters. The failed components are replaced when the antenna performance falls below the specified critical level. Therefore, the antenna MTBF determines the frequency of maintenance. However, the
7065_Agrawal_V6.indd 75
4/6/23 1:42 PM
76
Practical Aspects of Active Phased Array Antenna Development
maintenance cost is equal to the replacement of all failed components during the life of the antenna. The other factors affecting T/R modules failure include the environmental operating conditions such as humidity level, vibration, shock, presence of hot gases, operating and storage temperatures, manufacturing process defect level, and thermal interface quality; unfortunately, the impact of many of these factors on the T/R modules cannot easily be quantified.
3.12 T/R Module Cost As stated earlier, the T/R module is the heart of the active array system and is also the most significant cost contributor to the system cost. Here we discuss the cost of the MMIC and control module sections of the T/R module. Because of the large number of T/R modules in an active array, module production costs are critical to active array affordability. Module production costs can vary depending on performance, design complexity, production quantities, and other factors. Although the module cost breakdown can vary depending on the application, the typical cost breakdown of the T/R module, shown in Table 3.4, is representative of the X-band T/R module. The T/R module cost includes the MMICs, packaging, other components, assembly, and test. The MMICs often are the most significant cost element. Of the MMICs, the HPA typically is the most expensive, so the MMIC cost will increase with higher module transmit power. During the last several decades, the cost of MMICs has been reduced significantly; however, the cost of other parts of the Table 3.4 Typical T/R Module Cost Breakdown
7065_Agrawal_V6.indd 76
Component
Fraction of T/R Module Cost
MMICs
45%
Package/substrates
25%
Digital/analog circuitry
15%
Assembly
10%
Test
5%
4/6/23 1:42 PM
Transmit/Receive Modules
77
module, such as packaging, assembly, and test, has not come down in a similar proportion. Improvements in automation have reduced assembly costs to some degree. The system’s lifecycle cost consists of the initial production cost and the cost of maintaining the system through its lifetime, including the cost of replacement and repair of the modules. Hence the MTBF of the T/R modules impacts the total lifecycle cost. MMICs can have more than a million hours of MTBF, but T/R module MTBF includes contributions from other components and is affected by assembly processes. For example, the MTBF depends on the MMIC junction temperatures, cooling methods, hermetic package sealing, stress on the module packaging caused by the environment, and any defects in the manufacturing processes. A high T/R module MTBF is required to minimize maintenance time, maximize system availability, and keep the active array lifecycle cost affordable. The MTBF is also affected by the components in the control section. High reliability would also be desirable for power supplies. Since a power supply provides power to several T/R modules, its MTBF significantly impacts the system’s MTBF [11]. Antenna architectures for achieving high reliability are discussed in Chapter 8. The T/R modules and the power supplies must be easily removable from the line removable unit (LRU) assembly so they can be repaired or replaced separately since the MTBF of the power supplies is lower than the T/R modules. The cost of a military radar system is highly dependent on the associated production volume of the T/R modules. The cost of modules could be reduced by an order of magnitude if we could have volumes similar to commercial products, which exceed tens or even hundreds of millions. However, it is unlikely that we could have commercial factories produce military systems as the requirements for military systems are very different than the commercial systems. In particular, the packaging for most commercial electronics components has lower reliability as those components are not required to last for decades. The effort to design an MMIC to meet its requirements can vary in difficulty and development cost. However, once the MMIC design is complete, MMIC cost is usually expressed in dollars per square millimeter ($/mm2) of wafer area, which depends on the semiconductor material, the wafer processing cost, the wafer diameter, MMIC area, and MMIC yield to the test requirements. The
7065_Agrawal_V6.indd 77
4/6/23 1:42 PM
78
Practical Aspects of Active Phased Array Antenna Development
wafer processing cost for MMICs is highly dependent on the total number of wafers produced by a foundry. Higher-diameter wafers reduce the cost of MMICs as the cost of processing wafers changes a little with their size. Larger wafers contain more MMICs. Unfortunately, most wafer processing foundries providing MMICs for military applications are often underutilized, resulting in higher overhead costs. Higher volumes can be achieved if the foundries providing MMICs for military applications can produce a mix of low-volume MMICs foundries for military and high-volume commercial MMICs. The antenna’s lifecycle cost depends on the number of T/R modules and other components replaced during the antenna’s life. The lifecycle cost of the antenna includes the initial manufacturing and maintenance costs, including labor, to keep the antenna’s required performance. A detailed discussion of the lifecycle costs is provided in Chapter 12.
3.13 Performance Requirements of T/R Modules We provide a detailed list of typical T/R module requirements in the appendix. Since arrays tend to average T/R module performance, T/R module requirements usually include maximum, minimum, and lot average values to maximize production yields and minimize costs. The output power of the T/R modules varies in production from module to module due to random variations in the power amplifier MMIC’s output power and manufacturing process variations. For example, the output power of an HPA should be specified as maximum, minimum, and average over the frequency band. The total power radiated by an active array is the total combined power of the T/R modules. When T/R modules are manufactured in large quantities, their output power may vary by as much as 1 dB. If we only specify the minimum power, many T/R modules would be rejected, reducing the yield and increasing costs. A DC-to-DC converter provides the voltage and currents to the T/R modules. If the power amplifiers outputting near the maximum of the requirement are grouped in one lot, DC-to-DC converters capable of generating higher power would be required, resulting in a higher system cost. We need to meet the total power radiated by an active phased array
7065_Agrawal_V6.indd 78
4/6/23 1:42 PM
Transmit/Receive Modules
79
antenna, and variations from module to module have a small impact on the antenna performance. To achieve a high yield in production, lot averages should be defined, which permit the largest possible module-to-module variation. For an active phased array that contains thousands of T/R modules, using the entire array as the lot size would be impractical. A smaller lot size should be selected. For example, a lot size of 100 may be practical for a 5,000-element array. On the other hand, a smaller lot size would be preferable for an array containing fewer T/R modules. Similarly, if we specify only a minimum gain for the LNA, LNAs with excess gain would pass the gain requirement but might not be able to meet 1 dB compression or TOI requirements. Besides, higher gain LNAs are more likely to become unstable. Therefore, the LNA maximum gain must also be specified.
3.14 Application of Silicon Germanium (SiGe) BiCMOS Technology in T/R Modules1 Starting about 2000, two semiconductor technologies began to revolutionize phased array RF electronics design, in particular, T/R modules. Gallium nitride (GaN) made possible significantly higher output power, fractional bandwidth, and millimeter-wave operation, and silicon germanium (SiGe), through at least an order of magnitude higher level of functional integration, dramatically lowered array electronics costs and made feasible array design up to past W-band. Conventional T/R modules like the one in Figure 3.5 are expensive, costing about $1,000 each [12]. They are built from many discrete components, typically one MMIC per function, dedicated control components, and regulator ICs. These ICs are relatively large, given they are single-layer devices, with circuitry on the top side only. While the GaAs common leg circuit described in this chapter, which integrated a phase shifter, gain stage, and attenuator, along with a GaAs control circuit onto a single MMIC, was an advancement in integration, it did not significantly reduce the cost of T/R modules, which still contained only one or two full T/R channels. 1
7065_Agrawal_V6.indd 79
This section was contributed by Dr. Eric Holzman.
4/6/23 1:42 PM
80
Practical Aspects of Active Phased Array Antenna Development
In contrast to GaAs, GaN, and InP, SiGe BiCMOS (known as SiGe) is a unique multilayer technology introduced by IBM in 1989 [13]. Imagine shrinking a multilayer printed circuit board in size by two orders of magnitude by a SiGe radio frequency integrated circuit (RFIC) [14]. With multiple isolated layers of metal and dielectric (oxide) and a metalized via process, SiGe can accommodate RF, digital, and regulation functions within a single RFIC. Other benefits of SiGe include • High frequency transistors that allow operation up beyond 100 GHz; • Seven to nine metal layers for routing of RF, digital, and power; • Inductors, MIM capacitors, resistors; • Vector modulators that replace phase shifters and variable attenuators and achieve comparable RMS error levels (5°, 0.5 dB); • The high density of functions that allows 16 or more T/R channels, including all RF, control, and regulation, even built-in test (BIT) on a single RFIC; • Multilayer construction that allows 40+ dB isolation between adjacent RF channels and 60 dB isolation between RF output and input, as good as conventional, chip and wire T/R modules in metal boxes with isolation cavities; • Extremely high yields and very low channel-to-channel gain and phase differences. SiGe has one significant limitation: it is not a high voltage or high RF power process. Bias voltages typically range between 3 and 4V, and RF amplifiers are limited to a maximum of about 1W output. How can SiGe benefit phased array radar? In general, significant cost savings can be realized if a single SiGe RFIC can replace many MMICs in T/R modules. For very high-power arrays, transmitting more than 1W per element, one can use GaN MMICs for the HPAs and even the LNAs, and SiGe for all other functions (phase shift, attenuation, gain, RF switching, beam steering control, regulation). If the array can be made large enough or the transmit power requirements are low to moderate, then an all-SiGe array is possible [15–17].
7065_Agrawal_V6.indd 80
4/6/23 1:42 PM
Transmit/Receive Modules
81
3.15 Concluding Remarks This chapter provided an overview of T/R module technology, including a control circuit, RF architectures, packaging, and cost. T/R modules are considered the heart of an active array, for they determine all the performance parameters of an active array antenna, such as power radiated by the radiating elements, beam switching, pulse width, sidelobe levels, and phase settings for beam scanning. A common leg circuit provides the integration of attenuators, phase shifters, and other electronics into a single MMIC. GaN power amplifier technology has increased the output power by almost an order of magnitude over the GaAs amplifiers. T/R module performance parameters such as NF, TOI point, 1-dB compression point, and third-order intermodulation products were defined. T/R modules dissipate a large amount of heat that must be removed efficiently to maintain the device junction temperatures below a certain level to provide high reliability and lower cost. T/R module costs have also come down to levels that justify using solid-state technologies. However, there is a continuous desire to reduce the cost as it is still considered expensive. Finally, the application of emerging SiGe technology in T/R modules can increase the integration level and also reduce cost was introduced.
References [1] Komiak, J., and A. K. Agrawal, “Design and Performance of Octave S/C-Band T/R Modules for Multifunction Phased Arrays,” IEEE Trans. on Microwave Theory and Techniques, Vol. 39, December 1991. [2] Deluca, A., Gentry J., Thomas, D., Landry, N., and A. K. Agrawal, “Phased Array Antenna with Distributed Beam Steering,” U.S. Patent 5,339,086, issued August 16, 1994. [3] Agrawal, A. K., Clark, R., and J. Komiak, “T/R Module Architecture Tradeoffs for Active Phased Array Antennas,” IEEE MTT-S Int. Microwave Symp., San Francisco, CA, June 17–21, 1996. [4] Agrawal, A. K., and N. Landry, “Independent Control of Sum and Difference Patterns in Active Phased Array Antennas,” IEEE AP-S Int. Symp., June 1989. [5] Skolnik, M., Introduction to Radar Systems, McGraw-Hill, 1980. [6] Tsu, J. Bao-Yen, Microwave Receivers with Electronic Warfare Applications, John Wiley and Sons, 1986.
7065_Agrawal_V6.indd 81
4/6/23 1:42 PM
82
Practical Aspects of Active Phased Array Antenna Development [7] Agrawal, A. K., Clark, R., Komiak, J. J., and R. Browne, “Multilayer Interconnection and Packaging Using Multilayer Thin Film/Thick Film Technology,” IEEE MTT-S Int. Microwave Symp., Albuquerque, NM, June 1992. [8] Clark, R. Agrawal, A. K., and S. Miller, “Simulation of Multi-Chip Module Package Resonance Using Commercial Finite Element Electromagnetic Software,” IEEE MTT-S Int. Microwave Symp, Orlando, FL, June 15–19, 1996. [9] Holzman, E., Essentials of RF and Microwave Grounding Essentials, Norwood, MA: Artech House, 2006, pp. 149–154. [10] Agrawal, A. K., and E. Holzman, “Active Phased Array Design for High Reliability,” IEEE Int. Symp. on Phased Array Technology, Boston, MA, October 15–18, 1996. [11] Agrawal, A. K., and E. Holzman, “Active Phased Array Design for High Reliability,” IEEE Trans. on Aerospace and Electronic Systems, Vol. 5, No. 4, October 1999. [12] “Transmit/Receive Modules,” Microwaves101.com, https://www.microwaves101 .com/encyclopedias/transmit-receive-modules#:~:text=The%20classic%20 T%2FR%20module,the%20price%20down%20to%20%24100. [13] “Silicon Germanium Chips,” IBM’s 100 Icons of Progress, https://www.ibm.com /ibm/history/ibm100/us/en/icons/siliconchip/. [14] Rebeiz, G. M., “Advances in SiGe BiCMOS Technology with Chip-Scale Phased Array Applications,” IEEE Phased Array Symposium Short Course, October 2013. [15] “Silicon Germanium Chips,” IBM’s 100 Icons of Progress, https://www.ibm .com/ibm/history/ibm100/us/en/icons/siliconchip/. [16] Stander, T., “A Comparison of Basic 94 GHz Planar Transmission Line Resonators in Commercial BiCMOS Back-End-of-Line Processes,” International Conference on Actual Problems of Electron Devices Engineering (APEDE), 2014. [17] Rebeiz, G. M., “Millimeter-Wave SiGe RFICs for Large-Scale Phased-Arrays,” IEEE Bipolar/BiCMOS Circuits and Technology Meeting (BCTM), 2014.
7065_Agrawal_V6.indd 82
4/6/23 1:42 PM
4 Beamformer Architectures for Active Phased Array Antennas 4.1 Introduction Beamformers are passive power-combining networks that split signals on transmit, combine them on receive, and form amplitude distributions to shape the array beams. In combination with T/R modules, they perform beam steering. Radar systems often use monopulse techniques to derive angle tracking information from a single-echo received pulse, in which a separate receive beamformer forms each of the three monopulse beams, that is, sum and elevation difference (EL DELTA), and azimuth difference (AZ DELTA) [1]. Corporate beamformers are generally preferred in phased arrays for their wide instantaneous bandwidth, and good impedance match on both transmit and receive chains. The beamformers can be constructed either in a waveguide, coaxial cable or stripline for phased arrays. High transmit power passive phased array antennas use waveguide beamformers in the transmit mode and stripline/ microstrip line in the low power receive mode. In addition to splitting loss, manifold insertion loss from reflections and dissipation may range from 2 to 3 dB for an all-waveguide beamformer in the transmit mode and be as high as 8 dB in the receive mode for a stripline beamformer for large arrays. In active phased array antennas, transmit and receive functions are distributed using T/R 83
7065_Agrawal_V6.indd 83
4/6/23 1:42 PM
84
Practical Aspects of Active Phased Array Antenna Development
modules at the antenna aperture. The use of T/R modules near the radiating elements significantly reduces the sensitivity of transmit power and system NF to beamformer losses, thus allowing flexibility in the choice of array architectures. In this chapter, we only discuss corporate-fed active phased-array beamformers for radar applications. We also describe beamformer architectures for narrow and wide bandwidth arrays, including the choice of applying an amplitude taper in the T/R module or beamformer networks. As discussed in Chapter 1, active phased arrays use T/R modules to distribute transmit RF power at the aperture, eliminating the central high-power transmitter needed in passive phased-array antennas. Since the RF power and low-noise amplification in an active phased array are distributed close to the radiating aperture, the impact of losses in the transmit and receive beamforming networks are significantly reduced, since they occur before the transmit HPA and after the receive LNA. As discussed in Chapter 3, sufficient transmit gain in the T/R module allows the antenna to have very low RF input power since the T/R module generates the high transmit power that is radiated. In addition, a high-gain LNA reduces the noise contribution from secondary stages of the T/R modules and beamformers in the receive mode. By reducing the impact of transmit and receive beamformer losses, large active phased array antennas provide increased radar sensitivity in the range of 7 to 10 dB compared to passive arrays. A general discussion of beamforming feeds for various types of radiating elements is provided in [2], and an overview of beamformer architectures is provided in [3, 4]. First, we shall focus on beamformer architectures for active phased array antennas. High-performance radars use pulse-doppler detection modes to separate small signals from large levels of clutter received [1]. The system dynamic range from the T/R module through the analog-to-digital converter (ADC) must be wide enough to handle both the clutter return and the small target signal. Therefore, the receive chain architecture needs to consider the entire receiver chain. The gains of various components in the receive chain must be set to meet the system requirements. The transmit chain architecture is also affected by the system’s requirement to radiate a transmit beam with a sidelobe structure different from the receive beam sidelobes.
7065_Agrawal_V6.indd 84
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
85
High-power amplifiers usually operate compressed to maximize their efficiency and output power. Low sidelobes are not critical for the transmit beam for many applications, so a uniform distribution is chosen for maximum gain. Because low sidelobes are usually important in receive to mitigate jammers, separate transmit and receive beamformers may be required. The receive amplitude taper can be generated either by the variable attenuators in the T/R modules or in the receive beamformer [4]. Section 4.4 discusses the associated trade in system NF. We first review beamformer architectures for passive phased-array antennas used in radar applications. In passive arrays, the transmit power from a single transmitter is distributed to the array elements through a passive beamformer or splitting network. Separate transmit and receive beamformers are used for the transmit and receive beams. The receive beamforming network provides the amplitude taper for low receive sidelobes. The beamforming networks for passive phased arrays form the basis for the discussion of beamformer architectures for active phased- array antennas. As we discussed in Chapter 3, the receive beamformer architecture can be significantly simplified if the receive amplitude taper is formed in the T/R modules with a variable gain amplifier or attenuator (VGA). Section 4.3 presents a receive beamformer architecture with separate amplitude and phase control for each monopulse beam. Because amplitude and phase errors in the beamformer network are not critical, this architecture results in a simple, lowcost beamformer network. These errors can be corrected in the T/R module for the sum beam. However, to calibrate difference beams, separate beamformers are required for each monopulse beam. Section 4.4 presents a NF, input TOI, and dynamic range trade between two receive beamformer architectures. The first beamformer architecture applies the receive amplitude taper in the T/R modules, and the second beamformer architecture applies it in the passive beamformer. Since the transmit and receive amplitude tapers are different, the second architecture requires separate beamformers for the receive sum and transmit beams. On the other hand, a common transmit and sum receive beamformer can be used for the first architecture.
7065_Agrawal_V6.indd 85
4/6/23 1:42 PM
86
Practical Aspects of Active Phased Array Antenna Development
Beamforming networks for wide instantaneous bandwidth active phased array antennas are presented in Section 4.6. The instantaneous bandwidth of an array can be increased by dividing the aperture into subarrays and applying time-delay steering at each subarray.
4.2 Beamformer Networks for Passive Phased Array Antennas Radar systems often use monopulse techniques to derive angle-tracking information from a single transmitted pulse by generating two or more simultaneously formed receive beams [1] from the target return. High-performance phased array monopulse antennas in current use have several thousand radiating elements. Therefore, a receive monopulse beamformer for such an antenna array must include several thousand inputs and may require different amplitude weighting for signals applied to each input. The beamformer includes amplitude weighting to control the sidelobe levels in the receive beam patterns. A simplified block diagram of a typical receive beamformer for a monopulse beamformer for a linear array is shown in Figure 4.1.
Figure 4.1 Receive beamformer for a linear passive phased array antenna forms simultaneous sum and difference beams.
7065_Agrawal_V6.indd 86
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
87
In this example, the received signals from symmetrically located pairs of radiating elements are combined separately in 180° hybrid circuits to form in-phase (sum) and out-of-phase (difference) beams. The in-phase signals are combined and weighted to form a sum beam, and the out-of-phase signals are combined and weighted to form the difference beam. Figure 4.2 shows a simplified block diagram of a receive beamformer for a two-dimensional (planar) monopulse phased array radar antenna. Signals from the radiating elements in each column equidistant from the array centerline are fed to the column beamformers of sum and difference hybrids, where they are simultaneously added and subtracted. The sum and difference outputs are separately weighted and combined in column combiners to form one elevation sum and one elevation difference (Elevation Delta) signal per column of radiators. Next, the column sum and difference signals are combined symmetrically about the array centerline in the azimuth beamformers to form the three monopulse beams and
Figure 4.2 Receive beamformer for a two-dimensional (planar) passive phased array antenna forms up to four simultaneous receive beam outputs (sum, azimuth delta, azimuth elevation, and delta-delta).
7065_Agrawal_V6.indd 87
4/6/23 1:42 PM
88
Practical Aspects of Active Phased Array Antenna Development
one difference-difference (delta-delta) beam. A Taylor weighting is applied in the sum beamformers to obtain low sidelobes. In two dimensions, separable Taylor weightings are applied in the sum beamformer such that their multiplication results in the desired two-dimensional weighting function. Similarly, a Bayliss weighting is applied in each of the delta beamformers to provide low sidelobes in the difference beams. A passive array beamformer requires a large number of unique coupler designs to form the necessary array-level amplitude tapers. The overall beamforming network is likely to be very complex, and it requires careful design using high-performance numerical electromagnetic design tools to capture all coupler interactions to perform a thorough manufacturing tolerance analysis. Even so, the performing beamformers may require modification after testing to account for differences between the modeled and manufactured networks. Any amplitude and phase errors in the array tend to raise the sidelobe levels of the monopulse beams. The lower those sidelobes are required to be, the more sensitive their levels are to such errors. The design, manufacturing, and testing technology primarily set the random error contribution to the sidelobe level. To obtain low peak and RMS sidelobes, phase and amplitude errors must be reduced, and correlated or common errors among groups of radiating elements must be minimized. For a passive array, correlated phase errors can be minimized by measuring the phase errors in the beamformer network at each radiating element and using the electronic phase shifters that steer the array beam also to compensate for those errors. After correction, the residual phase error is a function of instantaneous bandwidth (since errors can only be corrected at one frequency at a time), random errors in the phase shifters, tracking, and alignment errors between beamformer channels. The effect of amplitude and phase errors on the array patterns depends on the degree of their correlation, meaning where they occur in the beamforming network. A group of elements that share a given error act together as a subarray. The impact of the subarray level error is greater than that of the same error at an individual element because the directivity of the subarray is larger than that of an individual element. In effect, a correlated error is multiplied by the number of elements affected by the error. For a monopulse network forming simultaneous beams, but having only one phase shifter per element, the phase shifters can correct the phase errors in the parts
7065_Agrawal_V6.indd 88
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
89
of the beamforming network that are common to the three beams; however, one must choose which errors to correct in unique circuits that generate the sum and difference beams. Typically, the errors in the sum beam are corrected. Since the phase errors for the difference beamformers are not fully corrected, the achievable sidelobe level for the difference beams usually is higher than for the sum beam. Besides the combining networks in the beamformer, any interconnections within the beamformer, such as between the radiating element inputs and the corresponding ports of the beamformer, must be carefully phase-controlled so as not to introduce errors. When a beamformer produces a sum, azimuth difference (AZ DELTA), and elevation difference (EL DELTA) beam, the insertion phase of the sum and delta portions of the beamformer must track each other closely so that the desired performance for monopulse operation is achieved. For example, phase errors between sum and delta beamformers can result in beam pointing angle errors. The complexity of the beamforming arrangement in Figure 4.2 is apparent. Additional complexity arises because of the amplitude weighting of the signals relative to each other in each column and from column to column to achieve the appropriate sidelobe level for both elevation and azimuth beams. Furthermore, even if the phase shifters are set correctly, assuming equal phase signals arrive at the phase shifters, phase and amplitude errors accumulate through the combiners, adversely affecting performance. In this regard, it should be noted that the actual physical lengths of interconnecting cables must nearly be equal for wide-instantaneous bandwidth signals, so as not to increase errors away from the frequencies at which error correction is performed. Corporate (equal path-length) beamformers can be used for both transmit and receive modes. Figure 4.3 shows a one-dimensional corporate beamformer constructed from hybrid couplers. In Figure 4.3, path lengths from the feed point to the radiating elements are equal, causing no progressive phase delay or frequency scanning, which enables a larger antenna instantaneous bandwidth. Power dividers are four-port hybrid couplers. In the transmit mode, spurious reflections are absorbed in the terminating loads of hybrids and do not degrade the antenna pattern. The beamformer for a low sidelobe receive array uses unequal power division couplers to form amplitude tapers. Mismatch reflections are absorbed in the hybrid loads.
7065_Agrawal_V6.indd 89
4/6/23 1:42 PM
90
Practical Aspects of Active Phased Array Antenna Development
Figure 4.3 Corporate beamformer constructed with hybrid couplers.
Energy loss in the coupler loads reduces the receive aperture (taper) efficiency; for a low sidelobe phased array with a 40-dB Taylor distribution, the aperture efficiency is about 2 dB [5].
4.3 Beamformer Networks for Active Phased Array Antennas Both transmit and receive functions in active phased array antennas are placed at the aperture using T/R modules. Each T/R module contains one or more transmit chains consisting of power amplifiers to provide power to each radiating element, a phase shifter, and receive chains comprising a low-noise amplifier, a phase shifter, and amplitude control. As in passive arrays, beamformers are required to form beams on receive and split the transmit exciter’s output to feed each of the T/R module inputs. Compared to passive arrays, T/R modules provide flexibility in amplitude weighting since they have variable attenuators. In Chapter 1, Figure 1.6 showed a simple block diagram of an active array. In the receive mode of operation, the outputs of the T/R modules are first combined using vertical column beamformers, and then the outputs of the column beamformers are combined using horizontal combiners to produce receive sum, delta elevation, and delta azimuth beams. A single beamformer can be
7065_Agrawal_V6.indd 90
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
91
used for both the receive sum and transmit modes by using T/R switch, as the active array transmit beamformer does not have to handle high power. The vertical column and horizontal combiners are uniformly weighted corporate designs realized from stripline or microstrip lines. The T/R modules provide the amplitude and phase distributions for the array, allowing the sum beamformer to be shared between receive and transmit. The delta elevation and azimuth beams are formed by simply setting one half of the array 180° out of phase with the other half, as shown in Figure 2.8. The active phased array antenna in Figure 1.6 uses a single uniformly weighted combining network to form all receive monopulse beams. If the T/R modules only have a single-phase shifter and attenuator for each receive beam, then only one of the monopulse beams, typically the sum beam, can have a low sidelobe taper. The sidelobes for the delta beams are not optimized. However, as we discussed in Chapter 2, reasonable low difference sidelobes can be achieved. The difference beam null depth, which determines how accurately the incident signal angle of arrival can be determined, is a function of the amplitude and phase errors in the two halves of the array that are subtracted to form the beam, as shown in Figure 2.8. Since there is no central high-power transmitter in active arrays, the transmit beamformer can be constructed using low-power transmission lines fabricated from printed circuit boards. It can also be shared with the receive sum beamformer. The receive amplitude taper can be formed either by the T/R modules (active weighted) or the beamforming network (passive weighted) for active phased array antennas. The resulting beamformers can be simple uniform weighted beamformers when the amplitude taper is generated in the modules. However, the amplitude taper can also be provided in the beamformers as in passive arrays. Section 4.4 will compare the two beamformer architectures for their dynamic range and NF performance. 4.3.1 Multiple Independent Receive Beams While monopulse beamformers generate multiple simultaneous beams, all are steered to the same angle in space, and some radar applications require multiple, independently steerable receive beams. Such independent receive beams can be generated using a
7065_Agrawal_V6.indd 91
4/6/23 1:42 PM
92
Practical Aspects of Active Phased Array Antenna Development
T/R module with multiple, parallel receive channels, each having its own amplitude and phase controls [6, 7]. Figure 4.4 shows a simplified block diagram of a monopulse active phased array antenna architecture in which each T/R module associated with the antenna element has three receive outputs. The architecture in Figure 4.4 consists of three beamformers, one each for sum, elevation delta, and azimuth delta receive beams. The transmit beamformer and sum receive beamformers are common, and these two functions are shared by this beamformer using a T/R switch on the beamformer (at left) and the T/R module. Figure 4.5 is a simplified block diagram of a T/R module with multiple receive outputs [4, 6]. The output of the LNA is divided equally using a three-way power divider. The divider outputs go through a phase shifter, low noise amplifiers, and a variable gain amplifier (VGA), and they are connected to three monopulse beamformers: sum, elevation delta, and azimuth delta. The sum receive channel is also shared with the transmit beamformer using a double-pole double-throw (DPDT) transfer switch, as shown in Figure 4.5. Independent amplitude and phase distributions can be applied in the three T/R receive channels to each monopulse beam to provide optimal sum and difference tapers so that the radiation
Figure 4.4 Beamformer architecture of an active phased array with three independent receive beams.
7065_Agrawal_V6.indd 92
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
93
pattern in either transmission or reception can be optimized separately for each sum, AZ Delta, and El Delta beams. An extra 180phase shift can be added to the elemental outputs from the selected half arrays for the difference beams. The VGAs can be adjusted to correct amplitude errors in the beamformer and provide the amplitude taper required to reduce the sidelobe level for each beam. The beamformers can be designed for uniform weighting rather than coupling values that are customized to provide the desired amplitude weighting. At each operating frequency of the array, the phase shifters and variable gain amplifiers can be programmed to correct the errors. Because the residual phase and amplitude errors in all three beamformers can be corrected using VGAs and phase shifters in the T/R module, a similar sidelobe performance can be achieved for all three monopulse beams, a significant improvement over the architecture of Figure 1.6 where the achievable sidelobe level for the difference beam is likely to be higher than for the sum beam. Accordingly, the limiting factors in the sidelobe performance of such a three-channel array are the amplitude and phase accuracy of the T/R modules over the instantaneous bandwidth and the longterm stability of the component RF insertion characteristics.
Figure 4.5 Block diagram of a T/R module with three receive outputs.
7065_Agrawal_V6.indd 93
4/6/23 1:42 PM
94
Practical Aspects of Active Phased Array Antenna Development
A disadvantage of the arrangements described in Figures 4.4 and 4.5 lies in the number of phase shifters and VGAs, requiring additional beamformers and increasing the beamformers’ complexity and the antenna cost. As an alternative, digital beamforming arrays may be used when large numbers of independent receive beams are needed. As described in Chapter 11, we can place receivers and A/D converters behind individual elements or subarrays and use a digital beamformer to generate as many beams as needed, with high-speed digital electronic circuits, subject to the limits of the array’s computer processors.
4.4 Impact of Beamformer Architecture on System Noise Figure In this section, two common receive beamformer architectures used in active phased arrays to generate low sidelobe sum patterns are compared for architectural complexity and their performance in terms of NF, input TOI, and dynamic range. In the first architecture, the active beamformer, the receive amplitude taper is applied in the T/R module. In the second architecture, referred to as the passive beamformer, the receive amplitude taper is provided in the beamformer [8]. Figures 1.6 and 4.6 show two popular corporate beamforming architectures; they differ in how they form the sum channel receive weighting. The transmit and receive sum beam amplitude tapers must differ to form the high gain transmit beam and the low sidelobe receive beam, so, in Figure 4.6 separate beamformers are required for the sum receive and transmit beams. On the other hand, a common transmit and sum receive beamformer can be used in Figure 1.6 since the T/R variable attenuators can be switched while the transmitted pulse propagates to the target and back to the array. In comparing these two architectures, we only consider the sum beamformers. In Figure 1.6, the weighting is implemented actively with the T/R modules, while in Figure 4.6, nonuniform (unequal) combiners are employed within the beamformer to create the low sidelobe weighting passively. The passive-weighted architecture provides the lowest system noise figure and highest output intercept point, while the active-weighted architecture is the easiest to construct and produce.
7065_Agrawal_V6.indd 94
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
95
Figure 4.6 Passive-weighted beamformer establishes the sum beam receive beamformer weighting in the passive combiners of the beamformer.
The active-weighted architecture of Figure 1.6 can use the same beamformer for transmit and receive. In Figure 1.6, an array of T/R modules is backed by a combiner for each array column, and a single horizontal combiner sums the columns to give the sum channel receive output. On transmit, the T/R modules are set to maximum gain and output power, and, on receive, the low sidelobe taper is applied by varying the gain of the receive VGAs in the T/R modules. The passive-weighted architecture of Figure 4.6 is more complex. It requires separate beamformers for transmit and receive operation. The transmit combiners have uniform weighting. However, on receive, all T/R module VGAs are set to maximum gain, and the low sidelobe taper is provided in the receiver combiners using nonuniform beamformers. The vertical combiners in each column are identical to a rectangular array with a separable amplitude distribution function; however, each column will require a different combiner for a circular array. The benefit of the passive-weighted array is that it has a lower receive NF and higher intercept point than the active-weighted array. A low antenna NF is critical if a radar is to detect small
7065_Agrawal_V6.indd 95
4/6/23 1:42 PM
96
Practical Aspects of Active Phased Array Antenna Development
returns, while a high antenna TOI is essential when there are big scatterers, such as clutter present at the same time. Lee [9] has analyzed the noise performance of active- and passive-weighted architectures, and Holzman [10] has derived equations for the output TOI for both array architectures. Table 4.1 summarizes beamformer signal gain, noise gain, input noise figure, and output TOI for both phased array architectures. Signal gain includes passive combining gain, active T/R module gain, and beamformer losses. The noise gain excludes the gain of the combiners. The input NF is the NF referenced to the input of the T/R module, and the output TOI is referenced to the output of the sum channel. Finally, the equations are normalized to the performance of a uniformly illuminated array, that is, an array with uniform combiners and all T/R modules with maximum receive gain. If we assume that reactive combiners are used in the passive-weighted array, the noise gain and input NF will be the same as those of the uniform array. The taper efficiency reduces the TOI and signal gain, typically 1 to 2 dB. The T/R modules are operated at maximum gain, so no reduction in noise gain occurs. On the other hand, for an active-weighted array, the T/R module gain varies with position in the array. For tapered aperture illuminations, the signal and noise gains of the active-weighted architecture can easily be 6 to 8 dB less than those of the passive-weighted array. However, the NFs of the two arrays are not much different since a VGA or attenuator sets the T/R module’s weighting at the back of the module’s receive chain. The module NF is higher for a module at the array edges where the attenuation is more severe; however, its contribution to the system NF is small because its NF is multiplied by the square of its voltage weight. The effect of the active-weighted array’s lower gain can be significant in a radar system when considering the stages that follow the phased array. As a vehicle to compare the two architectures, we use a rectangular array of 16,384 elements arranged on a triangular lattice with 256 rows and 256 columns. Applying a linear–linear 40-dB Taylor taper to the array, we obtain an aperture efficiency of 2.2 dB. Table 4.1 compares the resulting active- and passive-weighted performance. The passive-weighted array has 7.5 dB more signal and noise gain than the active architecture. For example, if the passive array has a 15-dB noise gain and a 5-dB input NF followed by a receiver with a 5-dB input NF, the system NF is 5.09 dB. On the other hand, an
7065_Agrawal_V6.indd 96
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
97
Table 4.1 Comparison of Active- and Passive-Weighted Arrays Parameter
Active-Weighted Array
Passive-Weighted Array
Signal gain
−9.7 dB
−2.2 dB
Noise gain
−7.5 dB
0.0 dB
5.5 dB
5.0 dB
Input NF Output TOI
−9.7 to −5.4 dB
−2.2 dB
active antenna with the same T/R module would have only 7.5 dB total noise gain and a system NF of 5.50 dB. The choice of providing amplitude taper in the beamformers or the T/R modules will depend on the required system noise performance, beamformer complexity, and cost.
4.5 Beamformer Architectures for High Reliability Since active phased-array antennas contain many active components, the antenna performance degrades gracefully as components fail over time. In addition, the choice of beamformer architecture and packaging of components will affect the system’s reliability. Consequently, the design of active phased arrays often involves a tradeoff among production cost, lifecycle cost, and performance. Strategies for designing active phased-array antenna architecture with high reliability are discussed in Chapter 8.
4.6 Beamformer Networks for Wideband Active Phased Array Antennas The beam of a phased array steered by phase shifters will scan as frequency changes, as discussed in Section 2.11.1. To overcome the limits of phase shift steering on the instantaneous bandwidth, we can use time delay steering instead. Rather than applying a progressive phase distribution, which is frequency dependent, across the array aperture, we apply a progressive delay, which, ideally, has no frequency dependence. To steer an array beam to angle θ scan with time delays, the required delay for the nth element is given by
7065_Agrawal_V6.indd 97
4/6/23 1:42 PM
98
Practical Aspects of Active Phased Array Antenna Development
ndsin(θ scan). If implemented using switched RF delay lines, the time delay at each element is prohibitively costly, bulky, heavy, and very lossy at microwave frequencies. However, in digitally beamformed arrays, time delays can be created digitally relatively easily. Rather than use time delay at every element in the array, the instantaneous bandwidth of a phase-steered array can be increased by dividing the aperture into subarrays and placing a time-delay network behind each subarray, as shown in Figure 4.7 for a linear array [11, 12]. An array’s instantaneous bandwidth is doubled by using only two time-delay units (TDUs), one for each half of an array. A planar array will require N2 time delays to increase its instantaneous bandwidth by a factor of N. With true time-delay steering for each subarray, the T/R modules provide phase compensation at the element level for the quantized time delays. The time delay at each subarray creates a correlated phase error for all subarray elements, generating grating lobes in the pattern. The subarrays are spaced at a distance equal to the distance between elements multiplied by the number of elements in the subarray. The resultant radiation pattern is the multiplication of the subarray and array patterns. The amplitude taper is applied in the T/R modules in the architecture of Figure 1.6 to produce the desired low sidelobe pattern. TDUs are combined in vertical columns, and the outputs of the vertical combiners are combined with a horizontal combiner to provide transmit input and receive the antenna outputs. Figure 4.9 shows the passive-weighted beamformer architecture with subarrays for time-delay steering. Figures 4.8 and 4.9 show active- and passive-weighted phased arrays with subarray time-delay steering. For the passive-weighted array, we must use two sets of TDUs or diplex the transmit and
Figure 4.7 A linear array divided into subarrays with a TDU behind each subarray.
7065_Agrawal_V6.indd 98
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
99
Figure 4.8 Wideband active-weighted array with phase shifters at the element level.
Figure 4.9 Passive-weighted beamformer architecture with subarray for time-delay steering.
7065_Agrawal_V6.indd 99
4/6/23 1:42 PM
100
Practical Aspects of Active Phased Array Antenna Development
receive signals through a single set; either option is more complex than the active-weighted architecture. In addition, the passive-weighted receive beamformer requires nonuniform combiners, so each subarray has a unique set of components. For an antenna with quadrantal symmetry, the number of unique subarrays can be reduced by a factor of four. Much like phase shifters, RF TDUs are divided into delay bits to provide different time delays for different angles. The total delay required to steer an array is determined by the formula Lsin(θ max), where θ max is the maximum array scan angle, and L is the length of the scan angle. So, a 10-foot-long array scanning to 45° would require up to 7 ns of delay. Delays of a few hundred picoseconds can be realized in MMICs. Nanosecond delays must be formed with a printed circuit or coax delay lines. Alternately, delays for even the largest arrays can be formed in the digital domain in increments of the digital rate. For example, a 1-GBit digital beamformer could form a 1 ns delay between two subarrays simply by delaying the clock of one subarray relative to the other by 1 bit. Smaller delays can be realized using digital interpolation. Typically, the same TDUs are used for all subarrays to simplify the production and maintainability of large-phased arrays that may have hundreds of subarrays. Overlapped subarrays can be used to reduce the grating lobes due to subarray quantization effects [13]. Section 4.4 stated that the active-weighted architecture has lower noise performance than the passive-weighted architecture. As a compromise between the more complex passive-weighted architecture and lower-performance active-weighted architecture, we can combine both characteristics in a hybrid architecture shown in Figure 4.10. For a beamformer with multiple levels of combining, some combiners can be uniform, while others can be nonuniform. In that way, we can reduce the number of unique nonuniform combiners in the antenna and have higher performance. In Figure 4.10, the vertical and horizontal combiners in the subarray are uniformly weighted and common to both transmit and receive paths. Therefore, partial amplitude weighting can be applied in the T/R modules in the receive mode. For a large-phased array, by making the first-level vertical receive combiners in the subarray uniform, we can significantly reduce the number of nonuniform combiners and
7065_Agrawal_V6.indd 100
4/6/23 1:42 PM
Beamformer Architectures for Active Phased Array Antennas
101
Figure 4.10 Hybrid subarray column-weighted architecture with active-weighted architecture.
achieve performance significantly better than the active-weighted architecture.
4.7 Concluding Remarks Active phased array antennas significantly improve performance over their predecessor passive phased array antennas. In addition to the reduction in transmit and receive losses, active arrays can provide wide instantaneous bandwidth, and high average radiated power. Active phased arrays offer many choices for beamformer architectures. This chapter examined the beamformer architectures for active phased array radar antennas for narrow and wide bandwidth. We compared passive-weighted and active-weighted architectures in terms of their beamformer complexity and dynamic range. The material presented in this chapter should aid the array RF design team in choosing a beamformer architecture for their application.
7065_Agrawal_V6.indd 101
4/6/23 1:42 PM
102
Practical Aspects of Active Phased Array Antenna Development
References [1] Skolnik, M., Introduction to Radar Systems, McGraw-Hill, 1980. [2] Ajioka, J. S., and J. L. McFarland, “Beam-Forming Feeds,” in Antenna Handbook, Van Nostrand Reinhold, 1993, pp. 179–300. [3] Mailloux, R. J., “Antenna Array Architecture,” IEEE Proc., Vol. 80, January 1992. [4] Agrawal, A. K., and E. Holzman, “Beamformer Architectures for Active Phased-Array Antennas,” IEEE Trans. on Antennas and Propagation, Vol. 47, No. 3, 1999, pp. 432–442. [5] Holzman, E., “A Different Perspective on Taper Efficiency for Array Antennas,” IEEE Trans. on Antennas and Propagation, Vol. 51, No. 10, October 2003, pp. 2963–2967. [6] Agrawal, A. K., and N. Landry, “Independent Control of Sum and Difference Patterns in Active Phased Array Antennas,” IEEE Trans. on Antennas and Propagation Int. Symp., San Jose, CA, June 1989, pp. 392–395. [7] Agrawal, A. K., and N. Landry, “Monopulse Phased Array Antenna with Plural Transmit–Receive Monopulse Phase Shifters,” U.S. Patent 5,017,927, issued May 21, 1991. [8] Holzman, E., and A. K. Agrawal, “A Comparison of Active Phased Array, Corporate Beamforming Networks,” IEEE Int. Symp. Phased-Array Technol., Boston, MA, October 1996, pp. 429–434. [9] Lee, J. J., “G/T and Noise Figure of Active Array Antennas,” IEEE Trans. on Antennas Propagation, Vol. 41, February 1993, pp. 241–244. [10] Holzman, E. L., “Intercept Points of Active Phased Array Antennas,” IEEE Microwave Theory and Techniques Soc. Int. Microwave Symp., San Francisco, CA, June 1996, pp. 999–1002. [11] Frank, J., “Bandwidth Criteria for Phased Array Antennas,” Phased Array Antennas, Artech House, 1972. [12] Mailloux, R., “Array Grating Lobes Due to Periodic Phase, Amplitude, and Time Delay Quantization,” IEEE Trans. on Antennas and Propagation, Vol. AP-32, No. 12, December 1984, pp. 1364–1368. [13] Azar, T., “Overlapped Subarrays: Review and Update,” IEEE Trans. on Antennas and Propagation, Vol. 55, No. 2, April 2013.
7065_Agrawal_V6.indd 102
4/6/23 1:42 PM
5 Radiating Elements 5.1 Introduction In this chapter, we discuss how to select a radiating element that best fits a given application and provide a design procedure for the radiator designer. The critical requirement for the array antenna or radiating element is that it radiates efficiently, with a good input impedance match to its feeding transmission line, over the operating frequency band and scan volume of the array. Unlike isolated antennas, the input match of an array radiating element is affected by its mutual coupling with other radiating elements, which varies with scan angle, and even polarization if the radiator is dual polarized. There are three types of radiating elements: (1) printed circuit types such as patch antennas, (2) waveguide-based, and (3) wire structures such as dipoles. The main requirements for phased array radiators include radiation pattern characteristics, including cross-polarization, over a scan volume in an array environment, RF bandwidth, efficiency, power handling, mechanical shock, vibration, weight, manufacturing cost, and electromagnetic environmental effects (EME) such as electromagnetic pulse (EMP), lightning, and gamma rays. The selection of a radiating element also depends on its integrability with the T/R module feeding network such that the array front-end RF losses are minimized.
103
7065_Agrawal_V6.indd 103
4/6/23 1:42 PM
104
Practical Aspects of Active Phased Array Antenna Development
This chapter aims to select a radiating element that best fits a given application and provide the design procedure for use in an active phased array antenna. In addition, a radome covers the radiating aperture for environmental protection against rain, snow, and ice. The choice of radiating elements is also a function of their integrability with the feeding networks to minimize the RF losses in the connection to the T/R modules. Since the input impedance of the radiating element is different from the isolated element, the element should be designed to provide the best impedance match with the feeding networks and space. The procedure for the radiating element design in the array environment is described in this chapter. The radiating elements in an active array are fed from the output of the circulator in the T/R module. Therefore, an RF connection between the T/R module and the radiating element is required. An RF coaxial cable provides flexibility within the array mechanical package but adds to the complexity of assembly and incurs an extra loss. Alternatively, the RF cable could be eliminated by including the radiating element as part of the T/R module assembly. We discuss packaging concepts for active phased arrays in Chapter 7.
5.2 Printed Circuit Radiating Elements Examples of printed circuit elements include microstrip patches, printed dipoles, and flared notch (Vivaldi) radiators [1–4]. Figure 5.1 shows several printed circuit radiator configurations. Figure 5.1(a) shows a typical metal patch radiator mounted on a dielectric substrate, such as the glass-filled, Teflon materials sold by Rogers Corporation. The patch radiator is fed by a microstrip line. A disadvantage of this patch antenna is that the feeding lines are exposed and radiate relatively high cross-polarization. To overcome this limitation, two printed circuit layers can be laminated together, as in Figure 5.1(b), which shows an aperture-coupled patch. The microstrip feed line and radiating patch are isolated by a ground plane containing a slot that allows the feed to excite the patch. Figure 5.1(c) shows a direct connection between a coaxial line attached to the back of the patch substrate and the radiating patch on the top. A patch can be excited to radiate circular polarization with a single feed, as shown in the upper left of Figure 5.1(d). Figure 5.1(e)
7065_Agrawal_V6.indd 104
4/6/23 1:42 PM
Radiating Elements
105
Figure 5.1 Printed circuit patch radiator: (a) microstrip fed patch; (b) aperture coupled patch; (c) microstrip patch fed by a coaxial line; (d) corner-fed circularly polarized patch; and (e) aperture coupled dipole.
shows a printed circuit dipole, which can be excited by a microstrip or stripline through a slot in the radiating conductor. Patches are easy to etch from low-loss printed circuit board laminates, backed by metal layers using low-cost, commercially available lithographic techniques. Complex, multifunction circuit boards, having a dozen or more layers joined together that integrate the radiator, the beamformer, logic, and power circuit layers, enable highly compact active array packaging. Patch antennas are widely used, and many books and hundreds of technical papers present detailed data for patch radiators [1–4]. The principal disadvantage of the patch radiator is
7065_Agrawal_V6.indd 105
4/6/23 1:42 PM
106
Practical Aspects of Active Phased Array Antenna Development
that it is not highly efficient, is relatively narrowband, and radiates relatively high cross-polarization. The bandwidth of patch radiators depends on the patch dimensions, substrate thickness, and dielectric constant. Adding substrate layers to allow the stacking of patches extends RF bandwidth by providing a double-tuned element. The narrow bandwidth of patch radiators has been overcome using more complex feeding structures, and tight coupling between the radiating structures. The PUMA element, invented by Holland, Schaubert, and Vouvakis, is one radiator that achieves up to a 6:1 bandwidth yet requires two circuit board dielectric layers [5, 6]. Once one has chosen the array element grid shape and dimensions, the first step in every patch array design should be a surface wave analysis of the dielectric substrate. Surface waves, when excited, transmit energy parallel to the array surface rather than into free space, as desired, causing a severe mismatch at the radiator. Pozar and Schaubert [7] describe an excellent, relatively simple procedure for determining the scan angles and frequencies at which a patch array will excite surface waves based on a highly accurate slab-mode model. 5.2.1 Printed Circuit Wideband Radiating Elements A Vivaldi flared notch or tapered slot radiating element is shown in Figure 5.2. A notch or flared slot transmission lines etched into one of the metal layers of a dielectric substrate is excited by an open-circuited orthogonally directed center conductor [8], as shown in Figure 5.2. The notch is terminated in a short circuit. The design of the notch taper and the impedance match between the open-circuited feed and short-circuited notch can achieve very wide bandwidth. Typically, the flared slot profile design uses an exponential, Dolph-Chebyshev, or Klopfenstein taper [9]. Traditionally, notch radiators have had high cross-polarization in the intercardinal planes. In 2018, Kindt and Logan [10] modified the notch radiator to achieve much improved cross-polarization radiation. The notch radiator can be made a part of the T/R module assembly. It can easily be inserted or removed from the array, eliminating connectors and an RF cable for connecting the radiating element and the T/R module.
7065_Agrawal_V6.indd 106
4/6/23 1:42 PM
107
Radiating Elements
Microstrip
Figure 5.2 (a) Vivaldi flared notch radiator fed by a stripline; (b) exploded view of the notch radiator assembly showing three substrate layers.
5.3 Waveguide Radiating Elements Waveguide elements achieve very high-power handling and thus are preferred for passive phased arrays with central high-power transmitters. Waveguide elements are typically machined into metal faceplates and thus tend to be heavy, but they have very low loss and graceful impedance degradation over the scan volume. They can also be used in active phased array antennas because of their low loss, roughly octave RF bandwidth, inherent high pass response from the dominant mode cutoff, low cross-polarization, and mechanical robustness. Waveguide elements also have less mutual coupling between the elements than patch elements. Cutoff frequency rejects radiation from external low-frequency electromagnetic (EMP) and lighting. The EMP radiation frequency content is mostly below 100 MHz, which is well below the cutoff frequency of the waveguide radiators used in radars operating in L-band and above. The waveguide further attenuates EMP and does not damage the receive circuits. Waveguide radiators can be machined directly into an array faceplate with automated precision machining of the face plate. Figures 5.3 shows the rectangular, circular waveguide, and slotted waveguide radiating elements. The slots in the narrow and wide walls are resonant narrowband structures. The Airborne Warning and Control System (AWACS) uses slotted waveguides and is rotated to scan the beam 360°.
7065_Agrawal_V6.indd 107
4/6/23 1:43 PM
108
Practical Aspects of Active Phased Array Antenna Development
Figure 5.3 Waveguide radiating elements: (a) rectangular waveguide; (b) circular waveguide; (c) slotted waveguide array with slots in the narrow wall; and (d) slots in the broad wall of rectangular waveguides.
Ridged waveguides can be used for wide bandwidth applications and also reduce the width of the waveguide to fit into the array grid. The ridges lower the dominant waveguide mode cutoff frequency without decreasing the second waveguide mode’s cutoff frequency, thereby increasing the dominant mode operating bandwidth. Double- and quad-ridged waveguide elements are shown in Figure 5.4. The quad-ridged waveguide elements provide dual and circular polarizations by exciting radiating fields between the ridges.
Figure 5.4 Double- and quad-ridge waveguide radiating elements.
7065_Agrawal_V6.indd 108
4/6/23 1:43 PM
Radiating Elements
109
5.3.1 A Wideband Tapered Double-Ridged Waveguide Element Fed by a Coaxial Probe Figure 5.5(a) shows a tapered double-ridged waveguide radiating element capable of operating over an octave frequency band [11]. The double-ridge loaded waveguide is excited by a coaxial probe from the rear. A ceramic window is affixed to the radiating aperture to seal the aperture from the outside environment. The ceramic window replaces a standard, separate radome and can be heated by a DC current in the waveguide structure to prevent ice formation on the antenna aperture. The ceramic window replaces the external radome, eliminating the cost of the radome and the complex heating mechanism of the radome. SPY-1 phased array antennas on the CG-47 cruisers and DDG-51 class destroyer ships use such ceramic windows and have been in service for more than forty years [12].
Figure 5.5 Double-ridged waveguide radiating element fed by an end-launched coaxial probe: (a) cross-section; (b) side view.
7065_Agrawal_V6.indd 109
4/6/23 1:43 PM
110
Practical Aspects of Active Phased Array Antenna Development
As shown in Figure 5.5(b), the ridges step down in height in three steps to lower the dominant mode cutoff frequency and match the impedance of the probe to the aperture impedance. This radiating element was used in a phased array antenna with a scan volume requirement of +/−45° in azimuth and −30° to +90° in elevation. The coaxial probe feeding the ridged waveguide element can be integrated with the T/R module assembly and inserted into the waveguide element. For example, Figure 5.6(a) shows the coaxial feeding probe attached to the T/R module line replaceable unit (LRU) assembly. Figure 5.6 shows a group of two T/R modules assembled as a LRU. An alternative to the tapered ridges is using a notch radiator, which can also be integrated with the T/R module, as shown in Figure 5.6(b). The notch radiator also can be inserted into the antenna structure or a waveguide. In addition, a gasket can be used to provide a waterproof seal at the RF interface. The T/R module and integrated radiating element assembly do not require an RF cable interface.
5.4 Radome Heating for Ice Inhibition Ground-based and shipborne radars must perform under all weather conditions, including icy conditions. Therefore, the radome’s front surface must be heated to prevent ice formation. Printed heating wires are embedded in the radome’s dielectric sheet to heat the front surface of the radome. Current flowing through the conductors heats the radome surface where the conductors are located. The geometry of the heating wires should be such that it does not significantly impact the antenna performance. First, Figure 5.7 shows parallel, horizontal heating wires passing through a radome. As long as the array radiated electromagnetic field is polarized perpendicular, vertical in this case, to the wires, little energy will be reflected by the wires. Furthermore, the spacing of the wires is a compromise between being sufficiently close to melt the ice, and sufficiently far apart to minimize reflections. It is possible for the heating wire arrangement to provide a lower reflection coefficient than the radome alone in the absence of the
7065_Agrawal_V6.indd 110
4/6/23 1:43 PM
Radiating Elements
111
Figure 5.6 (a) Coaxial probes integrated with T/R modules; (b) notch radiators assembled with T/R modules.
Figure 5.7 Parallel, horizontal heating wire arrangement for heating the radome in the presence of vertically polarized array radiation.
7065_Agrawal_V6.indd 111
4/6/23 1:43 PM
112
Practical Aspects of Active Phased Array Antenna Development
conductors. Figure 5.8 shows a heating wires geometry that can improve the antenna performance [13]. However, the conductive heating wires have an inductive reactance that can negatively interfere with the RF performance of the radome. To compensate, parasitic elements having a capacitive reactance can be interspersed or distributed among the heating wires. The arrangement in Figure 5.8 uses detuned dipoles as capacitive elements that can be aligned with an elongated length of the heating wires [13]. For manufacturing convenience, the heating wires and the capacitive elements can be printed on the dielectric substrate forming the radome.
5.5 Wideband Parallel Waveguide Phased Array Radiator We introduce a novel wideband parallel plate waveguide radiator fed by a microstrip to waveguide transition shown in Figure 5.9 [14]. The bandwidth of a rectangular waveguide phased array antenna is limited due to the cutoff frequency. Therefore, a parallel plate waveguide and a dominant mode without a cutoff can be a better choice for wide bandwidth applications. This radiator is capable of providing 3:1 bandwidth with scan coverage of at least ±45° in azimuth and elevation. An array can be constructed using continuous parallel-plate waveguides fed at multiple locations, eliminating the need for the assembly of a large number of individual elements. Unlike wideband tapered slot or Vivaldi arrays, the wideband parallel-plate waveguide array provides high
Figure 5.8 Radome with detuned, parasitic matching elements between the continuous heating wires.
7065_Agrawal_V6.indd 112
4/6/23 1:43 PM
Radiating Elements
113
cross-polarization rejection over the entire scan volume. The elements can be arranged in a rectangular or triangular lattice. Coaxial-fed and monopole-fed parallel plate waveguide phased array antennas have reported up to 40% bandwidth [15, 16]. Holzman describes a wideband TEM horn antenna that uses the horn flare to help match the element impedance to free space [17]. The parallel-plate waveguide is filled with air or foam for a dielectric constant approaching 1.0. The microstrip feed line is placed perpendicular to the parallel-plate waveguide walls. A relatively high dielectric constant is required to constrain the fields between the microstrip trace and ground over the initial portion of the transition. A portion of the wideband parallel-plate waveguide array is shown in Figure 5.10. Inside the waveguide, the microstrip trace (upper conductor) portion of the transition becomes an upward tilting taper that gradually increases in width until the top edge touches the parallel-plate waveguide top wall. After the top edge of the upper conductor contacts the waveguide top wall, the bottom edge continues on an upward slant until it also contacts the waveguide top wall. The transition’s microstrip ground (lower conductor) portion at the feed entry point extends from the waveguide top wall to the bottom wall. Along the length of the transition, the top edge of the lower conductor is gradually tapered away from the waveguide top wall. The bottom edge of the lower conductor stays in constant contact with the waveguide bottom wall until the end of the transition taper. The
Figure 5.9 Radiating element feed transition geometry.
7065_Agrawal_V6.indd 113
4/6/23 1:43 PM
114
Practical Aspects of Active Phased Array Antenna Development
Figure 5.10 Section of a wideband parallel-plate waveguide array.
microstrip substrate retains its original size throughout the transition and ends at the end of the transition. A linear taper is used in the final design. Exponential tapers were also investigated and may prove useful for increasing the bandwidth in future designs. The radiating element was modeled using Ansys High Frequency Structure Simulator (HFSS), a finite element electromagnetic analysis tool. Simulation results over the full scan coverage indicate a 3:1 bandwidth with less than 2.5:1 element VSWR and better than 28 dB cross-polarization rejection, as shown in Figure 5.11.
5.6 Mutual Coupling Between Radiating Elements The array pattern calculations in Chapter 2 assumed that the radiating elements in the array behave as isolated elements, and their excitation is proportional to their input currents. However, mutual coupling between the elements causes their impedance to vary as a function of the beam scan angle. In the literature, this scan-dependent impedance is referred to as the active impedance or the scan impedance. One of the radiator’s functions is to provide a good match between the transmitting RF source and free space. Therefore, the scan impedance looking into the radiator in the array environment must be matched to its generator impedance, typically the impedance at the output of the T/R module.
7065_Agrawal_V6.indd 114
4/6/23 1:43 PM
Radiating Elements
115
Figure 5.11 Radiating element VSWR performance: (a) E-plane scan; (b) H-plane scan; (c) intercardinal planes; and (d) cross-polarization.
The far-field array electric field for an infinite array can be written as E(θ, ϕ) = Ee(θ, ϕ) Ea(θ, ϕ) where Ee = the scan element pattern Ea = the array factor The scan or embedded element pattern is obtained when only one element is driven while all others are terminated into matched loads. The scan element pattern can be measured in a relatively small array and does not require an infinite array. This is because the mutual impedance between elements decays as 1/r2, where r is
7065_Agrawal_V6.indd 115
4/6/23 1:43 PM
116
Practical Aspects of Active Phased Array Antenna Development
the distance between the radiators. Holter and Steyskal [18] have shown that the center element of a five-wavelength square fractional array responds to the mutual coupling as if it is in an infinite array. For a standard rectangular grid array, with elements spaced half wavelength apart, a 10-by-10 array is sufficiently large. More elements will be required for wideband arrays because the electrical distance between elements is much closer at the bottom of the band. As the array is scanned, the peak of the array pattern will trace out the element pattern. Generally, a small array of 100–200 elements is sufficient to predict the scan element pattern. The effect of mutual coupling is quite localized. In general, to first order, all elements except the edge elements behave as if they are in an infinite array or a large array. Therefore, the effect of edge elements on array sidelobe performance in the transmit mode is negligible for very large arrays. For receive operation, a couple of rows and columns of dummy elements (nonradiating elements terminated into matched loads) can be added to the aperture so that the edge elements respond to the coupling environment much like center elements. However, dummy elements provide limited benefits on transmit since they do not radiate [19]. The design of the radiating element involves the following: • Element area less than the element lattice area to avoid grating lobes; • The element pattern provides appropriate aperture matching over the scan coverage; • Polarization and power-handling capability meet the requirements; • Withstand environmental requirements such as thermal, shock, and vibration; • Reliability, cost, manufacturability, and repeatability from unit to unit.
5.7 Selection of the Radiating Element Type Each dB of RF loss between the T/R module output and the array radiator input reduces the transmit power and increases the NF decibel for decibel. Therefore, the interconnection between the
7065_Agrawal_V6.indd 116
4/6/23 1:43 PM
Radiating Elements
117
output of the T/R module must be simple and minimize additional RF loss from coaxial cables and connectors. Eliminating a separate interconnection by making the radiating element part of the T/R module assembly such that it can be inserted directly into the array structure offers a near optimum solution. Waveguide and notch radiators are suitable for high-power ground-based and shipborne antennas that use the brick T/R module architecture, as they can be inserted from the back of the array. Printed circuit dipole and patch arrays often require a cable connection between the T/R module and the radiating elements. However, the printed circuit patch and dipole elements are a good radiator choice for the tile-style array, where the connection between different layers of components is made with coaxial connectors.
5.8 Radiating Element Design Process The goal of the radiating element engineer is to design an antenna that radiates efficiently, which requires a good input impedance match to the feed over the required operating frequency band and scan volume. Unlike isolated antenna design, an array radiating element always involves compromise because performance is affected by mutual coupling with other radiating elements. The design process is usually iterative and involves simulation using computational electromagnetic modeling software. Before building a full array, engineers may fabricate a waveguide simulator to verify the active match at a selected scan angles, followed by an assembly of a small test array (generally at least five wavelengths square) and measurement of mutual coupling between elements to verify performance over the scan volume. A well-designed element will radiate a power pattern on the order of cos1.25(θ scan) over the intended scan volume, where θ scan is the scan angle from the array broadside (normal to the aperture). The cosine factor represents the aperture projection in the direction of the scanned beam, and the 0.25 factor is due to the radiating element scan loss. So, one can write the scan loss of an array radiator in dB as Lscan = 10nLog10(cosθ scan) where n = the element factor, typically about 1.25.
7065_Agrawal_V6.indd 117
4/6/23 1:43 PM
118
Practical Aspects of Active Phased Array Antenna Development
The array geometry and the mutual coupling environment determine a radiator’s active input impedance and far-field pattern. The scan impedance of a radiating element in an infinite array can be measured at discrete scan angles by testing a single or a group of elements inside a waveguide [20, 21]. Imaging the structure in the waveguide walls perfectly replicates the infinite array mutual coupling environment at discrete scan angles. The dominant TE10 rectangular waveguide mode can be decomposed into two inclined plane waves traveling down the guide in directions ±θ, as shown in Figure 5.12 [22, 23]. The angle each plane wave makes with the broad wall is determined by the ratio of the operational wavelength to the TE10 mode cutoff wavelength, which is a function of the width of the waveguide. The effective scan angle is given by sin θ scan = λ/λ c where
θ scan = the scan angle λ = the free space wavelength at the test frequency λ c = 2a is the cutoff wavelength of the waveguide The waveguide dimensions are chosen to equal the array element spacing so that a radiating element placed in the waveguide sees mirror images in the waveguide wall that appear to be the same spacing as the array being simulated [21]. Finally, the active input impedance measurements are made by measuring the reflection coefficient of the waveguide simulator from the right with the radiating element structure on the left terminated in loads, which
Figure 5.12 A waveguide array simulator.
7065_Agrawal_V6.indd 118
4/6/23 1:43 PM
Radiating Elements
119
is equivalent to a plane wave impinging on an infinite array from free space at the scan angle given by the previous equation. Two simple H-plane waveguide simulators for triangular and rectangular lattices are shown in Figures 5.13 and 5.14. For the simulator in Figure 5.13, the simulator dimensions are: a = λ; b = λ/√3
λ c = 2a sinθ = λ/λ c = 1/2
θ = 30° The waveguide simulator has one full waveguide element and four-quarter elements for the triangular lattice. Since electromagnetic waves cannot propagate through quarter elements, they need not be terminated into matched loads. For a rectangular element lattice, the waveguide simulator in Figure 5.14 consists of a waveguide with a center element and two half elements. The simulator dimensions are a = λ; b = λ/2. The radiating element performance can be measured for various angles by building and testing several different simulators of different sizes and propagating different waveguide modes [20].
Figure 5.13 H-plane waveguide simulator for an array on a triangular lattice.
7065_Agrawal_V6.indd 119
4/6/23 1:43 PM
120
Practical Aspects of Active Phased Array Antenna Development
Figure 5.14 H-plane simulator boundaries for a rectangular array.
As mentioned in the previous section, tapered ridge waveguides and notch radiators are excellent choices for matching the radiating element impedance for various frequencies and scan angles. In addition, these simulators provide a low-cost solution for impedance matching of the radiating elements in an infinite array. However, this impedance matching using waveguide simulators is limited in the number of scan angles. The goal of the design is to match the impedance with a voltage standing wave ratio (VSWR) of less than 2.5:1. Waveguide simulators can be modeled and analyzed using computational electromagnetics solvers such as Ansys HFSS and Dassault CST Studio Suite. These solvers can analyze an infinite array unit cell by applying periodic boundary conditions at any scan angle. Although the simulated results using computer models have proven to be fairly accurate, building and testing at least one simulator to verify the design is a general practice. The next step in designing and verifying the radiating element’s scan performance is to measure a single element’s far-field radiation pattern in a fractional or small array with all other elements terminated into the matched loads. As mentioned earlier, the center element of a small array of five wavelengths square will have a far-field pattern very close in shape to the element pattern in an infinite array. Due to the small size of the test array, reflections from
7065_Agrawal_V6.indd 120
4/6/23 1:43 PM
Radiating Elements
121
the edges will cause the element pattern to show some ripple. As the array is scanned, the peak of the array pattern will trace out the element pattern. A ripple of about 1 dB is considered normal. Large mismatches at any scan angle will show dips in the pattern greater than 1 dB. Ripples greater than 1 dB or significant dips in the element pattern usually indicate other problems in the radiator design, such as surface waves or resonant modes. They must be investigated and resolved before building the full-size array. Once the single-element radiation pattern is known, the predicted antenna’s radiation pattern can be calculated by multiplying by the array factor and can be compared with the required pattern for verification.
5.9 Phased Array Radiation Pattern Calculation by Using the Mutual Coupling Between Elements in a Small Array The single or embedded element pattern can be calculated from the measured mutual coupling between all pairs of elements in the fractional array. For an N-element array, these mutual couplings can be gathered together in an N × N S-parameter matrix [b] = [S][a] where the a-waves represent the voltages exciting the inputs of the N elements of the array, and the b-waves represent the reflected waves at those same inputs. If we number the elements from 1 to N, then the amplitude of the reflected wave at element 1 is expressed by the equation
where
b1 = s11a1 + s12a2 + ⋯ . . . s1NaN
b1 = the reflected wave at element 1 a1 = the incident wave at the element 1 s11 = the self-impedance of element 1 s12 . . . s1n = the mutual coupling coefficients between element 1 and other elements
7065_Agrawal_V6.indd 121
4/6/23 1:43 PM
122
Practical Aspects of Active Phased Array Antenna Development
The active input or scan reflection coefficient of element 1 is given by Rin 1 =
b1 a1
Using a computer program, one can generate any desired excitation of the array by including the progressive phase coefficient for each radiator and compute the active reflection coefficient for all N elements Rin1 = s11 + s12ejφ + . . . s1nej(n–1)φ
5.10 Concluding Remarks This chapter discussed various radiating elements that can be used in active phased arrays. These radiating elements include printed circuit patches, waveguides and dipoles, and notch and tapered ridge waveguide elements for wideband arrays. The selection of the element depends on the array platform and architecture. The radiating element is designed to have a good impedance match in a mutual coupling array environment. We also described how heating wires can be included in the array structure or radome to prevent ice formation. Finally, the radiating element design process using waveguide simulators and computational electromagnetic software was described.
References [1] Munson, R., “Conformal Microstrip Antennas and Microstrip Phased Arrays,” IEEE Trans. on Antennas and Propagation, Vol. 22, 1974. [2] Behl, I. J., and P. Bhartia, Microstrip Antennas, Artech House, 1980. [3] Carver, K. R., and J. W. Mink, “Microstrip Antenna Technology,” IEEE Trans. on Antennas and Propagation, Vol. 29, No. 1, 1981, pp. 2–22. [4] Pandey, A., Practical Microstrip and Printed Antenna Design, Artech House, 2019. [5] Holland, S., Schaubert, D., and M. N. Vouvakis, “A 7–21 GHz Dual-Polarized Planar Ultrawideband Modular Antenna (PUMA) Array,” IEEE Trans. on Antennas and Propagation, Vol. 60, 2012, pp. 4589–4600.
7065_Agrawal_V6.indd 122
4/6/23 1:43 PM
Radiating Elements
123
[6] Logan, J., and M. Vouvakis, “On the Design of 6:1 mm-Wave PUMA Arrays,” IEEE Int. Symp. on Antennas and Propagation & USNC/URSI, 2013. [7] Pozar, D., and D. Schaubert, “Scan Blindness in Infinite Phased Arrays of Printed Dipoles,” IEEE Trans. on Antennas and Propagation, Vol. 32, 1984, pp. 602–610. [8] Franz, K. M., and P. E. Mayes, “Broadband Feeds for Vivaldi Antennas,” Antenna Applications Symp., University of Illinois, 1987. [9] Klopfenstein, W., “A Transmission Line Taper of Improved Design,” IRE Proc., Vol. 44, No. 1, 1956, pp. 31–35. [10] Logan, J. T., Kindt, R. W., and M. N. Vouvakis, “Low Cross-Polarization Vivaldi Arrays,” IEEE Trans. on Antennas and Propagation, Vol. 66, 2018, pp. 1827–1837. [11] Agrawal, A. K., Perry, M. S., and N. R. Landry, “Wideband Ridge Waveguide Radiating Element,” IEEE AP-S Int. Symp., Newport Beach, CA, July 1995. [12] McClane, J. L., and J. L. McClane, “The Ticonderoga Story: Aegis Works,” U.S. Naval Institute Proceedings, 1985, Vol. 111/5/987. [13] Munk, B., “Radome with Detuned Elements and Continuous Wires,” U.S. Patent 2007/0252775 A1, issued November 1, 2007. [14] Jablon, A., and A. K. Agrawal, “A Wideband Parallel-Waveguide Passed Array Radiating Element,” Unpublished article. [15] Chu, R. S., Lee, K. M., and J. U. Clark, “Analysis and Experiment of a Wideband Phased Array of Monopole Excited Parallel Plate Waveguides,” IEEE AP-S Symp. Digest, 1989, pp. 974–977. [16] J. B. L. Rao, Laxpati, S. R., and B. D. Wright, “Wideband Phased Array of Coaxially-Fed Probes,” IEEE AP-S Symp. Digest, 1987, pp. 290–293. [17] Holzman, E. L, “A Wide Band Horn Array Radiator with a Novel Microstrip Feed,” IEEE Int. Conference on Phased Array Systems and Technology, May 2000. [18] Holter, H., and Steyskal, H., “On the Size Requirement for Finite Phased-Array Models,” IEEE Trans. on Antennas and Propagation, Vol. 50, 2002, pp. 836–840. [19] Holzman, E., “On the Use of Dummy Elements to Match Edge Elements in Transmit Arrays,” IEEE Int. Symp. on Phased Array Systems and Technology, October 2013. [20] Hannon, P. W., and M. A. Balfour, “Simulation of a Phased Array Antenna in a Waveguide,” IEEE Trans. on Antennas and Propagation, Vol. AP-13, 1965, pp. 342–353. [21] Balfour, M. A., “Active Impedance of a Phased Array Antenna Simulated by a Single Element in a Waveguide,” IEEE Trans. on Antennas and Propagation, Vol. AP-15, 1967, pp. 313–314. [22] Mailloux, R., “Phased Array Antenna Handbook,” Artech House, 1996. [23] Cheston, T. C., and J. Frank, “Phased Array Radar Antennas,” in Radar Handbook, 2nd ed., M. I. Skolnik (ed.), McGraw-Hill, 1990.
7065_Agrawal_V6.indd 123
4/6/23 1:43 PM
7065_Agrawal_V6.indd 124
4/6/23 1:43 PM
6 Beam Steering and DC Power Distribution 6.1 Active Phased Array Antenna Beam Steering Controller An antenna beam steering controller (BSC) generates the phase shift digital circuit commands for all the T/R modules. In addition to phase shift commands, the BSC generates digital commands for each T/R module attenuator to set the low side lobe amplitude distribution across the antenna aperture on receive. The BSC also puts the array into transmit or receive mode by switching the appropriate RF switches in the T/R modules. Finally, BSCs collect circuit health status data from the T/R modules and serve as the interface between the array and the radar control computer (RCC) [1–3]. The BSC performs the following functions: • Generates the phase shift digital commands required by all the T/R modules to steer the beam; • Generates amplitude digital commands for each T/R module attenuator to produce the amplitude distribution across the antenna aperture during a receive operation; • BSC sets the antenna in transmit or receive mode by switching the appropriate RF switches in the T/R modules; • Sets beam switching time (number of beams per second, approximately few thousand);
125
7065_Agrawal_V6.indd 125
4/6/23 1:43 PM
126
Practical Aspects of Active Phased Array Antenna Development
• Performs fast beam switching between multiple targets within a dwell (pulse repetitive rate (PRF), approximately tens of thousands of hertz); • Transmit waveform from radar to antenna; • Switching of all drain and gate voltage of the amplifiers; • Generate commands for fault detection and fault isolation of elements; • Mutual coupling measurements between elements for calibration and fault detection, and fault isolation; • BSC provides the digital interface between the array and the RCC. When an array has many thousands of elements, generating the phase shift and attenuation commands may be necessary by multiple control circuits distributed throughout the array rather than in a single centralized location [1]. Hence, these two array control architectures are referred to as distributed BSC and centralized BSC. 6.1.1 Active Phased Array Distributed Beam Steering Controller In the distributed BSC architecture, shown in Figure 6.1, the central RCC generates simple commands such as scan angle, frequency, and timing, which are passed to the BSC central controller. The central controller converts scan angle and other information to amplitude and phase commands, which are distributed through to the T/R modules. Besides the central controller, each T/R module houses its own control electronics, which contains an application-specific integrated circuit (ASIC) or a FPGA, electrically erasable programmable read-only memory (EEPROM), field effect transistor (FET) switches, and storage capacitors. Each T/R module controller accesses its local memory to obtain data about the location within the array aperture of those elements it controls. The ASIC T/R controller then calculates the amplitude and phase settings for each T/R module, taking into account the amplitude and phase error corrections applied by the T/R module. In addition, EEPROMs may contain linearization amplitude and phase tables for each T/R module. Each module has its own tables derived from its unique factory test data. The EEPROMs can be erased and reloaded with new data when T/R modules are replaced. The amplitude and phase calculations
7065_Agrawal_V6.indd 126
4/6/23 1:43 PM
Beam Steering and DC Power Distribution
127
are performed separately for transmit and receive modes. As discussed in Chapter 3, a group of T/R modules may be controlled by a single T/R module controller. In addition to sending the scan angle and frequency, the BSC sends a command to set the array in either transmit or receive mode by setting RF switches appropriately in the T/R modules. A local crystal oscillator generates the clock for each local group of modules. The clock speed determines the time it takes for commands to reach all T/R modules. Generating the clock locally at the T/R module minimizes the addition of noise because the separate clocks are not synchronized. In addition, by storing many different beam-direction/gain-phase combinations in memory in advance of their use, beam switching can readily be accomplished without recalculation within a single dwell at a very high speed. As a result, the antenna can switch between multiple beam directions in microseconds or faster within the same dwell, allowing the array to interrogate many targets simultaneously. A control circuit in each T/R module LRU assists the central controller in commanding the array to steer and shape its transmit and receive beams. In Figure 6.1, the central controller distributes digital signals to each column simultaneously from the top and bottom, providing redundancy in the communication paths. Similarly, built-in test
Figure 6.1 Phased array distributed beam steering diagram.
7065_Agrawal_V6.indd 127
4/6/23 1:43 PM
128
Practical Aspects of Active Phased Array Antenna Development
equipment (BITE) information travels in two directions to provide a redundant path. The phase of the attenuator changes with the attenuator setting. Similarly, the amplitude of the phase shifter changes with the phase setting. Therefore, for an n-bit attenuator and n-bit phase shifter, there will be n2 data points for each T/R module. These data are stored in the EEPROMs or FPGAs of each T/R module. The input data to each T/R module from the central processor is distributed serially to minimize the number of digital connections. However, the T/R module’s buffer memory data is communicated in parallel to the phase shifters and attenuators. The advantages of distributed BSCs are: • Built-in data redundancy; • Very few I/Os from the central controller to the T/R modules; • Modest central computing capacity requirements; • T/R module ASIC or FPGA can provide increased functionality; • The number of digital I/Os and the central computing capacity are not a function of the number of elements in the array; • Fast data rate between ASIC and T/R module. The disadvantage of distributed BSCs is they require a custom ASIC for each new array. 6.1.2 Active Phased Array Centralized Beam Steering Controller In the centralized BSC architecture, all computations are performed centrally in a single control module. The data required by each T/R module are sent directly to the T/R modules or groups of T/R modules, such as the LRUs. The T/R modules do not contain any digital control electronics. The central BSC may contain many digital signal processor (DSP) cards and FPGAs; each card is responsible for a specific group of T/R modules and stores all calibration values (linearization tables) for that group of modules. Each DSP card performs beam steering calculations for each T/R module in its group within a minimum pulse repetition interval (PRI). The PRI and the number of modules determine the total computational processing requirement. The data containing the phase and attenuation bits sent over the control lines must be formatted for final delivery to
7065_Agrawal_V6.indd 128
4/6/23 1:43 PM
Beam Steering and DC Power Distribution
129
the T/R modules. For the centralized BSC architecture, this function can be implemented in field-programmable gate arrays that can easily be reprogrammed to meet the requirements of a new T/R module. Let’s review the advantages and disadvantages of centralized BSC. The advantages of a centralized BSC are: • Central computation does not require special hardware and can be commercial off the shelf (COTS); • It can be reconfigured for different radars without any significant design modifications; • It does not require custom ASICs. The disadvantage of a centralized BSC is it requires large computational capacity, which is a function of the number of elements in the array, and a high data transfer rate is required between the central controller and T/R modules; however, computer processing has become so fast that this is not an issue for the current phased arrays. In either BSC configuration, the latency or the time required to load the digital information into all T/R modules determines the maximum array beam switching rate.
6.2 Active Phased Array Power Distribution The three-phase 440V generator power is first converted to an unregulated DC voltage, usually 200V to 500V, by the AC-to-DC converters, as shown in Figure 6.2. Next, the unregulated DC voltage is switched at a high frequency, usually in the range of 15 kHz to 50 kHz and is converted to an AC voltage which is further stepped down to a lower voltage. Finally, rectifiers convert the AC output into DC voltage, which is then converted into three regulated voltages, 10 or 50, +5, and −5V for the T/R module. Linear power supplies are inefficient, expensive, bulky, and dissipate large amounts of heat. On the other hand, DC-to-DC converters using switched mode power supplies (SMPS) have higher power density, smaller size, lower cost, less weight, and efficiency, exceeding 90%. The size of the DC-to-DC converters is inversely proportional to their switching frequency. The voltage ripple of the
7065_Agrawal_V6.indd 129
4/6/23 1:43 PM
130
Practical Aspects of Active Phased Array Antenna Development
Figure 6.2 Active phased array antenna power conversion block diagram.
three voltages affects the RMS amplitude and phase errors. Therefore, one of the key requirements of the DC-to-DC converters is to have very low ripple or noise. 6.2.1 DC-to-DC Converter Key Requirements The specification for a DC-to-DC converter includes voltage and current required by the T/R modules, output voltage droop and ripple, random noise, efficiency, dynamic step response, and enough input and output energy storage capacitors required for the minimum droop due to load change during the longest transmit period meets requirements. The key requirements of the DC-to-DC converter are summarized here: • Minimum and maximum power requirements of the T/R modules; • Voltage droop and ripple; • Random noise; • Power added efficiency; • Voltage regulation; • Input and output energy storage capacitors to meet droop requirements during the longest pulse; • Heat dissipation; • Volume and weight; • Reliability. In an active phased array antenna system, the DC power can either be distributed throughout the system or be centralized. 6.2.2 Distributed Power System A single DC-to-DC converter feeds a small group of T/R modules or channels in a distributed power system, typically two to eight
7065_Agrawal_V6.indd 130
4/6/23 1:43 PM
Beam Steering and DC Power Distribution
131
[4]. DC-to-DC converters are distributed throughout the array using a high-voltage, low-current DC bus, as shown in Figure 6.3. The DC-to-DC converters often use the same heat sink or cold plate as the T/R modules. The T/R module and DC-to-DC converter assembly becomes the LRU. The DC-to-DC converters generate the three voltages required by the T/R module from the unregulated DC voltage from the AC-to-DC converters. The failure of a few DC-to-DC converters affects a relatively small number of T/R modules when power is distributed, so the impact on the peak and RMS sidelobes of the array is relatively modest, and the array far-field pattern degrades gracefully. The failed DC-to-DC converters are replaced at the next scheduled antenna maintenance event or when the array performance falls below a predefined critical level. Chapter 8 discusses how this built-in redundancy contributes to array reliability. Another way to limit the impact of failed DC-to-DC converters on array reliability is to daisy chain the converters in half columns and provide additional DC-to-DC converters for redundancy. In general, the MTBF of DC-to-DC converters is significantly lower than that of T/R modules; a higher number of the DC-to-DC converters would fail compared to the T/R modules. Therefore, to reduce an array’s maintenance and lifetime cost, its DC-to-DC converters should be accessible so that they can easily be replaced individually without requiring the replacement of a whole LRU.
Figure 6.3 Active phased array distributed power distribution.
7065_Agrawal_V6.indd 131
4/6/23 1:43 PM
132
Practical Aspects of Active Phased Array Antenna Development
The key features of the distributed power system are: • A single DC-to-DC converter feeds a small group of T/R modules (2 to 8). • DC-to-DC converters are distributed throughout the array. • A high voltage, low current bus delivers power from the AC-to-DC converter to the DC-to-DC converters. • Each DC-to-DC converter transforms the high voltage DC to regulated, lower DC voltages required by the T/R modules. • The distributed arrangement supports high clutter improvement (see Chapter 9) since noise from the converters is uncorrelated. • The distributed power system provides high reliability from graceful performance degradation. • The output DC voltage ripple of the DC-to-DC converters is offset by the amount of voltage regulation provided. 6.2.3 Centralized Power System In a centralized power distribution system, a group of DC-to-DC converters is combined to feed a large section of the antenna. Redundant power supplies are included in each group of power supplies to increase the group’s reliability. Figure 6.4 shows the centralized power distribution system. In Figure 6.4, a group of DC-to-DC converters feeds one-quarter of the array elements. A high-voltage, low-current DC bus delivers power to each group of DC-to-DC converters. The low DC voltages are distributed throughout the array, requiring low voltage, high current buses. The DC-to-DC converters convert high voltage DC to the low voltages required by the T/R modules, and voltage regulation is provided to generate voltages with very low ripple. The central power system also provides improved clutter performance because of the uncorrelated random power supply noise generated by multiple converters that are not synchronized. The key features of the centralized power distribution system are: • The voltage outputs of a group of DC-to-DC converters are combined in parallel to feed a large fraction of the array.
7065_Agrawal_V6.indd 132
4/6/23 1:43 PM
Beam Steering and DC Power Distribution
133
Figure 6.4 Active phased array centralized DC power distribution.
• Low voltage, high current buses deliver DC power from DC-to-DC converters to the T/R modules. • Redundancy is provided by including a few spare converters in each group of DC-to-DC converters: • Failures of a few DC-to-DC converters in a group will not result in the failure of a large section of the antenna since each group includes excess capacity. • The number of spare converters depends on the required reliability of the system and the periodic time interval between the replacement of failed converters. • Combining a group of converters provides low phase noise, supporting a high clutter improvement factor (see Chapter 9).
6.2.4 Average versus Peak DC-to-DC Converters The average power converter requires storage capacitors to maintain the desired voltage during the radar transmit pulse without exceeding the required droop over the course of the pulse. As the pulse width increases, the storage capacitance requirement increases, putting a physical limit on the length of the pulse. The other power
7065_Agrawal_V6.indd 133
4/6/23 1:43 PM
134
Practical Aspects of Active Phased Array Antenna Development
conversion approach uses a peak capacitor multiplying DC-to-DC converter. This converter has the unique advantage that the pulse energy can be stored on the converter’s primary high voltage side, resulting in a substantially smaller energy storage capacitor. Both converters are comparable, and the choice of converter depends on pulse width, cost, and volume. 6.2.5 Comparison of Distributed and Centralized Power Systems Both distributed and centralized power systems can be selected for active phased array antennas. The distributed power system provides significant reliability because the failure of a small number of DC-to-DC converters does not result in significant performance degradation of the array; the array would continue to operate until its performance falls below the critical level. In the centrally distributed power system, redundant DC-to-DC converters in each group provide adequate reliability. The major difference between the two distribution systems is the location of the DC-to-DC converters on the array. The DC-to-DC converters can be placed either behind the array or around the array. Placing the DC-to-DC converters behind the array increases the array’s depth. On the other hand, placing the DC-to-DC converters around the antenna increases the antenna size. The decision of where to place the converters often is platform specific. In addition to increasing the array size for the centralized power distribution, the antenna weight and its radar cross-section may increase significantly.
6.3 Concluding Remarks This chapter overviewed BSC and prime power distribution for active phased array antennas. The BSC function is either distributed throughout the array or centrally located. Most of the distributed steering controller calculations are performed at the element level. On the other hand, the centrally located BSC requires high data rates to send commands to all the elements, which is not a major issue for the current data processing speeds of the computers. Similarly, the DC power supplies can either be distributed throughout the array or centrally located in large groups that feed
7065_Agrawal_V6.indd 134
4/6/23 1:43 PM
Beam Steering and DC Power Distribution
135
the large sections of the array. DC-to-DC converters can either be located behind the array or around the array. In most ground-based and shipborne active phased array antennas, the distributed system may be preferable as it does not increase the array aperture size.
References [1] Deluca, A., Gentry, J., Thomas, D., Landry, N., and A. K. Agrawal, “Phased Array Antenna with Distributed Beamsteering,” U.S. Patent 5,339,086, issued August 1994. [2] Waldron, T. P., et al., “Distributed Beam Steering Control of Phased Array Antennas,” Microwave J., September 1986. [3] Gallagher, J. J., et al., “Radar System with Elevation-Responsive PRF Control, Beam Multiplex Control, and Pulse Integration Control Responsive to Azimuth Angle,” U.S. Patent 5,103,233, issued April 7, 1992. [4] Peil, W., “A Power Distribution System for a Phased Array Radar,” U.S. Patent 4,806,937A issued February 21, 1989.
7065_Agrawal_V6.indd 135
4/6/23 1:43 PM
7065_Agrawal_V6.indd 136
4/6/23 1:43 PM
7 Active Phased Array Antenna Packaging 7.1 Introduction Active phased array antennas include a radome, radiating elements, T/R modules, DC-to-DC converters, beamformers, and structure, as shown in Figure 7.1. The predominant packaging design considerations associated with active array antenna packaging include ease of maintenance, thermal management, DC power distribution system packaging, RF beamformers, radiating aperture design/ interface, and structural design. T/R modules generate about 70% of the dissipated heat in most high-power arrays. The T/R modules typically are mounted on air or, more often, liquid-filled metal cold plates, which provide a thermal path to remove the dissipated heat. The temperature across the T/R modules mounted behind the array aperture must be uniform as temperature variations cause the gain and insertion phase of the RF amplifiers to vary, resulting in amplitude and phase errors at the radiating elements. It’s no surprise then that thermal management drives the design of mechanical packaging of components. The array packaging also depends on the platform, whether it be airborne, ground, or shipborne. Environmental requirements such as shock, vibration, external temperature, wind loading, and shielding against external radiation such as EMP from nuclear blasts also significantly impact the packaging design. 137
7065_Agrawal_V6.indd 137
4/6/23 1:43 PM
138
Practical Aspects of Active Phased Array Antenna Development
Figure 7.1 Block diagram of an active phased array antenna.
The array operating frequency is the overriding driver in determining which packaging options are available to the designer. Consequently, packaging becomes more challenging as frequency increases, element spacing decreases, and power density increases. Fortunately for the designer, higher-frequency arrays tend to have more directivity and lower T/R module output power requirements; hence, the worst-case thermal design problems do not always correlate with the highest radiating element densities. The active components (T/R modules and power supplies) will fail during the antenna deployment and must be repaired or replaced. The active array components are packaged into LRUs to facilitate maintenance. The operator in the field should be able to remove LRUs easily and replace them with new LRUs in a relatively short time. The packaging design must incorporate measures that allow LRUs to meet their reliability requirements. Since the MTBF of electronic devices is a function of its junction temperature, LRU design must consider requirements for thermal management. For Silicon and GaAs devices, the maximum recommended junction temperature usually is 125°C, and for GaN, it is 200°C.
7065_Agrawal_V6.indd 138
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
139
Single points of failure that cause a large section of the array to stop operating must be avoided in the antenna architecture and the packaging design. Chapter 8 covers aspects of array reliability in detail. The following parameters impact the active array packaging methods. • Predominant factors in packaging design: • Ease of maintenance; • Radiating aperture design; • Thermal management; • DC power distribution; • RF beamformers; • Structural design; • Impact of frequency of operation: • Element spacing decreases as the frequency is increased, requiring tighter element spacing; • Available RF power decreases with an increase in frequency; • The worst thermal design does not correlate with the worst power density; • Ease of maintenance: • Active electronics configured as LRUs; • LRUs may include T/R modules, DC-to-DC converters, control circuits, and processors such as ASICs or FPGAs; • Fault isolation down to LRU is provided for replacement in case of failure; • Thermal management: • Cooling methods to provide uniform temperature across the aperture; • Maintain junction temperatures of MMIC devices to desired levels; • Tight control of temperature across the aperture is required; • Radome design; • Lifecycle cost; • Platform suitability; • Shock, vibration, thermal load, wind loading, nuclear blast, and so on; • Single points of failure.
7065_Agrawal_V6.indd 139
4/6/23 1:43 PM
140
Practical Aspects of Active Phased Array Antenna Development
7.2 Array Packaging Concepts There are two basic concepts of array packaging architectures: tile and brick, where the names refer primarily to how the main components (T/R modules, beamformers, power supplies, cold plates) are installed [1–5]. Tile arrays install T/R modules and other components in layers that lie parallel to the face of the array, which enables them to be very shallow, as shown in Figure 7.2. In general, the less volume an array requires, the lighter it will be. However, tile array modules are constrained to fit on the array grid, which makes heat transfer challenging with tile arrays because their compact form factor provides less space for thermal management. Consequently, tile array packaging is suited for moderate-power active arrays, such as arrays on airborne platforms, for which space is very limited.
Figure 7.2 Cross-section of the tile array layers.
7065_Agrawal_V6.indd 140
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
141
In a brick array package, the T/R modules, power supplies, and cold plates are installed perpendicular to the radiating aperture. Beamformers can be installed perpendicularly, or they can span the entire back side of the array. These arrays are not depth constrained, so they can be deeper than tile arrays and thus allow more space to power and thermally condition the high-power active phased arrays for ground-based and shipborne applications. 7.2.1 Tile Array Construction and Cooling Methods Tile array packaging is suitable for aircraft antennas because the space behind the array aperture is limited. Because tile arrays install T/R modules parallel to the face of the array, heat transfer is more challenging than brick arrays. The cold plates must also be parallel to the face of the array to make thermal contact with the T/R modules. In addition, all components are assembled in layers connected to the front aperture, as shown in Figure 7.2. Interconnections between many tiles are challenging due to the limited space for components on the individual layers [4]. The output power of the T/R modules is limited to about 10W at X-band because the surface area per element for cooling is limited by the radiator grid spacing. The cold plate is often fabricated from aluminum with internal channels for liquid coolant flow to provide uniform temperature across the cross-section of the cold plate. If the T/R modules and beamformers or radiators are on opposite sides of the cold plate, their RF interconnections will have to pass through the cold plate, complicating the cold plate design to allow the coolant channels in the cold plate to route around the holes passing through. Tile arrays often use printed circuit radiating elements that can be integrated into a multilayer printed circuit board with beamformer, control, and power distribution layers. With radiators on one side, RF transition structures are used to pass RF signals from the other side, where T/R modules connect through the circuit board. 7.2.2 Brick Array Packaging To optimize an array design for ease of maintenance, most active electronics are configured as LRUs to easily remove and replace them in the field. Reliability, cost, and performance impact in the
7065_Agrawal_V6.indd 141
4/6/23 1:43 PM
142
Practical Aspects of Active Phased Array Antenna Development
event of failure determine which and how many components should be included in the T/R LRU assembly. Fault isolation down to at least the LRU level is provided to minimize service time required during maintenance actions. The components to be included in an LRU include T/R modules, control modules, DC-to-DC converters, associated RF, logic, and DC circuits. Structural components, the coolant distribution system, RF beamformers, DC power distribution, and cabling are typically considered sufficiently reliable to be nonrepairable in the field. 7.2.3 Components of an LRU As discussed in Chapter 3, an LRU may contain several T/R modules, a control module, and a DC-to-DC converter to provide the power to the T/R modules. Alternatively, an LRU may only contain T/R modules and control modules with DC-to-DC converters placed around the array, as described in Chapter 6. Figure 7.3 shows
Figure 7.3 T/R LRU configuration: (a) T/R modules, control module, and DC-to-DC converter; (b) T/R LRU.
7065_Agrawal_V6.indd 142
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
143
an LRU configuration containing two T/R modules, one control module, and one DC-to-DC converter. The T/R modules are connected to the RF beamformer. Therefore, LRUs should be configured to remove the T/R module LRU without disconnecting RF cables. Push-on connectors facilitate repeated connect and disconnect actions between T/R modules and beamformers. Since the MTBF of the DC-to-DC converters is usually significantly less than that of the T/R modules and DC-to-DC converters drive multiple modules, the DC-to-DC converters would require more frequent replacement than the T/R modules. A single failure of a T/R module in an LRU would not require its replacement immediately. 7.2.4 Thermal Management The frequency band of the antenna determines the radiating element spacing at the aperture. Since the reliability of active components depends on temperature, thermal design is critical for maintaining the temperatures of array electronics at desired levels to support reliability requirements. In addition, thermal management systems must minimize the temperature gradients across the array to avoid module-to-module phase errors. The T/R modules account for about 70% of the heat generated within an array. Because of ever-increasing power density heat dissipation in modern shipborne and ground-based active phased arrays, liquid or even two-phase [5] cooling is normally required. The predominant cooling techniques in use today employ conduction away from the T/R module into liquid-filled cold plates. Referring back to Chapter 3, Figure 3.13 shows the component stack-up in an LRU. In Figure 3.13 the heat transfer path is down through the various layers from the junction of the integrated circuit to the coolant in the cold plate, with a temperature rise occurring at every interface. As shown in the figure, the high-power amplifier chips are eutectically attached to metal spacers to spread the heat. Thermal management of active array antennas is challenging because most of the dissipated heat is concentrated in a very small region in the power amplifier metal-semiconductor field-effecttransistor (MESFET) junction. As a result, the heat density in Watts per square centimeter can be very high. In addition, one of the main
7065_Agrawal_V6.indd 143
4/6/23 1:43 PM
144
Practical Aspects of Active Phased Array Antenna Development
factors determining long-term array reliability is the operating temperature of its electronic components, which depends primarily on thermal design. Therefore, the maximum RF power that can be transmitted depends on the thermal management subsystem and the array’s capability to remove heat efficiently. Moisture greatly impacts the devices’ reliability, and in most cases, a hermetic sealing may be required for systems that are deployed for long periods. DC-to-DC converters also generate a significant amount of heat, even though their conversion efficiency exceeds 90%. Because T/R modules exhibit RF insertion phase changes with temperature variations, tight thermal control of the modules must be maintained across an array. Phase errors can be actively compensated for if the MMIC temperatures are known. However, it isn’t easy to measure the actual transistor junction temperatures in realtime, report the temperatures to the BSC, and compensate during various operational modes and duty cycles. Therefore, most thermal management subsystem design focuses on maintaining uniform coolant temperatures across the array and calibrating the array at one or more operational temperatures. Maintenance requirements usually dictate that the LRUs must easily be replaceable. In addition, maintenance often must be performed from the back side of the array, so LRUs are inserted and extracted from the back of the array. Therefore, it is imperative that RF beamformers, power distribution systems, control signal distribution systems, and their associated cabling be installed to allow ready access to the LRUs. These passive components, generally high reliability, must be designed to fit between the T/R module LRUs, often requiring blind mate RF connections between LRUs and the radiating elements. Using coaxial cables between the LRUs and the radiating elements complicates the design and adds RF loss at the front end of the array and would affect transmit power and system noise figure. However, cables may be necessary for some applications.
7.3 Active Array Antenna Brick Packaging Schemes This section describes three active array brick packaging schemes using sliding and fixed liquid cold plates.
7065_Agrawal_V6.indd 144
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
145
7.3.1 Sliding Vertical Cold Plate Active Array Packaging T/R modules, control modules, DC-to-DC converters, and logic busses are mounted directly to vacuum-brazed liquid-cooled cold plates, as shown in Figures 7.4 [2] and 7.5 [3]. Each cold plate is a single structure that spans the height of the array aperture. These cold plates serve a dual purpose as an integral part of the antenna structure and manage component temperatures by efficiently removing dissipated heat during the array operation. Each cold plate has components mounted on both sides. Vacuum-brazed cold plates provide the flexibility to direct the coolant flow most efficiently, maximizing the heat removal from the critical high-power components. Vacuum-brazing manufacturing also yields design flexibility in that the brazing operation is fluxless, and there is no requirement to include large voided areas in the cold plates for flux washouts.
Figure 7.4 Sliding vertical cold plate active array antenna architecture [2].
7065_Agrawal_V6.indd 145
4/6/23 1:43 PM
146
Practical Aspects of Active Phased Array Antenna Development
In addition, the cooling channels in the cold plates are designed to achieve uniform temperatures across the cold plate surface. The array packaging design shown in Figure 7.4 uses standard slides mounted to the top and bottom of the cold plates in the array housing to facilitate the removal of the cold plates from the array housing to gain access to all components for repair or replacement. The figure shows a single cold plate in the repair position, pulled back from the antenna housing supported and locked in place. Access to all components mounted on the cold plate is possible, permitting component troubleshooting, repair, or replacement. Each cold plate is connected to the coolant pump with a coolant hose and cabling service loop, so disconnections are not required during normal array maintenance. When the column is inserted back into the array structure, dowel pins assembled into the cold plate front flange help guide the cold plate in position before the radiator probes enter the critical areas of the waveguide elements in the array faceplate. The vertical combiners are single, large contiguous circuits. The vertical combiner network RF interface to the T/R modules uses a microstrip to stripline flexible connection made with a custom-designed RF bellows and outer conductor connection. The bellows are connected to the RF microstrip on the T/R modules with an inherent spring contact. The beamformers are terminated at the end with standard SMA-style connectors to allow connection to the next level column-to-column combining network (horizontal combiners). The advantage of this array assembly approach is that it provides easy access to LRUs without the disconnection of RF, DC, logic, or fluid hoses, and ease of replacing components. Its main drawback is that it cannot be applied to very large apertures, as it is difficult for operators to handle large cold plates. In addition, if the aperture is not rectangular, cold plates of different sizes would be required, as shown in Figure 7.5. 7.3.2 Edge-Cooled, Horizontal Cold Plate Array Packaging Figure 7.6 shows the T/R module LRUs mounted vertically on the liquid-cooled horizontal cold plates. The array design shown in Figure 7.6 uses standard slides on the horizontal cold plates, and the T/R LRUs are inserted perpendicular to the slides. The T/R LRUs are held down to the cold plates using a type of clamp called a
7065_Agrawal_V6.indd 146
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
147
Figure 7.5 Sliding cold plate active array antenna structure with multiple cold plate sizes [3].
wedge lock, which provides a low thermal resistance contact between the base of each LRU and the cold plate surface. As a result, the T/R LRUs contact the horizontal cold plates at the top and bottom of the LRU. The slides on the horizontal cold plates facilitate the removal of the T/R LRUs to gain access to all components for repair or replacement. Each T/R LRU is cooled in this arrangement at the edges of its baseplate. Within the same LRU, the edge cooling results in a small temperature difference between the top, bottom, and central TR modules. This temperature difference can be
Figure 7.6 Horizontal cooling plates edge cooled packaging concept.
7065_Agrawal_V6.indd 147
4/6/23 1:43 PM
148
Practical Aspects of Active Phased Array Antenna Development
measured, and the resulting difference in the amplitude and phase errors of the top, bottom, and central modules can be corrected in the array control software. The inputs and outputs of the T/R modules in a row are combined using a horizontal beamformer attached to the cold plate. The outputs of the row combiners are then combined with the vertical combiners. Vacuum-brazed cold plates provide the flexibility to direct the coolant flow most efficiently, maximizing the heat removal on the critical high-power components. This design is similar to the technique employed in air-transportable rack equipment and allows components to be mounted to both sides of the baseplate. This technique offers a simple, easy-to-maintain packaging design that can accommodate tight element spacing. In addition, the cold plates become part of the antenna structure. 7.3.3 Vertical Fixed Cold Plate Packaging Concept In the configuration shown in Figure 7.7, the T/R modules are mounted on both sides of an LRU baseplate. The LRUs are mounted on a large vertical, fixed cold plate using wedge locks to press the baseplate against the cold plate. This configuration provides a large contact area to the cold plate and allows easy access to the LRUs; however, it tends to be better suited to larger element spacing, which allows LRU attachment by an insertion on one side. The primary difficulty in this approach comes from the LRU-to-cold plate
Figure 7.7 Vertical fixed cold plate packaging concept.
7065_Agrawal_V6.indd 148
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
149
interface. The interface must be free of thermal interface material to accommodate sliding insertion and extraction readily. Consequently, a good thermal interface between the LRU and cold plate must be achieved with tight tolerancing of the separation between both surfaces, making repeatable thermal resistance from module to module difficult to achieve. Alternatively, a phase-change–type interface material that would require heating the interface prior to extraction could be employed [5]. Since the cold plates are fixed and are part of the antenna structure, this packaging scheme can be used for larger antennas.
7.4 LRU to the Radiating Element RF Connections LRUs either must contain the radiating elements or must blind mate to the radiating elements on the array face. The latter is typically the case when the radiating elements are machined out of a faceplate to maintain tight element location tolerances. Overall mechanical alignment between the LRU and the radiating element requires tight control of large-tolerance stack-ups, and the resulting misalignment must be absorbed in the RF connection. There can easily be a dozen or more different interface tolerances, which combine to determine the radial float required in the RF connector. Because most floating RF connectors offer no more than 0.030 inches of radial float, it is necessary to use a combination of tight tolerances, statistical tolerancing methods, and occasionally specialized assembly fixtures to ensure proper alignment of RF connections between LRUs and radiators. Figure 7.8 shows a T/R LRU with attached radiating elements that can be inserted into the antenna structure.
Figure 7.8 T/R LRU with attached radiating elements.
7065_Agrawal_V6.indd 149
4/6/23 1:43 PM
150
Practical Aspects of Active Phased Array Antenna Development
7.5 Structural Design Structural design is a significant issue for shipboard active array radars [1]. Traditionally the military standard shipboard shock requirement (Mil-Std-901) has been the predominant structural design driver. For most array systems, floating platform barge testing is required to validate hardware compliance. This test imparts a shock pulse on the order of 70G at 11–14 Hz. A secondary design consideration involves maintaining adequate stiffness in the radiating aperture to ensure that the array surface remains flat during operational sea states and shipboard structural vibrations. This flexural stiffness directly contributes to the array error budget. In addition, error budgets must account for large-scale deflections of the ship’s structural movement, such as deckhouse or mast deflections. Maintaining adequate flatness, particularly in larger arrays, becomes increasingly difficult. A newer technique involves realtime, active measurement, and compensation for deflection.
7.6 Active Array Antenna Radome Design Phased array antennas use a radome to protect the radiating aperture from rain, wind, snow, and ice. Chapter 5 discussed using a radome with heating wires for ice inhibition. The radome is made from low RF loss and mechanically robust dielectric materials. Earlier in this chapter, we discussed techniques to remove the heat dissipated by the T/R modules. However, for high-power active phased array antennas, dissipated power in the radome can increase the radome temperature affecting the antenna aperture temperature and IR signatures. Therefore, the radome may have to be cooled to the ambient air temperature. In addition, forced air directed between the array aperture and the radome can cool the radome to the ambient temperature. In Chapter 5, we discussed waveguide radiating elements. Ceramic windows, one placed in front of each waveguide radiating aperture, serve as radomes for the waveguide elements so that no external radome over the entire array is required. In addition, since the ceramic widows become part of the antenna structure, cooling the ceramic windows is not required.
7065_Agrawal_V6.indd 150
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
151
Figure 7.9 shows the cross-section of a flat radome mounted on an array antenna frame. The antenna frame’s height and placement should be such that it does not obstruct the array radiation to the maximum scan angle of the array, typically up to 60°. As shown in Figure 7.9, a simple geometric calculation provides a design guideline. Air for radome temperature management is forced into the space between the array antenna and the radome through inlets and outlets. Figure 7.10 shows the integration of the radome with the antenna.
Figure 7.9 Active phased array radome cross-section of a flat radome.
Figure 7.10 Radome integration with the active phased array antenna flat radome.
7065_Agrawal_V6.indd 151
4/6/23 1:43 PM
152
Practical Aspects of Active Phased Array Antenna Development
Figure 7.11 shows the cross-section of an active array antenna with a curved fabric radome. The fabric radome can be installed under tension to reduce the radome movement from the wind and rain. RAYDEL® M26, produced by Saint Gobain, is a fabric composite designed specifically for use in RF applications. The composite is made of fiberglass and polytetrafluoroethylene (PTFE). The “M” group incorporates a flat, plain weave in the structural substrate. RAYDEL® has the following properties: • Maintenance free. RAYDEL® is self-cleaning due to its Teflon® surface. Normally the radome never requires maintenance, a critical factor on remote, tower-mounted installations. • Hydrophobic. Water will not film on Teflon®, maintaining a clear signal through the membrane even during heavy rains. This is becoming increasingly more important as communications equipment operates at higher frequencies. • Resists ice buildup. Ice has low adhesion to Teflon® so that low RF loss can be maintained, even during icing conditions. • Strength. RAYDEL® M26 has a tensile strength of 300 pounds per inch, far exceeding tent materials. • Incombustible. RAYDEL® fabric carries a Class-A fire rating and qualifies for use in all structures.
Figure 7.11 Active phased array cross-section with a curved, fabric radome.
7065_Agrawal_V6.indd 152
4/6/23 1:43 PM
Active Phased Array Antenna Packaging
153
7.7 Concluding Remarks This chapter overviewed the key parameters that dictate the packaging of active antenna components. In addition, cooling is critical for the thermal management of dissipated heat from T/R modules to control the operating temperatures of the solid-state devices for high reliability and maintain a uniform temperature across the antenna aperture. Due to the close spacing of array radiating elements, the heat load of most arrays is concentrated in a small area. Active arrays often use liquid-filled cold plates to remove the heat from the LRUs. One of the key packaging requirements is that the array components are easily replaceable in the field without many interface disconnections. Several packaging concepts for the tile and brick array architectures were presented. The extreme shock and vibration requirements of naval arrays can make the structural design more challenging than for ground-based radar. An array structure design for cooling the radome was also presented.
References [1] Agrawal, A. K., et al., “Active Phased Array Antenna Development for Modern Shipboard Radar Systems,” Johns Hopkins University Applied Physics Laboratory Technical Digest, Vol. 22, No. 4, October–December 2001. [2] Mattioli, J., Agrawal, A. K., and N. Landry, “Packaging Technique for Phased Array Antennas,” U.S. Patent 5,459,474, issued October 17, 1995. [3] Agrawal, A. K. K., Mattioli, J. A., and N. R. Landry, “An Active Phased Array Antenna Packaging Scheme,” Antennas and Propagation Soc. Int. Symp., AP-S, Digest, Vol. 3, 1996, pp. 1608–1611. [4] Carlson, D., “Tile Arrays Accelerate the Evolution to Next-Generation Radar,” Microwave J., January 13, 2017. [5] van Es, J., and H. J. van Gerner, “Benefits and Drawbacks of Using Two-Phase Cooling Technologies in Military Platforms,” Electronics Cooling, Vol. 19, No. 1, March 8, 2013.
7065_Agrawal_V6.indd 153
4/6/23 1:43 PM
7065_Agrawal_V6.indd 154
4/6/23 1:43 PM
8 Active Phased Array Antenna Design for High Reliability 8.1 Introduction Phased array antenna reliability is defined as antenna availability during the deployment period. We show how to consider fault tolerance in the design of the antenna architecture, so that replacement of failed components can be avoided for an extended period. First, we discuss active phased array architecture design for maximizing antenna MTBF. Then, we address the dependence of antenna lifecycle cost on component failure rates. The antenna MTBF is defined as when the receive peak sidelobes degrade by 3 dB [1]. The impact of a small number of failed elements on the RMS sidelobes is small. Although a small number of failed components can degrade the receive sidelobe level, the loss of the antenna transmit and receive gain and transmit sidelobes (uniform amplitude distribution) is negligible. Therefore, we can broadly say that an active phased array antenna has high reliability if its MTBF exceeds the deployment time. The choice of antenna architecture, including RF beamforming, power conversion and distribution, and beam steering control distribution, has a significant impact on the antenna MTBF, antenna lifecycle costs, and spare requirements. For example, antennas with large-element cluster beamformer architecture would require frequent maintenance and many spare parts during 155
7065_Agrawal_V6.indd 155
4/6/23 1:43 PM
156
Practical Aspects of Active Phased Array Antenna Development
the deployment period. The MTBF of the antenna can be made larger than its deployment period by using small element clusters and providing redundancy [1]. Another parameter, the frequency of repair, needs to be specified, for it impacts the repair and lifecycle costs. Phased array antennas fail gracefully, and maintenance is required when its performance falls below a defined threshold. Therefore, delaying maintenance until the antenna performance falls below that threshold, instead of repairing all failed components at the end of each mission or at fixed intervals to restore full performance, would result in increased system availability, reduced frequency of repairs, reduced labor hours, and a significant reduction in the lifecycle maintenance cost. The key enabler to implementing the reduced maintenance during mission concept is a built-in automated mechanism that assesses the system’s ability to perform each critical mission successfully using the system’s health map. The automated mechanism continuously monitors the numbers and types of failed components and predicts the antenna performance compared to the allowed threshold.
8.2 Antenna MTBF Solid-state T/R module technology brings many performance advantages to the phased array antenna, such as higher average transmit power and efficiency, lower system noise temperature, higher moving target indicator (MTI) performance, and broader bandwidth. However, as described in Chapter 7, an active phased array’s mechanical packaging is more complex than its passive predecessor since the transmit and receive functions are distributed at the aperture. Consequently, the design of active arrays often involves a trade-off between lifecycle cost, packaging, and performance. Another crucial parameter that must be considered when designing an active phased array antenna is reliability. Because active arrays often contain thousands of active components and must operate for long periods, we focus on reliability and its influence on active array design. This section examines the impact of component failure rates on active array lifecycle cost. Next, we discuss the design of a phased
7065_Agrawal_V6.indd 156
4/6/23 1:43 PM
Active Phased Array Antenna Design for High Reliability
157
array to minimize its failure rate. When components fail, the degradation in array performance often is measured as an increase in peak sidelobe level. So, eventually, after the array performance falls below a specified level of acceptability, the entire active array is considered failed and may need to shut down unless the array design allows the replacement of components while the antenna is operating. For a phased array deployed in the field for several decades, the lifecycle cost can be reduced to a minimum if the array can be designed so that its mean time between critical failures (MTBCF) exceeds the deployment time. Such a design minimizes the number of repairs required during deployment and reduces the number of spare components stored at the array location. For example, if the MTBCF is greater than the ship deployment time for shipborne antennas, the maintenance can be done more efficiently while the ship is in port instead of at sea. Our goal is to maximize the time between system shutdowns for maintenance by increasing the antenna MTBCF. This goal can be accomplished by minimizing the antenna critical failure rate by judiciously selecting the antenna architecture. We present simulated data showing the effect of random, single, and clustered element failures on the peak sidelobe level of a low sidelobe array antenna. From these data, it is evident that designing a phased array antenna that minimizes failures of large clusters of elements will significantly increase its MTBCF. The availability A0 of a system is defined as: A0 =
System MTBCF System MTBCF + System MTTR (8.1)
MTTR is the mean time to repair or replace components sufficient to return the antenna performance to an acceptable level. The antenna availability A0 is a function of both the MTBCF and the MTTR. Our goal is to achieve an availability as close to 1 as possible. The following section discusses array architecture designs that result in high reliability (availability). Not all failures are critical failures. Some component failures degrade the system performance but not below the required level. The system continues to operate with slightly degraded but acceptable performance adequate to complete its mission. A single point
7065_Agrawal_V6.indd 157
4/6/23 1:43 PM
158
Practical Aspects of Active Phased Array Antenna Development
of failure often is a critical failure that requires immediate repair. We want to minimize the number of single points of failure in a system.
8.3 Active Phased Array Antenna Architecture Description for High Reliability Figure 8.1 shows a typical block diagram of a T/R module [2]. Each T/R module contains a transmit path and a receive path. The transmit path consists of a phase shifter, a variable gain amplifier or attenuator, a driver amplifier, and a power amplifier. The power amplifier section may consist of several power amplifiers, typically two or four, where the output power of these amplifiers is combined to obtain the required output power for one radiating element. A circulator is used for duplexing and isolating the transmit and receive paths, including preventing power amplifier load pulling due to variations in the antenna element’s input impedance during beam scanning. The receive path consists of a limiter, a low noise amplifier, a phase shifter, and a variable gain amplifier or attenuator. In Chapter 1, as shown in Figure 1.6, RF architecture includes an active phased array antenna [3]. An active phased array system comprises radiating elements, radome and structure, T/R modules, associated control circuitry, RF beamformers, DC power distribution, DC-to-DC converters, and a BSC. Most active array electronics are configured as LRUs to optimize the array assembly, reduce cost, and ease maintenance. The LRU is the lowest level assembly that can be replaced during routine maintenance. An LRU can include T/R modules, control modules, DC-to-DC power converters, and various control/processor assemblies. For example, Figure 8.2(a) shows an LRU with two T/R modules driven by a single control module and a single power supply, and Figure 8.2(b) shows an LRU with four T/R modules driven by a single control module and a single power supply. Reliability and the magnitude of the system performance degradation in the event of failures determine the number of components in an LRU. For example, the number of T/R modules in an LRU directly impacts the antenna MTBF. In addition, the cost of components also determines how many are included within an LRU.
7065_Agrawal_V6.indd 158
4/6/23 1:43 PM
Active Phased Array Antenna Design for High Reliability
159
Figure 8.1 RF block diagram of a T/R module.
Fault detection/fault isolation (FD/FI) is provided at the LRU level to minimize the service time required during maintenance actions. Phased array antennas contain two types of components: (1) active components, which require DC bias, such as transistors and switches, and (2) passive components, such as transmission lines and power dividers, which do not require DC bias. The MTBF of a passive component is very high, and here we assume it to be infinite, so it does not contribute to the antenna MTBF.
Figure 8.2 T/R module LRUs: (a) LRU with two T/R modules; (b) LRU with four T/R modules.
7065_Agrawal_V6.indd 159
4/6/23 1:43 PM
160
Practical Aspects of Active Phased Array Antenna Development
8.4 Maximizing the Array MTBCF Over time, as an array’s components fail, the array’s performance degrades. Eventually, after the performance falls below a specified level of acceptability, the array fails and must be repaired. As defined in the previous section, the reliability of an array is defined by its MTBCF (i.e., the time between major or critical failures that cause its performance to fall below the acceptable level). This degradation in performance is generally defined in terms of the rise in peak sidelobes from the nominal level, the level without any component failures. We can calculate the MTBCF of an active phased array antenna based on this definition, given the array component types, quantities, and their MTBFs. In addition, we must know the antenna performance degradation per failure for each component type and the antenna’s minimum acceptable level of performance. Therefore, the reliability block diagram shown in Figure 8.3 includes
Figure 8.3 Reliability model of an active phased array antenna.
7065_Agrawal_V6.indd 160
4/6/23 1:43 PM
Active Phased Array Antenna Design for High Reliability
161
only the array components with finite MTBFs. The T/R module settings are commanded by the control module, which houses circuits such as drain switches, voltage regulators, digital controls, and memory chips. Although there is one T/R module per radiating element in this example, the control module and power supply often are configured to drive a group of more than one T/R module, called a cluster. For example, a single LRU per T/R channel, containing one power supply, one control module, and one T/R module, would have the maximum MTBF per channel but at the highest cost and packaging complexity since there would be the minimum number of shared parts within each control module and power supply. In general, increasing cluster size reduces the production cost but may have a bigger impact on the sidelobe level. Now, let’s consider how we define performance degradation with failed components in terms of array receive peak sidelobe levels. Table 8.1 shows the peak sidelobe level increase as a function of the number of failed elements for an 8,000-element circular array with a 40 dB Taylor amplitude taper. The element lattice is triangular with about half-wavelength spacing. Table 8.1 gives the percentage of failed elements allowed for a 3 dB and 6 dB increase in peak sidelobes versus the cluster size of one, two, four, and eight elements. The data in Table 8.1 is obtained through a Monte Carlo analysis, calculating the antenna patterns for an aperture with a different-sized cluster of failed elements removed from locations chosen randomly. The peak sidelobe degradation was determined for each trial as a function of cluster size. The results in Table 8.1 could be used as a rule of thumb for other arrays of comparable size or Table 8.1 Allowed Fractions of Element Failures for 3 dB and 6 dB Increases in Receive Peak Sidelobe Level as Function of Cluster Size for an 8,000-Element Circular Array with a Taylor 40 dB Amplitude Taper and a Triangular Element Lattice, Half-Wavelength Element Spacing
7065_Agrawal_V6.indd 161
Cluster Size
No. of Failures for 3 dB Sidelobe Increase
No. of Failures for 6 dB Sidelobe Increase
1
3.20%
6.50%
2
2.20%
4.40%
4
1.00%
2.60%
8
0.55%
1.10%
4/6/23 1:43 PM
162
Practical Aspects of Active Phased Array Antenna Development
larger. We can expect the fractions to decrease if the cluster sizes remain the same for smaller arrays since those clusters become a larger fraction of the total array size. Note that the data in Table 8.1 are independent of frequency. Although the 3-dB sidelobe increase was selected as the maximum acceptable performance degradation, the number of failures for the 6-dB increase is provided as a reference. Note that the number of allowed failures decreases with the increase in the cluster size. For a single component type, we can calculate the array MTBCF due to that component alone from the following equation:
MTBCFArray = F ⋅ MTBF/N (8.2) where N is the quantity of that component type in the array, MTBF is that component type’s MTBF, meaning the mean time before that component will fail critically, and F is the maximum allowable number of failures permitted for that component, as determined from Table 8.1. When there is more than one component type, as shown in Figure 8.3, the array MTBCF can be determined by paralleling the contributions from each component, as shown in the following:
MTBFAnt = [(FT/R ⋅ MTBFT/R/NT/R)–1 + (8.3) (FCM ⋅ MTBFCM/NCM)–1 + (FPS ⋅ MTBFPS/NPS)–1]–1 Where FT/R, FCM, and FPS are the maximum allowable number of failures for the T/R module, control module, and power supply, respectively. For (8.3) to apply, we assume that failure types independently increase the peak sidelobe levels for a relatively small number of failures. This equation may or may not be valid depending on the architecture of the array reliability block diagram. Other equations for other types of reliability diagrams are described in [4].
8.5 Antenna MTBF for Different Cluster Sizes This section provides guidelines for selecting an active array architecture that maximizes the array MTBCF. We begin with the phased array antenna architecture in Figure 8.4, in which each control
7065_Agrawal_V6.indd 162
4/6/23 1:43 PM
Active Phased Array Antenna Design for High Reliability
163
Figure 8.4 Reliability block diagram of the antenna with 8,000 T/R modules, 1,000 control modules, and 1,000 power supplies.
module and power supply independently drives eight T/R modules. For the 8,000-element array with the component reliability characteristics described in Table 8.2, one T/R module will fail every 125 hours, one control module will fail every 1,000 hours, and a power supply will fail every 200 hours, as shown in Table 8.2. We have chosen values for MTBFs of T/R modules, control modules, and power supplies that represent the numbers achievable with the current technologies; different MTBFs for these devices will result in a different array MTBCFs. In addition, the complexity of components would reduce their MTBFs. It is assumed here that the only types of active components in the array are the three components listed in Table 8.2.
7065_Agrawal_V6.indd 163
4/6/23 1:43 PM
164
Practical Aspects of Active Phased Array Antenna Development Table 8.2 Eight-Element Clusters: Calculation of Array MTBCF for 3 dB Rise in Peak Sidelobes (Assuming the Allowances of Table 8.1)
Failure Rate (Per hr)
Component Failures Allowed
Contribution to Array MTBCF
Cumulative Array MTBCF
Component
Quantity
MTBF (Khrs)
T/R module
8,000
1,000
0.008
256
32,000
32,000
Control module
1,000
1,000
0.001
5
5,000
4,324
Power supply
1,000
200
0.005
5
1,000
812
The column in Table 8.2 called “Contribution to Array MTBCF” shows the contribution to the MTBCF of the array for a 3-dB increase in peak sidelobes from that component alone, as calculated from (8.2). The column titled “Cumulative Array MTBCF” shows the calculation of the array MTBCF from (8.3). From the third row, we see that the aggregate MTBCF of the array is only 812 hours (i.e., the array must be shut down approximately once in approximately four weeks to replace failed components). If we only used T/R modules in the array, the array MTBCF would be 32,000 hours. Similarly, the MTBCF, due to the power supplies alone, is only 1,000 hours; for the control modules, it is only 5,000 hours. Control modules and power supplies drive the array MTBCF lower, and power supplies, which often have the lowest MTBF, determine the antenna MTBCF. Therefore, in this example, the power supplies dominate the array MTBCF. How can we increase the array MTBCF? We can increase it by (1) increasing the MTBF of the individual components, (2) reducing the number of T/R modules per control module and power supply, in other words, reducing the size of clusters, (3) adding redundancy in the reliability diagram, or (4) allowing the acceptable sidelobe level to increase. If we do not increase the MTBFs of the individual components, then the highest MTBCF can be achieved using smaller clusters. We will maximize the array MTBCF, if each element is driven by a single T/R module, control module, and power supply, as shown in Figure 8.5. Table 8.3 shows these calculations. We have increased the array MTBCF significantly to 4,571 hours or just over six months. However, this improvement has come at a higher production cost since we increased the number of LRUs in
7065_Agrawal_V6.indd 164
4/6/23 1:43 PM
165
Active Phased Array Antenna Design for High Reliability
Figure 8.5 Reliability block diagram for an array with no clusters (cluster size equals one).
the array. Tables 8.4 and 8.5 show MTBCF calculations for four-element and two-element clusters. Table 8.6 summarizes the MTBCF data for various cluster sizes. The MTBCF calculations for different MTBFs of control modules and power supplies are shown in Tables 8.7 and 8.8. Table 8.3 No Cluster Design: Calculation of Array MTBCF for 3 dB Rise in Peak Sidelobes Failure Rate per Hour
Component Failures Allowed
Contribution to Array MTBCF
Cumulative Array MTBCF
Component
Quantity
MTBF (Khrs)
T/R module
8,000
1,000
0.008
256
32,000
32,000
Control module
8,000
1,000
0.008
256
32,000
16,000
Power supply
8,000
200
0.040
256
6,400
4,571
Table 8.4 Four-Element Clusters: Calculation of Array MTBCF for 3 dB Rise in Peak Sidelobes
Component
Quantity
MTBF (Khrs)
Failure Rate per Hour
Component Failures Allowed
Contribution to Array MTBCF
Cumulative Array MTBCF
T/R module
8,000
1,000
0.008
256
32,000
32,000
Control module
2,000
1,000
0.002
20
10,000
7,619
Power supply
2,000
200
0.01
20
2,000
1,584
7065_Agrawal_V6.indd 165
4/6/23 1:43 PM
166
Practical Aspects of Active Phased Array Antenna Development Table 8.5 Two-Element Clusters: Calculation of Array MTBCF for 3 dB Rise in Peak Sidelobes Failure Rate per Hour
Component Failures Allowed
Contribution to Cumulative Array MTBCF Array MTBCF
Component
Quantity
MTBF (Khrs)
T/R module
8,000
1,000
0.008
256
32,000
Control module
4,000
1,000
0.004
80
20,000
12,308
Power supply
4,000
200
0.020
80
4,000
3,019
32,000
By increasing the array MTBCF, the repair frequency can be reduced so that maintenance can be performed at the end of a mission. Tables 8.6, 8.7, and 8.8 show that the array MTBCF is inversely proportional to the cluster size, and the lowest MTBF of the power supply has the biggest impact on the MTBCF. However, the small cluster size requires more parts, increases the antenna’s cost, and increases the complexity of the array architecture. Therefore, the cluster size should be optimized so that the array MTBCF is sufficient to meet the requirement at the lowest associated cost.
8.6 Increasing Array MTBCF with Redundant Power Supplies Redundancy cannot be provided in the T/R module since that would require a redundant T/R module (including the control module) for each T/R module or a cluster of T/R modules (LRU), Table 8.6 Comparison of Array MTBCFs for Eight-, Four-, Two-, and One-Element Clusters for 3 dB Sidelobe Increase for Control Module (MTBF = 1,000 Khrs and PS MTBF 200 Khrs) Cluster Size
No. of Array Elements
T/R Modules
Control Modules
Power Supplies
Array MTBCF
8
8,000
8,000
1,000
1,000
812
4
8,000
8,000
2,000
2,000
1,584
2
8,000
8,000
4,000
4,000
3,019
1
8,000
8,000
8,000
8,000
4,571
7065_Agrawal_V6.indd 166
4/6/23 1:43 PM
167
Active Phased Array Antenna Design for High Reliability Table 8.7 Comparison of Array MTBCFs for Eight-, Four-, Two-, and One-Element Clusters for 3 dB Rise in Peak (Sidelobes) Control (MTBF = 500 Khrs and PS MTBF 200 Khrs) Cluster Size
No. of Array Elements
T/R Modules
Control Modules
Power Supplies
Array MTBCF
8
8,000
8,000
1,000
1,000
699
4
8,000
8,000
2,000
2,000
1,368
2
8,000
8,000
4,000
4,000
3,019
1
8,000
8,000
8,000
8,000
4,000
as would increase the number of T/R assemblies by a factor of 2. In addition, it would be nearly impossible to package additional modules. However, redundancy can be provided in the power supplies as they are in a separate assembly. Since power supplies often have the lowest MTBF, adding redundancy can increase the array MTBCF significantly. Power supplies (DC-to-DC converters) can typically be divided into high current or DC power and low current types. High-power DC-to-DC converters usually are required to bias the drains of highpower RF amplifiers, which consume the bulk of the T/R module’s power. The low-power DC-to-DC converters are required for gate voltages and the drain voltages of lower-power amplifiers, such as the low-noise amplifiers and switches. The MTBF of a low-power power supply is commonly significantly greater than that of a highpower supply. Furthermore, by splitting the power supplies into
Table 8.8 Comparison of Array MTBCFs for Eight-, Four-, Two-, and One-Element Clusters for 3 dB Rise in Peak (Sidelobes) Control (MTBF = 500 Khrs, PS MTBF = 500 KHrs) Cluster Size
No. of Array Elements
T/R Modules
Control Modules
Power Supplies
Array MTBCF
8
8,000
8,000
1,000
1,000
1,203
4
8,000
8,000
2,000
2,000
2,319
2
8,000
8,000
4,000
4,000
4,324
1
8,000
8,000
8,000
8,000
8,000
7065_Agrawal_V6.indd 167
4/6/23 1:43 PM
168
Practical Aspects of Active Phased Array Antenna Development
high- and low-voltage power supplies, redundancy needs to be provided only in the low power supplies since the loss of a small number of the transmit elements has a negligible effect on the transmit gain and sidelobes. For example, the power supplies in a half-column are daisy-chained, and secondary power supplies are added for redundancy (see Figure 8.6). As a result, the array MTBCF can be increased to more than six months. For example, for shipboard phased array antennas, the maintenance could be deferred if the antenna MTBF is greater than the deployment period until the ship goes to a port or depot. The next generation of Navy ships, such as DDG-1000 and CVN-78 Carrier, have significantly fewer personnel onboard the ships. If the MTBF of the antenna can be made larger than the deployment period by providing redundancy, the need for maintenance personnel can be reduced significantly. Similarly, the number of spare parts to be carried onboard could be reduced considerably or even eliminated. Similarly, the frequency of repair can be minimized for the ground-based radars deployed in remote locations.
Figure 8.6 Redundant power supplies in half columns.
7065_Agrawal_V6.indd 168
4/6/23 1:43 PM
Active Phased Array Antenna Design for High Reliability
169
8.7 Driver Amplifier Boosters in the Active Phased Array Beamformers The output power of the exciter that generates the radar waveform usually is insufficient to overcome the transmit beamformer splitting and dissipative losses and provide the input RF power required to drive the T/R modules into compression. Therefore, RF power amplification is required in most active array beamforming networks. To accomplish this, the array is divided into N subarrays, and a driver amplifier is used to drive each subarray, as shown in Figure 8.7. The T/R module drivers provide amplification in both the transmit and receive modes. However, since each T/R module driver in the beamforming network provides RF power to a large number of T/R modules in the subarray, it becomes a single point of failure of the group of modules, resulting in unacceptable degradation in the array performance. To avoid failure of a group of many array elements, the exciter output power can be first divided into N plus one (or more redundant), where N is the number of subarrays, as shown in Figure 8.8. Then, the output power of N plus one driver is combined passively. The combined output is then divided into N outputs for N subarrays groups of T/R modules, as shown in Figure 8.8. The redundant
Figure 8.7 Active array with T/R module used as transmit drivers in the beamforming networks.
7065_Agrawal_V6.indd 169
4/6/23 1:43 PM
170
Practical Aspects of Active Phased Array Antenna Development
Figure 8.8 Active array T/R module driver amplifiers configuration for high reliability.
T/R module driver(s) can fail before the total driver power of the groups of T/R modules is no longer sufficient. The number of redundant amplifiers can be greater than one depending on the size of the subarrays.
8.8 Lifecycle Maintenance Cost Estimation of an Active Phased Array Antenna The lifecycle cost of an active phased array antenna consists of two parts: (1) the initial production cost, which includes all items related to the design, purchase, and fabrication of parts, assembly, and testing of the antenna; and (2) the maintenance costs over an antenna’s operational lifetime to replace or repair all failed components in the antenna. An antenna that has been in service for 40 to 50 years could have a maintenance cost over its operational period that exceeds its initial production cost. Given the MTBF of a component, we can calculate the contribution of that component to the array maintenance cost (PLCC) using the following:
7065_Agrawal_V6.indd 170
4/6/23 1:43 PM
171
Active Phased Array Antenna Design for High Reliability
PLCC = N ⋅ T ⋅ C/MTBF (8.4)
where N is the quantity of that component in the array, T is the duration of the array lifecycle, and C is the cost to repair or replace the component. We can lower an array’s lifecycle cost by reducing component cost or increasing the component MTBF. The lifecycle maintenance cost of the array architecture shown in Figure 8.5 can be calculated from the following: Array Lifecycle Cost = T ⋅ (NTR ⋅ CTR/MTBFTR (8.5) + NCM ⋅ CCM/MTBFCM + NPS ⋅ CPS/MTBFPS)
For example, consider four-element clusters array architecture in Table 8.4, which shows a portion of an array with 8,000 elements, 8,000 T/R modules, 2,000 control modules, and 2,000 power supplies. The allowed number of failures can be determined from Table 8.2. We assume that the T/R module has an MTBF of 1,000 hours, a control module with an MTBF of 1,000 hours, and a power supply with an MTBF of 2,000 hours. The cost to repair or replace components per year is shown in Table 8.9. The cost to maintain an active phased array antenna over its entire lifetime is the per-year maintenance cost multiplied by the array’s lifetime in years. For example, from Table 8.9, for an antenna with a 40-year lifetime, we multiply the number of failed components by 40, resulting in the failure of 2,800 T/R modules, 720 control modules, and 3,520 power supplies. Notice that the number of failed power supplies exceeds the total number in the array, meaning that some supplies are replaced more than once. The array maintenance cost can be reduced if components can be designed for higher MTBF. However, designing a component for Table 8.9 Maintenance Cost per Year for Four-Element Clusters Architecture Failure Rate per Year
Failure Rate per Year
Cost to Replace
Cost to Repair
Array Production Cost
Component
Quantity
MTBF (Khrs)
T/R module
8,000
1,000
0.008
70
70 ⋅ CTR
35 ⋅ CTR
8,000 ⋅ CTR
Control module
2,000
1,000
0.002
18
18 ⋅ CCM
9 ⋅ CCM
2,000 ⋅ CCM
Power supply
2,000
200
0.01
88
88 ⋅ CPS
44 ⋅ CPS
2,000 ⋅ CPS
7065_Agrawal_V6.indd 171
4/6/23 1:43 PM
172
Practical Aspects of Active Phased Array Antenna Development
higher MTBF may increase the manufacturing cost of that component. The number of a particular component that fails per year is inversely proportional to the MTBF of that component. Moreover, the antenna architecture determines how often the components need to be replaced per year. For example, for the no-cluster design, failed components could be replaced or repaired at the depot after each six-month deployment, while for the eight-element cluster design, components must be repaired five times during six months of deployment. The critical reliability parameter that drives maintenance cost is the component’s MTBF. Phased array antennas contain two types of components: (1) active components, which require DC bias, such as transistors and switches; and (2) passive components, such as transmission lines and power dividers, which do not require DC bias. Typically, the MTBF of a passive component is very high, and here we have assumed it is infinite so that its contribution to the lifecycle cost is negligible. However, passive components’ contributions can easily be included in our analysis. On the other hand, the MTBFs of active components such as T/R modules and power supplies are finite; these two components drive the lifecycle cost of an array. Here we only consider the cost of replacing the T/R modules and power supplies, as they represent most of the maintenance costs of active phased array antenna components. Delaying maintenance until the array performance falls below a threshold, compared to repairing all failed components at periodic intervals to restore full performance, would result in increased system availability, reduced frequency of repairs, and reduced work hours, resulting in a significant reduction in the lifecycle maintenance cost. For example, if we repair all failed components at fixed intervals, we might be repairing a large number of T/R assemblies, even though a single channel failure in the T/R module assembly does not cause significant sidelobe degradation. However, if an array comprises many four-channel LRUs, and there are single failures in one hundred of those LRUs, we would need to repair or replace 400 channels, even though only 100 had failed. Similarly, the cost of manufacturing small numbers of T/R modules in several batches or lots would be higher than the production cost of manufacturing the same number of T/R modules in a single lot. The key enabler to implementing the reduced maintenance concept is a built-in automated mechanism called a FD/FI subsystem
7065_Agrawal_V6.indd 172
4/6/23 1:43 PM
Active Phased Array Antenna Design for High Reliability
173
that assesses the system’s ability to perform each required operation mode successfully. The array system continuously monitors the failed components through built-in test (BIT) methods such as mutual coupling and drain current monitoring, from which it creates a health map of its components. The health map describes which components have failed. The array performance periodically is calculated using the health map. Combining the known amplitude and phase weighting across the array with the locations of the failed elements, performance can be calculated using far-field radiation pattern software. The decision as to when to perform maintenance would be embedded in the software and automatically reported to the system operator on board the ship.
8.9 Active Phased Array Antenna Availability and Sparing We can compare different array designs by calculating the availability of each array and the number of spare components required during a deployment period. Safe replacement of active components often requires shutting down a major portion of the array. Therefore, we want to maximize the length of time between shutdowns. An antenna’s availability describes the fraction of time it is operational during a given period and is defined by (8.1). Let us return to our four array architectures: (1) eight-element cluster design (Table 8.2), (2) four-element cluster design (Table 8.4), (3) two-element cluster design (Table 8.5), and (4) no-cluster design (Table 8.3). Consider the MTBFs for 3 dB peak sidelobe degradation shown in Table 8.6 for various array architectures. Consider an array deployment period of 6 months or 4,320 hours. The eight- element cluster design will require the replacement of components on average every 812 hours; the four-element cluster design will require the replacement of components on average of 1,584 hours; the two-element design will require the replacement of components on average every 3,019 hours; and the no-cluster design can wait until after the deployment period is over. For the eight-element cluster design, replacement of failed components will be required six times during the deployment period. On average, 7 T/R modules, two control modules, and nine power supplies will be replaced during each maintenance period. Therefore, if the MTTR to replace all failed components is eight hours for
7065_Agrawal_V6.indd 173
4/6/23 1:43 PM
174
Practical Aspects of Active Phased Array Antenna Development
each occurrence, the array availability would be 0.98. Similarly, for the four-element cluster design, replacement of failed components will be required three times during the deploying period, and the array availability will be 0.994. For smaller than 8 hours MTTR, the availability will be slightly higher.
8.10 Concluding Remarks This chapter has covered the fundamentals of array reliability and discussed ways to increase availability through proper design of the array architecture. Two means to increase the MTBCF of an array are to reduce the size of element clusters or provide redundancy. Another way to increase the array MTBCF is to increase the MTBF of the active components in the array. Today, T/R modules and control modules have high reliability. Typically, the lower MTBF of power supplies has the most significant impact on the array MTBCF. We also discussed single points of failure and how to avoid them by adding driver amplifiers in the beamforming network; the need to carry along spare components to maintain the array’s performance can either be eliminated or reduced to a minimum. Following the analysis presented here, a trade-off between antenna MTBCF, availability, and cost can be made to arrive at the architecture that minimizes the array’s lifecycle cost.
References [1] Agrawal, A. K., and E. Holzman, “Active Phased Array Design for High Reliability,” IEEE Trans. on Aerospace and Electronics, Vol. 35, No. 4, October 1999. [2] Komiak, J., and A. K. Agrawal, “Design and Performance of Octave S- /C-Band T/R Modules for Multifunction Phased Arrays,” IEEE Trans. on Microwave Theory and Techniques, Vol. 39, December 1991. [3] Agrawal, A. K., and E. Holzman, “Beamforming Networks for Active Phased Array Antennas,” Special Issue on Phased Array Antennas, IEEE Trans. on Antennas and Propagation, Vol. AP-47, March 1999. [4] Romeu, J., “Understanding Series and Parallel Systems Reliability,” Selected Topics in Assurance Related Technologies, DoD Reliability Analysis Center, Vol. 11, No. 5, 2004, pp. 1–8.
7065_Agrawal_V6.indd 174
4/6/23 1:43 PM
9 Active Phased Array Design for High Clutter Improvement Factor 9.1 Introduction Radars are required to detect targets in the presence of large clutter, the reflections of the transmitted pulse by the earth’s surface. Because radar echoes from clutter can overwhelm a receiver, one of the challenges for ground-based and shipborne radars is detecting small, moving targets against high-clutter backgrounds [1]. Examples of clutter include reflections from land, mountains, sea, rain, and atmospheric turbulence. Some clutter is distributed spatially, in that it is usually much larger in physical size than the radar resolution cell. There also are point-clutter echoes, such as towers, poles, and other highly reflective objects. Echoes from the land or sea are known as surface clutter, and echoes from three-dimensional phenomena such as rain are known as volume clutter. Moving target indication (MTI) is a technique radar systems use to extract the reflection from the target in the presence of a clutter return. Typically, the radar performs MTI by transmitting a burst of at least two pulses. From pulse to pulse, the return from the clutter remains constant in amplitude, while the target return or signal amplitude varies at a rate depending on the Doppler frequency. The returns from successive pulses are subtracted in the signal processor, an operation that cancels the clutter but leaves a residual signal 175
7065_Agrawal_V6.indd 175
4/6/23 1:43 PM
176
Practical Aspects of Active Phased Array Antenna Development
from the target. The level of cancellation MTI can achieve is quantified by its clutter improvement factor (CIF): the signal-to-clutter ratio at the output of the radar processor divided by the signal-to-clutter ratio at the input to the antenna. CIF is usually averaged over all target velocities of interest for an MTI process. The CIF of a radar is limited by pulse-to-pulse instabilities in the radar’s active components, such as the transmitter and receiver and the signal processor. Typically, these instabilities are thermally induced and thus random in behavior. For instance, any pulse-topulse variations in the amplitude or phase of the transmitter signal will not be canceled when the signal processor’s returns from successive pulses are subtracted. This uncanceled noise can exceed the signal for small targets even after MTI processing. One way to improve a radar’s MTI performance is to use an active phased array antenna, which has the potential for extremely high CIF because the amplitude and phases are distributed at the aperture. In addition, the pulse-to-pulse amplitude and phase instabilities of the T/R modules tend to be randomized between modules and, thus, combine in a noiselike fashion, while target returns add coherently. We will discuss how an active phased array antenna can be designed to achieve high CIF by examining the benefits of the distributed architecture, the placement of components in the beamformer, and the layout of the power distribution network [2].
9.2 Centralized Phased Array Architecture Figure 9.1 shows a simplified block diagram of a passive phased array antenna [3]. N radiators are steered by N phase shifters driven by a central transmitter. To simplify our analysis, we let the transmitter be DC biased by a power supply at a single voltage Vr. Since the phase shifters are high-power passive components that do not provide RF amplification, we ignore them in the following analysis. We are concerned with two main limitations to the antenna’s CIF: thermal RF jitter generated in the active RF components and power supply voltage ripple. Since CIF depends on pulse-to-pulse variations, we can ignore characteristics such as transmit power droop that affects an individual pulse’s quality. Jitter is essentially amplitude and phase noise on top of the signal. The standard equations for the maximum CIF, Iaj and Iφj, due to the amplitude and phase jitter of a single active component such as the transmitter, is
7065_Agrawal_V6.indd 176
4/6/23 1:43 PM
Active Phased Array Design for High Clutter Improvement Factor
I aj =
1
(d )
2
177
(9.1)
aj
I jj =
1
(d ) jj
2
(9.2)
where
δ aj = RMS RF voltage jitter δ φj = the pulse-to-pulse RMS phase jitter in radians There would be two separate jitter contributions for a component that operates in both transmit and receive. Typically, CIF is expressed in dBs as CIF = l0Log10(I) (9.3)
Besides jitter, power supply voltage variations from pulse to pulse limit the maximum CIF of an antenna because small changes in the DC bias voltages applied to RF components cause changes in the amplitude and phase of the RF signal. Typically, a supply voltage is specified in terms of its ripple δV, while an RF component’s sensitivity to the ripple is specified in terms of its RF amplitude and phase pushing factors, δ as, volts/volt, and δ φs radians/volt. For a single component biased by a single voltage supply, the maximum CIF is given by I as =
I fs =
1
( dVdas )2 1
( dVd ) fs
2
(9.4)
(9.5)
If we return now to the antenna in Figure 9.1, we can write its total CIF as the contributions from the transmitter jitter and power supply ripple as
7065_Agrawal_V6.indd 177
4/6/23 1:43 PM
178
Practical Aspects of Active Phased Array Antenna Development
Figure 9.1 Generic, passive phased array architecture.
⎡ 1 1 1 1 ⎤ CIFpassive = −10Log 10 ⎢ + + + ⎥ (9.6) ⎢⎣ I aj I fj I as I fs ⎥⎦ For example, let the CIF due to amplitude and phase jitter be 50 dB and the CIF due to power supply ripple be 50 dB. Then the antenna CIF is 47 dB. If we want to increase the antenna’s CIF, we can build a power supply with a lower ripple and a transmitter with lower jitter and pushing factors. However, let us assume that these parameters are already as low as possible. Then, our only option is to change the antenna architecture, as discussed in the next section. As a first choice, we can distribute the transmitter and receiver at the aperture among the N array elements using solid-state T/R modules, whose outputs we combine to produce the same output power as the tube amplifier. Figure 9.2 shows a transmitter constructed from multiple parallel, smaller amplifiers whose outputs are combined to form the output that drives the passive array. Suppose we assume that the jitter is distributed among the N amplifiers
7065_Agrawal_V6.indd 178
4/6/23 1:43 PM
Active Phased Array Design for High Clutter Improvement Factor
179
Figure 9.2 Amplifier module that supports high CIF by combining many lower CIF amplifiers.
like white noise. In that case, the summation should result in a reduction of the pulse-to-pulse residual power due to the jitter by a factor of N since the noise power is incoherent and does not add in space or the combiners. We rewrite (9.1) and (9.2) as I aj =
N
(d )
2
(9.7)
aj
I fj =
N
(d ) fj
2
(9.8)
9.3 Distributed Array Architecture Referring back to Chapter 1, Figure 1.6 shows a simplified block diagram of an active phased array antenna. An active phased array distributes the transmit and receive functions at the aperture by placing a T/R module at each radiating element.
7065_Agrawal_V6.indd 179
4/6/23 1:43 PM
180
Practical Aspects of Active Phased Array Antenna Development
The array in Figure 1.6 has N T/R modules, and we assume that the noise generated by each T/R module is uncorrelated. As a result, the noise from the T/R modules does not add in space, while the transmitted signal has a gain = N. Therefore, the clutter improvement factor of the active array is 10Log10(N) greater than the clutter improvement factor of the passive array in Figure 9.1, assuming the CIF of one T/R module is the same as the CIF of the transmitter. For a 5,000-element active array, the improvement in CIF would be 37 dB. Transmit and receive driver amplifiers are generally needed in the beamformers to boost the signal both in transmit and receive modes. Figure 9.3 shows an active array block diagram with N T/R modules and M driver amplifiers. A phased array may have hundreds or even thousands of T/R modules. On receive, the jitter contributions of the T/R modules are combined in the beamformer while they are summed spatially on transmit. Thus, we can apply (9.7) and (9.8) from the previous section:
Figure 9.3 Active array architecture with driver amplifiers in the beamformers.
7065_Agrawal_V6.indd 180
4/6/23 1:43 PM
181
Active Phased Array Design for High Clutter Improvement Factor
I aj =
N
(d )
2
(9.9)
aj
I fj =
N
(d ) fj
2
(9.10)
However, since power supply voltage ripple can be caused by effects that are not entirely random, such as the AC-to-DC rectification process, an increase in CIF of N/f may result, where f is a correlation factor greater than one. In addition, often, more than one T/R module may be biased by a single power supply. Such clustering further correlates with the power supply ripple between T/R modules. Thus, we write the maximum CIF for N T/R modules, with nT/R modules biased by the same voltage supply as I as = I js =
N/f
( dVdas nT/R ) N/f
( dVd
n
js T/R
2
)
2
(9.11)
(9.12)
As an example, the phased array shown in Figure 9.3 has a maximum CIF given by a contribution IT/R from the T/R module and a contribution IBA from the driver amplifier: ⎡ 1 1 ⎤ CIFpassive = −10Log 10 ⎢ + ⎥ (9.13) ⎣ NIT/R MI BA ⎦
where
IT/R
(
)
(
)
⎡ d 2+d 2 + d aj 2 + djj 2 jjT/R ⎢ ajT/R T/R T/R transmit receive =⎢ 2 2 ⎢ + dV d + dT Vdjs nT fTR + dVR d as T as ⎢⎣ transmit
{(
) (
)} (
I BA = ⎡⎢ d aj 2 RFTx + djj 2 + dVBA d as BA ⎣ BA
7065_Agrawal_V6.indd 181
)
2
{(
) + ( dV d ) 2
R js
(
nBA f BA RFTx + dVBA d as
)
2
2
}
⎤ ⎥ ⎥ nR fTR ⎥ ⎥⎦ receive
nBA f BA ⎤⎥ ⎦
−1
−1
4/6/23 1:43 PM
182
Practical Aspects of Active Phased Array Antenna Development
where RFTx is a factor (in watts/watt) that accounts for the linearity of the T/R module on transmit. If the T/R module operates in the linear mode, RFTx is equal to one. However, T/R modules are usually compressed on transmit so that RFTx, is less than one, which reduces the contribution of the driver amplifier’s amplitude residual power to the antenna CIF. From Section 9.2, recall our example passive array with a centralized transmitter having a CIF of 47 dB. Consider a 10,000-element active phased array with 10,000 T/R modules, each with a CIF from a jitter of 50 dB, CIF from a power supply ripple of 50 dB, and each T/R module driven by its own power supply. Then, the CIF from the T/R modules is 47 dB + 40 dB = 87 dB, even though we have not improved the performance of the individual components. If the array has 1,250 driver amplifiers, each powered by its own power supply, and the CIF of the driver amplifier from jitter is 50 dB and 50 dB from ripple, then the CIF from the driver amplifiers alone is 47 dB + 31 dB = 78 dB. The CIF for the entire array then is 77.5 dB, which clearly is dominated by the CIF of the driver amplifiers since there are only one-eighth as many driver amplifier modules as T/R modules. To reduce the impact of the driver amplifiers on the array CIF, we could build each amplifier from eight smaller amplifiers, as shown in Figure 9.2. Then, the driver amplifier CIF would increase to 87 dB, and the array CIF would be 84 dB. To increase the array CIF further, we would have to increase the number of driver amplifiers and reduce the size of the cluster of T/R modules (T/R LRU) driven by a single supply.
9.4 Concluding Remarks Several general conclusions are evident from this discussion: 1. An active phased array will tend to have higher CIF than a passive array, given that all other performance parameters are equal. 2. The maximum possible CIF will increase as the number of elements in an active phased array increases. 3. The CIF will increase if we reduce the number of T/R modules or driver amplifiers biased by a single power supply. We are better off with many small power supplies than one large
7065_Agrawal_V6.indd 182
4/6/23 1:43 PM
Active Phased Array Design for High Clutter Improvement Factor
183
supply because more supplies have a smaller correlated ripple contribution. 4. The CIF will increase as the number of driver amplifiers increases. Figure 9.2 shows one way to design the driver amplifier to maximize its CIF. By paralleling many identical amplifier channels, we can achieve jitter decorrelation, as is done for the T/R modules in the phased array. However, we must be careful not to combine too many amplifiers; otherwise, the losses in the combining networks will become too high. Eliminating the driver amplifier entirely would give an even higher CIF. 5. We can increase the CIF by reducing the power supply ripple, pushing factors, and jitter.
References [1] Skolnik, M., Introduction to Radar Systems, McGraw-Hill, 1980. [2] Holzman, E., Agrawal, A. K., and J. Ferrante, “Active Phased Array Design for High Clutter Improvement Factor,” IEEE Int. Symp. on Phased Array Technology, October 15–18, Boston, MA, 1996. [3] Agrawal, A. K., and E. Holzman, “Beamformer Architectures for Active Phased Array Antennas,” IEEE Trans. on Antennas and Propagation, Vol. 47, March 1999.
7065_Agrawal_V6.indd 183
4/6/23 1:43 PM
7065_Agrawal_V6.indd 184
4/6/23 1:43 PM
10 Active Phased Array Antenna Calibration 10.1 Introduction An active phased array antenna requires amplitude and phase calibration of each element to maintain compliance with their specified performance throughout the deployment period. Initial array calibration in the factory is generally accomplished by measuring the phase and amplitude insertion characteristics from the exciter through the beamformer to each radiating element on a near-field range. The array’s variable attenuators and phase shifters can minimize element-to-element differences in amplitude and phase. This initial calibration is usually called array alignment or tuning. Nearfield range array alignment involves placing an antenna probe in front of every array element, one at a time, with that element either in the transmit or receive mode and measuring the element’s nearfield. In this way, transmit and receive modes accurately measure each radiating element’s amplitude and phase. The required array amplitude and phase distributions are achieved by adjusting the T/R module amplitudes and phases as determined by the nearfield measurements. Calibration is performed over all frequencies, amplitude, and phase settings. The phase and amplitude characteristics of passive and active RF devices depend on frequency, and those of active devices also depend on temperature and can drift over time while they warm up or cool down.
185
7065_Agrawal_V6.indd 185
4/6/23 1:43 PM
186
Practical Aspects of Active Phased Array Antenna Development
Because an active phased array antenna may be deployed for many years or even decades, its performance may deteriorate over time due to changes in the solid-state components and DC-to-DC converters. In addition, failed T/R modules and DC-to-DC converters must be repaired or replaced, often while the array is in the field. As a result, the array amplitude and phase settings must be readjusted periodically to compensate for active component drift and after components are replaced or repaired. This measurement and readjustment of the element amplitude and phase settings during array deployment is sometimes called field calibration. In contrast to calibration in the factory, a near-field range rarely is available, so other calibration techniques, often automated, are used. Several techniques for field calibration have been proposed and implemented, each having various advantages and disadvantages [1–13]. All calibration techniques use coupling between the array elements or the array elements and external elements. The next section describes four active array calibration techniques that rely on the mutual coupling between the radiating elements: 1. Coupling between array elements and external elements; 2. Mutual coupling between array elements; 3. Mutual coupling between one calibration element and all array elements; 4. Mutual coupling between a few dedicated array elements and the remaining array elements.
10.2 Active Array Calibration Using Mutual Coupling Between Array and External Elements The calibration technique described in [1] employs a set of external elements, not part of the array, such as horns, that provide transmit and receive couplings to each active array element. These external elements are located around the perimeter of the array, as shown in Figure 10.1. The array is first aligned in the near-field range using the standard method. Then, the coupling between the external and each of the array elements is measured, with the element either in the receiver or transmit mode. These factory coupling measurements are saved with each array as its reference standard. Field
7065_Agrawal_V6.indd 186
4/6/23 1:43 PM
Active Phased Array Antenna Calibration
187
calibration is accomplished by comparing mutual coupling measurements between the array and external elements taken during a deployment with the factory standard data set. Any differences are assumed to be caused by changes in the array rather than the external environment. The amplitude and phase settings of the array elements are adjusted based on the differences between the coupling values measured in the field and the factory coupling values. Such a mutual coupling measurement field calibration subsystem should have enough receive dynamic range to handle the transmit power levels of elements across the array, from edge to center. The dynamic range can be increased by increasing the size of the external calibration horns. Each external element calibrates a group of elements in the array. In addition, these measurements should be repeatable to the amplitude and phase accuracy of 0.1 dB and 1°, respectively. A downside to this calibration technique is that it increases the array system’s total footprint and increases its RCS. Furthermore, the dynamic range requirements may require external elements with high gain.
Figure 10.1 Active array and external horns for field calibration using coupling between array elements and external elements.
7065_Agrawal_V6.indd 187
4/6/23 1:43 PM
188
Practical Aspects of Active Phased Array Antenna Development
10.3 Active Array Calibration Technique Using Mutual Coupling Between Array Elements The second calibration method, proposed by Aumann, Fenn, and Wilworth [2] and by Shiply [3], relies on the mutual coupling between array elements to measure the relative insertion phase and amplitude of the array elements by transmitting from one element while simultaneously receiving from another. This calibration method can be used in the field without factory-standard data. Theoretically, this technique is sufficient to eliminate the near-field calibration, so it is sometimes called the self-calibration or autocalibration technique. For self-calibration, one first uses the mutual coupling measurements between adjacent elements to align the array in the near-field range at the factory. After the array is deployed, it can be calibrated again using the same technique. At the factory, the initial calibration is usually performed in a near-field range to achieve the lowest elemental residual errors. The calibration procedure involves sequentially transmitting from a single array element and sequentially receiving from only the adjacent elements using separate array beamformers (transmit and receive) with all other elements turned off, as shown in Figure 10.2. In a large planar array of thousands of elements, the mutual coupling between pairs of adjacent elements is assumed to be invariant to the pair position. The measured coupling between two adjacent elements includes effects of feed lines, power combiners/dividers, and T/R modules. How one applies the mutual coupling method to a two-dimensional array depends on the array geometry (i.e., hexagonal, triangular, rectangular) [3]. Figure 10.2 shows a hexagonal or equilateral triangular array. In a hexagonal lattice, all elements can be calibrated relative to each other. The center element can be
Figure 10.2 Calibration of a hexagonal lattice array using the mutual coupling method between array elements.
7065_Agrawal_V6.indd 188
4/6/23 1:43 PM
Active Phased Array Antenna Calibration
189
calibrated by measuring the mutual coupling between the center element and the six adjacent elements since the coupling between equidistant elements is assumed to be equal. This process can be repeated for all elements to calibrate the array. Mutual coupling between two neighbor elements depends on the antenna element pattern and the distance between the elements. Typically, adjacent element coupling is between −15 dB and −25 dB, depending on the element type. Printed circuit elements usually have higher mutual coupling than waveguide elements. For the mutual coupling calibration to be accurate, the amount of mutual coupling should be high enough to be detected and low enough not to saturate the receiving elements. Nonadjacent elements can be chosen to avoid the saturation of the receiving element. High isolation is required between transmit and receive paths so that when one element is turned on to transmit and the other to receive with all the other elements turned off, the internal coupling from all the elements through the beamforming network does not alter the measurement. However, it is assumed that all internal radiating elements see the same environment, including beamformers. A couple of rows and columns of dummy elements will be required so that the edge element coupling is the same as the other element coupling. This technique can easily be extended to the triangular and rectangular array lattices [2–3]. The advantage of this technique is that it does not require mutual coupling measurement data at the factory for comparison in the field. The array can be independently calibrated at any time using the mutual coupling between elements. Disadvantages of this calibration technique include the need for two independent transmit and receive beamformers, dummy elements for calibrating the edge elements, and the requirement that the receiving element must not be saturated. In addition, for T/R modules transmitting high RF power, the coupling to the adjacent elements may be large enough to saturate the LNA, resulting in erroneous results. Therefore, elements at farther distances should be used to measure the mutual coupling to avoid saturating the low noise amplifier. Post calibration, residual RMS amplitude and phase errors in an array are determined by the number of bits and random errors in the attenuator and the phase shifter. In turn, the RMS errors determine the RMS sidelobe level of the array. Therefore, this calibration technique assumes that the array’s RMS amplitude and phase errors will remain the same between calibrations. It is quite possible that
7065_Agrawal_V6.indd 189
4/6/23 1:43 PM
190
Practical Aspects of Active Phased Array Antenna Development
the RMS amplitude and phase errors could increase after the array is calibrated many times in the field due to the random errors in the mutual coupling measurements. However, no such information is available in the published literature. To verify the initial accuracy of this calibration technique, the array can be aligned in the near-field range, and the array pattern can be calculated from the amplitude and phase settings. Then the array can be calibrated using this technique, and the calculated array pattern can be compared with the pattern calculated from the near-field aligned amplitude and phase settings.
10.4 Active Array Calibration Technique Using Mutual Coupling Between One Calibration Element and All Array Elements Smaller phased arrays of 1,000 to 2,000 elements can be calibrated with a single center, dedicated calibration element [4], as shown in Figure 10.3. The central calibration element replaces one of the array elements, but it is passive (i.e., it does not have a T/R module behind it), and is not beamformed with the other array elements. Instead, it is connected directly to its calibration transmit/receive circuit. The calibration element aligns all array elements by measuring the mutual coupling between the array elements and the calibration element in both transmit and receive modes. This series of tests is performed in the near-field range after the array has been initially aligned. The mutual coupling data measured in the near-field range are stored as the factory data that will be used to calibrate
Figure 10.3 Array calibration with one center element.
7065_Agrawal_V6.indd 190
4/6/23 1:43 PM
Active Phased Array Antenna Calibration
191
the array in the field. In the field, the mutual coupling data is measured and compared to the factory data. The elemental amplitude and phase settings are adjusted until the mutual coupling measurements match the factory settings. In addition, when elements fail in the field and are replaced, the new elements can be calibrated by measuring the mutual coupling between the replaced and the calibration elements. The advantage of the single-element calibration technique is that it can be used with a single array transmit/receive beamformer, as shown in Figure 1.6 in Chapter 1. To provide redundancy, a second array element can be chosen as a calibration element or a single calibration element can have redundant transmit and receive circuitry. The main disadvantage of the single calibration element technique is that it cannot be used for very large arrays. The calibration unit’s amplitude and phase settings are subject to drift from changes in the environmental conditions, particularly temperature, much like the array elements. The calibration unit itself can also be recalibrated periodically using its own selftest process. Since the calibration element is not part of the array, the RMS amplitude and phase errors after recalibration over a long period should remain stable. The electrically largest array that can be calibrated with a single calibration element depends on the magnitude of the mutual coupling between the central and edge elements and the dynamic range of the calibration circuit. To avoid saturating the elements nearest the calibration element, the transmit power output of the calibration unit can be adjustable using a variable attenuator. In addition, when elements fail in the field and are replaced, the new elements can be calibrated by measuring the mutual coupling between the replaced and the calibration elements.
10.5 Active Array Calibration Technique Using Mutual Coupling Between a Few Dedicated Internal Elements and the Array Elements To overcome some of the limitations of the external calibration source technique, the complexity of the mutual-coupling–based autocalibration technique, and calibration with a single passive element, we describe a calibration technique that uses a few dedicated passive internal elements as calibration elements [4]. Using passive elements significantly increases the accuracy of mutual coupling
7065_Agrawal_V6.indd 191
4/6/23 1:43 PM
192
Practical Aspects of Active Phased Array Antenna Development
measurements because there are no T/R modules behind them. A notional phased array aperture utilizing multiple calibration elements is shown in Figure 10.4. The elements are chosen from the set of array elements. They are directly connected to a dedicated calibration T/R unit via a switch or coupler network. These calibration elements are not part of the array. Calibration is achieved by sequentially measuring the mutual coupling between each passive calibration element and its own group of active array elements. A switch network is used to route the calibration signals from the T/R calibration unit to one calibration element at a time. A block diagram of an array with a calibration unit and switch network is shown in Figure 10.5. As shown in Figure 10.4, each calibration element calibrates a group of elements situated at a distance sufficiently far from the calibration element so that the elements in the group are not saturated when they are in receive mode. In addition, more than one calibration element can be used to provide redundancy and improve accuracy with multiple measurements. The number of passive calibration elements required to calibrate an active array is a small percentage, typically much less than 1%,
Figure 10.4 Active phased array antenna calibration scheme using multiple internal array elements.
7065_Agrawal_V6.indd 192
4/6/23 1:43 PM
Active Phased Array Antenna Calibration
193
Figure 10.5 Active phased array architecture.
of the total array elements. It is proportional to the size of the array. For example, nine passive elements are sufficient for calibrating a large array of 10,000 elements. As we will show later, not having the calibration elements radiate with the array has a negligible effect on the gain and sidelobe levels of the array. 10.5.1 Calibration Procedure The array is divided into several blocks, with a single calibration element located near the center of each block, as shown in Figure 10.4. The mutual coupling measurements are taken after array alignment in the near-field antenna range. To collect the calibration data in the near-field range, the transmit signal from the calibration unit is routed through the switch network to one of the calibration elements. Every other active element in the array is initially turned off. Then a single array element in a block adjacent to the passive calibration element block is turned on and receives the transmit signal from the calibration element via mutual coupling. When the measurement is completed, the first active element is turned off, and an adjacent active element in the same block is turned on to receive the calibration signal from the same calibration
7065_Agrawal_V6.indd 193
4/6/23 1:43 PM
194
Practical Aspects of Active Phased Array Antenna Development
element. This measurement process is repeated for every active array element in the block associated with the chosen calibration element. When all the elements in the block are measured, the elements in other blocks adjacent to the same calibration element block are calibrated. Then, a new calibration element is chosen, and the active elements in its adjacent blocks are calibrated. This process continues until all active elements in the array have been calibrated. With the calibration elements transmitting, the previous procedure calibrates the elements in the array in the receive mode. The receive mode calibration process is reversed to calibrate array elements in the transmit mode. The signal is transmitted through the active array element and received in the calibration element. The nearfield range calibration process produces a factory standard mutual coupling data set, which is stored with the active array. Since one calibration element can calibrate active elements in more than one adjacent block, it provides redundancy in the calibration elements. It is possible to calibrate the active elements in most blocks using more than one calibration element. Calibration of active elements using multiple calibration elements provides additional redundancy to protect the system in the event of a failure of a calibration element. When the array is deployed in the field, the mutual coupling measurements between calibration elements and array elements are repeated exactly as they were done at the factory near-field testing. The field mutual coupling data are then compared to the stored factory standard data. Finally, the amplitude and phases of the attenuators and phase shifters in the T/R modules are adjusted to bring the antenna performance to the factory standard. As mentioned in the previous section, the calibration unit’s amplitude and phase settings are subject to drift from changes in the environmental conditions, particularly temperature, much like the array elements. The calibration unit itself can also be recalibrated periodically using its own self-test process. Since the calibration element is not part of the array, the RMS amplitude and phase errors after recalibration over a long period should remain stable. This calibration procedure can also be used to detect failed elements. The radiating elements’ health can be monitored continuously by running the calibration software in the background using a small fraction of the system timeline. When failed elements are replaced with new elements, the array is recalibrated to bring
7065_Agrawal_V6.indd 194
4/6/23 1:43 PM
Active Phased Array Antenna Calibration
195
the replaced elements and the array performance to the factory standard. 10.5.2 Required Number of Calibration Elements Ideally, the number of elements dedicated to calibration should be small to avoid undue antenna gain and sidelobe degradation from the gaps they create in the array aperture distribution. Less than 1% of the elements should be calibration elements. The following example shows that nine calibration elements (0.09%) are sufficient for calibrating a 10,000-element active array antenna. Figure 10.4 shows the location of the calibration elements in the array aperture. A relatively small number of passive elements dedicated to calibration minimizes the EIRP loss and sidelobe degradation. For instance, in an array of 10,000 elements, if nine elements (0.09%) are used as dedicated calibration sources, the effective isotropic radiated power (EIRP) loss is only 0.004 dB. A planar array analysis code evaluated peak and RMS sidelobe degradation for a 10,000-element array. The results are plotted in Figure 10.6 for −40 dB peak and −58 dB RMS sidelobe levels, with 1.0 dB amplitude and 5° phase errors. The peak and RMS sidelobe degradation are each less than 0.2 dB.
Figure 10.6 Peak and RMS sidelobe degradation versus percent of array used as calibration elements (for −40 dB peak and −58 dB RMS sidelobe levels).
7065_Agrawal_V6.indd 195
4/6/23 1:43 PM
196
Practical Aspects of Active Phased Array Antenna Development
The previous example of a 10,000-element array is divided into nine square subarrays of 1,111 elements, as shown in Figure 10.4. Generally, the subarrays do not have to be square in shape and can take any appropriate regular shape. For example, the edge subarrays of a circular or elliptical array can be partial blocks. Each subarray has a dedicated calibration element near its center. Each calibration element is used to calibrate all array elements in the adjacent subarray. As shown in Figure 10.4, a calibration element can be used to calibrate elements in more than one subarray or block. The block dimensions determine the total number of calibration elements in the array and depends on mutual coupling levels between the calibration elements and the nearest and farthest active elements in the next block, the calibration unit’s input TOI, and dynamic range, and its transmit power. The calibration unit output power level is chosen so that the next block’s nearest element is not saturated and the farthest array element has a sufficient coupling for an accurate measurement. Even if we increase the number of calibration elements by a factor of 2 or 4, the degradation in array EIRP, peak, and RMS sidelobes is not very significant. 10.5.3 Calibration Accuracy Near-field antenna measurement ranges typically can provide a high degree of amplitude and phase accuracy (errors on the order of less than one degree in phase and 0.1 dB in amplitude). Consequently, the calibration technique must provide a measurement accuracy comparable to a near-field range to avoid degrading the array performance. An important consideration in a calibration system is the phase and amplitude drift that inevitably occurs in the active components of the array and the calibration unit elements over time. If left uncorrected, this drift can eventually result in unacceptable errors in calibrating the array. Therefore, the calibration unit should be calibrated periodically to reset it to the factory amplitude and phase settings. The calibration unit can be calibrated using directional couplers to insert signals at its input and output ports in transmit and receive modes, as shown in Figure 10.5 [4]. The calibration technique presented here would provide the desired measurement accuracies using standard test equipment
7065_Agrawal_V6.indd 196
4/6/23 1:43 PM
Active Phased Array Antenna Calibration
197
configured for this purpose. Because this calibration system would be periodically calibrated, the accuracy of the mutual coupling measurements will not degrade over time and will maintain factory-level performance. This is a unique feature of this calibration technique. Calibration of the calibration unit would be required less often than calibration of the array elements and can be performed at several weeks’ intervals. 10.5.4 Effect on Array Packaging For ease of maintenance, most of the active array electronics are configured as LRUs, including T/R modules, DC-to-DC converters, and various control/processor assemblies. The LRUs are mounted on both sides of a cold plate in groups of two to eight on each side. LRU RF outputs typically interconnect with the inputs of the passive array radiators. In cases where those radiators are used for calibration instead, the radiating element is connected directly (coax, for example) to the calibration unit via the switch network. Therefore, two types of LRUs are required: (1) LRU with active array T/R modules only and (2) LRU with a T/R channel replaced by the interconnection between the radiator and calibration unit, as shown in Figure 10.7. The beamformer port corresponding to the missing
Figure 10.7 Array architecture showing a regular LRU and one connected to a calibration element.
7065_Agrawal_V6.indd 197
4/6/23 1:43 PM
198
Practical Aspects of Active Phased Array Antenna Development
T/R module will be terminated in a matched load. Since the number of calibration elements is small, the cost impact of producing a variation of LRUs for calibration will be minimal.
10.6 Concluding Remarks We have presented four techniques used to calibrate active arrays in the field. All of the calibration techniques use mutual coupling between internal array elements and calibration elements. The mutual coupling measurements are first performed at the factory, and the data are stored with the array as a reference set to compare with the measurements of the same coupling in the field. Then, the array phase shifters and attenuator settings are adjusted to regain the factory coupling values by comparing the field- and factory-measured values. Since the calibration elements are passive, measurement accuracy is not affected by variations in the active elements and isolation between elements. As a result, the calibration accuracy is high and does not degrade over time, as the calibration unit can be calibrated periodically using standard test equipment.
References [1] Sarcione, M., et al., “The Design, Development, and Testing of the THAAD (Theater High Altitude Area Defense) Solid-State Phased Array (Formerly Groundbased Radar),” IEEE Int. Symp. on Phased Array Systems and Technology, Boston, MA, October 15–18, 1996, pp. 260–265. [2] Aumann, H. M., Fenn, A. J., and F. G. Willwerth, “Phased Array Antenna Calibration and Pattern Prediction Using Mutual Coupling Measurements,” IEEE Trans. on Antennas and Propagation, Vol. AP-37, No. 7, July 1989, pp. 844–850. [3] Shipley, C., and D. Woods, “Mutual Coupling-Based Calibration of Phased Array Antennas,” IEEE Int. Symp. on Phased Array Systems and Technology, Danapoint, CA, May 21–25, 2000. [4] Agrawal, A. K., and A. Jablon, “A Calibration Technique for Active Phased Array Antennas,” IEEE Int. Symp. on Phased Array Systems and Technology, Oct. 13–16, Boston, MA, 2003. [5] Şeker, I., “Calibration Methods for Phased Array Radars,” Proc. SPIE 8714, Radar Sensor Technology XVII, 87140W, May 31, 2013. [6] Sanchez, R. H. M., “Calibration Technique for Phased Array Antennas,” U.S. Patent 8,199,048 B11, issued June 12, 2012.
7065_Agrawal_V6.indd 198
4/6/23 1:43 PM
Active Phased Array Antenna Calibration
199
[7] Van Werkhoven, G. H. C., and A. K. Golshayan, “Calibration Aspects of the APAR Antenna Unit,” IEEE Int. Conference on Phased Array Systems and Technology, 2000, pp. 425–428. [8] Pawlak, H., Charaspreedalarp, A., and A. F. Jacob, “Experimental Investigation of an External Calibration Scheme for 30 GHz Circularly Polarized DBF Transmit Antenna Arrays,” European Microwave Conference, Manchester, U.K., September 10–15, 2006, pp. 760–767. [9] Lee, K.-M., Chu, R.-S. S.-C. and Liu, “A Built-in Performance-Monitoring Fault Isolation and Correction (PM/FIC) System for Active Phased-Array Antennas,” IEEE Trans. on Antennas and Propagation, Vol. 41, No. 11, 1993, pp. 1530–1540. [10] Andersson, S., Forssén, U., Ovesjö, F. B., and S. O. Petersson, “Antenna Array Calibration,” U.S. Patent 6,339,399 B1, 2002. [11] Steyskal, H., and J. S. Herd, “Mutual Coupling Compensation in Small Array Antennas,” IEEE Transactions on Antennas and Propagation, Vol. 38, Issue 12, 1990, pp. 1971–1975. [12] Gao, T., Wang, J., Guo, Y., and X. Chen, “Large Phased Array Antenna Calibration Using Mutual Coupling Method,” CIE International Conference on Radar, 2001, pp. 223–226. [13] Neidman, Y., Shavit, R., and A. Bronshtein, “Diagnostic of Phased Arrays with Faulty Elements Using the Mutual Coupling Method,” IET Microwaves, Antennas & Propagation, Vol. 3, No. 2, pp. 235–241.
7065_Agrawal_V6.indd 199
4/6/23 1:43 PM
7065_Agrawal_V6.indd 200
4/6/23 1:43 PM
11 Digital Beamforming for Active Phased Array Antennas 11.1 Introduction So far, in the previous chapters, we have only discussed analog beamformers. In an all-analog phased array antennas, the received signals from all the radiating elements are combined using an analog beamformer into a few monopulse outputs, each of which is converted to a digital signal using an analog to digital converter (ADC), as shown in Figure 11.1. A limitation of analog beamforming is that only one receive beam can be formed and steered by a set of variable attenuators and phase shifters in the T/R modules and analog beamformer. Each additional receive beam requires another set of circuitry in the T/R module and its beamformer hardware in the array. There is only one degree of freedom. Digital beamforming (DBF) overcomes this one-beam-per-beamformer limitation by increasing the number of simultaneous and independently controllable receive beams that can be formed by orders of magnitude. In addition, DBF can improve receive sensitivity and dynamic range, including in the presence of interference, and reduce the performance requirements of the receiver. Digital beamforming’s primary benefits are achieved on receive; however, the analog beamformer used to form the transmit array excitation can be replaced by a set of digital synthesizers, each followed by a digital-to-analog converter 201
7065_Agrawal_V6.indd 201
4/6/23 1:43 PM
202
Practical Aspects of Active Phased Array Antenna Development
(DAC), as many as one per element, which provides tremendous flexibility in transmit beam formation. While people have been publishing papers about DBF for decades, only since 2000 has digital circuit technology advanced to the state where DBF is becoming relatively common [1–6]. In a digital beamformer, analog-to-digital converter (ADC) and digital receiver are placed behind each element, as shown in Figure 11.2, and the digital data stream output from each ADC can be processed by a digital signal processor. Given enough processing capability, multiple simultaneous, independently steerable, and shapeable receive beams can be formed in the digital domain. The ultimate goal of placing a digital beamformer at the element level (DBF) leads to active array capabilities far beyond conventional, analog beamformed active arrays, which can, at best, form a handful of simultaneous receive beams, given that a separate beamformer and receive circuitry in the T/R module is required for each beam [7]. The electromagnetic environment is becoming increasingly dense with unintentional and intentional sources of interference proliferate, including other radars, high levels of clutter from land and sea, and jammers. To mitigate this interference, digital beamformers can place nulls in the direction of several jammers by applying adaptive array techniques through complex element-level weighting.
Figure 11.1 The RF output of an active phased array analog beamformer is converted to a digital signal prior to signal processing in the radar processor.
7065_Agrawal_V6.indd 202
4/6/23 1:43 PM
Digital Beamforming for Active Phased Array Antennas
203
Figure 11.2 Active array digital beamforming architecture.
In addition to producing multiple, simultaneous receive beams, digital beamformers have many other features that are difficult to achieve with analog beamformers, including multiple closely spaced receive beams that can reduce surveillance time and communicate with other systems, very low element level amplitude and phase errors capable of achieving ultra-low sidelobe levels, improved array element patterns [3], increased system dynamic range, angle measurement accuracy, reduced system phase noise, and flexible management of the radar’s power and timeline. Digital beamforming concepts have evolved over a long period, and it has taken several decades to mature and become affordable. The digital beamforming array, or the digital array, is the latest generation of active phased array antennas. In the 1990s, digital beamforming active arrays were considered the antenna engineer’s ultimate dream. Further improvements will come from digital processors, A/D converters, and adaptive techniques. Two recently developed radar systems—air and missile defense radar (AN/SPY-6 [V]) at S-band [Raytheon] and Space Fence [Lockheed] also at S-Band [5]—use digital beamforming phased arrays.
11.2 Dynamic Range Improvement Analog active phased arrays use only one receiver for each beamformer output, as shown in Figure 11.1. In a large array, say one with 10,000 elements, this receiver requires a high dynamic range. Instantaneous dynamic range is defined as the ratio of the maximum to
7065_Agrawal_V6.indd 203
4/6/23 1:43 PM
204
Practical Aspects of Active Phased Array Antenna Development
minimum possible signals that can be resolved by the ADC [6] supported by the receiver and, ultimately, the ADC dynamic range is also defined as the spurious free dynamic range (SFDR), where all receiver-generated intermodulation (IMD) products must be maintained at a certain level below the maximum signal. The maximum signal-to-noise ratio of an ADC is roughly determined by the formula [6] SNR max = (6.02n + 1.76)dB, where n is the number of ADC bits. So, a 12-bit ADC will have about 74 dB of dynamic range if the system sensitivity is noise-floor limited, meaning that the gain of the analog receiver is chosen so that the minimum discernable signal at the input to the ADC is at the noise floor. In general, the minimum signal needs to be some number of decibels above the noise floor. A comparison of the dynamic range of analog and digital beamformers is shown in Figure 11.3 on page 205. In the analog array, the receiver handles the signal at the output of the passive beamformer at point B in Figure 11.3(a). In contrast, in an elemental digital beamformed array, the receiver handles the signal at point A in Figure 11.3(b) after each element. Therefore, the dynamic range required for a receiver for an elemental digital beamformed array of N elements is reduced by approximately 10Log10N as compared to the analog beamformer receiver. In other words, if the same receiver is used in the elemental digital beamformed array as in the analog beamformed array, the digital beamforming array’s dynamic range would be 10Log10N higher, assuming spurs are decorrelated [2].
11.3 Digital Beamforming at the Subarray Level Elemental digital beamformed (eDBF) arrays require receivers at each element and extremely high digital processing capability. Digital data throughput is a major challenge in DBF systems. A 1,000-element array, with eDBF, operating at 1 GHz, with a direct digital sampling of the RF requires 1,000 × 1 GB/s = 1 TB/s of data streaming from the array to the digital processor. So, it’s not surprising that their implementation in fielded systems has been slowed
7065_Agrawal_V6.indd 204
4/6/23 1:43 PM
Digital Beamforming for Active Phased Array Antennas
205
Figure 11.3 Comparison of the analog and digital beamformer: (a) analog beamformer; (b) digital beamformer.
by A/D converter’s performance limitations, digital processing circuit complexity, DC power consumption, weight, and high costs. To overcome these challenges, digital beamforming can be applied at a subarray level instead of an element level. Subarray digital beamforming places receivers and A/D converters behind groups of analog combined elements, as shown in Figure 11.4. The reduced number of receivers in the array limits the dynamic range compared to elemental DBF. More importantly, analog phase shifters at the element level must be included in each subarray to prevent the formation of grating lobes when the array is steered, as shown in Figure 11.4. The phase shifters do most of the work of steering the beam to a commanded angle, while the digital
7065_Agrawal_V6.indd 205
4/6/23 1:43 PM
206
Practical Aspects of Active Phased Array Antenna Development
Figure 11.4 Digital beamforming at the subarray level.
beamformer creates multiple simultaneous beams whose steering angles can be varied about the analog steering angle to the point that grating lobe levels become unacceptable. Digital beamforming at the subarray level reduces the number of receivers and complexity of the digital beamforming circuit but requires analog amplitude and phase weighting w at the element level.
11.4 Digital Beamforming of Multiple Simultaneously Independent Receiver Beams The digital signals from each A/D converter are processed by a digital beamformer before being passed on to the digital signal processor. Digital beamformers can apply amplitude, phase, and even time delay weighting to those digital signals in many ways, such that multiple independently steered received beams of arbitrary weighting and shape can be formed simultaneously in the digital domain. Chapter 4 discussed forming multiple beams in an analog active phased array. Figure 4.4 shows an active array architecture to form three independent sum, delta azimuth, and delta elevation monopulse beams. A similar array architecture would form three receive independent receive beams at arbitrary scan angles. The threebeam array architecture requires three independent beamformers
7065_Agrawal_V6.indd 206
4/6/23 1:43 PM
Digital Beamforming for Active Phased Array Antennas
207
and three separate channels in the T/R modules, as shown in Figure 4.5, the extra channels require separate attenuators and phase shifters for reach channels. The additional receive channels in the T/R module increase the packaging complexity. Since the width of the T/R modules is fixed by the element spacing at the aperture, it becomes increasingly difficult to package extra channels in the same width. The extra channels can be accommodated in the module package at the lower frequencies, such as L- through C-bands. However, it becomes increasingly difficult at higher frequencies. Similarly, it becomes difficult to package too many receive beamformers in the array structure. Therefore, if the requirement for the number of beams is limited to three beams, an analog array may be cost-effective instead of a digital beamformer.
11.5 Angle Tracking Accuracy Radar systems often use monopulse techniques to derive angle-tracking information from a single target return pulse, which requires three monopulse beams, that is, sum, Azimuth Delta, and Elevation Delta. The target tracking angle information is acquired by processing the outputs of the three monopulse beams simultaneously. Typically, the delta beams are divided by the sum beam, and the tracking angle accuracy is a function of the null depth in the two difference channels. In an analog array, a single beamformer can form three monopulse beams for active phased arrays, as shown in Figure 1.6 in Chapter 1. However, the three monopulse beams are not completely independent in pointing or beam shape. The null depth depends on how well the sum and difference channels track each other and the residual amplitude and phase errors at the aperture. Forming multiple independent monopulse beams is complex and expensive to implement in an analog beamformer. On the other hand, an elemental digital beamformer can easily generate multiple independent monopulse beams in the digital domain with high angle accuracy because it can apply amplitude and phase weighting in the digital domain with far less error than analog phase shifters and variable attenuators.
7065_Agrawal_V6.indd 207
4/6/23 1:43 PM
208
Practical Aspects of Active Phased Array Antenna Development
11.6 Adaptive Digital Beamforming Digital beamforming arrays can adapt to the external electromagnetic environment and place nulls in the array pattern to attenuate signals from jammers, as shown in Figure 11.5. An unadapted array provides protection against jammers by achieving very low sidelobe levels at a cost in array directivity. The digital beamforming array can use adaptive techniques [8, 9] in real time to place nulls on demand in the receive pattern at jammer angles. The sidelobe weighting can be less severe and the corresponding taper efficiency higher since the nulls can be pushed much deeper than the lowest sidelobe peaks and provide excellent protection against jammers. Digital array technology is conducive to adaptive cancellation techniques in which the array automatically adjusts its weights to the signal environment to minimize jammer and interference power at the array output. The fundamental problem in adaptive beamforming is to determine the adaptive weights (Figure 11.6) that minimize jammer or interferer contributions without significantly degrading the desired signal contribution. However, in practice, hardware errors, particularly channel matching, limit the achievable cancellation performance.
Figure 11.5 Receive array low sidelobe pattern: (a) without nulling suppresses jammer (J) signals down to the level of the sidelobe peaks; (b) with nulling provides additional jammer signal (J) suppression.
7065_Agrawal_V6.indd 208
4/6/23 1:43 PM
Digital Beamforming for Active Phased Array Antennas
209
Figure 11.6 Adaptive digital array processing.
In digital arrays, the adaptive weighting is applied to the target returns before they reach the radar signal processor (see Figure 11.6). Many algorithms have been developed to determine the weights and require high data processing speeds because the algorithms are very complex [8, 9]. 11.6.1 Adapting Nulling in Analog Arrays In an analog array, the sidelobe level is a function of the number of amplitude and phase shifter bits. For example, the errors from 6-bit variable attenuators and phase shifters would result in a −57 dB RMS (2.38) below the main beam. Analog device technology and manufacturing errors limit the smallest bit size. Analog arrays use coherent auxiliary sidelobe cancelers (SLC) to reduce the sidelobe level in the interference direction. The sidelobe cancelers consist of a small number of radiating elements with broad beams placed around the periphery of the array, as shown in Figure 11.7. The outputs of the SLCs are adaptively weighted and combined with the main array output to cancel interfering signals that enter through the sidelobes. When there are a small number of jammers, sidelobe cancellers are the least costly solution. However, digital beamforming can place many more and much deeper nulls in the jammer directions.
7065_Agrawal_V6.indd 209
4/6/23 1:43 PM
210
Practical Aspects of Active Phased Array Antenna Development
Figure 11.7 Analog beamformed arrays suppress jammer signals by using auxiliary arrays with broad beams to cancel unwanted signals received in the main array.
11.7 Exciter Noise and Clutter Attenuation Radars are required to detect targets in the presence of large clutter backgrounds, such as the Earth, ocean, and mountains. One means for performing this function is MTI. Typically, the radar performs MTI by transmitting a burst of at least two pulses. From pulse to pulse, the return from the clutter remains relatively constant in amplitude, while the target return or signal amplitude varies at a rate depending on the Doppler frequency. The returns from successive pulses are subtracted in the signal processor, an operation that cancels the clutter to a low level below the residual signal from the target. As discussed in Chapter 9, the level of cancellation MTI can achieve in a particular array system is quantified by its CIF: the signal-to-clutter ratio at the output of the radar processor divided by the signal-to-clutter ratio at the input to the array. Chapter 9 discussed that active phased arrays provide high clutter rejection compared to passive phased arrays. In passive phased arrays, all elements have the same noise as the central exciter, and the output noise from the radiating elements is correlated. On the other hand, in analog active arrays, the transmitter and receiver are distributed at the aperture, and the phase noise from array elements is uncorrelated. Therefore, the clutter improvement factor of active phased arrays is improved by a factor of N, where N is the number of elements in the array. In an analog array, there typically is a single transmit exciter, which provides a signal that is split to feed all the T/R modules. Since its noise is correlated across the entire array, its phase noise must be very low not to limit the radar CIF.
7065_Agrawal_V6.indd 210
4/6/23 1:43 PM
Digital Beamforming for Active Phased Array Antennas
211
A digital array with M uncorrelated exciters producing noise at the subarray or element level benefits from a 10Log10M reduction in exciter noise at the transmit array beam output compared to an analog array with a single exciter. Alternatively, we can reduce the noise requirements of the M exciters [2]. Even in an active analog array, it may be more cost-effective to have many exciters with higher noise figures than having only one exciter with very low noise. In some cases, an exciter with very low noise may be difficult to achieve. In those cases, distributing the exciter may be the only choice.
11.8 Concluding Remarks Active phased array radar performance can benefit significantly from digital beamforming. Digital beamforming arrays distribute receivers and A/D converters at the subarray or element level. We provided an overview of the many advantages of digital beamforming, including simultaneous independent beams, improved receiver dynamic range, angle tracking accuracy, and adaptive nulling. For moderate jammer requirements, sidelobe cancelers in analog phased arrays can provide a cost-effective nulling solution. However, digital beamformed arrays can handle far more interferers with much deeper nulls. For clutter attenuation, distributing the exciter at the element or subarray level can relax the exciter’s output noise requirements. Some of the digital beamforming array challenges include packaging large numbers of receivers and exciters and digital processing data at very high speeds in real time.
References [1] Steyskal, H., “Digital Beamforming Antennas,” Microwave J., January 1987, pp. 107–124. [2] Telisa, S., et al., “Benefits of Digital Phased Array Radars,” Proc. of the IEEE, Vol. 104, No. 3, March 2016, pp. 530–543. [3] Herd, J., “Array Element Pattern Correction in a Digital Beamforming Array,” URSI National Radio Science Meeting, Canada, 1985. [4] Simonangeli, L., and A. K. Agrawal, “A C-Band Digital Beamforming Array,” IEEE AP-S Int. Symp., Syracuse, NY, June 1988.
7065_Agrawal_V6.indd 211
4/6/23 1:43 PM
212
Practical Aspects of Active Phased Array Antenna Development [5] Haimerl, J. A., et al., “Overview of the Large Digital Arrays of the Space Fence Radar,” IEEE Int. Symp. on Phased Array Systems and Technology Boston, MA, 2016. [6] “How to Calculate ENOB for ADS Dynamic Performance Measurement,” Application Note 6854, Analog Devices, Maxim Integrated Products, Inc., June 24, 2019, downloaded from: https://www.maximintegrated.com/en/design/technical-documents/app-notes/6/6854.html. [7] Agrawal, A. K., and E. Holzman, “Beamforming Networks for Active Phased Array Antennas,” Special Issue on Phased Array Antennas, IEEE Trans. on Antenna and Propagation, Vol. AP-47, March 1999. [8] Allen, B., and M. Ghavami, Adaptive Array Systems: Fundamentals and Applications, John Wiley & Sons, 2005. [9] Applebaum, S. P., “Adaptive Arrays,” IEEE Trans. on Antenna and Propagation, Vol. AP-24, No. 5, 1976.
7065_Agrawal_V6.indd 212
4/6/23 1:43 PM
12 Cost Reduction Strategies for Active Phased Array Antennas 12.1 Introduction Active phased array radar technology continues to evolve to meet new and more demanding requirements. Concurrent with this need for increased performance, there is a practical need to reduce acquisition, operation and support costs, improve reliability, and reduce manning requirements [1, 2]. System users want the cost of the next generation of active phased array radars to be similar to the previous generation. Manufacturing technologies and automation advancements have enabled new systems to achieve this cost goal in many industries. This chapter discusses design and manufacturing techniques that enable active phased array antennas to achieve the lowest possible initial production and lifecycle costs. We concentrate our attention on land and shipborne platforms; however, the same cost-reduction techniques should also apply to other platforms. In the past, the primary focus for military array systems under development was on achieving the best performance with a cost of secondary importance. It is well known that most department of defense (DoD) development programs significantly overrun their award costs [3–5]. As the number of requirements has increased, systems have become more complex, requiring technology that 213
7065_Agrawal_V6.indd 213
4/6/23 1:43 PM
214
Practical Aspects of Active Phased Array Antenna Development
advances the state-of-the-art, yet acquisition and lifecycle costs are desired to be the same or less. This is a tough challenge indeed, and it will require paradigm shifts in design and production methods. The industry has successfully applied six sigma and other techniques to improve quality and reduce cost. However, we need to go further by addressing why systems cost so much in the first place and then make a paradigm shift in design and production methods. Changing how military systems are developed by learning from the commercial industry is necessary to achieve our cost goals. To produce products at a lower cost, one must treat cost like a performance requirement and address it during the development process, starting from the beginning of the program. Many of the most complex and costly systems are designed and produced by large corporations, but we cannot be limited by their intractable bureaucratic processes. It is possible to reduce the cost of systems that have already been developed and are in production, as illustrated by a couple of examples that follow. First, a discussion of the reasons for the high cost is provided, followed by a description of various components of an active phased array antenna. Next, we discuss how the cost of antenna systems that have already been developed can be reduced by 15% to 20%. Finally, a description of cost-reduction strategies is provided. The story of the SPY-1 phased array antenna, deployed on the Navy’s destroyers and cruisers, shows how engineers can conceive very innovative methods to meet the challenge of reducing the antenna production cost by 25% to 30%, even after the system already is in full rate production. These savings can be huge when the active array is produced in large quantities, as is the case for ships and aircraft. However, since many costs are locked in because of decisions made early in program development, much larger cost savings can be achieved if the cost-reduction strategies are applied from the beginning of the development program while production is still years away. We aim to provide general guidance to engineers and program managers to help them reduce the cost of active phased array antennas. First, we discuss the reasons for the high cost of military systems, followed by a description of various components of an active phased array antenna. Next, we discuss how the cost of antenna systems that have already been developed can be reduced by replacing some components with form, fit, and function and performance
7065_Agrawal_V6.indd 214
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
215
components and negotiating with suppliers to lower the costs of the components they provide. Finally, we describe cost-reduction strategies.
12.2 High Cost of Current Active Phased Array Antennas We describe the development process from contract award through design and engineering model demonstration to production and deployment to highlight the reasons for the high cost of active phased array antennas. The cost reduction strategies would be applicable to the fixed-price contracts. Almost all phased array antenna development contracts for the U.S. government involve significant technical and schedule risks, as they require advancing the current state of the art. Consequently, these contracts typically are cost-plus contracts, meaning they allow the prime contractor to recoup any costs over the original award. Generally, prior to the design and development of the first fully compliant engineering model, a technology risk retirement contract is awarded to one or more contractors to demonstrate the key technologies, such as the power amplifier and radiating element, that the program will require. After that, through a competitive bidding process [6], the government selects a contractor to develop the active phased array antenna and its subcomponents and build an engineering development model (EDM) to validate its performance against the requirements. After the development phase, nonrecurring engineering (NRE) and production costs are allocated for various subsystems or components. The NRE and production costs are generally estimated using a “bottoms-up” process described later in this chapter. The subsystems or components for an active phased array antenna include the radiating aperture, radome, T/R modules, beamformer, DC-to-DC converters, power and logic distributions, structure, cold plates, cables and interfaces, receivers, and signal processor. The components are procured outside or manufactured by the contractor and assembled into an active array system as part of integration and testing. The lead engineers for each subsystem are responsible for developing its NRE and recurring engineering (RE) cost estimates. During the development phase, the lead engineer supports his team in designing the subsystem to meet its requirements. The
7065_Agrawal_V6.indd 215
4/6/23 1:43 PM
216
Practical Aspects of Active Phased Array Antenna Development
critical elements of the development process are performance, schedule, and budget. Manufacturing costs are estimated during the development cycle using a cost model of the system, but the cost is secondary to a successful demonstration of a prototype system. The cost model depends on many assumptions regarding the future production of the system. Assumptions usually are optimistic and predict the system production cost to be significantly lower than it ends up being. The contractor always feels pressure to keep the cost model predictions in agreement with the program’s target costs. Therefore, it is not surprising to see a cost model that predicts production costs in line with the initial estimates. Since production occurs years in the future, there is little accountability. It is not unusual to see a program’s development and production costs increase by as much as 20% to 30% over the original estimate. After many years of development, often past the original schedule, the increased costs are accepted, as there is no more time and funding to change the design. And since the contract is cost-plus, the contractor never had a strong incentive to complete the work within the original award price since the customer pays the additional costs. Defense contractors have become accustomed to developing systems for the U.S. government on cost-plus programs. As a result, the government usually pays increased development and production costs, and neither the government sponsoring department nor the contractor is held accountable unless laws are broken. To reduce production costs, the U.S. government and its contractors must change the system development and production paradigm. Next, we describe some impediments to achieving lower costs in military array radar systems.
12.3 SPY-1 Array Antenna Cost Reduction In the early 1990s, not pleased with the production cost of the SPY-1 array antenna, the Navy decided to hold a competition to find a second source. As a result, a contract was awarded to Westinghouse Defense. However, the second source contract was canceled after General Electric (GE), the prime contractor, promised to reduce the manufacturing costs by approximately 25%. This example highlights one of the government’s primary problems—there is only
7065_Agrawal_V6.indd 216
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
217
one supplier for many of its most important and expensive military systems. Single suppliers, in effect, have a monopoly and, unless pressured by the threat of losing business, do not try harder than necessary to satisfy their customers. Fearing it might lose its SPY-1 antenna contract to Westinghouse, GE spent its own internal research funding and had its engineers rise to the challenge and develop innovative methods to reduce the SPY-1 array production cost by 25% to 30%, even as the array was in production. The GE engineering team decreased the array cost by reducing the cost of components and subassemblies purchased from external suppliers. GE engineers helped the external suppliers improve their manufacturing processes, so many components were replaced with parts with the same form, fit, and function without requiring any changes to the production array system produced in GE’s factory. As a result, the suppliers’ manufacturing processes were improved to lower their labor costs and thus the costs charged to GE. Similarly, the cost reductions for the SPY-3 phased array antenna was achieved by getting better prices from suppliers and making small changes in the design without affecting the antenna performance. Some methods used include alternate T/R module packaging, optimizing the test and manufacturing processes, and working with vendors to reduce costs by asking them what small changes can reduce the cost.
12.4 Improvements in Technology and Manufacturing Processes Improvements in technology and manufacturing processes have enabled products to be produced with improved performance for about the same price as their predecessors. For example, iPhones are upgraded with new features; however, the next generation with significantly higher performance sells for about the same price as the previous generation. Active array electronics technology is always improving. Even as the performance of GaN MMIC HPAs has improved as the technology has matured, the price per MMIC has not increased because foundry manufacturing processes continue to improve. A single SiGe RFIC can replace a dozen GaAs phase shifters, attenuator and gain amplifier MMICs, and cost less than one of the GaAs MMICs it replaces.
7065_Agrawal_V6.indd 217
4/6/23 1:43 PM
218
Practical Aspects of Active Phased Array Antenna Development
The performance requirements for phased array antennas continue to evolve. Therefore, it is not unreasonable for the government to expect the next generation of antennas to cost similar to their predecessors.
12.5 Paradigms This section describes the paradigms that are impediments to the development of active phased array antennas at a lower cost. 12.5.1 Legacy Systems Like other publicly traded corporations, defense companies are in the business of generating profits for their stockholders. The biggest defense contractors have developed sophisticated, high-performance radar systems that they have produced successfully. These legacy systems act as yardsticks for comparison with the cost of all future systems. It is natural to reuse designs and technologies that have been successful since design reuse is generally less risky than new development. An expert can often tell which company made a particular radar system and when by examining its architecture and packaging. There is absolutely nothing wrong with using proven techniques; however, these may not be the best approach for developing future systems at a lower cost. Using newer methods, materials, and architectures can lower production costs significantly. Engineers tend to be risk-averse because the reward for completing a system design and placing it into production generally is small, but the punishment for failure can be severe. 12.5.2 Commercial Parts and Processes Are Not Adequate for Military Applications Defense contractors often cannot use commercial parts because of environmental, reliability, or performance requirements, but they can learn from the commercial industry. Because they are screened to less stringent requirements [7], commercial parts cost an order of magnitude less than military parts, but they can fail catastrophically. In most cases, catastrophic failures are not acceptable for
7065_Agrawal_V6.indd 218
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
219
military systems. However, the commercial industry has come a long way in producing durable and quality products. Therefore, if the defense industry wants to lower costs, it must learn from the commercial industry and adapt its practices as much as possible. 12.5.3 Cost-Plus Contracts A majority of development contracts are cost-plus incentive fee types. Generally, the penalty is relatively small for cost overruns and schedule delays, as consideration is given to the hard work that people have put in. Since most new active phased array antenna requirements are technically challenging and may require inventions or a quantum leap in technology, contractors are not willing to accept the entire risk of overrunning fixed-price contracts. However, there is almost always a risk reduction program that addresses high-risk items. The best time to consider cost is during the initial development phase, when all designs, materials, and production techniques are selected. The defense industry is used to thinking that the U.S. government will pay for a cost and schedule delay, whatever its cost is. Such thinking is in the past. Today, the government demands a specified performance and price. When a contractor fails to meet performance or schedule, resulting in a significant system cost overrun, the system may not be deployed, or the production quantity may be drastically reduced. For example, due to higher costs, the Zumwalt class destroyer DDG-1000 ship production was cut to three from the original twenty-seven [5]. Similarly, cost overrun and schedule delays resulted in fewer F-35 aircraft (4). 12.5.4 Lack of Incentives As previously discussed, there is little incentive to reduce development and production costs for cost-plus contracts. The culture at all defense companies is not conducive to lowering costs. An engineer may get an extra pay increase and recognition for developing a successful product. On the other hand, the penalty for failure can be very severe. It is then apparent why engineers and program managers choose conservative approaches and do not take unnecessary risks. It is not suggested that they should take extreme risks to reduce cost, but there are calculated risks that one can take.
7065_Agrawal_V6.indd 219
4/6/23 1:43 PM
220
Practical Aspects of Active Phased Array Antenna Development
12.5.5 Schedule Limitations Do Not Permit Any Design Changes Most radar programs have challenging schedules since there is a need to deploy the system by a fixed date. When funding is delayed, the schedule may become even more challenging. During the early stages of the design, preliminary designs become more or less final as long as they meet requirements. Due to the impact on schedules and costs, there is little or no time to make design changes. 12.5.6 The Benefits of Competition to the Buyer: An Automobile Industry Example In the 1960s and 1970s, American automobile manufacturers claimed that quality products would cost much more than their customers would pay. They made money by maintaining vehicles with mediocre reliability over their lifetime. If it were not for the Japanese automobile manufacturers producing highly reliable cars for the U.S. market, we would still be driving poor-quality automobiles and paying a much higher price for them. In the commercial industry, competition is the key to improving product quality and driving prices lower. The Japanese made higher quality and more reliable automobiles for the same price as American manufacturers. American car producers lost significant market share to the Japanese before they could make improvements. The DoD acknowledges defense program competition is very limited [6]. 12.5.7 Use the Best Available Technology There is a temptation to use the best technology for a new radar system design. For example, active phased array antennas improve performance over passive phased array and reflector antennas that use high-power tube transmitters. Sometimes, the requirements could be satisfied with a reflector or a passive phased array antenna. However, the system cost is much less if the requirements can be satisfied with older and more mature technology. For example, to meet limited scan requirements, a small phased array antenna can be used to feed a reflector that provides a higher gain for the antenna. RF components are available in different qualities, and their prices vary accordingly. For example, RF connectors cost between $5 and $30, with the price difference partly related to the insertion loss of the connector. If a slight increase in the insertion loss does
7065_Agrawal_V6.indd 220
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
221
not adversely impact system parameters, the resulting cost savings could be significant because many arrays require thousands of RF connectors. Every part of the system does not need to be gold-plated. 12.5.8 Changes Will Increase Program Costs and Schedule Delays The majority of DoD programs have challenging schedules. Early designs quickly become final during the design phase, and alternate methods that could result in lower costs are not considered due to lack of time. If a better strategy is discovered, it is generally rejected as it may delay the schedule. During the preliminary design phase, the design team should conduct a thorough design trade study and evaluate all potential design solutions in terms of their schedule impact versus their potential cost savings to make appropriate decisions in a risk versus payoff trade. It takes a mindset from the beginning of a program to focus on the final production cost.
12.6 Design Philosophy Regarding system cost, there are two design philosophies for active phased arrays: bottom-up and top-down. 12.6.1 Bottom-Up The design and its performance requirements are based on available technologies and components in the bottom-up approach. Sometimes the customer’s desired system requirements are adjusted to what the technology can deliver. This approach may or may not meet the cost target since the system cost depends on the prices of available components. Because the system cost is incidental to the design and not a driver of the design, this approach is often more expensive than the top-down approach. The drawback of the bottom-up approach is that the prices of existing technology may cause the estimated system cost to be significantly higher than the goal. However, engineers can justify the cost by comparing it to other programs and adding a factor called relative complexity. Unfortunately, it isn’t easy to find details of each task in terms of labor hours and component cost. Contractors are
7065_Agrawal_V6.indd 221
4/6/23 1:43 PM
222
Practical Aspects of Active Phased Array Antenna Development
required to provide the basis of estimate (BOE) for each task. In the real situation, the total costs are estimated and later justified with a BOE. The other drawback of this method is that the inefficiencies of the previous programs continue in the new program. As discussed earlier, we must have a new mindset to control costs at every program stage. It is difficult to implement changes later as the cost and schedule may not permit any changes. 12.6.2 Top-Down The system is designed to meet a cost goal in the top-down approach. A cost model is developed to assign a cost for components and subassemblies, including assembly and testing. The first iteration of the cost model is based closely on the array antenna system architecture. Various trades are performed to determine the optimum array architecture to meet the cost goals while meeting the performance requirements. Sometimes, cost and performance are traded. Suppose the cost can be reduced significantly by giving up a small amount of performance. In that case, the systems engineer may reallocate the subsystem components’ requirements while still meeting the overall system performance. For example, using components with some commonality with commercial parts can reduce the cost of volume production. Environmental and reliability requirements also drive the system cost. For example, hermeticity requirements increase T/R module costs significantly. Rather than make each T/R module hermetic to meet the system requirements, the hermeticity requirement can also be met by controlling the environment inside the array housing instead and using nonhermetic modules. The end-item costs should be reviewed continuously and update the cost model. Once costs for all components and subassemblies have been allocated, the challenge is to continuously monitor the cost and make tweaks to the cost model against the goal of meeting the challenge at every design stage. As with other array requirements, like weight and prime power, monitored closely with budgets, if new model estimates indicate costs are exceeding their allocations, the design team must work to bring the costs in line with the allocated costs by making design changes while watching the schedule. The top-down approach can result in lower overall costs. It is reasonable to assume that we can reduce the system cost by at least
7065_Agrawal_V6.indd 222
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
223
15% to 20% compared to the bottom-up approach by focusing on cost at every stage and applying other methods described in this chapter.
12.7 Cost Reduction Strategies Prime contractors use the standard procedure of getting quotations from three vendors for array components and subassemblies and then choosing one of the vendors. The selection criteria are the vendor’s cost, capability to meet requirements, and associated risks. If all three vendors qualify, the contract is awarded to the lowest bidder. After that, there are no additional negotiations with the selected vendor to reduce the cost further. However, since we know the winning bidder must be making a profit, the contractor should be able to work with the vendor to reduce his product’s cost further, such as by focusing on cost drivers and evaluating the cost benefit of relaxing requirements. In addition, some manufacturing processes could be automated. The system program manager works with a limited budget allocation for every vendor. Sometimes small changes in performance margin can be traded with other parts of the array by the systems engineering team. At the end of the system development program, after an array EDM is manufactured, the costs of the system are well understood and become set going into the array production program. At this point, there are opportunities to reduce the costs by working with the vendors. Sometimes, the prime contractor may seek a second source for a critical component to force a vendor into accepting a reduced price for fear of not getting the contract. 12.7.1 Optimizing T/R Module RF Output Power Levels for Phased Array Antenna Cost, Size, Prime Power, and Dissipated Heat Let us examine the impact of T/R module RF output power level on active phased array antennas’ electrical and mechanical characteristics such as sensitivity, cost, size, prime power, and cooling system. The prime power capacity available at the array deployment site often is a critical requirement in determining the antenna architecture. For example, ground-based missile defense radars are deployed at remote locations and have to generate their own
7065_Agrawal_V6.indd 223
4/6/23 1:43 PM
224
Practical Aspects of Active Phased Array Antenna Development
power. Similarly, ships must generate all electrical power required by the electrical systems. Often, a ship’s power generation capacity is defined by a budget with allocations for all the subsystems, including active arrays, being powered. We describe trade-offs that assist radar systems engineers in making an optimum choice between T/R module power, cost, and prime power. Several alternate array architectures are described in this section for selecting an antenna size, power, and architecture. The radar range equation for a tracking radar can be written as follows [8]: 4 Rmax =
( 4p )
2
PtGt Ae s
(
kTo Bn Fn So /N o
)
(12.1) min
And the radar sensitivity is defined by the following: PtGt Ae s P S = r = N kTB ( 4p )2 kTB R 4 (12.2) n
where
Pt = total transmitted power Gt = transmit antenna gain Ae = receive antenna effective aperture
σ = radar cross-section Bn = noise bandwidth k = Boltzmann’s constant = 1.38 × 10–23 J/deg To = standard temperature of 290°K (approximately room temperature) kToBn = input noise power FN = noise figure (So/No)min = minimum signal-to-noise ratio The figure of merit (FoM) used for tracking and discriminating radars from the radar range equation is the transmit power-receive aperture-transmit gain PtAeGt product [8]. We use this figure
7065_Agrawal_V6.indd 224
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
225
of merit to compare the aperture size, the prime power required, and the cost for a range of T/R module peak power levels. Given that Pt = NP0 Ae ∼ N Gt ∼ N the FoM can be written as, PtAeGt ∝ P0N3
(12.3)
where P0 is the output power of the T/R module and N is the number of modules in the array. We have ignored front-end losses between the T/R module output and the radiating element. The larger the FoM, the farther the array radar can detect a target with a given radar cross-section at a specified signal-to-noise ratio. From (12.3), we note that sensitivity is proportional to PtAeGt, so radar sensitivity is proportional to the cube of the number of radiating elements but only directly proportional to the RF power output from the T/R modules. In other words, the FoM can be increased by a factor of eight by either increasing the T/R module output power by a factor of eight or by doubling the number of modules, as shown in the following: 8PtAeGt ∝ (8P0)N3 (8 times the module power, same number of elements)
(12.4)
8PtAeGt ∝ P0(2N)3 (same module power, twice the number of elements)
(12.5)
Doubling the array size increases the prime power requirement by a factor of two to power the array, but increasing the module power by a factor of eight, grows the required prime power by nearly an order of magnitude. Table 12.1 shows antenna costs, prime power requirements, and dissipated heat for obtaining eight times the reference power-aperture gain (PAG) for a notional system. The entries in Table 12.1 were calculated using a notional antenna cost model by allocating the costs of different array components as a percentage of the total array cost. For other systems’ cost models, these numbers will vary. However, Table 12.1 provides first-order estimates for comparison. For example, the prime power
7065_Agrawal_V6.indd 225
4/6/23 1:43 PM
226
Practical Aspects of Active Phased Array Antenna Development Table 12.1 Comparison of Array Cost and Prime Power for Array FoM of PAG and 8PAG as a Function of the Module Power and the Number of Modules Module Power
P0
8P0
P0
Number of Elements
N
N
2N
PAG
PAG
8∗PAG
8∗PAG
Item
Relative Cost
Relative Cost
Relative Cost
T/R/Control modules
0.5
0.85
1
DC-to-DC converters
0.1
0.5
0.2
AC-to-DC converters
0.05
0.3
0.1
Radiating aperture
0.05
0.05
0.1
Structure
0.05
0.05
0.1
Coldplates
0.03
0.03
0.06
RF combiners
0.04
0.04
0.08
Power/logic distribution
0.03
0.03
0.06
BSC
0.05
0.05
0.1
Cables
0.03
0.03
0.06
Assembly and test
0.06
0.06
0.12
Miscellaneous
0.01
0.01
0.02
Total cost
1
2
2
Relative prime power
1
8
2
Relative dissipated power
1
8
2
requirement for double the number of modules increases by a factor of two, while increasing the T/R module power by a factor of eight increases the power requirements by a factor of eight. Similarly, the dissipated power will be proportional to the prime power. Therefore, the prime and dissipated power for the 8P0 modules is four times that in the antenna with 2N elements. However, the cost of increasing the PAG by a factor of eight is about the same for both alternatives. On the other hand, the array could be built with 2N modules and one-eighth of the module power for about the same cost, while the prime power and heat dissipation would be reduced by a factor of four. The antenna weight with 2N lower power modules may not be significantly higher since the low power modules
7065_Agrawal_V6.indd 226
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
227
are lighter than the higher power modules; the weight of the power generator and cooling system is reduced significantly. For array systems deployed in remote areas, the cost of providing prime power over the lifetime of the antenna can be significantly higher. Sometimes, the module power and the number of modules can be optimized for prime power and cost. However, increasing the array size to reduce prime power may not be an option if installation space is limited, as on a ship. 12.7.2 Trading the Number of Array Faces for a Hemispherical Field of View A 360-degree hemispherical scan coverage can be provided by either three or four antenna apertures. For example, for a four-face array system, each face scans over a ±45 azimuth field of view. In contrast, each face must scan over a ±60° azimuth region for a threeface system, as shown in Figure 12.1. The problem of determining the optimum number of antenna faces to cover a hemisphere has been addressed by several authors [9–11] using various figures of merit. The figure of merit in [9–11] was chosen to be horizon or volume scan time with equal detection performance in all beam positions. It was concluded that the threeface system had the minimum scan time in all cases for passive or active arrays and whether the array was operating all faces simultaneously or sequentially. For phased array radar radars performing the tracking and discrimination functions, the important figure of merit for antenna performance is the PAG product. Jablon and Agrawal [12] determined the number of elements required for three- and four-face arrays to have comparable PAG performance for ground and shipboard active phased arrays. Two criteria were used to optimize the array design: (1) minimization of the total number of elements, and (2) minimization of the maximum scan loss in the coverage area. Therefore, the PAG performance was compared for the minimum number of elements and minimizing the loss, so the arrays had the same scan loss at extreme angles. We concluded that the three-face system had significantly fewer elements than the four-face system. However, when extra scan loss for the three-face system at the maximum angle was included, the
7065_Agrawal_V6.indd 227
4/6/23 1:43 PM
228
Practical Aspects of Active Phased Array Antenna Development
Figure 12.1 Three- and four-face array system deckhouse configurations for naval radar.
three- and four-face systems had the similar number of elements. However, when we add the cost of the structure and the receivers and signal processors for each face, the three-face system cost is lower than the four-face system. Therefore, shipborne SPY-3 and SPY-4 antennas have selected the three-face system for the cost advantage. 12.7.3 Band-Aid Solutions It is common to encounter many technical issues during the development, assembly, and testing of active phased array antennas,
7065_Agrawal_V6.indd 228
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
229
because arrays are complex systems with many components and complicated control systems. Such issues can surface during the validation testing of components or subassemblies or during the assembly and testing of the array. Technical problems arise at the array level for many reasons, including a component’s performance when installed in the array environment differing from that measured when the component was tested alone; design problems with the array, such as oscillating amplifiers, caused by electromagnetic interference and coupling; and manufacturing tolerance errors preventing proper fit between components and subassemblies. Technical problems that are uncovered during the preliminary design review (PDR) are much less costly to repair on paper than technical issues at the final assembly and test. During the development of an active array, problems that arise with the design are often fixed with Band-Aid solutions since parts have been procured and assembled, and the integration and test schedule rarely permits any significant redesign or modifications of components or the array architecture. The cost of fixing technical problems depends on when they occur in the array development process. Program management makes decisions to fix problems based on the cost and schedule of the program. Experience shows that ad hoc solutions often lead to more technical problems requiring their own ad hoc solutions. Boeing’s experience with the 737-Max aircraft is an example of how ad hoc solutions to problems with software resulted in a catastrophic system failure. Unfortunately, many similar examples can be found in almost all industries. 12.7.4 Antenna Architecture A key to cost control during active phased array antenna development is making the design and architecture as simple as possible. In addition, the array should be designed so that maintenance is as convenient as possible, as described in Chapter 8. Integrating components in LRUs reduces manufacturing costs by reducing the number of subassemblies. Finally, the lead systems engineer should have a clear vision of the end product, including the structure, manufacturability, maintainability, power distribution, and thermal management. Issues often occur during the final assembly of the array, and it is nearly impossible to change the design cost effectively at that stage. Sometimes too many other components
7065_Agrawal_V6.indd 229
4/6/23 1:43 PM
230
Practical Aspects of Active Phased Array Antenna Development
have to be disconnected to remove a particular LRU, making it difficult to maintain an array in the field. For example, access to components, such as T/R modules and power converters, which tend to fail more often than power and RF distribution, is critical for fast maintenance in the field. For example, an active phased array antenna is designed using printed circuit patch radiating elements. A cable connection is required to connect the T/R modules and radiating elements. Part of the beamformer network is installed between the radiating elements and the cold plates. Although the beamformer network consists of passive elements only, failure of any single connection would disable many radiating elements. Since the failed connections can not be replaced in the field, the array would have to be returned to the factory to fix that problem. As described in Chapter 7, the predominant packaging considerations associated with the mechanical design of active phased array antennas include designing for ease of maintenance, thermal management, DC power distribution, RF beamforming, radiating aperture design/interface, and structural robustness. A lowest-cost mechanical design can be achieved by following these general guidelines: • Most straightforward architecture for a particular application; • Use standard manufacturing processes; • Minimize the number of interfaces; • Minimize the number of parts; • Avoid a large number of connectors; • Integrate as much as possible; • Use COTS parts as much as possible; • Avoid requiring very high-precision machining; • Make designs tolerant to a certain number of system errors and failures; • Test 100% of components before higher-level assembly; • Maximize automatic testing; • Minimize touch labor. 12.7.5 Minimize the Number of Interfaces One of the array cost drivers is the number of RF, power, and logic interfaces between components. An array with many RF cables
7065_Agrawal_V6.indd 230
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
231
and connectors tends to be expensive and mechanically complex. In addition, meeting environmental requirements often requires an RF cable with a stable insertion phase over temperature and bending, which adds further to the cost. RF connectors that can be mated and demated during maintenance increase insertion loss and array cost. Therefore, innovative array mechanical designs that minimize interfaces are needed to minimize cost. Semirigid coaxial cables that are less than one-third the cost of phase-stable cables can be used where cable positions remain fixed after initial installation. Integration of components into LRUs also reduces the number of interfaces and costs. 12.7.6 LRU Size versus Cost Most T/R electronics are configured in subassemblies as LRUs to optimize the array design for ease of maintenance. A LRU can include T/R modules, DC-to-DC power converters, and control/ processor assemblies. A LRU may consist of 2, 4, 8, or 16 T/R modules and associated electronics. A control module and power supply can control and power condition a group of T/R channels in an LRU. The reliability and the impact of a failure on system performance determine whether a particular assembly can be an LRU. The size of the LRU also is determined by its maintainability, weight limitation for LRU removal, and cost per T/R channel. The impact of failures on sidelobe levels and reliability requirements [2] tends to drive the LRU size to be smaller, while the cost per channel tends to drive the LRU size to be larger. The reliability requirements for a large LRU can be met by providing redundancy in power supplies but may result in slightly higher array cost and more performance sensitivity to failures than if a smaller LRU is chosen. In summary, the LRU size is a compromise between reliability, sidelobe levels, cost per T/R channel, and maintainability. The mechanical layout of an LRU is an integral part of its design. An LRU may contain blind-mate connections to the radiating element, logic, power, and RF distribution. Various cooling techniques are used to keep the MMIC junction temperature low. Schemes involving exotic thermal materials to reduce the temperature drop between MMICs and cold plates should be avoided. We should keep the antenna array mechanical configuration as simple as possible at every design stage.
7065_Agrawal_V6.indd 231
4/6/23 1:43 PM
232
Practical Aspects of Active Phased Array Antenna Development
12.7.7 Radiating Element As discussed in Chapter 5, active phased array antennas may use several different kinds of radiating elements, including waveguides, patches, dipoles, and Vivaldi notch elements. Since all arrays require a metal housing to shield the RF, digital, and power electronics they contain, radiating elements must operate in the presence of a ground plane. Because the RF outputs of the T/R modules must connect to the radiating element, and many arrays are maintained from their aft end, blind-mate connections between the T/R modules and the radiating elements are common. Radiating elements are passive and usually do not fail, but the blind-mate connections can fail if too many are attempted simultaneously, or the mechanical design does not include an adequate apparatus for aligning and making the connections. Furthermore, as the array frequency band increases, the connectors become smaller, and the tolerances on blind-mate connections become tighter. Also, as the array operating frequency increases, it becomes more difficult to keep the spacing of the T/R module RF outputs lined up with the spacing of the elements. Sometimes, the radiating element spacing may not be the same as the T/R module spacing, requiring a transition or aperture dilation layer. The dilation layer introduces additional losses, complexity, and thus cost to the array.
12.7.8 T/R Modules The cost of a T/R module comprises the costs of MMIC devices, packaging, and labor for assembly and testing. Semiconductors are typically the most significant cost element within the T/R module, accounting for almost half of the T/R module cost. The packaging consists of multilayer substrate(s), metal housing, heat spreaders, and input and output connections. Packaging may be 35% to 40% of the module cost, and the remainder is for assembly and testing. The wafer processing cost for MMICs highly depends on the size and volume of wafers produced by a foundry. Current T/R module production rates do not require a large enough volume of semiconductor wafers to provide high foundry loading. Commercial volumes are several orders of magnitude higher than military volumes. Consequently, the key is to design MMICs using commercial processes and suppliers to provide MMICs for radar use. SiGe
7065_Agrawal_V6.indd 232
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
233
RFICs provide integration and processing on large wafers and provide huge cost savings compared to individual MMICs. For highpower RF, the combination of a GaN HPA/LNA MMIC and SiGe for all other T/R module functions may be the lowest-cost option. As discussed in Section 12.7.1, increasing the number of elements and reducing the module power reduces the prime power and cooling system requirements and the cost of those subsystems. In addition, the power amplifier junction temperatures would be lower, resulting in higher module reliability and lower lifecycle costs. If we use a large enough array, we can change semiconductor technologies. Recall our figure of merit, (12.3), is proportional to N3Pt, so we can replace a 500-element GaN array, transmitting 100 watts/ element with a 3,000-element SiGe array, transmitting 0.5 watts/ element, assuming the narrower beamwidth of the larger array is acceptable. If our SiGe RFICs have 16 T/R channels each, we need less than 200 SiGe RFICs for our 3,000-element array. Besides the 500 GaN MMICs, which will surely cost more than the 200 SiGe RFICs, the high-power array will also need LNAs, driver amplifiers, phase shifters, and variable attenuators. Those functions could be implemented with SiGe. Internally matched FET amplifiers [14] are sometimes used for higher-power commercial and military applications. Internally matched FET amplifiers consist of unmatched FET semiconductor devices with input and output matching circuits on a separate substrate. Since they lack distributed matching networks, the FET devices are very small compared to MMICs. They have a very high yield since semiconductor device yield is proportional to the device’s area. The GaAs or GaN wafers are thinned to 1 mil, and a layer of gold is applied on the backside for strength. By thinning the substrate to 1 mil, the temperature drop in the substrate is significantly reduced compared to the 3 mil or 4 mil thickness of a standard impedance-matched MMIC amplifier. A common misconception is that the input and output matching circuits used on internally matched FETs must be hand-tuned, thus adding additional labor costs. This is a misconception as Japanese companies sell internally matched power amplifiers as packaged devices, requiring no additional tuning. A computer generates the matching circuits, and automated wire bonding machines do the necessary tuning, resulting in a relatively small additional cost. Since this product is marketed for base stations, these devices can be obtained
7065_Agrawal_V6.indd 233
4/6/23 1:43 PM
234
Practical Aspects of Active Phased Array Antenna Development
at commercial prices. This technique of externally matching FETs is applicable up to X-band [14], where the spacing between elements allows for such components. The point is to select features with some commercial applications to benefit from the higher volumes. The packaging for the control circuit of the T/R module requires a high number of interconnections for the phase shifter and attenuator. So multilayer substrates are attractive because they minimize the module footprint and reduce coupling between the microwave and control signals. The number of interconnects can be reduced using MMICs with higher integration levels. As discussed in Chapter 4, a T/R module can be made with only three MMIC chips. And with SiGe, all control, RF and regulation circuitry for many T/R channels can be included in a single RFIC [15]. The requirements for individual module components play an important role in minimizing cost. For example, the yield of an HPA MMIC depends on the output power, gain, and efficiency requirements specified for acceptance. For an active array containing hundreds or thousands of HPAs, the MMIC requirements should be defined as lot averages, with the widest ranges acceptable, and the lot size should be as large as practical. As long as the average of the HPA output power in an array exceeds the requirement, the array will meet its EIRP requirement, even though some devices perform better than average and some worse than average. 12.7.9 Module Packaging As the cost of MMIC devices has decreased significantly over time, the module packaging costs have not decreased proportionately because the packages consist of metals and substrates. The cost of the module packages can be reduced significantly by using alternate packaging technologies and providing MMIC chip environmental protection by controlling the environment and applying a very thin layer of organic coating over the chip that does not significantly degrade the chip’s RF performance. In addition, cheaper organic substrates can be used for signal routing. However, most high-volume commercial applications use plastic packaging. Commercial modules also use flip chip mounting for low-power devices that can be surface mounted. The interconnect parasitics can change the performance of the devices. High-power amplifiers
7065_Agrawal_V6.indd 234
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
235
often need to be mounted on heat sinks to provide a direct thermal path to the coolant (as shown in Figure 7.4). We can control the internal environment of the array housing to compensate for the lower hermeticity of plastic-packaged devices. Humidity inside the enclosure can easily be controlled using dehumidifiers and desiccants. Moreover, providing environmental conditioning can be much cheaper than using hermetic packages. 12.7.10 DC Power Distribution As discussed earlier, DC power can be distributed using either a distributed or a centralized scheme. The distributed scheme includes power supplies as part of T/R module assemblies, making them larger and increasing the antenna depth. However, high voltage, low current bus bars distribute DC power to these supplies. Centralized power supplies can be placed around the array aperture in groups, requiring a larger array package footprint. However, larger, low voltage, high current bus bars must distribute the DC power to the modules. The distributed and centralized schemes are relatively equivalent in cost. Power supplies for radar applications are relatively expensive. On the other hand, commercial power supplies made in large volumes are cheaper by order of magnitude. If commercial supplies cannot be used as is, the next best way to reduce an array’s power converter cost is to convince a commercial manufacturer to add any additional custom circuitry and manufacture the power supply using the commercial manufacturing process. This modified power supply would cost more than the commercial power supply, but it would still be much cheaper than those specifically designed for radar applications. A factor of four or more reduction in cost is possible. 12.7.11 Beamformers, Cables, and Connectors Since the transmit and receive beamformers for active arrays have to handle very low power, low-cost printed circuit organic boards can be used for beamforming. The cost of beamformers for a monopulse active array antenna is generally less than 7% of the total array cost. Coaxial connectors are the most expensive components in beamformer assemblies. Therefore, the key to reducing beamformer cost
7065_Agrawal_V6.indd 235
4/6/23 1:43 PM
236
Practical Aspects of Active Phased Array Antenna Development
is to design a beamformer with a minimal number of connectors. The cost of the connectors can vary widely. Lower-cost connectors have a slightly higher loss than higher-cost connectors. Other contributors to beamformer cost are the number of circuit board layers required to route all signals and the choice of circuit board dielectric substrate, which tends to be a trade between RF loss and cost. In an active array, the power amplifier and LNA are placed at the aperture; therefore, a slight increase in beamformer loss will not significantly impact the antenna performance. 12.7.12 Power-Added-Efficiency and Cost T/R modules and power supplies account for 70% to 80% of the prime power and dissipated heat for an active phased array antenna. The power-added efficiency of the active devices in T/R modules directly impacts the power-added efficiency of the system. The prime power requirements for shipborne radars should be minimized to reduce the quantity of fuel that has to be supplied at sea and minimize the number of port calls. It is even more critical for land-based radars deployed in remote locations. For example, the Space Fence Active Array radar is located on Kwajalein island [16–17], very far from other United States major military installations in the Pacific: 1,600 miles from Guam and more than 2,000 miles from Hawaii and Japan. It takes a dedicated supply chain to keep it fully operational. 12.7.13 Active Phased Array Antennas for Wide Bandwidth Operation A phased array antenna’s instantaneous bandwidth (IBW) is inversely proportional to the antenna size. The bandwidth restriction due to beam squint and grating lobes can be eliminated if the true-time delay at each element replaces the phase shift steering [13]. However, time delay devices tend to be large, so it is not practical to package true-time delays at each element. Instead, we can divide an array into subarrays and place a true-time delay circuit behind each subarray, as shown in Figure 12.2. Dividing an array into N subarrays increases its IBW by a factor of N. Since the spacing between time delays exceeds the element spacing, grating lobes are only partially suppressed by the subarray architecture. Further
7065_Agrawal_V6.indd 236
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
237
Figure 12.2 Wideband subarrayed architecture with TDUs.
reduction of grating lobe peaks is possible if overlapped subarrays are used. The packaging of a wideband array is more challenging than that of a narrowband array. From the cost point of view, larger subarrays are preferred; and smaller subarrays are preferred from the grating lobe point of view. Wideband active phased arrays are more expensive than narrowband active phased arrays because of the additional cost of overlapped subarray beamformers and TDUs. 12.7.14 Antenna Assembly and Test Transmit and receive modules, LRU assemblies, and DC-to-DC converters make up 70% to 80% of active phased array costs. The key to reducing their costs is focusing on productivity and eliminating touch labor during assembly. Due to the highly redundant nature of the array architecture, many LRUs are produced in sufficient quantities to warrant volume production. The key to reducing
7065_Agrawal_V6.indd 237
4/6/23 1:43 PM
238
Practical Aspects of Active Phased Array Antenna Development
assembly costs is to make the designs of components, subassemblies, LRUs, and antenna structures as simple as possible. The processes for fabricating components and subassemblies should not be highly specialized and component specific. Where possible, the components should be designed so they can be produced using commercial processes. The array assembly costs include assembling all components and subassemblies, including logic circuits, DC power distribution, input and output interfaces, liquid cold plates, and RF beamformers, into the array structure. Components should be integrated into subassemblies and LRUs that can easily be installed and replaced to reduce assembly costs. Avoid complex designs that make the maintenance of the array on the ship difficult and that require highly trained personnel. The array test costs include pretesting subassemblies and LRUs before they are integrated with the array, followed by the near-field array test for performance verification. Array tests include array alignment, sidelobe pattern measurements for all four monopulse beams, EIRP, gain, scan loss, calibration constants, and verification of the calibration method over the instantaneous bandwidth. Pretesting of active components avoids problems arising after the array is assembled during the system test. At that point, access to failed components can require the time-consuming disassembly of large portions of the array. For example, all T/R modules’ performance parameters, such as output power, NF, and receive gain, for all amplitude and phase settings, should be measured before the modules are assembled into the array. The amount of T/R module testing can be reduced by testing major amplitude and phase bits and generating just one pair of amplitude versus phase and phase versus amplitude linearization tables, assuming all T/R modules behave the same way.
12.8 Concluding Remarks This chapter overviewed strategies for reducing active phased array antennas’ acquisition and lifecycle costs. With advances in technologies and manufacturing methods and applying cost-reduction strategies, new active phased array antennas can be developed at a
7065_Agrawal_V6.indd 238
4/6/23 1:43 PM
Cost Reduction Strategies for Active Phased Array Antennas
239
cost similar to their predecessors. A top-down design approach is preferred over the bottom-up approach to achieve the lowest costs. It takes a mindset at every development stage and overcoming paradigms that impede achieving lower costs. Finally, we must learn from the commercial industry, explore synergies, and use commercial production methods as much as possible to develop affordable active phased array antenna systems.
References [1] Agrawal, A. K., Kopp, B., Luesse, M., and K. O’Haver, “Active Phased Array Antenna Development for Modern Shipboard Radar Systems,” APL Technical Digest, Vol. 22, No. 4, 2001. [2] Agrawal, A. K., and E. Holzman, “Active Phased Array Design for High Reliability,” IEEE Trans. on Aerospace and Electronic Systems, Vol. 35, October 1999, pp. 1204–1210. [3] “Can We Afford Our Own Future? Why A&D Programs Are Late and Over-Budget—and What Can Be Done to Fix the Problem,” Deloitte Development, 2009, https://www2.deloitte.com/content/dam/Deloitte/us/Documents/manufacturing/us-ad-canweaffordourownfuture-0127.pdf. [4]
Woody C., https://www.businessinsider.com/us-reduces-F35-buys-as-other-militaries -buy-stealth-fighters-2022-3.
[5] Department of Defense, “DDG 1000 Zumwalk Class Destroyer (DDG 1000),” December 2021 Selected Acquisition Report, April 2022, https://www.esd.whs .mil/Portals/54/Documents/FOID/Reading%20Room/Selected_Acquisition_ Reports/FY_2021_SARS/22-F-0762_DDG_1000_SAR_2021.pdf. [6] Department of Defense, “State of Competition within the Defense Industrial Base,” https://media.defense.gov/2022/Feb/15/2002939087/-1/-1/1/state-of -competition-within-the-defense-industrial-base.pdf, February 2022. [7] Cidley, A., “Use COTS Parts to Cut Costs in Military and Aerospace Systems,” Electronic Design, accessed 1/20/2023, https://www.electronicdesign .com/technologies/components/article/21799319/use-cots-parts-to-cut-costs -in-military-and-aerospace-systems. [8] Skolnik, M., Introduction to Radar Systems, McGraw-Hill, 1980. [9] Trunk, G. V., and D. P. Patel, “Optimal Number of Phased Array Faces for Horizon Surveillance,” IEEE Int. Symp. on Phased Array Systems and Technology, 1996, pp. 214–216. [10] Trunk, G. V., and D. P. Patel, “Optimal Number of Phased Array Faces and Signal Processors for Horizon Surveillance,” IEEE Trans. on Aerospace and Electronic Systems, Vol. 33, No. 3, July 1997, pp. 1002–1006.
7065_Agrawal_V6.indd 239
4/6/23 1:43 PM
240
Practical Aspects of Active Phased Array Antenna Development [11] Waters, W. M., Patel, D. P., and G. V. Trunk, “Optimal Number of Faces of a Volume-Scanning Active Array Radar,” IEEE Trans. on Aerospace and Electronic Systems, Vol. 34, No. 3, July 1998, pp. 1032–1037. [12] Jablon, A., and A. K. Agrawal, “Optimal Number of Array Faces for Active Phased Array Radars,” IEEE Trans. on Aerospace and Electronics, January 2001. [13] Frank, J., “Bandwidth Criteria for Phased Array Antennas,” in Phased Array Antennas, A. Oliner and G. Knittel, eds., Artech House, pp. 243–253, 1972. [14] Kimura, M., et al., “GaN X-Band 43% Internally-Matched FET with 60W Output Power,” IEEE Asia-Pacific Microwave Conference, 2008. [15] Rebeiz, G. M., et al., “Highly Dense Microwave and Millimeter-Wave Phased Array T/R Modules and Butler Matrices Using CMOS and SiGe RFICs,” IEEE Symp. on Phased Arrays and Technology, 2010. [16] “Unprecedented Space Situational Awareness with Kwajalein Space Fence Radar,” The Kwajalein Hourglass, Vol. 61, No. 2, January 11, 2020, p. 4. https://www.smdc. army.mil/Portals/38/Documents/Publications/Hourglass/2020/01-11-20Hourglass.pdf. [17] “On-Site Owner Representation of Space Fence,” Nationwide Consulting LLC, https://www.nationwideconsultingllc.com/on-site-owner-representation/.
7065_Agrawal_V6.indd 240
4/6/23 1:43 PM
Appendix
T/R Module Requirements and Flow Down to the Components T/R Module Requirements Flow Down (Transmit Channel) • RF Frequency Band • Operating band • Instantaneous bandwidth • RF Transmit Output Power • Peak and average power output in watts • Minimum over the frequency band • Lot average • RMS variation • RF Transmit Input Power • Minimum input RF power needed over the frequency band • Maximum input RF power needed over the frequency band • Lot average • RMS variation • Maximum Duty Cycle in Percentage • Determines average power • Determines total dissipated heat 241
7065_Agrawal_V6.indd 241
4/6/23 1:43 PM
242
Practical Aspects of Active Phased Array Antenna Development
• Maximum Pulse Width • Affects device junction temperature • Junction temperature and dissipated heat removal capacity limit the pulse width • Minimum Pulse Width Determined by the T/R Switching Time • Pulse Fall and Rise Time • Waveform related • Limits minimum pulse width • Maximum Pulse Repetitive Frequency • Maximum number of times T/ R switching operations per second • Affects control circuitry • Buffer storage capacity • Communication between BSC and T/R module • Power-Added Efficiency • Dissipated heat • Junction temperatures • Power supply (DC-to-DC converter power requirements) • Total system power requirement • Amplitude Pulse Droop • Time Sidelobes • Phase Pulse Droop • Time sidelobes • Pulse-to-Pulse Random Amplitude and Phase Error (Amplitude and Phase Noise) • Clutter improvement factor • RMS Amplitude and Phase Error • RMS sidelobe level • Amplitude (Gain) Control • Range determined by the amplitude taper • Number of bits determined by the RMS amplitude error • Normally uniform taper, but some range is required to equalize module gains • Phase control • 360° phase control required • Number of bits determined by the RMS phase error and RMS sidelobe level • Harmonic Output • Allowed radiated harmonic power
7065_Agrawal_V6.indd 242
4/6/23 1:43 PM
T/R Module Requirements and Flow Down to the Components
243
• Interference with other systems • May use a harmonic filter to reduce out-of-band harmonics • Spurious Output • Allowed radiated spurious power • Interference with other systems • Power Pushing Factor • Variation of output power with input power, DC voltages (drain, gate), and temperature • Determines the amount of voltage regulation required • Contributes to amplitude error • Input Voltage Standing Wave Ratio (VSWR) • VSWR at the RF input port, typically connected to the beamformer • Load VSWR, typically connected to the radiating element • VSWR at the output port • Load pull effect T/R Module Requirements Flow Down (Receive Channel) • Frequency Band • Frequency band set by the antenna requirements • Instantaneous Bandwidth • The frequency bandwidth of the input signal • Affects module amplitude and phase errors • Errors are defined at the center of the instantaneous frequency band • Module Receive Gain • Determined by the NF and TOI requirements • Maximum and minimum gain • Lot average • RMS variation • NF • Determined by the antenna noise temperature (NF) requirements • Lot average and maximum • RMS variation • Phase Control • 360° phase control required • Number of bits determined by the RMS phase error and RMS sidelobe level
7065_Agrawal_V6.indd 243
4/6/23 1:43 PM
244
Practical Aspects of Active Phased Array Antenna Development
• Gain Control • Range determined by the amplitude taper • Number of bits determined by the RMS amplitude error • Pulse-to-Pulse Random Amplitude and Phase Error • Clutter improvement factor • RMS Phase Error • Antenna RMS sidelobe level • RMS Amplitude Error • Antenna RMS sidelobe level • Gain Pushing Factor • Variation of gain with input power, DC voltages (drain, gate), and temperature • Determines the amount of voltage regulation required • Contributes to amplitude error • Phase Pushing Factor • Variation of phase with input power, DC voltages (drain, gate), and temperature • Determines the amount of voltage regulation required • Contributes to phase error • Input TOI • Establishes the receiver dynamic range • Lot average • Receive Protection • Interference signals from external sources • Reflections from the antenna • Must handle the total reflection from the antenna • Transmit power • Power Handling • From external sources • Reflections from the antenna • Greater than or equal to the module output power • Limited recovery time • Input VSWR, typically connected to the radiating element • Load VSWR, typically connected to the beamformer • Spectral purity • Number of parallel receive channels • Switching time (T/R, R/T) • MTBF • Input voltages and power • Manufacturing cost
7065_Agrawal_V6.indd 244
4/6/23 1:43 PM
T/R Module Requirements and Flow Down to the Components
245
• Size • Determined by the array beamwidth, the element grid spacing determines • Weight • From antenna budget
7065_Agrawal_V6.indd 245
4/6/23 1:43 PM
7065_Agrawal_V6.indd 246
4/6/23 1:43 PM
List of Acronyms A/D Analog-to-digital ADC
Analog-to-digital converter
ASIC
Application-specific integrated circuit
AWACS
Airborne Warning and Control System
AZ Azimuth EL Elevation BIT
Built-in test
BSC
Beam steering controller
BW Bandwidth CIF
Clutter improvement factor
COTS
Commercial off the shelf
CTE
Coefficient of thermal expansion
CVN
Nomenclature for aircraft carrier
DAC
Digital-to-analog converter
dB Decibel DBF
Digital beamformer
DDG
Nomenclature for destroyer ships 247
7065_Agrawal_V6.indd 247
4/6/23 1:43 PM
248
Practical Aspects of Active Phased Array Antenna Development
DSP
Digital signal processor
DoD
Department of Defense
DPDT
Double pole double throw
DR Driver eDBF
Elemental digital beamformer
EEPROM Electrically erasable programmable read-only memory EL Elevation
7065_Agrawal_V6.indd 248
EMP
Electromagnetic pulse
FD/FI
Fault detection/fault isolation
FPGA
Field programmable gate array
GaAs
Gallium arsenide
GaN
Gallium nitride
G/ATOR
Ground/air task oriented radar
GLB
Grating lobe boundary
HEXFET
Trademark for power
HFSS
High-frequency structure simulator
HPA
High-power amplifier
HTCC
High temperature cofired ceramic
IBW
Instantaneous bandwidth
IM3
Third-order intermodulation product
InP
Indium Phosphite
LNA
Low noise amplifier
LRDR
Long range defense radar
LRU
Line or least replaceable unit
LTCC
Low-temperature cofired ceramic
LSB
Least significant bit
4/6/23 1:43 PM
List of Acronyms
MCM
Multichip module
MESFET
Metal-semiconductor field-effect transistor
MMIC
Monolithic microwave integrated circuit
MSSL
Mean-squared sidelobe level
MTBCF
Mean time between critical failures
MTBF
Mean time between failures
MTI
Moving target indicator
NF
Noise figure
NRE
Nonrecurring expenser
P1 dB
1-dB compression point
PAE
Power added efficiency
PAG
Power aperture gain
PDR
Preliminary design review
249
PFTE Polytetrafluoroethylene PRF
Pulse repetitive frequency
PS
Power supply
RCC
Radar central computer
RCS
Radar cross-section
REX
Receiver exciter
RFIC
Radio frequency integrated circuit
RMS
Root mean square
RSB
Required scan boundary
SiGe
Silicon germanium
SLC
Sidelobe canceller
SLL
Sidelobe level
SPDT
Single pole double throw
T/R Transmit/receive
7065_Agrawal_V6.indd 249
4/6/23 1:43 PM
250
7065_Agrawal_V6.indd 250
Practical Aspects of Active Phased Array Antenna Development
TDU
Time delay unit
THAAD
Theater high altitude area defense
TOI
Third-order intercept
VGA
Variable gain amplifier
4/6/23 1:43 PM
About the Author Ashok K. Agrawal received MS and PhD degrees in electrical engineering from the University of New Mexico in 1976 and 1979, respectively. He has more than 35 years of experience developing and manufacturing active phased array antennas. He also provided oversight of the Navy’s two active phased array antennas, SPY-3 and SPY-4, for the Zumwalt class DDG-1000 ships and Ford Class CVN-78 aircraft carriers. For several years, he taught a three-day short course on active phased array antenna development and manufacturing. He has published numerous articles on the subject and holds five U.S. patents.
251
7065_Agrawal_V6.indd 251
4/6/23 1:43 PM
7065_Agrawal_V6.indd 252
4/6/23 1:43 PM
Index A Accuracy angle of arrival, 27 angle tracking, 207 beam pointing, 40 calibration, 190, 196–97 Acronyms, this book, 247–50 Active array technology, 2, 38 Active impedance, 114 Active phased array antennas about, 4–5 availability and sparing, 173–74 bandwidth criteria for, 44–47 beamformer architectures for, 83, 90–94 block diagram, 5, 138 calibration, 185–98 components, xv, 4–5, 138 cost reduction strategies, 213–39 design for high reliability, 155–74 digital beamforming for, 201–11 directivity, 38–40 with driver amplifiers in beamformers, 180 environmental requirements, 137 gain, 40 high cost of, 215–16 implementation of monopulse beams for, 26–27 inputs, 4 key radar system-level advantages, 10–12 lifecycle maintenance cost estimation, 170–73 linear, 9 multiple independent receive beams, 91–94
operating frequency, 138 orders of magnitude, xv–xvi packaging, 137–53 performance improvement, 59–61 power distribution, 129 practical aspects of development, 1–18 RF block diagrams, 9 two-dimensional, 9 for wide bandwidth operation, 236–37 Active-weighted architecture, 95 Adaptive cancellation techniques, 208 Adaptive digital beamforming about, 11–12, 208–9 nulling in analog arrays, 209–10 processing diagram, 209 Amplitude errors effect on array patterns, 88 effect on performance, 40–44 VGA in correcting, 93 See also Errors; Phase errors Analog to digital converter (ADC), 17, 201, 203, 204 Antenna architecture, in cost reduction, 229–30 Aperture distributions Bayliss, 26 Dolph-Chebyshev, 23–24 low sidelobe, 23–27 Taylor, 25–26 Aperture effects, 46 Application-specific integrated circuit (ASIC), 56, 126 Array alignment, 16, 185, 193, 238 253
7065_Agrawal_V6.indd 253
4/6/23 1:43 PM
254
Practical Aspects of Active Phased Array Antenna Development
Array factor about, 21 maximum, 21–22 planar array with, 28 rectangular lattice, 28 triangular lattice, 31 Array maximum, 22 Array steering, 22 Assembly costs, 237–38 Attenuation array edges and, 96 bits, 128 clutter, 210–11 quantization errors, 40–41 Autocalibration, 188 Automated maintenance, 172–73 Availability and sparing, 173–74 Azimuth difference (AZ DELTA), 8, 14, 83, 89, 93 B Band-aid solutions, 228–29 Bandwidth criteria, 44–47 instantaneous (IBW), 44, 45–46, 236 operating, 44, 46–47 wide, operation, 236–37 Beamformer architectures for active phased array antennas, 90–94 active-weighted, 95 complexity of, 89 hybrid couplers, 90 impact on system noise figure, 94–97 introduction to, 83–86 for passive phased array antennas, 86–90 passive-weighted, 95–96 reliability and, 97 for wideband active phased array antennas, 97–101 Beamformers about, xv, 4 active array architecture with driver amplifiers in, 180 cost reduction and, 235–36 driver amplifier boosters in, 169–70 for linear passive phased array antenna, 86 passive array, 88 for two-dimensional passive phased array antenna, 87 Beam scanning, aperture effects, 46
7065_Agrawal_V6.indd 254
Beam steering controller (BSC) about, xv, 4, 14, 125 centralized, 128 distributed, 126–28 functions, 125–26 Bottom-up design, 221–22 Brick packaging about, 141 edge-cooled, horizontal cold plate, 146–48 fault isolation, 142 heat transfer and, 141 schemes, 144–49 sliding vertical cold plate, 145–46, 147 vertical fixed cold plate, 148–49 See also Packaging Built-in test (BIT), 14, 173 Built-in test equipment (BITE), 127–28 C Calibration array alignment, 185 autocalibration, 188 of calibration unit, 197 conclusion, 198 dedicated elements, 192–98 element, mutual coupling and, 190–91 field, 186–98 initial accuracy of, 190 introduction to, 185–86 mutual coupling and, 186–98 phase errors, 191 single-element, 190–91 techniques, 186–98 Centralized BSC, 128–29 Centralized phase array architecture, 176–79 Centralized power system, 132–33 Changes, cost reduction and, 220, 221 Circular polarization, 69–70 Clutter attenuation, 210–11 surface, 175 volume, 175 Clutter improvement factor (CIF) about, 16 centralized phase array architecture and, 176–79 as dBs, 177 distributed array architecture and, 179–82 driver amplifiers and, 182, 183 high, amplifier module supporting, 179 high, design for, 175–83
4/6/23 1:43 PM
Index introduction to, 175–76 maximum, 177, 181, 182 number of T/R modules and, 180, 182–83 power supply ripple and, 183 pulse-to-pulse instabilities and, 176 summary, 182–83 total, 177–78 Common leg T/R module architecture, 58–59 Competition benefits, 220 Component failure rates, 156–57 Construction methods, 141 Control module about, 56 cost of, 76–78 integration, 73–74 LRU, 73–74 packaging, 234 PRF rate and, 56–57 reliability, 75 See also T/R modules Cooling methods, 141 Correlated errors, 40 Cost, reduced, 11 Cost-plus contracts, 219 Cost reduction about, 17 antenna architecture, 229–30 antenna assembly and test, 237–38 band-aid solutions, 228–29 beamformers, cables, connectors and, 235–36 changes and, 221 commercial parts and accessories and, 218–19 competition benefits and, 220 conclusions, 238–39 cost-plus contracts and, 219 DC power distribution and, 235 design philosophy, 221–23 guidance, 214–15 hemispherical field of view, 227–28 high current costs and, 215–16 impediment paradigms, 218–21 introduction to, 213–15 lack of incentives and, 219 legacy systems and, 218 LRU size versus cost, 231 manufacturing process improvements, 217–18 minimizing number of interfaces and, 230–31 module packaging and, 234–35 optimizing T/R module RF output, 223–27 power-added efficiency and, 236
7065_Agrawal_V6.indd 255
255
radiating elements and, 232 schedule limitations and, 220 SPY-1 array antenna example, 214, 216–17 strategies, 223–38 technology improvements and, 217–18, 220–21 T/R modules and, 232–34 wide bandwidth operation and, 236–37 Cross-polarization radiation, 104, 106 D DC power distribution, 235–36 DC-to-DC converters about, xv, 4 average versus peak, 133–34 centralized power system, 132–33, 134 distributed power system, 130–32 failure of, 131 heat generation, 144 high-power generation, 78 integration, 73–74 key parameters, 130 MTBF, 131, 143 peak capacitor multiplying, 134 redundancy, 131 T/R module integration with, 57–58 voltages, 57 Dedicated element calibration about, 191–93 accuracy, 196–97 effect on packaging, 197–98 passive calibration elements, 192–93 procedure, 193–95 required number of calibration elements, 195–96 See also Calibration Design philosophy, 221–23 Difference patterns, Bayliss distribution for, 26 Digital attenuators, 40 Digital beamforming (DBF) about, xvi, 11 adaptive, 208–10 angle tracking, 207 architecture, 203 clutter attenuation, 210–11 concepts, 203 dynamic range improvement, 203–4 exciter noise, 210–11 introduction to, 201 of multiple receiver beams, 206–7 primary benefits, 201–2 at subarray level, 204–6
4/6/23 1:43 PM
256
Practical Aspects of Active Phased Array Antenna Development
Digital signal processor (DSP), 128 Digital-to-analog converter (DAC), 202 Directivity about, 38 active arrays, 40 higher-frequency arrays and, 138 max, 39–40 sidelobe levels and, 23–24 Distributed array architecture, 179–82 Distributed BSC, 126–28 Distributed power system about, 130–31 centralized power system comparison, 134 failure of DC-to-DC converters and, 131 illustrated, 131 key features of, 132 See also Power distribution Dolph-Chebyshev aperture distribution, 23–24 Double-pole double-throw (DPDT), 92 Driver amplifier boosters, 169–70 Dynamic range, DBF and, 203–4 E Edge-cooled, horizontal cold plate packaging, 146–48 Effective isotropic radiated power (EIRP), 195–96 Electrically erasable programmable read-only memory (EEPROM), 126 Electromagnetic environmental effects (EME), 103 Electromagnetic pulse (EMP), 103 Elemental digital beamformed (eDBF) arrays, 204–5 Elevation difference (EL DELTA), 8, 14, 83, 89, 93 Engineering development model (EDM), 215 Errors amplitude, 40–44, 88, 93 correlated, 40 effect on performance, 40–44 phase, 40–44, 88–89, 191 quantization, 40–42 random, 40 RMS sidelobe level due to, 42–44 Exciter noise, 210–11 F Fabric radome, 152 FD/FI subsystem, 172–73 FET devices, 233 Field calibration about, 16–17, 186
7065_Agrawal_V6.indd 256
active array and external horns for, 187 with mutual coupling between array and external elements, 186–87 with mutual coupling between array elements, 188–90, 192 with mutual coupling between calibration element and array elements, 190–91 with mutual coupling between dedicated elements and array elements, 191–98 self-calibration, 188 techniques, 186 See also Calibration Field-programmable gate array (FPGA), 56, 126 Figure of merit (FoM), 224 G Gain active arrays, 40 maximum, T/R modules, 96 power aperture (PAG), 12–13, 225–27 Gallium arsenide (GaAs) devices, 3 Gallium nitride (GaN) about, 3 advantages, 60 power generation, 60–61 thermal conductivity, 61 wide bandgap power amplifiers, 60–61 GE engineering, 216–17 GRAFOIL, 72 Grating lobe boundaries (GLB), 35, 37 Grating lobe circles rectangular lattice, 29 triangular lattice, 33 Grating lobes, 22–23, 29, 32, 34–37, 47–50 H Hemispherical field of view, 227–28 High Frequency Structure Stimulator (HFSS), 114 High-power amplifiers (HPAs) about, 54–56 antenna MTBF and, 156–58 Moly spacer conductivity, 72 voltage distribution, 71 High reliability architecture description for, 158–59 design, 155–74 driver amplifier boosters and, 169–70 introduction to, 155–56 MTBCF and, 160–62 MTBCF with redundant power supplies, 166–68
4/6/23 1:43 PM
Index MTBF and, 156–58 MTBF for different cluster sizes, 162–66 High-temperature cofired ceramic (HTCC), 73 Hybrid couplers, 90 I Ice inhibition, 110–12 Incentives, lack of, 219 Instantaneous bandwidth (IBW), 44, 45–46, 236 Interfaces, limiting number of, 230–31 L Legacy systems, 218 Lifecycle maintenance cost estimation, 170–73 Linear arrays with amplitude and phase control, 20 analysis of, 19–23 brick packaging, 143–44 grating lobes, 22–23 of isotropic point sources, 20 low sidelobe aperture distributions, 23–27 low sidelobes, 23 radiation pattern, 21, 22 Taylor distribution for, 25–26 uniformly illuminated, 21 Linear frequency modulated (LFM), 46 Line replaceable units (LRUs) about, 15, 110, 138 calibration and, 197–98 components of, 142–43 configuration illustration, 142 control module, 73–74 discontinuities and, 58 edge-cooled, horizontal cold plate packaging and, 146–47 maintenance requirements and, 144 radiating element RF connections, 149 size versus cost, 231 sliding vertical cold plate array packaging and, 146 vertical fixed cold plate packaging and, 148–49 Low-noise amplifiers (LNAs), 54–56, 79 Low sidelobes, aperture distributions, 23–27 Low-temperature co-fired ceramic (LTCC) substrates, 73 M Maintenance, delaying, 172 Maintenance cost, 170–73 Matched FETs, 233 Mean time between critical failures (MTBCF) about, 157, 174
7065_Agrawal_V6.indd 257
257
aggregate, 164 calculations, 165 cluster comparison, 166–67 four-element clusters, 165 increasing with redundant power supplies, 166–68 maximizing, 160–62, 164 MTBF and, 174 no cluster design, 165 two-element clusters, 166 Mean time between failures (MTBF) about, 3, 15–16, 155 DC-to-DC converters, 131, 143 determining, 162 for different cluster sizes, 162–66 equation, 162 greater than deployment period, 168 high reliability and, 155–56 increasing, 156 maintenance cost and, 172 MTBCF and, 174 T/R modules, 75–76, 77 values, 163 Mean time to repair (MTTR), 157 Metal-semiconductor field-effect transistor (MESFET) junction, 143 Microstrip trace, 113 Military radar requirements, 2 Mil-Std-901 (shipboard shock requirement), 150 MMIC module about, 54–56 cost of, 76–78 illustrated, 54 integration, 73–74 reliability, 75 See also T/R modules Moly spacer, 72 Monolithic microwave integrated circuits (MMICs), 54, 232–33 Moving target indication (MTI), 175–76 Multiple independent receive beams, 91–94 Mutual coupling between array and external elements, 186–87 between array elements, 188–90 between calibration element and array elements, 190–91 between dedicated elements and array elements, 191–98 field calibration with, 186–98 between radiating elements, 114–16 as small array, 121–22
4/6/23 1:43 PM
258
Practical Aspects of Active Phased Array Antenna Development
N Noise figure (NF) beamformer architecture impact on, 94–97 of cascaded network, 62–63 defined, 61 spurious-free dynamic range and, 67 T/R module, 61–62 Noise temperature, 11, 63–64 Nonrecurring engineering (NRE), 215 O 1-dB compression point, 64 Operating bandwidth, 44, 46–47 Organization, this book, 13–17 P Packaging brick, 141–42, 144–49 concepts, 140–44 construction and cooling methods, 141 control module, 234 cost reduction and, 234–35 dedicated element calibration effect on, 197–98 edge-cooled, horizontal cold plate, 146–48 introduction to, 137–39 LRUs and, 142–43 parameters impacting, 139 radome design, 150–52 schemes, 144–49 sliding vertical cold plate, 145–46, 147 structural design, 150 summary, 153 thermal management, 143–44 tile array layers, 140 vertical fixed cold plate, 148–49 Parallel-plate waveguide, 113 Passive phased array antennas about, 5–6 active phased array antenna advantages over, 10–12 beamformer architectures for, 86–90 beamformer networks for, 86–90 limitations, 8 linear, 7 receive beamformer for, 7 RF block diagrams, 6–7 See also Active phased array antennas Passive-weighted architecture, 95–96 Patch antennas, 105–6
7065_Agrawal_V6.indd 258
Peak grating lobes, 48–49 Peak sidelobes, 50 Performance effect of amplitude and phase errors on, 40–44 moving target indication (MTI), 176 radiating elements, 119, 120–21 T/R module receive, 68–69 T/R module requirements, 78–79 Phased array antennas about, 3 amplitude and phase errors, 40 beam steering diagram, 127 centralized architecture, 176–79 component types, 159 phase shifter bits, 42 radiation pattern calculation, 121–22 T/R modules, 180 See also Active phased array antennas Phase errors effect on array patterns, 88 effect on performance, 40–44 phase shifters and, 88–89 recalibration, 191 See also Amplitude errors; Errors Phase shifters, 54, 58, 88–89 Planar arrays about, 27–28 analysis and synthesis of, 27–33 rectangular and triangular grids comparison, 33–34 rectangular lattice, 28–31 sum and difference beams of, 27 triangular lattice, 31–33 Power-added efficiency, 61–62, 236 Power-aperture gain (PAG), 12–13, 225–27 Power distribution about, 129–30 average versus peak DC-to-DC converters, 133–34 centralized power system, 132–33, 134 conversion block diagram, 130 DC, cost reduction and, 235 DC-to-DC converter key parameters, 130 distributed power system, 130–32 system comparison, 134 Power supplies high reliability and, 77 redundant, 166–68 ripple and, 183
4/6/23 1:43 PM
Index switched mode (SMPS), 129 types of, 167 Preliminary design review (PDR), 229 Pretesting, 238 Prime power requirement, reduced, 11 Printed circuit radiating elements about, 104–6 illustrated, 105 wideband, 106 See also Radiating elements Pulse repetition frequency (PRF), 56–57, 128 PUMA element, 106 Q Quantization errors, 40–42 R Radar control computer (RCC), 125 Radar sensitivity, 224–25 Radiating elements cost reduction and, 232 design, 116 design process, 117–21 engineering goal, 117 feed transition geometry, 113 introduction to, 103–4 LRU connections, 149 mutual coupling between, 114–16 performance, 119, 120–21 printing circuit, 104 scan impedance, 118 T/R modules and, 104 type, selection of, 116–17 types of, 103 VSWR performance, 115 waveguide, 107–10 waveguide array simulator and, 118 waveguide dimensions and, 118 Radiation pattern linear arrays, 21, 22 phased array calculation, 121–22 ripples, 121 single-element, 121 Radome about, 150 cross-section, 151 design, 150–52 fabric, 152 integration, 151 temperature management, 151
7065_Agrawal_V6.indd 259
259
Radome heating, 110–12 RAYDEL, 152 Rectangular lattice array factor, 28 grating lobe boundaries, 30 grating lobe circles, 29 triangular grid comparison, 33–34 Redundant power supplies, 166–68 Reflection coefficient, 122 Reliability beamformer architectures and, 97 block diagram, 183 high, 155–74 improved, 11 maintenance cost, 172 model, 160 T/R module, 75–76 See also High reliability Required scan boundaries (RSB), 35, 37 RF technology, 4 Ridged waveguides, 108 RMS sidelobe level antenna size and, 44 due to amplitude and phase errors, 42–44 VGA and, 55 S Scan impedance, 114, 118 Self-calibration, 188 Sensitivity DBF and, 201 enhanced, 2 improved, 10 radar, defined, 224–25 ripple and, 177 Sidelobe cancellers (SLC), 209 Signal-to-clutter ratio, 176, 210 Signal-to-noise ratio, 225 Silicon carbide, 72 Silicon germanium (SiGe), 79–80 Single-element calibration, 190–91 Single pole double throw (SPDT) switch, 54–55 Sliding vertical cold plate packaging, 145–46, 147 SPY-1 array antenna cost reduction, 214, 216–17 Stability, T/R module, 74–75 Subarrays about, 45 digital beamforming and, 204–6
4/6/23 1:43 PM
260
Practical Aspects of Active Phased Array Antenna Development
Subarrays (Cont.) overlapped, 100 spacing, 98 vertical and horizontal combiners in, 100 wideband performance and, 47 Switched mode power supplies (SMPS), 129 T Target detection, xv, 10 Taylor distribution, 25–26 Taylor weighting, 88 Technical problems, 229 Technology improvements, 217–18, 220–21 Test costs, 238 Thermal management, xv, 143–44 Thermal stack-up, T/R module, 71–73 Third-order intercept point (TOI) about, 65 input, 85, 94 requirements, 79 spurious-free dynamic range and, 66, 67 Tile packaging, 140–41 Tilt angle, 35–38 Tilted arrays about, 34 backward, geometry, 35 grating lobe technique, 34–35 ground-based, 38, 39 optimized grid design, 37 shipboard, 38 triangular grid design, 36 Time-delay units (TDUs), 98–100 Top-down design, 222–23 Tracking radar, performance metric, 12–13 Triangular lattice array factor, 31 equilateral, 32 grating lobe circles, 33 planar array with, 32 rectangular grid comparison, 33–34 T/R module architecture about, 54–56 brick, 117 for circular polarization, 69–70 common leg, 58–59 control module, 56–57 illustrated, 68 integration with DC-to-DC converter, 57–58 MMIC module, 54–56 tradeoffs, 67–69
7065_Agrawal_V6.indd 260
T/R modules about, xv–xvi, 4, 53–54 amplitude taper and, 98 block diagram, 55 construction, 70–71 cost, 54, 76–78, 222 cost reduction and, 232–34 design of, 13 driver amplifiers, 169–70 efficiency, 3 factors affecting failure, 75–76 failed, 16 function of, 2 heat generation, 137 key performance parameters, 61–66 LRUs, 159 at maximum gain, 96 MTBF, 77 noise figure, 62–63 noise temperature, 63–64 1-dB compression point, 64 optimizing RF output power levels, 223–27 output, 90, 116–17 packaging cross-section, 71 performance improvement and, 59–61 performance requirements, 78–79 phase compensation, 98 power-added efficiency, 61–62 power tradeoffs, 224 radiating element connection, 104 receive performance, 68–69 reliability, 75–76 requirements flow down (receive channel), 243–45 requirements flow down (transmit channel), 241–43 RF block diagram, 159 RF insertion phase changes, 144 S-/C-band, 59 SiGe BiCMOS technology, 79–80 stability, 74–75 system dynamic range, 84 technology, 54 thermal stack-up of, 71–73 third-order intercept point, 65–66 weighting, 95–96 V Vacuum-brazed cold plates, 148 Variable gain amplifier (VGA), 55, 85, 93
4/6/23 1:43 PM
Index Vertical fixed cold plate packaging, 148–49 Vivaldi radiators, 106–7 Volume clutter, 175 W Waveform and pattern flexibility, 10 Waveguide array simulator, 118–19, 120 Waveguide dimensions, 118
7065_Agrawal_V6.indd 261
261
Waveguide radiating elements about, 107 double-ridged, 108, 109 illustrated, 108 ridged, 108 See also Radiating elements Wideband operation, improved, 10 Wideband parallel-plate waveguide array, 114 Wide bandwidth operation, 236–37
4/6/23 1:43 PM
7065_Agrawal_V6.indd 262
4/6/23 1:43 PM