Software
A Modern Approach
to Radio Engineering
The first comprehensive guide
to software radio design
and implementation
Multirate DSP, RF front-ends, direct
digital synthesis of modulated
waveforms, A/D and D/A
conversion, and more
Enhancing performance through
smart antennas and other
adaptive array algorithms
Techniques for building more
flexible, object-oriented
real-time software
Jeffrey
H. Reed
Prentice Kail Communications Engineering and Emerging Technologies Series
Theodore S. Rappaport, Series Editor
Software Radio:
A Modern Approach
to Radio Engineering
ISBN
a-i3-rjflii5fl-a
Prentice Hall Communications Engineering
and Emerging Technologies Series
Theodore S. Rappaport, Series Editor
DOSTERT Powerline Communications GARG Wireless Network Evolution: 2G to 3G
GARG 7 5 - 9 5 CDMA and cdma.2000: Cellular/PCS Systems Implementation
GARG & WILKES Principles and Applications of GSM
HAC Multimedia Applications Support for Wireless ATM Networks KlM Handbook of CDMA System Design, Engineering, and Optimization
LlBERTI & RAPPAPORT Smart Antennas for Wireless Communications: IS-95 and Third
Generation CDMA Applications
PAHLAVAN & KRISHNAMURTHY Principles of Wireless Networks: A Unified Approach RAPPAPORT Wireless Communications: Principles and Practice, Second Edition RAZAVI RF Microelectronics
REED Software Radio: A Modern Approach to Radio Engineering
STARR, ClOFFl & SILVERMAN Understanding Digital Subscriber Line Technology TRANTER, SHANMUGAN, RAPPAPORT, & KOSBAR Principles of Communication Systems
Software Radio:
A Modern Approach
to Radio Engineering
Jeffrey H. Reed
Prentice Hall PTR
Upper Saddle River, New Jersey 07458 www.phptr.com
A cataloging-in-publication data record for this book can be obtained from the Library of Congress.
Production Supervision: Laura Burgess Publisher: Bernard Goodwin
Editorial Assistant: Michelle Vincente Marketing Manager: Dan DePasquale Manufacturing Manager: Alexis Heydt-Long Cover Design Director: Jerry Votta
Cover Design: Nina Scuderi Art Director: Gail Cocker-Bogusz
Production Services/Compositor: Lori Hughes
• M B ©2002 Prentice Hall PTR
^RH^H A division of Pearson Education, Inc. • E l Upper Saddle River, NJ 07458
Prentice Hall books are widely used by corporations and government agencies for training, marketing, and resale. The publisher offers discounts on this book when ordered in bulk quantities. For more information, contact:
Corporate Sales Department Prentice Hall PTR
One Lake Street
Upper Saddle River, New Jersey 07458 Phone: 800-382-3419
Fax: 201-236-7141
Email: [email protected] All products mentioned herein are trademarks or registered trademarks of their respective owners.
All rights reserved. No part of this book may be reproduced, in any form or by any means, without permission in writing from the publisher.
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
ISBN 0-13-081158-0 Pearson Education LTD.
Pearson Education Australia PTY, Limited Pearson Education Singapore, Pte. Ltd. Pearson Education North Asia Ltd. Pearson Education Canada, Ltd.
Pearson Educacion de Mexico, S.A. de C.V. Pearson Education—Japan
Contents
PREFACE xv ACKNOWLEDGMENTS xix
1 INTRODUCTION TO SOFTWARE RADIO CONCEPTS 1
1.1 The Need for Software Radios 1 1.2 What Is a Software Radio? 2 1.3 Characteristics and Benefits of a Software Radio 3
1.4 Design Principles of a Software Radio 6
1.5 Questions 8
2 RADIO FREQUENCY IMPLEMENTATION ISSUES 11
2.1 The Purpose of the RF Front-End 11 2.2 Dynamic Range: The Principal Challenge of Receiver Design 13
2.3 RF Receiver Front-End Topologies 14 2.3.1 Characteristics of the Topologies 14
2.3.2 Topologies 15 2.4 Enhanced Flexibility of the RF Chain with Software Radios 21
2.5 Importance of the Components to Overall Performance 21
2.5.1 Antennas 21 2.5.2 Duplexer and Diplexer 26
2.5.3 RF Filter 26 2.5.4 Low Noise Amplifier 27
2.5.5 Image Reject and IF Filters 27
2.5.6 RF Mixer 27 2.5.7 Local Oscillator 28 2.5.8 Automatic Gain Control 29 2.5.9 Analog to Digital Converter 32 2.6 Transmitter Architectures and Their Issues 33 2.7 Noise and Distortion in the RF Chain 35
2.7.1 Noise Characterization 35 2.7.2 Distortion Characterization 38 2.8 ADC and DAC Distortion 42
viii Contents
2.9 Predistortion 43 2.10 Flexible RF Systems Using Microelectromechanical Systems 49
2.11 Conclusion 52 2.12 Questions 53
3 MULTIRATE SIGNAL PROCESSING 55
3.1 Introduction 55 3.1.1 Cost 55 3.1.2 Flexibility 56 3.1.3 Overview of the Chapter 57
3.2 Sample Rate Conversion Principles 57
3.2.1 Decimation 58 3.2.2 Interpolation 62 3.2.3 Two Multirate Identities 77
3.2.4 Non-Integer-Rate Conversion 78 3.2.5 Sampling Rate Conversion by Stages 79 3.2.6 Cascaded Integrator Comb Filter 87
3.3 Polyphase Filters 96 3.3.1 Polyphase Decimation 96
3.3.2 Polyphase Interpolation 98
3.4 Digital Filter Banks 101 3.4.1 Implementation 103 3.4.2 DFT Filter Banks 103 3.4.3 Transmultiplexers 112 3.5 Timing Recovery in Digital Receivers Using Multirate Digital Filters 114
3.5.1 Timing Recovery in a Classical Analog Receiver 117 3.5.2 Timing Recovery in the Digital Domain Only 117
3.5.3 Early-Late Gate Synchronizer 120 3.5.4 Timing Offset Control Using the Early-Late Gate Principle 120
3.6 Conclusion 124 3.7 Questions 125
4 DIGITAL GENERATION OF SIGNALS 127
4.1 Introduction 127 4.2 Comparison of Direct Digital Synthesis with Analog Signal Synthesis 129
4.3 Approaches to Direct Digital Synthesis 131 4.3.1 Pulse Output Direct Digital Synthesis 131 4.3.2 ROM Look-Up Table Approach 133 4.3.3 Phase Truncation Distortion 133 4.3.4 Analysis of the Output Sequence 139
4.4 Analysis of Spurious Signals 140 4.5 Spurious Components due to Periodic Jitter 143
4.6 Bandpass Signal Generation 144 4.7 Performance of Direct Digital Synthesis Systems 146
Contents ix
4.7.1 Experimental Findings 146 4.7.2 Use of Hybrid Systems 147 4.8 Hybrid DDS-PLL Systems 148 4.9 Applications of Direct Digital Synthesis 148
4.10 Generation of Random Sequences 150 4.10.1 Types of Sequences and Their Properties 150
4.10.2 Randomization with the Wheatley Procedure 153
4.11 ROM Compression Techniques 157 4.11.1 Interpolation Using Taylor's Series Expansion 158
4.11.2 Interpolation Using Trigonometric Identities 160 4.12 Sine-Phase Difference Algorithm Approach 163 4.13 Modified Sine-Phase Difference Approach (Parabolic Approximation) 164
4.14 Conclusion 166 4.15 Questions 167
5 ANALOG TO DIGITAL AND DIGITAL TO ANALOG CONVERSION 169
5.1 Introduction 169 5.2 Parameters of Ideal Data Converters 171
5.2.1 Sampling Process 171 5.2.2 Quantization 184 5.3 Parameters of Practical Data Converters 195
5.3.1 Generic Data Converter Physical Models 195 5.3.2 Practical Transfer Characteristic Considerations 199
5.3.3 Dynamic Range Considerations 202 5.3.4 Practical Timing Issues 208 5.3.5 Analog Bandwidth 217 5.3.6 Power Consumption 217 5.3.7 Impact of Noise and Interference on Dynamic Range
Requirements 221 5.4 Techniques to Improve Data Converter Performance 224
5.4.1 Dithering 225 5.4.2 Automatic Gain Control 228
5.5 Common ADC and DAC Architectures 232 5.5.1 Parallel Structures: Flash ADCs, String DACs, and Binary
Structures 232 5.5.2 Segmented Structures: Folding and Interpolating ADC
and Segmented Ladder DAC 236 5.5.3 Iterative Structures: Subranging/Pipelined/Half-Flash ADC,
Successive Approximation ADC 242 5.5.4 Sigma-Delta Structures: ADC and DAC 246
5.6 Conclusion 256 5.7 Questions 259
X Contents
6 SMART ANTENNAS 263
6.1 Introduction 263 6.2 Vector Channel Modeling 264
6.2.1 Array Steering Vectors 265 6.2.2 Multipath Channel Models 270 6.2.3 Multi-User Channel Models 273 6.3 Benefits of Smart Antennas 274
6.3.1 Beamforming 275 6.3.2 Space-Time Equalization 276
6.3.3 Diversity 276 6.4 Structures for Beamforming Systems 276
6.4.1 Multiple Fixed-Beam Antenna Array 278
6.4.2 Fully Adaptive Array 278 6.4.3 Relative Benefits and Trade-Offs of Switched Beam
and Adaptive Array Systems 282 6.5 Smart Antenna Algorithms 282
6.5.1 Diversity Combining Techniques 283 6.5.2 Adaptation Algorithms Using Training Sequences 289
6.5.3 Blind Algorithms 292 6.6 Diversity and Space-Time Adaptive Signal Processing 300
6.6.1 Algorithms for Receiver STAP 301 6.6.2 Overloaded Array Processing 308 6.7 Algorithms for Transmit STAP 309
6.7.1 Space-Time Pre-Filtering 309 6.7.2 Space-Time Trellis Coding 310 6.7.3 A Simple Transmit Diversity Scheme 314
6.8 Hardware Implementation of Smart Antennas 316 6.8.1 Digital Beamforming Receiver Implementation 317 6.8.2 Digital Beamforming Transmitter Implementation 317
6.8.3 Component Issues 319 6.9 Array Calibration 321
6.9.1 Remote Transmitter Approach 321
6.9.2 Test-Tone Approach 322 6.10 Virginia Tech Space-Time Adaptive Radio Case Study 323
6.10.1 Overview of the VT-STAR Architecture 324
6.10.2 RF Design of VT-STAR 325 6.10.3 Software Issues for VT-STAR 327 6.10.4 Key Design Issues of VT-STAR 333
6.11 Conclusion 333 6.12 Questions 335
7 DIGITAL HARDWARE CHOICES 339
7.1 Introduction 339 7.2 Key Hardware Elements 340
Contents xi 7.3 DSP Processors 342 7.3.1 DSP Core 342 7.3.2 DSP Architectures 342 7.3.3 Numeric Representation 347 7.3.4 Addressing 350 7.3.5 Pipelining 353 7.3.6 Peripherals and Additional Features 355
7.3.7 Multi-Processing 355 7.3.8 Multi-Processing Using a Real-Time Operating System 357
7.3.9 The Software Design Cycle 358 7.3.10 Maximizing Performance 360 7.3.11 Benchmarks and Performance Evaluation 368
7.3.12 Case Study: TMS320C54x Series DSPs 370
7.4 Field Programmable Gate Arrays 371 7.4.1 Operation of an SRAM-Based FPGA Cell 371
7.4.2 Implementing DSP Functions in FPGAs 373
7.4.3 FPGA Architectures 373 7.4.4 Applications of FPGAs to Software Radios 377
7.4.5 Design Principles using FPGAs 378 7.5 Trade-Offs in Using DSPs, FPGAs, and ASICs 379
7.6 Power Management Issues 379 7.6.1 DSP Power Management 380 7.6.2 Low-Power VLSI Design 381 7.6.3 Architectural-/System-Level Approaches 383
7.7 Using a Combination of DSPs, FPGAs, and ASICs 386
7.8 Conclusion 387 7.9 Questions 388
8 OBJECT-ORIENTED REPRESENTATION OF RADIOS AND NETWORK
RESOURCES 391 8.1 Introduction 391 8.2 Networks 392 8.2.1 System Layers 395 8.2.2 Switching 396 8.2.3 Quality-of-Service 396 8.2.4 Internet Protocol 397 8.2.5 Asynchronous Transfer Mode 403
8.2.6 Networks and Software Radios 404 8.3 Object-Oriented Programming 405
8.3.1 Objects 406 8.3.2 Java 411 8.3.3 Java and Software Radios 414
8.3.4 The Radio Virtual Machine 415 8.3.5 Object-Oriented Software and Software Radios 415
xii Contents
8.4 Object Brokers 416 8.4.1 Common Object Request Broker Architecture 419
8.4.2 Software Radio Implementation Issues 422 8.4.3 Object Brokers and Software Radios 424
8.5 Mobile Application Environments 424
8.5.1 MExE 427 8.5.2 Service Discovery 432
8.5.3 Mobile Application Environments and Software Radios 432
8.5.4 Security in Software Radio 433 8.6 Joint Tactical Radio System 434
8.6.1 Hardware Classes 434 8.6.2 SCA Structure 435
8.7 Conclusion 440 8.8 Questions 441
9 CASE STUDIES IN SOFTWARE RADIO DESIGN 4 4 3
9.1 Introduction and a Historical Perspective 443 9.1.1 Architectural Characteristics Intrinsic to a Software Radio 445
9.1.2 Architectural Characteristics Important to a Software Radio 446 9.1.3 Architectural Characteristics of Practical Software Radios 448
9.2 SPEAKeasy 450 9.2.1 SPEAKeasy Phase I 451
9.2.2 SPEAKeasy Phase II 455 9.2.3 SPEAKeasy Summary 462
9.3 JTRS 462 9.3.1 Goals of the SCA 465
9.3.2 Attributes of the SCA Developed from the PMCS Guidance
Document 465 9.3.3 SCA Architectural Details 466
9.3.4 JTRS Summary 475 9.3.5 SDR Forum Architecture Details 476
9.3.6 Summary 481 9.4 Wireless Information Transfer System 481
9.4.1 Architecture Goals 483 9.4.2 Architecture Overview 483 9.4.3 Software Architecture 484 9.4.4 Hardware Architecture 484 9.4.5 Architectural Details 486 9.4.6 WITS Summary 490 9.5 SDR-3000 Digital Transceiver Subsystem 492
9.6 SpectrumWare 494 9.6.1 SpectrumWare System Description 495
9.6.2 Input/Output 495 9.6.3 Programming Environment 496
Contents xiii
9.7.2 Layered Radio Architecture Implementation Example 507
9.7.3 CHARIOT Summary 509
9.8 Conclusion 509 9.9 Questions 512 A RF ENGINEERING BOOKS AND TRADE PUBLICATIONS 515
A.l Electronics 515 A.2 RF Circuit Design and S-Parameters 515
A.3 Filters 516 A.4 Microwaves 516 A.5 Oscillators 516 A.6 Phase Locked Loops and Frequency Synthesizers 516
A.7 Receivers and Systems 517
A.8 PSpice 517 A.9 Trade Publications and Periodicals 517
A. 10 Web-Accessible Tutorial Materials 518
B THE COORDINATE ROTATION DIGITAL COMPUTER ALGORITHM 519
B.l Introduction 519 B.2 CORDIC Overview 520 B.3 Derivation of the CORDIC Algorithm 520
B.3.1 Translating a Point Along a Circle of Radius R 520 B.3.2 Rotation Through Iterative Subrotations 522 B.3.3 Computationally Simplifying the Iterative Rotations 523
B.3.4 Putting the Equations in Final Form 524
B.3.5 Vectoring Mode 525 B.4 CORDIC Algorithm Performance 526
B.5 Extensions to the CORDIC Algorithm 527
REFERENCES 531 GLOSSARY OF ACRONYMS 549
ABOUT THE AUTHOR 557 ABOUT THE CHAPTER CO-AUTHORS 559
P R E F A C E
Software radios represent a major change in the design paradigm for radios in which a large portion of the functionality is implemented through programmable signal processing devices, giving the radio the ability to change its operating parameters to accommodate new features and capabilities. A software radio approach reduces the content of radio frequency (RF) and other analog components of traditional radios and emphasizes digital signal processing to enhance overall receiver flexibility.
This change in the design paradigm for new radios has occurred so rapidly that it has left a significant void in the educational material available to train new radio engineers. Traditional radio engineering textbooks emphasize analog component-level design with little mention of the increasingly important role of digital signal processing in performing the central functions of the radio transceiver. Individual references covering the key analog and digital subsystems tend to be insufficient in that they fail to provide a full understanding of the interaction between these subsystems.
I became acutely aware of this void when conducting research into the development of novel high-performance radios for the Defense Advanced Research Projects Agency (DARRA). While constructing radio prototypes, I found there was no comprehensive re-source to which I could point my students for information on how to build DSP-based radios. This experience, combined with similar frustrations voiced by my colleagues from both academia and industry, has led me to write this book on modern radio design princi-ples. My goal in developing this book was to provide this necessary understanding of the interaction of key subsystems.
Software radios are emerging in commercial and military infrastructure. This growth is motivated by the numerous advantages of software radios.
1. Ease of design—Traditional radio design requires years of experience and great care on the part of the designer to understand how the various system components work in conjunction with one another. The time required to develop a marketable product is a key consideration in modern engineering design, and software radio implemen-tations reduce the design cycles for new products, freeing the engineer from much of the iteration associated with analog hardware design. It is possible to design many different radio products using a common RF front-end with the desired frequency and bandwidth in conjunction with different signal processing software.
xvi Preface
2. Ease of manufacture—No two analog components have precisely identical perfor-mance, necessitating rigorous quality control and testing of radios during the man-ufacturing process. However, given the same input, two digital processors running the same software will produce identical outputs. Thus, the move to digital hardware reduces the costs associated with manufacturing and testing the radios.
3. Multimode operation—The explosive growth of wireless has led to a proliferation of transmission standards, and in many cases, it is desirable that a radio operates according to more than one standard. For example, wireless carriers throughout the U.S. are deploying systems that make use of the GSM (Global System for Mo-bile Communications) standard in some markets and the IS-95 Code Division Mul-tiple Access (CDMA) standard in other markets. Furthermore, the advent of third-generation wireless has introduced a number of standards within that framework. Traditionally, multimode operation has required multiple complete sets of hardware, increasing the size and cost of the radio. However, a software radio can change modes by simply loading appropriate software into the memory.
4. Use of advanced signal processing techniques—The availability of high speed sig-nal processing on board the radio allows implementation of new receiver structures and signal processing techniques. Techniques such as adaptive antennas, interference rejection, and strong encryption, previously deemed too complex, are now finding their way into commercial systems as the performance of digital signal processors continues to increase. The impact will be enhanced range and quality of service to the consumer while reducing overall infrastructure cost for the service provider. 5. Fewer discrete components—A single high-speed digital processor may be able to
implement many traditional radio functions such as synchronization, demodulation, error detection, and decryption of data, reducing the number of required components and decreasing the size and cost of a radio.
6. Flexibility to incorporate additional functionality—Software radios may be mod-ified in the field to correct unforeseen problems or upgrade the radio. For example, it may even be possible to transmit software upgrades to the radio, such as a new vocoder to handsets, to improve overall system performance. Another important improved functionality is the capability of self-diagnosis of the radio and network operations, which means improved reliability with less human intervention.
Given these clear advantages and the increasing processing power available in commer-cial digital signal processing devices, I anticipate that radio engineers that software radios will become the standard approach for radio design.
The challenge in creating the software radio is the broad scope of knowledge necessary, including digital signal processing algorithms, RF circuits, software methodologies, and digital circuits. The approach in this text is to provide an understanding of key areas in radio design for the digital signal processing engineer. For example, a digital signal processing engineer must know the ramifications of the choices in RF parameters and the resulting limitations to be able to understand the appropriate subsequent signal processing to account
Preface xvii
for these limitations. This book reviews critical and interdependent radio subsystems from the perspective of the DSP engineer.
Chapter 1 provides a basic introduction to software radio concepts, discusses the bene-fits of software radios, and sets the stage for discussing software radio design.
Digital signal processing engineers tend to know very little about RF engineering and, likewise, RF engineers tend to know very little about digital signal processing. However, to take full advantage of the software radio approach, these subsystems cannot be treated separately. Chapter 2 provides the digital signal processing engineer with fundamentals in constructing RF front-ends and describes processing that can be performed in the digital domain to overcome problem areas in RF design.
Multirate digital signal processing uses different sample rates, and this is the topic of Chapter 3. This approach to signal processing is particularly important in software radios where bandwidths and sample rates are high initially and must be reduced for efficient sub-sequent processing. Multirate digital signal processing is commonly used to channelize the operating band into distinct communication channels. Multirate digital signal processing is also the foundation for modern synchronization techniques.
Much of the flexibility of a software radio comes from being able to create arbitrary modulation types directly within the digital domain. In many cases, the direct digital syn-thesis methods used to generate these signals are more than just digitized realizations of analog techniques and afford the designer greater freedom in design signal waveforms. Chapter 4 surveys the topic of direct digital synthesis of modulated waveforms.
Analog to digital converters and digital to analog converters, along with the power amplifier, are the most critical components in software radio design. The demands on these components can be very high. A rigorous understanding of the conversion process and the trade-offs between the resolution, sample rate, and dynamic range of the resulting system are the focus of Chapter 5.
An important benefit of software radios is the ability to incorporate sophisticated algo-rithms, such as smart antennas, into the radio to enhance performance. Chapter 6 reviews the wide variety of adaptive array algorithms and hardware implementation issues.
The basics of digital signal processing microprocessors, Field Programmable Gate Ar-rays (FPGAs), and Application Specific Integrated Circuits (ASICs) and how one would choose one these alternatives for constructing a software radio are discussed in Chapter 7.
A systematic design approach to creating software is essential to enable expandability of the radio capability. Furthermore, as new applications are created to run over the soft-ware radio, the radio itself must become transparent to the new applications. Chapter 8 examines object-oriented programming approaches, including JAVA and Common Object Request Broker Architecture (CORBA) for creating software radios.
Chapter 9 examines some examples of software radios that have been built. The Software-Defined Radio (SDR) Forum, a consortium of companies, universities, and research or-ganizations, has defined guidelines and standards for the creation of software radios. A description of this standardized software radio is provided in this chapter.
If this book is being used for a course, there is much flexibility in selecting chapters to create a customized course. For the one semester class, I recommend covering Chapters 1 -3, Sections 4.1-4.8, 5.1-5.4, 6.1-6.6, 7.1-7.-3, and Chapters 8-9. For a class on the quarter
xviii Preface
system, I recommend Chapters 1, 2, and 5, Sections 6.16.5, Chapter 8, and Sections 9 . 1 -9.3. Although there is much latitude in mixing and matching sections to customize the course to the instructors objectives, I do recommend that Chapters 8 and 9 be covered together as a single unit. Students who have an electrical engineering background in basic circuit analysis (typically junior-level), computer architecture (junior- or senior-level), and communications (senior-level) have a sufficient background for all chapters in this book.
URLs included in the text and in citations were correct when the book was written. However, due to the dynamic nature of the World Wide Web, URLs may no longer be ac-tive. Periodic updates, information for instructors, and errata to the book can be found at h t t p : / / w w w . m p r g . o r g / p u b l i c a t i o n s / p u b s . s h t m l # B o o k s . Additional in-formation about software radios can be found at h t t p : / /www. s o f t r a d i o s . com.
A C K N O W L E D G M E N T S
I am deeply indebted to many individuals for their efforts, without which this book would not have been possible. Special thanks go to Bernard Goodwin, Acquisitions Editor at Prentice Hall, and Lori Hughes, Freelance Production Editor, for not losing faith. Without Lori's tenacity and hard work, this book would not have been possible.
I would like to thank all the students in Virginia Tech's software radio classes for their valuable input. In particular, Prinya Atiniramit, John Davies, Patrick Herhold, Suk Won Kim, Oliver Praetor, and Uwe Ringel offered numerous suggestions. Many of my col-leages, both at Virginia Tech and at other educational institutions, have offered their advice and knowledge. Thanks goes to Dennis Sweeney of Virginia Tech for offering his expertise and his bibliography on RF design. I would also like to acknowledge James Neel, Kim Phillips, P. Max Robert, Sujayeendar Sachindar, and Srikathyayani Srikanteswara, all of whom co-authored chapters. In addition, my students, particularly Jason Aron, Yigang Fan, Ran Gozali, Ariful Hannon, James Hicks, Seshagiri Krishnamoorthy, Vikas Kukshya, Shakheela Marikar, Raqibul Mostafa, Ramesh Chembil Palat, and Kazi Abu Zahid, offered assistance with various chapters.
The reviewers of various chapters of the book offered valuable perspectives. My thanks goes to Brad Brannon, Paul Smith, and Zoran Zvonar of Analog Devices, Inc.; Michael A. Komara of Airnet; Robert Ulman of the Army Research Office; Andrew Park, Bob Watson, and Mark Cear of BAE Systems; Bala Ramachandran of Conexant Systems; Mar-cus Bronzel, Gerhard Fettweis, and Tim Hentschel of Dresden University of Technology; Neiyer Correal, Dave Dohse, and Bruce Fette of General Dynamics; Bill Newhall and Steve Thompson of Grayson Wireless; Albert Garrett of Harris Corp.; Irving E. Hodnett, indepen-dent consultant in wireless communications engineering services; Ding Qi of Huawei; Yash Vasavada of Hughes Network Systems; Ron Oliver of Impinj, Inc.; Peter Curry, Jim Isaacs, and Jerry Sanders of ITT Industries; Young-Soo Kim of Kyung Hee University School of Electronics and Information in Kyungki-Do, Korea; JeongHo Kim of LG Electronics, Inc.; David Murotake of Mercury Computer Systems, Inc.; Jerzy Kirrander of Mid-Sweden University; Joseph Mitola, III, of The MITRE Corp.; Pascal Renucci of National Semicon-ductor Corp.; Milap Majmundar of Southwestern Bell; Carl Panasik of Texas Instruments, Inc.; Dennis Ferguson of Theseus; Peter Athanas, Louis Beex, Bob Boyle, Ted Rappaport,
XX Acknowledgments
Warren Stutzman, Dennis Sweeney, and Bill Tranter of Virginia Tech; and Brian Fox of Xircom Technology Group.
I offer my appreciation to the following for allowing me to use their copyrighted ma-terial: lOlcommunications, Altera Corp., Analog Devices, Inc., Tom Biedka, Cadence Design Systems, Inc., e-tenna Corp., Richard B. Ertel, Brian Fox, HRL Laboratories, L L C , Institute of Electrical and Electronics Engineers, Inc., International Engineering Consortium, Motorola, Pearson Education, Inc., Qualcomm, Sarnoff Corp., SDR Forum, Spectrum Signal Processing, Inc., United Feature Syndicate, and the U.S. Department of Defense.
C h a p t e r 1
I N T R O D U C T I O N T O
S O F T W A R E R A D I O
C O N C E P T S
1.1 T h e N e e d for Software Radios
With the emergence of new standards and protocols, wireless communications is develop-ing at a furious pace. Rapid adoption of the wireline-base Internet has led to demand for wireless Internet connectivity but with added capabilities, such as integrated services that offer seamless global coverage and user-controlled quality of service (QoS). The challenge in creating sophisticated wireless Internet connectivity is compounded by the desire for future-proof radios, which keep radio hardware and software from becoming obsolete as new standards, techniques, and technology become available. The concept of integrated seamless global coverage requires that the radio support two distinct features: first, global roaming or seamless coverage across geographical regions; second, interfacing with dif-ferent systems and standards to provide seamless services at a fixed location. Multimode phones that can switch between different cellular standards like IS-95 and Global System Mobile (GSM) fall in the first category, while the ability to interface with other services like Bluetooth or IEEE 802.11 networks falls in the second category. Further, the rate of technology innovation is accelerating, and predicting technological change and its ramifi-cations to business is especially problematic. As a result, to keep their systems up to date, wireless systems manufacturers and service providers must respond to changes as they oc-cur by upgrading systems to incorporate the latest innovations or to fix bugs as they are discovered. Many manufacturers tell horror stories of releasing hundreds of thousands of defective phones that had to be recalled and discarded. Since frequent redesign is expen-sive, time-consuming, and inconvenient to end users, interest is increasing in future-proof radios.
Existing technologies for voice, video, and data use different packet structures, data types, and signal processing techniques. Integrated services can be obtained with either a single device capable of delivering various services or with a radio that can commu-nicate with devices providing complementary services. The supporting technologies and
2 Introduction t o Software Radio Concepts Chapter 1
networks that the radio might have to use can vary with the physical location of the user. To successfully communicate with different systems, the radio has to communicate and de-code the signals of devices using different air-interfaces. Furthermore, to manage changes in networking protocols, services, and environments, mobile devices supporting reconfig-urable hardware also need to seamlessly support multiple protocols, such as IP (Internet Protocol) and MExE (Mobile Execution Environment). Such radios can be implemented efficiently using software radio architectures in which the radio reconfigures itself based on the system it will be interfacing with and the functionalities it will be supporting.
Second-generation (2G) wireless technology consists of a handful of incompatible stan-dards, and the goal behind the development of third-generation (3G) standards is compat-ibility among these standards within and between different generations' standards. Even if cellular standards globally converge, 3G systems require multimode operation and au-tomatic mode selection. With fourth-generation (4G) and possibly 3G systems, the user's application will likely have the ability to control the quality of service and obtain a higher QoS for a higher cost. Higher QoS can be achieved through priority scheduling of packets, changes in data packaging, improved protection coding, better channel equalization tech-niques, implementation of smart antennas, and so on. The mobile subscriber must have the ability to select the network provider as well as the services needed.
1.2 W h a t Is a Software Radio?
The term software radio was coined by Joe Mitola in 1991 to refer to the class of repro-grammable or reconfigurable radios [1]. In other words, the same piece of hardware can perform different functions at different times. The SDR Forum defines the ultimate soft-ware radio (USR) as a radio that accepts fully programmable traffic and control information and supports a broad range of frequencies, air-interfaces, and applications software. The user can switch from one air-interface format to another in milliseconds, use the Global Positioning System (GPS) for location, store money using smartcard technology, or watch a local broadcast station or receive a satellite transmission.
The exact definition of a software radio is controversial, and no consensus exists about the level of reconfigurability needed to qualify a radio as a software radio. A radio that includes a microprocessor or digital signal processor (DSP) does not necessarily qualify as a software radio. However, a radio that defines in software its modulation, error cor-rection, and encryption processes, exhibits some control over the RF hardware, and can be reprogrammed is clearly a software radio. A good working definition of a software radio is a radio that is substantially defined in software and whose physical layir behavior can
be significantly altered through changes to its software. The degree of reconfigurability is
largely determined by a complex interaction between a number of common issues in ra-dio design, including systems engineering, antenna form factors, RF electronics, baseband processing, speed and reconfigurability of the hardware, and power supply management.
The term software radio generally refers to a radio that derives its flexibility through software while using a static hardware platform. On the other hand, a soft radio denotes a completely configurable radio that can be programmed in software to reconfigure the physical hardware. In other words, the same piece of hardware can be modified to perform
Section 1.3 Characteristics and Benefits of a Software Radio 3
different functions at different times, allowing the hardware to be specifically tailored to the application at hand. Nonetheless, the term software radio is sometimes used to encompass soft radios as well.
The functionality of conventional radio architectures is usually determined primarily by hardware with minimal configurability through software. The hardware consists of the am-plifiers, filters, mixers (probably several stages), and oscillators. The software is confined to controlling the interface with the network, stripping the headers and error correction codes from the data packets, and determining where the data packets need to be routed based on the header information. Because the hardware dominates the design, upgrading a conventional radio design essentially means completely abandoning the old design and starting over again. In upgrading a software radio design, the vast majority of the new content is software and the rest is improvements in hardware component design. In short, software radios represent a paradigm shift from fixed, hardware-intensive radios to multi-band, multimode, software-intensive radios.
1.3 Characteristics and Benefits of a Software Radio
Implementation of the ideal software radio would require either the digitization at the an-tenna, allowing complete flexibility in the digital domain, or the design of a completely flexible radio frequency (RF) front-end for handling a wide range of carrier frequencies and modulation formats. The ideal software radio, however, is not yet fully exploited in commercial systems due to technology limitations and cost considerations.A model of a practical software radio is shown in Figure 1.1. The receiver begins with a smart antenna that provides a gain versus direction characteristic to minimize interference, multipath, and noise. The smart antenna provides similar benefits for the transmitter. Most practical software radios digitize the signal as early as possible in the receiver chain while keeping the signal in the digital domain and converting to the analog domain as late as pos-sible for the transmitter using a digital to analog converter (DAC). Often the received signal is digitized in the intermediate frequency (IF) band. Conventional radio architectures em-ploy a super heterodyne receiver, in which the RF signal is picked up by the antenna along with other spurious/unwanted signals, filtered, amplified with a low noise amplifier (LNA), and mixed with a local oscillator (LO) to an IF. Depending on the application, the number of stages of this operation may vary. Finally, the IF is then mixed exactly to baseband. Digitizing the signal with an analog to digital converter (ADC) in the IF range eliminates the last stage in the conventional model in which problems like carrier offset and imaging are encountered. When sampled, digital IF signals give spectral replicas that can be placed accurately near the baseband frequency, allowing frequency translation and digitization to be carried out simultaneously. Digital filtering (channelization) and sample rate conversion are often needed to interface the output of the ADC to the processing hardware to imple-ment the receiver. Likewise, digital filtering and sample rate conversion are often necessary to interface the digital hardware that creates the modulated waveforms to the digital to ana-log converter. Processing is performed in software using DSPs, field programmable gate arrays (FPGAs), or application specific integrated circuits (ASICs). The algorithm used to modulate and demodulate the signal may use a variety of software methodologies, such
Smart Antenna Flexible RF Hardware IF ADC DAC Channelization and Sample Rate Conversion Processing Software Hardware - Algorithms -FPGAs - Middleware - D S P s -CORBA - ASICs
- Virtual Radio Machine
Output
Input
Control
Section 1.3 Characteristics and Benefits of a Software Radio 5
as middleware, e.g., common object request broker architecture (CORBA), or virtual radio machines, which are similar in function to JAVA virtual machines. This forms a typical model of a software radio.
The software radio provides a flexible radio architecture that allows changing the radio personality, possibly in real-time, and in the process somewhat guarantees a desired QoS. The flexibility in the architecture allows service providers to upgrade the infrastructure and market new services quickly. This flexibility in hardware architecture combined with flex-ibility in software architecture, through the implementation of techniques such as object-oriented programming and object brokers, provides software radio with the ability to seam-lessly integrate itself into multiple networks with wildly different air and data interfaces. In addition, software radio architecture gives the system new capabilities that are easily im-plemented with software. For example, typical upgrades may include interference rejection techniques, encryption, voice recognition and compression, software-enabled power mini-mization and control, different addressing protocols, and advanced error recovery schemes. Such capabilities are well-suited for 3G and 4G wireless requirements and advanced wire-less networking approaches. In summary, five factors are expected to push wider accep-tance of software radio.
1. Multifunctionality—With the development of short-range networks like Bluetooth and IEEE 802.11, it is now possible to enhance the services of a radio by leverag-ing other devices that provide complementary services. For instance, a Bluetooth-enabled fax machine may be able to send a fax to a nearby laptop computer equipped with a software radio that supports the Bluetooth interface. Software radio's recon-figuration capability can support an almost infinite variety of service capabilities in a system.
2. Global mobility—A number of communication standards exist today. In the 2G alone, there are IS-136, GSM, IS-95/CDMA1, and many other, less well known stan-dards. The 3G technology tried to harmonize all the stanstan-dards. However, there are many standards under the 3G umbrella. The need for transparency, i.e., the ability of radios to operate with some, preferably all, of these standards in different geo-graphical regions of the world has fostered the growth of the software radio concept. Military services also face a similar issue with incompatible radio standards existing between as well as within branches of the military.
3. Compactness and power efficiency—Multifunction, multimode radios designed using the "Velcro" approach of including separate silicon for each system can be-come bulky and inefficient as the number of systems increases. The software radio approach, however, results in a compact and, in some cases, a power-efficient design, especially as the number of systems increases, since the same piece of hardware is reused to implement multiple systems and interfaces.
4. Ease of manufacture—RF components are notoriously hard to standardize and may have varying performance characteristics. Optimization of the components in terms
6 Introduction to Software Radio Concepts Chapter 1
of performance may take a few years and thereby delay product introduction. In general, digitization of the signal early in the receiver chain can result in a de-sign that incorporates de-significantly fewer parts, meaning a reduced inventory for the manufacturer.
5. Ease of upgrades—In the course of deployment, current services may need to be updated or new services may have to be introduced. Such enhancements have to be made without disrupting the operation of the current infrastructure. A flexible archi-tecture allows for improvements and additional functionality without the expense of recalling all the units or replacing the user terminals. Vocoder technology, for exam-ple, is constantly improving to offer higher quality voice at lower bit rates. As new vocoders are developed, they can be quickly fielded in software radio systems. Fur-thermore, as new devices are integrated into existing infrastructures, software radio allows the new devices to interface seamlessly, from the air-interface all the way to the application, with the legacy network.
Users/customers expect service regardless of the geographical areas in which they travel and the wireless technologies that are in use in different regions in the world, but carrying several devices that cover the broad range of technology alternatives is impractical. Users expect one device to utilize services in all regions, which is possible only by reconfiguring the receiver to the air-interface standards used in the respective regions. By dynamically downloading the software to cover the needed air-interface standard, perhaps through trans-mission of the software configuration to the remote terminal, such over-the-air updates will allow for speedy implementation of software upgrades and new features.
1.4 Design Principles of a Software Radio
Radio design has always required a broad set of design skills. Although one might initially
assume that software radios would require simply a higher level of digital signal processing programming skill than conventional radio design, this is not the case; a higher skill level
is needed for almost all aspects of the radio design because of the dependency of the radio subsystems.
Software radios derive their benefits from their flexibility, complete and easy reconfig-urability, and scalability. It is important to ensure that these characteristics are present in the final product. A generic design procedure for software radios follows and demonstrates the interaction between the various subsystems of the radio design. Subsequent chapters in this book focus on the details of these design procedures.
• Step 1: Systems engineering—Understanding the constraints and requirements of
the communication link and the network protocol allows the allocation of sufficient resources to establish the service given the system's constraints and requirements. For instance, constraints on the range and transmit power constrain the modulation types and data rate that can be supported. For a well-defined standard, the systems engineering aspects, such as the routing protocol, are to a great extent predetermined. However, as additional flexibility is allowed in defining the network, systems engi-neering and optimization becomes a more complex task. In an ideal software radio
Section 1.4 Design Principles of a Software Radio 7
with the ability to change a number of system parameters in real-time, optimizing an active communications session is a major challenge.
• Step 2: RF chain planning—The ideal RF chain for the software radio should
in-corporate simultaneous flexibility in selection of power gain, bandwidth, center fre-quency, sensitivity, and dynamic range. Achieving strict flexibility is impractical and trade-offs must be made. If the communication system is constrained to selected commercial or military bands, this optimization problem is simplified. Nevertheless, with a software radio design, it is possible to compensate for some of the inadequa-cies of the RF components in the digital domain. Compensations for power amplifier distortion or power management of the RF circuitry, for example, can be accom-plished in the digital domain.
• Step 3: Analog to digital conversion and digital to analog conversion selection—
Analog to digital conversion and digital to analog conversion for the ideal software radio is difficult to achieve, and in practice, the selection requires trading power con-sumption, dynamic range, and bandwidth (sample rate). Analog to digital conversion and digital to analog conversion selection is closely tied to the RF requirements for dynamic range and frequency translation. Channelization requirements also impact the selection of the analog to digital conversion and digital to analog conversion. Current conversion technology is very limited and is often the weak link in the over-all system design. There are post-digitization techniques based on multirate digital signal processing that can be used to improve the flexibility of the digitization stage.
• Step 4: Software architecture selection—The software architecture is an important
consideration to ensure maintainability, expandability, compatibility, and scalability for the software radio. Ideally, the architecture should allow for hardware indepen-dence through the appropriate use of middleware, which serves as an interface be-tween applications-oriented software and the hardware layer. The software needs to be aware of the capabilities of the hardware (both DSP and RF hardware) at both ends of the communications link to ensure compatibility and to make maximum use of the hardware resources. Furthermore, given that the software radio will operate in an existing data infrastructure, it must interface quickly and efficiently with this infrastructure. This means that the software radio needs to control issues such as attribute naming, error management, and addressing, regardless of the protocol used in the infrastructure. Partitioning the radio functions into objects can help with these issues as well as aid in portability and maintenance of the software. Example objects might include the blocks of the model software radio shown in Figure 1.1. Security is an important issue to ensure that software downloads are legitimate. Finally, given that higher-layer protocols such as TCP have constraints inherent to the way in which they manage a session, the software architecture should consider latency and timing for the whole protocol stack.
• Step 5: Digital signal processing hardware architecture selection—The core
dig-ital signal processing hardware can be implemented through microprocessors, FP-GAs, and/or ASICs. Typically microprocessors offer maximum flexibility, highest
8 Introduction to Software Radio Concepts Chapter 1
power consumption, and lowest computational rate, while ASICs provide minimal flexibility, lowest power consumption, and highest computational rate. FPGAs, on the other hand, lie somewhere between an ASIC and a DSP in these characteristics. The selection of the core computing elements depends on the algorithms and their computational and throughput requirements. In practice, a software radio will use all three core computing elements, yet the dividing line between the implementation choices for a specific function depends on the particular application being supported.
• Step 6: Radio validation—This step is perhaps the most difficult. It is essential
to ensure not only that the communicating units operate correctly, but also that a glitch does not cause system-level failures. Interference caused by a software radio mobile unit to adjacent bands is an example of how a software radio could cause a system-level failure, and this is of great concern to government regulators [2]. Given the many variable parameters for the software radio and the desire for an open and varied source of software modules, it is very difficult to ensure a fail-proof system. Testing and validation steps can be taken to help minimize risk. Structuring the software to link various modules with their limitations and deficiencies can help in testing compatibility of software modules.
As you can see from the cartoon in Figure 1.2, Dilbert is skeptical of the ideal software radio. This skepticism is understandable; software radios require a much higher level of systems-level engineering than today's products. To carry out this cooperative interdisci-plinary design, engineers must understand the ramifications of their design on the overall system and be willing to have their subsystem control and be controlled by other subsys-tems, and they must be knowledgeable in a variety of technical disciplines.
OUR DEVICE CONFORMS TO ALL INTERN ATION/KL STANDARDS FOR COfWUNICATIONS I N OTHER WORDS, I T DOESN'T DO ANYTHING USEFUL AND I T ' S NOT YOUR FAULT. I S THERE SOMEBODY LESS EXPERIENCED I COULD TALK T O7
;
DO YOU HAVE W BOSS'S NUMBER?Figure 1.2: Dilbert's View of Software Radios.
S O U R C E : S. Adams, "Dilbert," 4/11/1994. © United Feature Syndicate, 1994. Used by Permission.
1.5 Questions
Fill in the design matrix in the following table to show how one design step may be related to another design step. For the sake of illustration, some examples are given.
Step 1: Systems Planning
Step 2: RF Chain Planning
Step 3: ADC and DAC Selection Step 4: Software Architecture Selection Step 5: Digital Signal Processing Hardware Architecture Selection Step 6: Radio Validation Step 1: Systems Engineering Not Applicable Step 2: RF Chain Planning
Not Applicable Dynamic range of the ADC or DAC should match the RF chain. Step 3: ADC and
DAC Selection
ADC and DAC may operate at baseband or IF. Not Applicable Step 4: Digital Signal Processing Hardware Architecture Selection
Not Applicable Test the compatibility of hardware modules needed for various parameters. Step 5: Software Architecture Selection Responsive software control of the RF can help compensate for imperfections in the RF. Not Applicable Step 6: Radio Validation Not Applicable
C h a p t e r 2
R A D I O F R E Q U E N C Y
I M P L E M E N T A T I O N ISSUES
A good communication system design must strike the right balance in its components. For example, having an expensive ADC capable of more than 90 dB of dynamic range makes little sense when the dynamic range of the RF front-end is only 60 dB. To properly balance subsystem performance, the digital signal processing engineer must be aware of the limitations of the analog RF front-end so that some compensations can be made via digital signal processing.
A software radio poses an even greater challenge for the RF designer because it is very difficult to create an RF front-end that is applicable to a variety of signals with dis-parate parameters, such as bandwidth and center frequency. This chapter addresses these concerns from the perspective of the digital signal processing engineer and highlights the alternatives and challenges in the front-end design process. The problems associated with designing the RF front-end and the methods of compensating for some of the inadequa-cies of the RF front-end using digital signal processing are described to provide the digital signal processing engineer with an awareness of the overall system limitations and the re-sources and alternatives available. More detailed information regarding specific areas of RF engineering can be found in the annotated bibliography in Appendix A.
2.1 T h e Purpose of the R F Front-End
A transceiver consists of a receiver section and a transmitter section. Of the two sections, the receiver is by far the most complex and is the primary focus of this chapter. The purpose of the receiver is to isolate the desired signal from interference and noise for demodulation and further processing. The key is to reject undesired signals and condition the desired signal for digital signal processing by taking the signal from the antenna, filtering it to remove undesired signals, converting the signal to a center frequency with an amplitude compatible with the analog to digital conversion process, and finally performing the analog to digital conversion process. This process is depicted in Figure 2.1, and its objectives are summarized in the following list.
12 Radio Frequency Implementation Issues Chapter 2 A n t e n n a Selection D o w n c o n v e r s i o n A n a l o g to Digital C o n v e r s i o n Digital O u t p u t R e c e i v e r B a c k - E n d
• X
-R F Front-EndFigure 2.1: Processing Steps in the Receiver RF Front-End.
• Reject as many undesired signals as possible.
• With minimal distortion, convert the desired signal's center frequency to a range compatible with the ADC. The process of mixing or frequency translation may lead to undesirable non-linear distortion and introduction of additive noise.
• With minimal distortion, amplify the desired signal to the level required by the ADC. • Minimize additive noise.
• Achieve a dynamic range that is compatible with that of the ADC. Dynamic range is defined as the difference in power between the weakest detectable signal (usually limited by noise) and the strongest signal (often the interference).
The RF front-end must separate a desired signal, typically in the picowatt range (—160 to - 1 0 0 dBW, or - 1 3 0 dBm to - 7 0 dBm), from a background RF environment that may be in the milliwatt range ( - 2 0 dBm to 0 dBm). In many systems, the RF front-end also sets the system signal-to-noise ratio (SNR) and should be designed to add minimal noise.
Thus the overall system must have a considerable dynamic range to accommodate both the high-power background signals and the lowest-power desired signal. The wider the bandwidth of the receiver, the more potential for interference and noise, and thus, the more difficult it is to achieve high dynamic range. Achieving adequate dynamic range is one of the central issues in RF design.
The purpose of the transmitter RF section is to convert the digital representation of the analog signal into a radiated analog signal. The process is nearly reverse that of the receiver and includes converting a digital signal (via the DAC to its analog representation), upcon-verting the analog signal to the desired RF center frequency, amplifying the signal to the appropriate power level, and limiting the signal's bandwidth before it is radiated. The pro-cess is depicted in Figure 2.2. In practice, multiple stages of conversion and amplification may occur before the signal is radiated.
Section 2.2 Dynamic Range: The Principal Challenge of Receiver Design
13
A n t e n n a Digital Input Digital to A n a l o g C o n v e r s i o n Conversion Selection a n d AmplificationFigure 2.2: Processing Steps in the Transmitter RF Front-End.
2.2 D y n a m i c Range: T h e Principal Challenge
of Receiver Design
Dynamic range is the key design challenge in building an RF front-end since it provides a measure of the highest- and lowest-level signals that can be simultaneously accommodated by the radio. Furthermore, there is a strong tie between battery consumption and dynamic range, which is a particularly important trade-off for mobile systems. Numerous physical limitations of the radio's components impact the dynamic range of the system, but various approaches to radio design can help improve the dynamic range. Downstream digital signal processing can sometimes improve the dynamic range of the system, but the improvement provides minimal compensation for non-ideal characteristics of the RF front-end. Here we describe how the limitations on dynamic range occur. For more details about RF design techniques for improving the dynamic range, see [3-8].
Dynamic range is limited at the bottom of the range by noise that enters the system through thermal effects of the components or through non-idealities of the ADC, such as quantization noise or sampling aperture jitter (see Chapter 5). Low-level signals can be masked by this noise. Dynamic range is limited at the high-end by interference. The source of this interference could be co-channel, adjacent channel, or self-induced by the transceiver. High levels of interference may cause the receiver to become more non-linear and introduce cross-products (spurious components), which may inhibit the detection of low-level signals or reduce the desired signal bit error rate (BER). Simply attenuating the high-level signals before they drive the receiver into a non-linear operating region is insuffi-cient since low-level desired signals that are also present will be attenuated until masked by system induced noise and thus will fall below the sensitivity of the receiver. The dynamic range limitation is depicted in Figure 2.3.
Non-linear distortion can be characterized as intermodulation, cross-modulation, and reciprocal mixing. Intermodulation distortion creates signal energy at the sum and differ-ence frequencies due to products of all signals present. Cross modulation occurs when the modulation of the stronger signal is imparted onto the weaker desired signal. Reciprocal mixing is the cross-product of an adjacent channel signal and a noisy oscillator signal and results in a higher noise floor in the region of the desired signal.
14 Radio Frequency Implementation Issues Chapter 2 B E R Noise Limited ( t h e r m a l effects, quantization, s a m p l i n g jitter) S p u r i o u s Signal Limited ( c o - c h a n n e l , a d j a c e n t c h a n n e l , a n d self-induced interference p r o d u c e receiver overload) 0.5 U s a b l e D y n a m i c R a n g e
Input Signal Level Figure 2.3: Limitation of Dynamic Range.
SOURCE: M. McMahan, A. Khatzibadeh, and P. Shah, "Wireless Systems and Technology Overview" [9]. ©Texas Instruments, Inc. Used by Permission.
Proper selection of components and good circuit design techniques, particularly in the active components, can improve the dynamic range of the RF front-end. Providing good initial selectivity at the RF front-end can reduce the interference levels, thereby reducing the need for high dynamic range in subsequent sections. Direct current (DC) bias in the signal can consume the dynamic range. This bias is possibly produced by self-mixing in the mixer circuits (the local oscillator signal is picked up at the other mixer input and produces a DC output) or a warping of the signal constellation from phase mismatch of the inphase and quadrature (I&Q) stages due to circuit drift. This bias can be estimated within software, and a DAC and analog subtracters can be used to remove the bias before the signal is digitized [ 1 0 ] [ 1 1 ] .
2.3 R F Receiver Front-End Topologies
A number of different R F front-end topologies are appropriate for software radios. Each has its own advantages and disadvantages. The most common types of RF front-ends are dual conversion, single conversion, and tuned radio frequency receivers.
2 . 3 . 1 C h a r a c t e r i s t i c s o f t h e T o p o l o g i e s
The suitability of a particular receiver topology depends on a number of parameters that may include the following.
• Sensitivity defines the weakest signal level that a receiver can detect and is usually determined by the various noise sources in the receiver.
Section 2.3 RF Receiver Front-End Topologies 15
• Selectivity represents the ability of the receiver to detect the desired signal and reject
all others.
• Stability indicates the lack of change in the receiver gain and operating frequency
with temperature, time, voltage, etc.
• Dynamic range is the difference in power between the weakest signal that the receiver
can detect and the strongest signal that can be supported (either in-band or out-of-band) on the receiver without detrimental effects.
• Spurious response is a receiver's freedom from interference due to internally
gener-ated spurious signals or to their interaction with external signals.
2.3.2 Topologies
The tuned radio frequency (TRF) receiver, shown in Figure 2.4, consists of an antenna connected to an RF bandpass filter (BPF). The BPF selects the signal and the LNA with the automatic gain control (AGC) raises the signal level for compatibility with the ADC. This BPF bandwidth relative to the carrier frequency can be quite narrow, while in absolute bandwidth, it may be quite broad. For example, a second-order inductor and capacitor filter would require a filter quality factor1 of 1 07 to extract a 30 kHz signal at 900 MHz with 60
dB of attenuation for a channel 60 kHz away, which is highly impractical [12]. The pri-mary difficulty in creating a practical TRF receiver is the limitation of the ADC, which must handle high-frequency signals. In addition, given the bandwidth and roll-off limitations of the RF filter, the sampling rate of the ADC must be very high to avoid significant aliasing. High power consumption is inevitable with high sampling rate conversion. The ADC must accommodate multiple signals over the wide bandwidth of the RF filter (potentially tens of megahertz or more) with high dynamic range of approximately 100 dB. Achieving this sampling characteristic is difficult, expensive, and power-intensive, and extreme demands are made of the tunable RF filter to remove interference signals that consume the dynamic range of the ADC. Non-idealities of the ADC, such as jitter and finite aperture size, lead to distortion of the signal (see Chapter 5). In practice, the RF filter can select only a general band of interest; subsequent filtering within the DSP is required to extract the desired chan-nel. The AGC adjusts its gain to accommodate varying signal levels to utilize the full range of the ADC without overloading it. However, the especially high gain required for a single-stage AGC in this application may be difficult to control. Nevertheless, the advantage of this approach is the minimal number of analog parts required.
A very popular topology for low-power application is the single conversion receiver (also known as homodyne, direct conversion, or zero-IF receiver), which uses a single mixing stage to convert the signal to baseband or near baseband. This receiver architecture, shown in Figure 2.5,2 has one stage of downconversion. In the case of a phase or frequency
' Filter quality factor reflects the sharpness of the magnitude and phase response and is denned as the ratio of the filter's center frequency to the filter's bandwidth.
2 It may be necessary to supplement the AGC with another stage after the lowpass filter (LPF) to prevent the mixer from being saturated by high levels of in-band interference. The AGC attempts to maintain full-scale ADC input swing, and if the AGC precedes the BPF, interferers are amplified along with the desired signal.
BPF RX Filter
ADC
Binary Output
Figure 2.4: TRF Digital Signal Processing Receiver.
\
/
B P F f ' r B P F R X Filter R X Filter L P F A D C L P F V A D C B a s e b a n d Digital O u t p u t L O (a) > B P F R X F i l t e r B P F A G C > ^ B P F A G C R X F i l t e r L P F A D C L O B a s e b a n d D i g i t a l O u t p u t 90° L P F A D C (b)Figure 2.5: (a) Single Conversion Receiver for Binary Phase Shift Keying (BPSK) and Amplitude Modulation (AM), (b) Single Conversion for Frequency and Phase Modulated Signals.
Section 2.3 RF Receiver Front-End Topologies 17
modulated signal, I&Q downconversion is required since the upper and lower sidebands of these signals contain different information and the sidebands would overlap for a real downconversion. Mixers tend to have high power consumption, and since only one mixer stage (possibly I&Q) is used in the single conversion receiver, the receiver potentially offers good power consumption characteristics. Typically, improved power consumption at the mixer can be traded for dynamic range.
LO leakage has the potential of creating leakage across input ports, causing the mixer to downconvert a received version of itself (self-mixing), which may result in a large DC bias at the mixer output. Isolation between the LO and input to the mixer or other components is very desirable but difficult to achieve. An alternating current (AC) coupling capacitor helps but may remove important DC information in the signal. A more effective though more costly approach is to track the DC error after digitization and feed back a correcting bias signal using a DAC and subtractor.
A non-ideal I&Q downconversion may result if the phase and amplitude of the branches are not matched and cause a warping of the received signal constellation diagram as shown in Figure 2.6 for the case of a quadrature phase shift keying (QPSK) signal. Furthermore, the phase stability on the local oscillator is extreme given the high and precise frequency needed to convert the signal to baseband since phase noise falls within the baseband. Good circuit design with digital signal processing based compensation can help mitigate these problems. Note that these problems are absent when using the TRF receiver.
In some cases, rather than directly downconverting the signal to baseband, it may be more convenient to downconvert to some low intermediate frequency at which the signal may be digitized and downconverted by subsequent digital signal processing operations. A more complex LPF with better roll-off characteristics can help reduce out-of-band interfer-ence and thus lessen the dynamic range requirement of the ADC, but it could also allow more noise to enter the system (less sensitivity,) resulting in non-linear distortion products from the filter.
The most common RF front-end for radios is the heterodyne receiver. This receiver, shown in Figure 2.7a, is commonly used in analog radios. A heterodyne receiver works by frequency translating the incoming signal to an IF that is fixed and independent of the de-sired signal's center frequency. When this IF frequency is lower than the center frequency of the received signal's carrier frequency and higher than the bandwidth of the desired signal, the receiver is called a superheterodyne receiver. The desired signal that is now frequency translated to a fixed IF can be more easily filtered, amplified, and demodulated. Plenty of good quality RF components are available for standard IF frequencies. Often a superheterodyne receiver involves using two stages of downconversion. Such a dual con-version receiver has the advantage of relaxed filtering requirements. Because the filtering occurs in stages, the filtering requirements at each stage can be more relaxed than in a sin-gle conversion receiver. That is, by lowering the center frequency of the signal using the first stage of downconversion, the filter quality factor can also be relaxed because the ratio of center frequency to filter bandwidth is reduced. Gain can also be achieved in stages, re-ducing the LO power on the mixers and relaxing the isolation needed between the LOs and the mixer inputs. The distribution of this gain throughout the front-end impacts the overall dynamic range. DC offset is of no concern in this architecture since the LO frequency UJLO
18 Radio Frequency Implementation Issues Chapter 2
is not equal to the center frequency of the desired signal at the input of the mixer. The additional mixer and LO result in higher power consumption and a larger circuit than that of a single conversion receiver, and often the second filter can be expensive and may exist off-chip. The characteristics of the I&Q mixers need to be matched to prevent distortion
O symbol states without noise • symbol states with noise O • • o .
• • •
•• o ;
• o .(a). Ideal QPSK Constellation. (b). DC Bias due to Self-Mixing. DC is present in both l&Q.
. ' O • • o : / o
• • o •
• • •
(c). Gain Mismatch. Q is amplified more than I.
(d). Phase Mismatch. The phase between l&Q branches is not 90°, and amplitude is distorted.
Figure 2.6: Impact on Constellation Due to Imperfect Mixing Process. (Note that with noise and distortion, a symbol is more likely to lie in the wrong quandrant, leading to a bit error.)
