COMMUNICATION SYSTEM DESIGN USING DSP ALGORITHMS WITH LABORATORY EXPERIMENTS FOR THE TMS320C6701 AND TMS320C6711

Transcription

1 COMMUNICATION SYSTEM DESIGN USING DSP ALGORITHMS WITH LABORATORY EXPERIMENTS FOR THE TMS320C6701 AND TMS320C6711

2 Information Technology: Transmission, Processing, and Storage Series Editor: Editorial Board: Jack Keil Wolf University of California at San Diego La Jolla, California Robert J. McEliece California Institute of Technology Pasadena, California John Proakis Northeastern University Boston, Massachusetts University of California at San Diego La Jolla, California William H. Tranter Virginia Polytechnic Institute and State University Blacksburg, Virginia Coded Modulation Systems John B. Anderson and Arne Svensson Communication System Design Using DSP Algorithms: With Laboratory Experiments for the TMS320C6701 and TMS320C6711 Steven A. Tretter A First Course in Information Theory Raymond W. Yeung Multi-Carrier Digital Communications: Theory and Applications of OFDM Ahmad R. S. Bahai and Burton R. Saltzberg Nonuniform Sampling: Theory and Practice Edited by Farokh Marvasti Principles of Digital Transmission: With Wireless Applications Sergio Benedetto and Ezio Biglieri Simulation of Communication Systems, Second Edition: Methodology, Modeling, and Techniques Michel C. Jeruchim, Philip Balaban, and K. Sam Shanmugan A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact the publisher.

3 COMMUNICATION SYSTEM DESIGN USING DSP ALGORITHMS WITH LABORATORY EXPERIMENTS FOR THE TMS320C6701 AND TMS320C6711 Steven A. Tretter University of Maryland College Park, Maryland Kluwer Academic / Plenum Publishers New York, Boston, Dordrecht, London, Moscow

4 Library of Congress Cataloging-in-Publication Data Tretter, Steven A. Communciation system design using DSP algorithms: with laboratory experiments for the TMS32OC6701 and TMS320C6711. p. cm. - (Information technology: transmission, processing, and storage) Includes bibliographical references and index. ISBN -13: Signal processing-digital techniques-data processing. 2. Texas Instruments TMS320 series microprocessors. I. Title. II. Series. TK T dc Additional material to this book can be downloaded from ISBN -13: DOl: e-isbn -13: Kluwer Academic/Plenum Publishers, New York 233 Spring Street, New York, New York A CI.P. record for this book is available from the Library of Congress All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work.

5 To Teresa, Anne, Jeffrey, Max, and in memory of David

6 Preface The first edition this book began in January 1993 when I was "elected" by Dr. William Destler, who was then Chairman of the Electrical Engineering Department and is now Provost and Vice President for Academic Affairs of the University of Maryland's College Park Campus, to set up and write experiments for a new senior elective laboratory course, ENEE418c Communications Laboratory, scheduled to be given for the first time in the Fall 1993 semester. At that time, we chose to use the state-of-the-art Texas Instruments TMS320C30 EVM (evaluation module) DSP board. In January 2001 we upgraded the lab to use the new state-of-the-art Texas Instruments TMS320C6701 EVM DSP board and their Code Composer Studio software interface. Some discussion of how to use the much cheaper TI TMS320C6711 DSK is included in this book also. Initially, the lab used Intel 486 based host PC's running Windows 3.1. These have been replaced by modern Pentium based PC's running Windows NT which are all connected to the campus network. Each lab group is given a private workspace on the lab server. Students are given only read/execute privileges for the standard utility and development software so they do not inadvertently alter these files. In 1993, books for hardware based laboratory courses with standard digital signal processing and filter design experiments existed, but no book focusing on analog and digital communications techniques was available. This is still largely true. Laboratories in the Department of Electrical and Computer Engineering at the University of Maryland are separate courses. Each week they have a one hour lecture given by a regular faculty member to introduce the theory and explain the experiments followed by a three hour laboratory period run by a graduate teaching assistant. Students in this lab work in pairs. We have found that this works well because both group members actively participate. With groups of three or more, some members just sit and watch. Students are given the combination to the lock on the Communications Lab door and can work whenever they please outside of regular class hours. One section of the lab was first offered in the Fall 1993 semester and two in the Spring 1994 semester. Then five sections a week were offered for several years until about a year ago when a new communications capstone design course was offered in addition. The students have been highly enthusiastic and often spend extra hours working on the experiments because they find them to be very interesting and challenging. They also have realized that this course will help them get jobs and provide them with the skills required to perform well in their future jobs. The lab was designed for seniors, but 1/4 to 1/3 of the class is now graduate students who want to learn some real-world practical skills in addition to the purely theoretical concepts presented in the typical graduate communications and signal processing courses. When asked why they are taking this senior class, the graduate vii

7 viii Preface students often say they think it will help them get jobs. The goal of this set of experiments is to explore the digital signal processing and communication systems theoretical concepts presented in typical senior elective courses on these subjects by implementing them with actual hardware and in real time. In the process, students will gain experience using equipment commonly used in industry, such as, oscilloscopes, spectrum analyzers, signal generators, error rate test sets, digital signal processors, and analog-to-digital and digital-to-analog converters. They will also learn about typical software development tools. In addition, they will learn that there is a big step in going from an equation on paper to a real working system. This book differs from any others on the market in that its primary focus is on communication systems. Fundamental digital signal processing concepts like digital filters and FFT's are included because they are required in communication systems. Approaches that are particularly useful for DSP implementations are presented. While the experiments, particularly the earlier ones, are described for the TMS320C6701 Evaluation Module and TMS320C6711 DSK, they can be easily modified for any PC DSP board with an AID and D / A converter. There are several books on digital signal processing experiments for stable software packages like MATHCAD and MATLAB. In my view, one of the purposes of a laboratory course is to help prepare students for industrial jobs. Off-line software simulation is no substitute for making actual hardware work in real-time. It does not present students with the strange unexpected and often frustrating things that occur when using actual hardware in real-time which can not be explained by nice equations, nor does it teach them how to use standard lab equipment. The prerequisites for this course are an understanding of linear systems and transform methods at a level that is often presented in a junior required course on Signals and Systems and a working knowledge of PC's and C programming. Students who have programmed in other languages like BASIC, PASCAL, or FORTRAN can quickly learn enough C to do the experiments if they are willing to make the effort. Corequisite are a senior level elective course in Digital Signal Processing and/or Communication Systems. Ideally, both courses should be taken before the Communications Laboratory. However, this is not usually possible for our seniors. We wanted our students to have the opportunity to take this lab, so we made just one of them a corequisite. With the engineering background of a senior, the presentation of the necessary theory in the text, and the one hour lab lecture to explain the theory, students have quickly learned the signal processing and communication system concepts required for the experiments. In addition, it can be argued that a lab course should help prepare students for the work world where they will have to figure out new things for themselves so the experiments should have some uncertainty and require students to fill in some of the details. There is a large initial hurdle for the students to get over while learning the details required to use the lab's hardware and software tools. Chapters 1 and 2 gradually introduce them to these tools and the architecture of the TMS320C6701 and TMS320C6711 fioatingpoint DSP's. An attempt has been made to reduce this hurdle by including some basic programs on the program disk for initializing the EVM and DSK that can be used as a starting point for the experiments. FIR and IIR filter design and implementation are explored in Chapter 3. Filters are re-

8 Preface ix quired in many communication system signal processing algorithms. Experiments comparing the relative merits of C and assembly language implementations are performed. In particular, TI's linear assembly is briefly discussed. Modern DSP applications in industry are often written primarily in C with only numerically intensive critical functions written in assembly to reduce development time and improve portability to new platforms. TI's optimizing C compiler generates somewhat efficient software pipelined executable code that is adequate for the experiments in this course. Therefore, assembly programming is not emphasized. Chapter 4 investigates the FFT and power spectrum estimation. A simple spectrum analyzer is made. Chapters 5 through 8 explore the classical analog communication methods of amplitude modulation, double-sideband suppressed-carrier amplitude modulation, single-sideband modulation, and frequency modulation. Transmitters and receivers are built using DSP techniques. Noncoherent receivers using envelope detectors and coherent receivers using phase-locked loops are implemented. The use of Hilbert transforms and complex signal representations in modulation systems are explored. Chapters 9 through 16 introduce some modern digital communication techniques. These experiments focus on methods used in high-speed wire-line data moderns where DSP's have been extensively used. Topics covered include linear shift register scramblers, the RS232C interface, pulse amplitude modulation (PAM), variable phase interpolation, and quadrature amplitude modulation (QAM). The experiments lead up to building almost a complete V.22bis transmitter and receiver. Symbol clock recovery and tracking, carrier tracking, and adaptive equalizer receiver functions are implemented. The echo canceling technique used in V.32, V.34, V.90, and V.92 modems is studied in Chapter 16. Enough details are included so that this set of experiments could form a good practical guide to engineers in industry interested in modem design. I learned many of these techniques while consulting since 1970 for companies that build high-speed wire-line moderns and have seen them employed in hundreds of thousands of modems. Chapter 17 briefly presents some ideas for additional projects related to high-speed wireline modem design, error-control coding, and speech codecs. These ideas can be expanded to satisfy the capstone design project requirements of the ABET accrediting committee. Appendix C contains a complete list of the equipment used for this laboratory at the University of Maryland. It has been included as a guide to others setting up a similar lab and is not intended to be an endorsement of any specific manufacturer. Clearly, any equipment with equivalent capabilities can be substituted for items in the list. There are many more experiments in this book than can be performed in one semester. Based on our experience, an ambitious goal is to have all students do Chapters 1, 2, and 3 followed by a choice of any three additional experiments. In some semesters, we have limited the choice to three of the classical analog modulation chapters and in others to three of the digital communication chapters. It would be nice if students could continue in the lab for a second semester for additional credit and build on their earlier experiments. Utility programs, software updates, text corrections, lab lecture slides, and supplementary material can be found on my web site www. ee. umd. edurtretter. Steven A. Tretter

9 Acknowledgements I would like to thank Mark Kohler and Sonjai Gupta for helping to unpack equipment and install and debug hardware and software in the PC's during the Spring 1993 semester when the lab was initially being set up. The students in the original Communications Laboratory section offered in the Fall 1993 semester as well as their Teaching Assistant, Yifeng Cui, deserve a great deal of thanks for being extremely patient and enthusiastic Guinea pigs. They helped correct and significantly improve the first few experiments. In particular, I want to thank Mike Barr who, with his partner Brian Silverman, forged well ahead of the rest of the students and was so enthusiastic by the end of the semester that he asked about helping with the lab in the Spring 1994 semester. The Electrical Engineering Department was able to make an exception and assign Mike as a senior to be the TA for one section and he did an outstanding job. Mrs. Tahereh Fazel was also an excellent Graduate TA for a second lab section in the Spring 1994 semester. I also want to thank those who have been TA's for the lab since that time and have all done excellent jobs as well as Dr. Adrian Papamarcou and Dr. Jerome Gansman who have shared in teaching the lab. I also thank Dr. Gansman for collecting information about the 'C6701 EVM and 'C6701 DSP and generating a set of very detailed and useful lecture slides for a capstone design course. Dr. Brian Evans of the University of Texas at Austin also deserves thanks for reviewing preliminary versions of this manuscript and making good suggestions for improvements. I would also like to thank my friends from RIXON (which was in Silver Spring, MD, but has long since dissipated in a chain of corporate take-overs), from Penril Datability Networks (which was in Gaithersburg, MD, and was bought by Bay Networks, which was in turn bought by Nortel Networks and is now the Signal Processing Group of Nortel in Germantown, MD), and at Telogy Networks (which is a Texas Instruments company in Germantown, MD) for helping me stay at the state-of-the-art in DSP applications to wire-line moderns by using me as a consultant. I also want to thank Texas Instruments for designating our lab as one of ten Elite Digital Signal Processing Labs in the country. TI has been very generous in supporting our lab with TI hardware and software tools. In particular, Torrence Robinson with University Programs at TI in Houston deserves many thanks for his constant and cheerful support. Figures in the book with captions marked by t have been reprinted from TI documents with their approval. xi

10 Contents 1 Overview of the Hardware and Software Tools 1.1 Some DSP Chip History and Typical Applications 1.2 The TMS320C6701 and 6711 Floating-Point DSP's The 'C6000 Central Processing Unit (CPU) Memory Organization for the TMS320C Memory Organization for the TMS320C Direct Memory Access Controllers Serial Ports Other Internal Peripherals Brief Description of the TMS320C6000 Instruction Set Parallel Operations and Pipelining The TMS320C6701 Evaluation Module (EVM) The CS4231A Stereo Codec Interface The TMS320C6711 DSP Starter Kit (DSK) The Audio Interface Onboard the 'C6711 DSK The PCM3003 Audio Daughter Card Software Support for the EVM and DSK Boards and 'C6x DSP's Software Support for the 'C6701 EVM Software Support for the 'C6711 DSK. 1.6 Code Composer Studio Project Files and Building Programs The Optimizing Compiler and Assembler The Linker Building Programs from Command Line Prompts The Archiver Additional Code Composer Studio Features 1. 7 Other Software Digital Filter Design Programs Commercial Software 1.8 Introductory Experiments Learning to Use the Hardware and Software Tools 2.1 Getting Started with a Simple Audio Loop Through Program A Beginning C Program and Linker Command File xiii

11 xiv Contents Creating a CCS Project for evmstart. c Testing evmstart. out.... More Details on the McBSP Serial Ports and Codecs Basic McBSP Transmitter and Receiver Operation Useful Functions in the C6x Peripheral Support Library Example C Code for Reading and Writing the EVM Codec The 'C6000 Timers.... Method 1: Generating a Sine Wave by Polling DXR Instructions for the Polling Experiment... Method 2: Generating a Sine Wave Using Interrupts The CPU Interrupt Priorities and Sources What Happens When an Interrupt Occurs TI Extensions to Standard C Interrupt Service Routines Using the Peripheral Support Library for Interrupts Instructions for the Experiment of Generating Sine Waves by Using Interrupts Method 3: Generating a Sine Wave by DMA From a Table The DMA Control Registers Major Steps for Setting Up a DMA Transfer DMA Autoinitialization Triggering DMA Transfers Split-Channel Operation Priorities Peripheral Support Library Functions for DMA Instructions for the Experiment of Generating Sine Waves by Using the DMA Controller Digital Filters Discrete-Time Convolution and Frequency Responses Finite Duration Impulse Response (FIR) Filters Block Diagram for Most Common Realization Two Methods for Finding the Filter Coefficients to Achieve a Desired Frequency Response Using Circular Buffers to Implement FIR Filters in C Circular Buffers Using the 'C6000 Hardware How the Circular Buffer is Implemented Indirect Addressing Through Registers Interfacing C and Assembly Functions Responsibilities of the Calling and Called Function Using Assembly Functions with C Linear Assembly Code and the Assembly Optimizer A Linear Assembly Convolution Function that Uses a Circular Buffer and Can be Called from C Infinite Duration Impulse Response (IIR) Filters

12 Contents xv Realizations for IIR Filters A Program for Designing IIR Filters Two Methods for Measuring a Phase Response Laboratory Experiments for Digital Filters Laboratory Experiments for FIR Filters Entirely in C Laboratory Experiments for FIR Filters Using C and Assembly Laboratory Experiments for IIR Filters Additional References The FFT and Power Spectrum Estimation 4.1 The Discrete-Time Fourier Transform Data Window Functions The Discrete Fourier Transform and its Inverse. 4.4 The Fast Fourier Transform Using the FFT to Estimate a Power Spectrum 4.6 Laboratory Experiments FFT Experiments III III ll Experiments for Power Spectrum Estimation Additional References Amplitude Modulation Theoretical Description of Amplitude Modulation Mathematical Formula for an AM Signal Example for Single Tone Modulation The Spectrum of an AM Signal Demodulating an AM Signal by Envelope Detection Square-Law Demodulation of AM Signals Hilbert Transforms and the Complex Envelope Laboratory Experiments for AM Modulation and Demodulation Making an AM Modulator How to Capture EVM Output Samples with CCS for Plotting Making a Square-Law Envelope Detector Making an Envelope Detector Using the Hilbert Transform Additional References DSBSC Amplitude Modulation and Coherent Detection Mathematical Form for a DSBSC-AM Signal The Ideal Coherent Receiver The Costas Loop as a Practical Approach to Coherent Demodulation Exercises and Experiments for the Costas Loop Theoretical Design Exercises Hardware Experiments Additional References

13 xvi Contents 7 Single-Sideband Modulation and Frequency Translation 7.1 Single-Sideband Modulators Coherent Demodulation of SSB Signals 7.3 Frequency Translation Laboratory Experiments Making an SSB Modulator Coherent Demodulation of an SSB Signal Additional References. 8 Frequency Modulation 8.1 The FM Signal and Some of its Properties l.1 Definition of Instantaneous Frequency and the FM Signal. 8.l.2 Single Tone FM Modulation... 8.l.3 Narrow Band FM Modulation The Bandwidth of an FM Signal FM Demodulation by a Frequency Discriminator An FM Discriminator Using the Pre-Envelope A Discriminator Using t.he Complex Envelope 8.3 Using a Phase-Locked Loop for FM Demodulation Laboratory Experiments for Frequency Modulation Experimentally Measuring the Spectrum of an FM Signal FM Demodulation Using a Frequency Discriminator Using a Phase-Locked Loop for FM Demodulation. 8.5 Additional References Pseudo-Random Binary Sequences and Data Scramblers Using Shift Registers to Generate Pseudo-Random Binary Sequences l.1 The Linear Feedback Shift Register Sequence Generator l.2 The Connect.ion Polynomial and Sequence Period l.3 Properties of Maximal Length Sequences Self Synchronizing Data Scramblers The Scrambler The Descrambler Theoretical and Simulation Exercises Exercises for a Shift Register Sequence Generator with a Primitive Connection Polynomial Exercises for a Shift Register Sequence Generator with an Irreducible but not Primitive Connection Polynomial Exercises for a Shift. Register Sequence Generator with a Reducible Connection Polynomial Additional References

14 Contents xvii 10 The RS-232C Protocol and a Bit-Error Rate Tester 10.1 The EIA RS-232C Serial Interface Protocol Error Rate for Binary Signaling on the Gaussian Noise Channel The Navtel Datatest 3 Bit Error Rate Tester Bit-Error Rate Test Experiment Additional References Digital Data Transmission by Pulse Amplitude Modulation Description of a Baseband Pulse Amplitude Modulation System Baseband Shaping and Intersymbol Interference The Nyquist Criterion for No lsi Raised Cosine Baseband Shaping Filters Splitting the Shaping Between the Transmit and Receive Filters Eye Diagrams Implementing the Transmit Filter by an Interpolation Filter Bank Symbol Error Probability with Additive Gaussian Noise Symbol Clock Recovery Simulation and Theoretical Exercises for PAM Generating Four-Level Pseudo-Random PAM Symbols Eye Diagram for a PAM Signal Using a Raised Cosine Shaping Filter Eye Diagram for a PAM Signal Using a Square-Root of Raised Cosine Shaping Filter Theoretical Error Probability for a PAM System Hardware Exercises for PAM Generating a PAM Signal and Eye Diagram Testing the Square-Law Symbol Clock Frequency Generator Optional Team Exercise Additional References Variable Phase Interpolation Continuously Variable Phase Interpolation Computing the Least-Squares Fits Quantized Variable Phase Interpolation Closing the Tracking Loop Changing the Sampling Rate by a Rational Factor Experiments for Variable Phase Interpolation Open Loop Phase Shifting Experiments Making a Symbol Clock Tracking Loop Additional References Fundamentals of Quadrature Amplitude Modulation A Basic QAM Transmitter Two Constellation Examples The 4x4 16-Point Constellation A 4-Point Four Phase Constellation. 221

15 xviii Contents 13.3 A Modulator Structure Using Passband Shaping Filters Ideal QAM Demodulation QAM Modulator Experiments Steps to Follow in Making a Transmitter Testing Your Transmitter Optional Exercise - Testing Your Transmitter by Sending to a Commercial Modem Additional References QAM Receiver I - Clock Recovery and Other Front-End Subsystems Overview of a QAM Receiver Details About the Receiver Front-End Subsystems Automatic Gain Control The Carrier Detect Subsystem Symbol Clock Recovery Experiments for the QAM Receiver Front-End Additional References QAM Receiver II - Equalizer and Carrier Recovery System The Complex Cross-Coupled Passband Adaptive Equalizer The LMS Method for Adjusting the Equalizer Tap Values Theoretical Behavior of the LMS Algorithm Adding Tap Leakage to the LMS Algorithm The Phase-Splitting Fractionally Spaced Equalizer Decision Directed Carrier Tracking Blind Equalization Blind Equalization with the Complex Cross-Coupled Equalizer Blind Equalization with the Phase-Splitting Equalizer Complex Cross-Coupled Equalizer and Carrier Tracking Experiments Implementing the Slicer Making a Demodulator and Carrier Tracking Loop Making a Complex Cross-Coupled Adaptive Equalizer Bit-Error Rate Test Optional Experiment - Receiving the 16-Point V.22bis Constellation Optional Experiment - Ideal Reference Training Optional Phase-Splitting Fractionally Spaced Equalizer Experiment Optional Blind Equalization Experiment Additional References Echo Cancellation for Full-Duplex Modems The Echo Sources in a Dialed Telephone Line Circuit The Data-Driven, Nyquist, In-Band Echo Canceler General Description The Near-End Echo Canceler The Far-End Echo Canceler. 271

16 Contents xix Far-End Frequency Offset Compensation Echo Canceler Experiments Making a Near-End Echo Canceler Making a Far-End Echo Canceler with Frequency Offset Correction Additional References Suggestions for Additional Experiments 17.1 Elementary Modem Handshake Sequence Make an ITU-T V.21 Frequency Shift Keyed (FSK) Modem 17.3 Fast Equalizer Training Using Periodic Sequences 17.4 Trellis Coded Modulation V.34 Constellation Shaping by Shell Mapping Nonlinear Precoding for V Multi-Tone Modulation for ADSL and Wireless Systems 17.8 Speech Codecs.... A Generating Gaussian Random Numbers B A TTLjRS-232C Interface for McBSP1 C Equipment List for Each Station References I. List of Manuals. II. Selected Reference Books and Papers A. DSP Laboratory Books Using DSP Hardware. B. DSP Laboratory Books Using Software Simulation C. Books and Papers on Digital Signal Processing D. Books and Papers on Communications E. Books and Papers on Error Correcting Codes III. Interesting Web Sites Index

17 COMMUNICATION SYSTEM DESIGN USING DSP ALGORITHMS WITH LABORATORY EXPERIMENTS FOR THE TMS320C6701 AND TMS320C6711