Communication System Design Using DSP Algorithms

The purpose of this initial chapter is to introduce you to the main features of the hardware and software tools that will be used in this course. A variety of signal processing and communication system components will be implemented by writing C and/or assembly language programs for the TMS320C6701 floating-point DSP in our lab. The TMS320C6701 resides on a board Texas Instruments (TI) calls the TMS320C6701 Evaluation Module (EVM) which plugs into a PCI slot inside the PC. The TMS320C6701 communicates with the analog world through a CRYSTAL CS4231A stereo codec on the EVM board and the host PC thorough the PCI bus. The experiments in this book can be easily modified for a much cheaper external board using a TMS320C6711 floating-point DSP which TI calls the TMS320C6711 DSP Starter Kit (DSK) and connects to the PC’s parallel printer port. The DSK has a TI TLC320AD535 codec which is configured with a mono audio input, stero speaker outputs, and a fixed sampling rate of 8 kHz. An inexpensive audio daughter card can be bought for the DSK which has a Burr Brown PCM3003 stereo audio codec with a sampling rate that can be varied up to 48 kHz. It is important to have a general picture of the hardware platform so you will understand how to write programs to accomplish the desired tasks.

The goal of this experiment is to learn how to use the hardware and software tools available at each station. The hardware includes a PC, TMS320C6701 EVM or TMS320C6711 DSK, a signal generator, and an oscilloscope. The main software tool you will use is Code Composer Studio (CCS) which contains an editor and a project building facility which automates calling the C compiler, assembler, and linker. You will also use CCS to load programs into the DSP boards, run them, and monitor their execution. You will gradually learn about the DSP’s architecture including the McBSP serial ports, interrupt controller, and DMA controllers by generating a sine wave by using polling, by using interrupts, and by using DMA from a table.

The major goal of this chapter is to learn how to implement the discrete-time filtering techniques usually presented in a typical required Electrical Engineering undergraduate course on signals and systems and a Senior elective course on digital signal processing by using real-time hardware. Discrete-time filtering is often called digital filtering. In the process, you will learn more about the TMS320C6701 DSP and its EVM, and the TMS320C6711 DSP and its DSK.

In this chapter, you will review and implement some important techniques for digital signal processing. In particular, you will build a spectrum analyzer using the Fast Fourier Transform (FFT). It is assumed that the reader is taking or has had a course on the theory of digital signal processing, so the presentation is brief. It sets the notation and summarizes important results. Comprehensive developments of the theory can be found in the books on digital signal processing listed in the references. References for specific topics are suggested at the end of this chapter.

The standard AM modulated signal contains a sinusoidal component at the carrier frequency which does not convey any of the baseband message information. This component is included to create a positive envelope which allows demodulation by a simple, inexpensive envelope detector. From an information theory point of view, the power in the sinusoidal carrier component is wasted. In this experiment, you will see that it is not necessary to transmit the carrier component and that the baseband message can be recovered by a coherent demodulator. In fact, it can be shown that a coherent demodulator performs better than an envelope detector when the received signal is corrupted by additive noise. The type of modulation that will be studied in this chapter is called double-sideband suppressed-carrier amplitude modulation (DSBSC-AM). A close approximation to an ideal coherent demodulator called a Costas loop will be implemented.

AM and DSBSC-AM modulation do not use the frequency spectrum efficiently. Their spectral components equal distances above and below the carrier frequency contain identical information because they are complex conjugates of each other. The portion above the carrier frequency is called the upper sideband and the portion below the lower sideband. In this experiment you will see how a baseband message can be transmitted by using only one of the sidebands and, consequently, half the bandwidth of AM or DSBSC-AM. This type of modulation is called single-sideband (SSB) modulation. It has been extensively used in many radio transmission systems and in the telephone network.

Frequency modulation (FM) was invented and commercialized after amplitude modulation. Its main advantage is that it is more resistant to additive noise than AM. In addition to commercial radio, it is used as a component of television signals, for satellite and microwave communications, and for digital data transmission. In this chapter the basic theory of FM modulation and demodulation will be presented and you will implement two types of demodulators, the frequency discriminator and the phase-locked loop.

This chapter begins a series on digital communications. DSP technology has made a dramatic impact on digital communications, particularly narrow band systems like voice-band telephone line modems and cellular telephones. In 1970, a plain 9600 bps telephone line modem was the size of a big microwave oven and cost at least $15,000. It was basically just a data pump with no extra features. Now a state-of-the-art V.90 56 kbps modem can be bought for less than $100 and fits in a small box or on a small card. In addition, this modern modem has many features like data compression, error detection and correction, trellis coded modulation, fax modes, automatic dialing, network management functions, a secondary channel, and the ability to do most of the past popular modem standards ranging from speeds of 300 bps up to 33,600 bps. It is now possible to concurrently run at least 12 full duplex V.90 modems in a single state-of-the-art DSP core and chips with multiple cores are in development. These high-end chips will be used in remote access servers (RAS) by Internet service providers for voice over IP (VOIP) and modem pools. Because of the flexibility of the software approach to implementing signal processing algorithms with DSP’s, new theoretical developments have almost instantaneously been included in commercial telephone line modems. These techniques have later found their way into higher speed systems that use greater channel bandwidths like high speed digital subscriber lines (DSL), microwave systems, and satellite communications.

The receiver in a digital communication system usually knows the nominal symbol rate used by the transmitter. Since the receiver is at a distance from the transmitter, has slightly different components, and is at a different temperature, the locally generated symbol clock in the receiver will differ in phase and slightly in frequency from the transmitter’s clock. Therefore, the receiver must synchronize its symbol clock to the clock in the signal received from the transmitter. It must do this just using information derived from the received signal. Some codecs designed for modem front ends have built-in hardware capability for changing their sampling phase by small increments as directed by commands from the DSP they are connected to. The clock tone generator discussed in Chapter 11 can be used to determine the needed phase increments. However, the codecs for the TMS320C6701 EVM and TMS320C6711 DSK do not have this capability and run with a fixed phase. In this chapter, we will see how to implement the phase shifting in the DSP by a variable phase interpolator. First, a continuously variable phase shifter will be presented. Then a phase shifter using fine quantized steps will be discussed.

Quadrature amplitude modulation (QAM) is a widely used method for transmitting digital data over bandpass channels. It can be viewed as a generalization of PAM to bandpass channels. All current telephone line modems based on the ITU-T V series recommendations for transmission at rates of 2400 bps or more use QAM or include it as an option. These include recommendations V,22 through V.92. This series includes FAX modems. Recommendation V.90 modems normally use PAM in the downstream direction from the server to the client modem and always use QAM in the upstream direction from the client to the server. V.90 modems can choose to use QAM downstream if a digital link from the server to the codec in the local office on the client side does not exist. V.92 modems normally use PAM in the downstream and upstream directions but can choose to use QAM based on line conditions. QAM is also used in high speed cable, multi-tone wireless, microwave, and satellite systems. It is a popular choice because it uses bandwidth efficiently and linear channel distortions can be corrected by adaptive equalization at the receiver. In addition, QAM fits in nicely with a popular combined coding and modulation scheme called trellis coded modulation (TCM).

In this chapter and the next you will make a QAM receiver. You should not do these experiments until you have completed Chapter 13 and have made a working QAM transmitter. First, the basic subsystems required in the receiver are briefly described. Then the receiver front-end components, in particular a symbol clock recovery method, are described in detail. These front-end subsystems are what you will implement in the experiments for this chapter.

An important milestone in high speed data transmission over narrow-band channels like the voice-band telephone channel was the invention and commercialization of the FIR adaptive equalizer by R.W. Lucky at AT&T Bell Laboratories in the early 1960’s [II.D.30]. The purpose of the adaptive equalizer is to remove the intersymbol interference caused by the amplitude and phase distortions of the channel. Adaptive filters are used because the frequency response of the channel is not known accurately in many situations. Lucky’s original equalizers used the zero forcing algorithm. Other people soon replaced this algorithm by Widrow’s [II.D.43] more powerful least-mean-square (LMS) algorithm. Another major influence has been the remarkable advances in VLSI technology. This has led to ever more powerful DSP’s which allow complex algorithms to be implemented very inexpensively. For example, modems that include data rates of 300 bps, 1200 bps, and 2400 up to 56000 bps, as well as error correction, data compression capabilities, and FAX modes can be bought for less the $100.

An important advance in the design of high speed voice-band telephone line modems for the dial network was the introduction of echo cancelers to achieve full-duplex data transmission over 2-wire circuits. This technique was studied in the early 1980’s and then widely introduced in commercial products in the mid 1980’s when the CCITT V.32 recommendation for a 9600 bps modem was approved. A few years later, the V.32bis recommendation for 14.4 kbps modems was approved, and the V.34 recommendation for rates up to 33.6 kbps was approved in June 1994. These also use echo cancelers. The recent V.90 and V.92 modems that use PCM downstream use echo cancelers to. Echo cancelers are also used in some high speed digital subscriber lines at data rates of 64 kbps or more. Line echo cancelers were used with analog voice transmission to eliminate annoying talker echo prior to the inclusion of echo cancellation in digital data modems. The voice echo cancelers are placed at different points in the telephone circuit than the ones for data transmission and are disabled during data transmission by a special signal in the modem handshake sequence. The technique is also used in speaker phones to eliminate annoying acoustic reflections from the speaker to the microphone and then back to the far end talker.

It would take a long time (maybe a couple of years?) to complete all the experiments in this book. However, in light of the recent ABET requirement in engineering education for capstone design projects, several additional topics for experiments related to current communication techniques are very briefly described below. References to get you started on each project are included.