Ultimately the design of a digital communication system depends on the properties of the channel. The channel is typically a part of the digital communication system that we cannot change. Some channels are simply a physical medium, such as a wire pair or optical fiber. On the other hand, the radio channel is part of the electromagnetic spectrum, which is divided by government regulatory bodies into bandlimited radio channels that occupy disjoint frequency bands. In this book we do not consider

more »

... the design of the transducers, such as antennas, lasers, and photodetectors, and hence we consider them part of the channel. Some channels, notably the telephone channel, are actually composites of multiple transmission subsystems. Such composite channels derive their characteristics from the properties of the underlying subsystems. Section 18.1 discusses composite channels. Sections 18.2 through 18.4 review the characteristics of the most common channels used for digital communication, including the transmission line (wire pair or coaxial cable), optical fiber, and microwave radio (satellite, point-to-point and mobile terrestrial radio). Section 18.5 discusses the composite voiceband telephone channel, which is often used for voiceband data transmission. Finally, Section 18.6discusses magnetic recording of digital data, as used in tape and disk drives, which has characteristics similar in many ways to the other channels discussed. The most prevalent media for new installations in the future will be optical fiber and microwave radio, and possibly lossless transmission lines based on superconducting materials. However, there is a continuing strong interest in lossy transmission lines and voiceband channels because of their prevalence in existing installations. Thus all the media discussed in this chapter are important in new applications of digital communication. Physical media as well as the composite channels derived from them impose constraints on the design of a digital communication system. Many of these constraints will be mentioned in this chapter for particular media. The nature of these constraints usually fall within some broad categories: • A bandwidth constraint. Sometimes this is in the form of a channel attenuation which increases gradually at high frequencies, and sometimes (particularly in the case of composite channels) it is in the form of a very hard bandwidth limit. • A transmitted power constraint. This is often imposed to limit interference of one digital communication system with another, or imposed by the inability of a composite channel to transmit a power level greater than some threshold, or by a limitation imposed by the power supply voltage of the digital communication system itself. This power constraint can be in the form of a peak-power constraint, which is essentially a limit on the transmitted voltage, or can be an average power constraint. Example 18-4. An FDM system such as that in Fig. 18 -2, where there are perhaps thousands of channels multiplexed together, is designed under assumptions on the average power of the channels. From this average power, the total power of the multiplexed signal can be deduced; this power is adjusted relative to the point at which amplifiers in the system start to become nonlinear. If a significant number of VF channels violate the average power constraint, then the multiplexed signal will overload the amplifiers, and the resulting nonlinearity will cause intermodulation distortion and interference between VF channels. Example 18-5. The PCM system of Fig. 18-3 imposes a bandwidth constraint on the data signal, which must be less than half the sampling rate. A peak power constraint is also imposed by the fact that the quantizer in the PCM system has an overload point beyond which it clips the input signal.