Frequency division multiplexing

2 Many people find it difficult to imagine that a large number of conversations may coexist on a single line without any effect upon each other. It may be worth pointing out that exactly the same thing happens with radio broadcasting. All the various programmes arrive simultaneously to our radio set without affecting each other (well, sometimes they do!) and the selection of the right station takes place in our own set by turning a knob.

Time division multiplexing was introduced in order to transmit several telegraph signals simultaneously on the same line. The trick was to intersperse the signals in time and then separate them at the receiving end (see Chapter 4). Frequency division multiplex does the same thing in the frequency domain. The information is put on a number of carrier waves of different frequencies which propagate entirely independently of each other, and then at the receiving end they are separated.2

How did the idea of frequency division multiplexing originate? It started in the telegraphy business which was big business in the 1870s. If by some means it was possible to send ten messages simultaneously on one single wire, that would lead to tremendous economies. Two men, Alexander Graham Bell and Elisha Gray, worked on the problem independently. Both hit on the idea of using tuning forks.

Tuning forks, as it was well known at the time, emit a well-defined sound frequency, or conversely they can be brought into vibration if sound of the same frequency is incident upon them. Thus a possible way of sending a telegraph message is by interrupting the sound emitted according to the telegraphic code, and then converting the interrupted sound into electrical form. The remarkable thing is that, if one modulates a number of forks tuned to different frequencies in a similar manner, then they can travel on the same wire entirely independently of each other. When they arrive at the receiving end they can be separated according to their frequencies. It was a good idea but it just did not work out in practice. On the other hand once the idea of turning sound into electrical form took root, the obvious next step was to invent the telephone, and that's exactly what both Bell and Gray did.

The road to the practical application of frequency division multiplexing opened up when it became technically possible to (i) generate electromagnetic waves with well specified frequencies, (ii) modulate them by voice, (iii) amplify them without distortion and (iv) separate the different frequencies at the receiver. All these technical problems were solved by the middle of the 1930s. In 1938 the London-Birmingham trunk line was opened to service. It could carry as many as 40 simultaneous telephone conversations.

What will determine, at least in principle, the number of telephone conversations which can be carried on a single line? Is the answer the result of terribly complicated calculations which only the greatest experts can understand? The answer is no: for single sideband AM transmission the calculation can be performed by anyone with some dexterity in the four basic mathematical operations. If one conversa-

Fig. 75

tion is carried between 200000 Hz and 203 500 Hz then, leaving a little margin o^oHz, the next conversation could be modulated on to a carrier at 204000 Hz, and the next carrier at 208000 Hz, and so on. The number of telephone channels clearly depends on the available bandwidth. If a transmission line can carry frequencies from o to 25 MHz then the number of telephone channels available is equal to

25000,000/4000 = 6250.

Clearly, going up to 50 MHz would provide 12500 simultaneous channels, and so on. Can this go on indefinitely? Can we have millions of channels? The answer is yes, but for that optical fibres are needed. They will be discussed in Chapter 12.

0 0

Post a comment