Crosstalk in amplitude-modulated time-division-multiplex systems

Crosstalk arises in a time-division-multiplex system using amplitude modulation when some of the energy proper to one channel arrives during the time allocated to another channel. Networks inserted in the transmission path will almost always introduce crosstalk, the more so as their bandwidths are increasingly restricted. Analyses of the crosstalk produced by those networks of amplifiers most commonly met with in the transmission path show that no one method of analysis is satisfactory for all networks. Coupling and decoupling networks, producing attenuation distortion at low frequencies, are conveniently tackled by Fourier methods. The use of both networks simultaneously can result in a reduction of crosstalk. The simplest anode loads and those improved by inductance compensation are much better treated by the operational calculus. Experimental results confirm the analyses. Cable problems can also be analysed by operational methods. Because practical low-pass filters introduce serious difficulties in analysis most of the data presented for them are experimental. The oscillatory response found with filters brings new problems. The crosstalk depends on the shape of the transmitted pulse and the duration of the demodulating pulse. Practical systems show little advantage from the use of pulse shapes other than rectangular when the number of networks in the transmission path which distort the pulse shape is small; when the number is large, however, the improvement can be worth-while. The demodulating pulse should not normally be much wider than the transmitted pulse; if it is narrower there may be an increase in crosstalk attenuation.