GENERATION AND ACCUMULATION OF TIMING NOISE IN PCM SYSTEMS - AN EXPERIMENTAL AND THEORETICAL STUDY

Three sources of timing noise in a self‐timed regenerative PCM repeater, namely, tank circuit mistuning, amplitude to phase conversion, and pulse shape, were studied both experimentally and theoretically. We discuss how these noises accumulate and combine along a chain of repeaters. The theoretical work is from the viewpoint of frequency analysis which leads easily to the spectrum of the timing noise. We first give a simple form of this theory applicable in a number of cases, and then a more general form useful in other cases, which shows the approximations and limitations of the simple theory. We found that the spectrum of timing noise caused by tank circuit mistuning has no energy at zero frequency and because of this fact, timing noise from this source does not build up indefinitely along a chain of repeaters but soon reaches a limit. On the other hand, the spectrum of timing noise caused by amplitude to phase conversion does have energy at zero frequency, thus, timing noise from this source increases indefinitely along a repeater chain. Some of the timing noise is attributable to pulse shape alone and in some cases may include a very low frequency part. This latter comes about through the small energy near the harmonics of the pulse rate in the tuned circuit response and the aliasing of this energy down to very low frequencies by the sampling process used in measuring the phase deviation or in generating the retiming pulses. © 1969 The Bell System Technical Journal