Jitter Sampling of Deterministic Signals and Noise

Abstract

In the implementation of any digital signal processing system, noise can be introduced due to hardware limitations. Some examples of noise are aliasing and amplitude quantization noise. Another noise source that is often ignored is the result of jitter, or random fluctuations of the sample period. Since clock jitter is present in almost all oscillators, a digital signal processing system rarely has perfectly timed samples. In this paper, an approximate autocorrelation function of the noise introduced by jitter sampling a deterministic function is derived. The results are applied to the specific case when the sampled function is periodic. In addition, closed-form expressions for the signal-to-noise ratio of the jittered samples are obtained. These expressions can be used to determine how stable the system clock has to be to reduce jitter noise to acceptable levels. Computer simulation was used to check the validity of the results.