Jitter in the square synchroniser on direct detection optical communications. Part 1: theoretical analysis

Very high bit rates (multigigabit per second) are being increasingly used on direct detection optical systems. The influence of clock recovery circuit (CRC) inaccuracies is of special significance in these systems since a small timing error may represent a high fraction of the bit period. Accordingly, it is quite convenient to develop analytical results to compute the jitter arising in the most commonly employed CRC. In this paper, the performance of the square synchroniser in a direct detection optical receiver using the generalised Gaussian approach model for the photodetection and subsequent filtering is evaluated theoretically: a general expression for dynamic timing jitter valid for any type of photodetector (PIN or APD), optical pulse shape, pulse shape at the input to the squarer prefilter, narrow bandpass filter, noise type dominance and signal to noise ratio is presented. A first qualitative analysis of dynamic timing jitter is provided, stressing several differences between these systems and those with additive signal-independent noise