Eye diagram reconstruction using asynchronous imperfect sampling, application to ber estimation for fiber-optic communication systems

In this paper we present a new and robust method to construct the eye diagram from asynchronous samples of a digital communication signal. No a priori knowledge of the bit period is needed. The method uses an approach based on periodogram estimation which is suitable even for (highly) under-sampled signals. Random shifts due to clock errors are also being corrected. We apply this method to the Bit Error Rate estimation of an optical signal in an experimental setup.

1 Introduction

The main motivation for this work comes from a problem related to the monitoring of optical networks. In next generations of such networks, optical technology will be more and more used not only for transmission but also for switching (in replacement of present electrical cross-connects) of signals. The objective is to alleviate the bottlenecks due to capacity and cost of electronic solutions and to provide a versatile “transpar-ent” (without opto-electronic conversions) optical network able to carry client signals independently of the various formats of their electrical frame. However, the capacity to monitor the quality of the signal along its path is a required feature to build a manageable net-work. It is clear that transparent optical networks can exist only if transparent monitoring methods are developed. This is the aim of the technique presented in this paper which successfully measures the main quality indicator of a digital communication system, namely the Bit Error Rate (BER), without accessing neither the electrical frame of the corresponding signal nor even knowing its bit rate.