Performance Estimation of PolMux-QPSK Modulated Signals by Modeling Phase Variations

In this thesis the behaviour of phase variances in a polarization multiplexed 45.8 Gb/s QPSK system was analysed. The influence of the number of spans, the fibre type, the launch power, the span length, the dispersion compensation scheme and the number of channels were investigated. Tools were developed to extract the phase information from the optical and the electrical signal and to analyse the influence of the different link parameters.

The results show that the optical phase jitter increases continuously with growing numbers of spans. fibres with low dispersion like TWC and LEAF exhibit much stronger phase jitter compared to SSMF fibres. I-line dispersion compensation generally results in higher phase jitter. Especially the combination of a nine channel WDM-system with DCM management leads to strong phase jitter, because the optical bit patterns from neighbour channels cannot walk-off from each other. The distribution of the bits in the optical constellation diagram is not always Gaussian, this is due to an interaction between nonlinear effects and dispersion.

Analysis of the receiver characteristics showed that the electrical phase jitter is in general lower than the optical phase jitter, because of signal processing in the receiver. For large optical phase jitters the receiver is not able to recover the bits correctly, which leads to an increased bit-error ratio. The receiver has an intrusive noise level due to signal processing.

The simulation results were used to derive simple rules that can be used to predict the system performance based on the link parameters and amplified spontaneous emission. The predictions give good results in reasonable power regions. The agreement of the results with recent paper publications was demonstrated. If the power is too high the estimation deviates more from simulation results, because not all fiber parameters were included in the estimation.


Future works might overcome these limitations. The effect of the number of channels, the dispersion and the interactions between dispersion and nonlinear effects has to be studied in more detail. The performance prediction in this thesis is only based on the estimation of the phase jitter. Next steps have to incorporate the Kerr-induced amplitude jitter to achieve higher accuracy.