The use of distributed Raman amplification in fibre spans can improve the optical signal-to-noise ratio (OSNR) and decrease the effects of nonlinearity in transmission systems. However, Rayleigh backscattering occurring within the transmission fibre degrades the performance of distributed Raman amplified systems in two ways: OSNR degradation arising from Rayleigh backscattering of the backward-propagating amplified spontaneous emission noise; and in-band crosstalk or equivalently, multi-path interference, caused by double-Rayleigh backscattering of the signal. The dependence of receiver sensitivity penalties on in-band crosstalk has been studied theoretically and experimentally. However, the direct measurement of in-baud crosstalk within a distributed Raman amplifier is difficult owing to the distributed characteristics of Rayleigh scattering and Raman gain. We have previously investigated the use of a self-homodyne technique in the measurement of double-Rayleigh backscattering noise within a distributed Raman amplifier t61. We found the technique to be well suited for the measurement of a wide range of values of the in-band crosstalk ratio, R{sub}c the ratio of the scattered power (with signal polarisation) to the average signal power. In this paper, we study the relationship between R{sub}c and system penalties. The crosstalk was generated by Roman gain together with either one or two reflections. The addition of a single reflection adds crosstalk caused by discrete-Rayleigh scattering to the double-Rayleigh scattering crosstalk. A third category of in-band crosstalk, discrete-discrete crosstalk, is added when a second reflection is added. The amount of crosstalk generated by each mechanism depends on system properties such as fibre attenuation, Raman gain coefficient, signal mode Si/C, connector reflectance, and fibre length. We find that R{sub}c provides a convenient means for predicting system penalties from discrete-Rayleigh and discrete-discrete crosstalk occurring under a variety of system parameters.
展开▼
