Polarization dependent loss induced gain ripple statistics in erbium-doped fiber amplifiers

A Monte-Carlo model is used to simulate the wavelength-dependent polarization dependent loss (PDL) statistics and PDL induced gain-ripple statistics for a chain of concatenated optical elements in an erbium-doped fiber amplifier (EDFA). Simulations indicate that unlike polarization-mode dispersion (PMD), the quadrature sum of PDL overestimates the mean of the PDL distribution. We show that EDFA total PDL variance scales linearly with the individual component PDL variance. Moreover, the relative impact of PDL on the gain ripple is reduced in the presence of the PDL independent gain ripple in the EDFA.