STATISTICAL FIBER-OPTIC CONNECTOR LOSS BUDGETS: PROBABILITY DISTRIBUTIONS AND CASCADED LOSSES

摘要

The log-normal was by far the best fit for cases where the mean was at or below 1μ. At mean values above la, but less than 1.8μ, the log-normal continued to be the best fit. The Rayleigh_σ exhibited a few good predictions, but not consistently. At mean values above l.Scr the log-normal was still the best fit. The Gaussian distribution also fit the data well with respect to shape, but the Gaussian density function peak was at a higher IL than the actual data.