TIA TIA-455-198

Abstract

This procedure can be applied to any single-mode passive component, including connectors, splices, couplers, attenuators, isolators, switches, multiplexers and demultiplexers, optical amplifiers (non-operating), circulators and filters. It is used to measure the total range of insertion loss as PDL due to changes in polarization of the launch state. For branching devices, it can also be used to measure the total range of coupling ratio. It cannot be used to measure polarization-maintaining components or to measure the polarization dependence of return loss nor is this method applicable to measuring the polarization dependence of higher order attributes such as center wavelength and bandwidth of filters.. This procedure could be used to measure the polarization dependence of non-diagonal transmission coefficients, e.g. wavelength isolation and directivity. <br>This method differs from that described in FOTP-157, which is based on manipulation of the state of polarization of light either continuously or in small increments in order to measure maxima and minima of the attenuation of transmitted light. This method involves the measurement of the behavior of Specimen when illuminated by a small set of well-defined states of polarization of input light. These measurements are followed by a matrix calculation to determine the polarization dependent loss (PDL) of the Specimen. It may be considered an efficient alternative to FOTP-157 when a result is required from a small, discrete number of measurements, such as in automated testing. <br>Generally, PDL techniques based on matrix methods fall into two categories, i.e. those based on Mueller calculus (hereafter referred to as Mueller/Stokes methods) and those based on the Jones calculus. While the two techniques are mathematically equivalent for completely polarized light, only the Mueller/Stokes method is applicable in the case of partially polarized light and is therefore to be considered to be more general. In addition, methods based on the Jones matrix generally require known polarization states on both input and output while the Mueller/Stokes methods require only known input states. Precise prior characterization of the input states is required in either case.