Insertion-loss Measurement Accuracy for Fiber-optic Components – an Analysis

This paper attempts to provide an answer to the question: To what accuracy can insertion-loss measurements be made on low-loss, multi-mode fiber-optic cables, patch cords, and modules having a variety of terminations? To answer this question, and to provide insight regarding the factors that dominate the measurement accuracy, a thorough analysis of the measurement process is required. Relevant terminology is defined for purposes of this analysis to minimize confusion. In addition, to limit the discussion to those factors affecting the question at hand, a list of underlying assumptions is provided. An analysis of a measurement system using custom-built power measurement hardware showed an overall insertion-loss measurement accuracy ranging from ± 0.18 to ± 0.32 dB, depending on the measurement system configuration and the calibration procedure. Two separate factors contribute to the overall accuracy: an instrument measurement accuracy of about ± 0.092 dB, and a calibration accuracy which ranges from ± 0.15 dB to ± 0.30 dB. The analysis included the effects of two connector types having maximum (manufacturer specified) insertion losses of ± 0.25 to ± 0.30 dB. While power measurement factors (such as wavelength variability, amplifier gain variability, receiver nonlinearity, and finite A/D converter resolution) contribute to the overall measurement accuracy, calibration uncertainty due to connector loss variability dominates the overall measurement accuracy and can be two to three times larger than these factors. An analysis of a measurement system composed of commercial optical power measurement equipment, fiber-optic switches, and LED sources showed an overall insertion-loss measurement accuracy ranging from ± 0.33 to ± 0.59 dB, depending on the measurement system configuration, the calibration procedure, and power meter measurement accuracy. The instrument measurement accuracy ranged from about ± 0.13 dB to ± 0.35 dB, depending on the power meter. The calibration accuracy again ranged from ± 0.15 to ± 0.30 dB, as the same connector types were used in this analysis, as before. For systems using the more accurate (± 0.10 dB) power measurement equipment, the overall accuracy was dominated by the connector loss variability. However for systems using the less accurate (± 0.25 dB) power measurement equipment the overall insertion-loss measurement accuracy was dominated by the power measurement system accuracy.