The effect of preamplifier compression on measured noise power ratios

Noise power ratio (NPR) measurements are becoming widely used to characterize components (particularly optical components) for use in hybrid fiber/coax networks. Choice of this measurement technique is based on the assumption that the composite signal resulting from actual subcarrier loading will exhibit statistical characteristics that tend toward Gaussian. However, care must be taken to ensure that any test signals used in making NPR measurements are also Gaussian in nature. This paper examines the errors that may be produced in an NPR measurement if preamplifier compression results in a test signal that is non-Gaussian. Analytical as well as empirical results are presented.