13-TH4 Statistical Principles and Trends in Mask Testing

摘要

In recent years mask testing added hit ratios, margins and auto-fit. In parallel, jitter and noise analysis produced statistical models identifying categories of impairments and allowing extrapolation of results to lower BER (bit error ratio) levels. Mathematically the models are presented as PDF eyes and BER contours. Tests expanded from measurements on simple acquired data to measurements on data generated by models built in statistical space. Mask testing on these models provides more accurate and repeatable results, possibly extrapolated to lower BER levels. This paper relates mask testing on these models with that ultimate figure of merit, the system BER as measured against a mask. It includes a new method that determines the BER of a mask from an eye diagram without access to the pattern.