An Automatic, Statistical-based Detection of Outliers in an Inter-laboratory Comparison of Nonlinear Measurements

This paper describes the comparison of calibrated nonlinear Radio-Frequency (RF) measurements performed on a nonlinear device by five different measurement laboratories. The device-under-test is a nonlinear active semiconductor device that is designed to generate a maximum number of harmonic tones. The goal is to obtain a simple, automated method that allows detecting if some laboratory had measurement problems during the measurement campaign. The developed comparison method is based on the Analysis of Variance (ANOVA) technique.