Investigation of Reverberation Chamber Measurements Through High-Power Goodness-of-Fit Tests

Abstract

This paper aims to improve the analysis of distribution functions of a rectangular component of the electric field (E <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</sub> ) and the power received in an overmoded reverberation chamber. All data and analysis were achieved in the Institute of Electronics and Telecommunications of Rennes (IETR) mode-stirred chamber. For the power received on a large antenna, tests are consistent with the exponential probability density function assumption. However, high-power goodness-of-fit tests modify the determination of the lowest frequency from which the ideal underlying theoretical distributions can be associated with measurements. For the electric field in an overmoded regime, a Weibull distribution is proposed to model E <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</sub> measurements, instead of the Rayleigh distribution hypothesis, which is rejected by statistical tests. Furthermore, Weibull distribution provides better agreement with standard deviation of samples. An additional experiment with a monopole-like antenna illustrates that the exponential distribution is rejected when the monopole is small with respect to the wavelength, but is accepted when the antenna length is roughly over lambda/4. Experimental results are provided by a large number of goodness-of-fit tests. The paper highlights that the use of adapted critical values is necessary for testing a distribution function whose parameters are estimated.