Correlation and Spectral Density Functions in Mode-Stirred Reverberation -- III. Measurements

Abstract

Experimental auto- and cross-correlation functions and their corresponding spectral density functions are extracted from measured sweep data of mode-stirred fields. These are compared with theoretical models derived in part I, using estimated spectral moments from part II. The second-order Pad\'{e} approximant based model accounts for the main features of the spectral density function, including its slope near stir DC, corner frequency, stir DC-to-Nyquist level drop, and asymptotic spectral density. Ensemble averaging across secondary tune states offers a reduction of spectral bias and RMS spectral fluctuation, compared to spectral densities for individual stir sweeps or their concatenation. Periodogram- and correlation-based methods produce near-identical results. Distinctive theoretical features between power-based vs. field-based spectral densities are experimentally verified. Interchanging the roles of stirrer and tuner demonstrates the effect of stir efficiency on correlation and spectral density. The spectral characterization allows for stirrer diagnostics, which is demonstrated through detection and identification of EMI caused by mains power harmonics in the measured stir spectrum at low frequencies.