Equivalent source localization for stochastic electromagnetic fields

Modern electronic devices, consisting of complex digital and analog circuitry may unintentionally radiate electromagnetic interference. Noisy electromagnetic fields can be characterized by their correlation spectra. In this work, we present a method, based on field-field correlations, for localizing equivalent dipole sources, from electromagnetic field data obtained by near-field scanning. From data sampled in a plane above the device under test, we calculate auto- and cross-correlation functions, representing the second order moments of the random process describing the electromagnetic field. By subsequent forming of ensemble averages and fast Fourier transform, we obtain auto- and cross-correlation spectra. We present a method for localizing equivalent dipoles, resembling the actual source distribution of the device under test.