Abstract
Predicting the propagation of electromagnetic (EM) emissions by using a method based on the Wigner transformation requires source data from near-field measurement. The near-field data can be recorded in a time domain via a two-probe measurement system. Before the measurement is performed, the chosen device under test must be investigated to understand how the programs running in the background could affect the characteristic of the EM emissions. Three different programs, namely local number crunching, filling random numbers into sequential addresses and filling random numbers into random addresses, are tested and analysed. Time-frequency plot and autocorrelation function analysis are used as tools to study the effect of the device's mode of operations on the characteristic of the radiated emissions. Results from the time-frequency plot analysis prove that different programs can produce EM emissions of different characters and, at the same time, affect the ensemble average of the autocorrelation function.