Near field program-dependent EMI measurement and data reduction for IOMarking method

IOMarking method can characterize program-dependent conducted EMI (cEMI). This paper firstly proposes a direct extension for program-dependent near-field EMI (rEMI) behaviors under a complex HW/SW environment. For the program-dependent EMI analysis by IOMarking, sample rate is a major parameter that affects the bandwidth. Due to high sample rate and long program execution time, large amounts of sample points are acquired for the analysis. The second contribution of this paper is to overcome the difficulty by finding a size-reduction general rule. In this paper, we find a partial IOMarking data for EMI characterizing can be done without too much reducing the analysis accuracy. This means the processing and storage efficiency are greatly improved. About the reduction percentages, our case study shows that, for a high speed data acquisition system (1.5G sample per sec), the IOMarking EMI characterization can be reduced up to 30% of the total data size without too much affecting the EMI analysis accuracy. When the reduced size is below 20%, the analysis still remains but with less accuracy.