Optimized sequential sampling algorithm for EMI near-field scanning

In this paper, a recently proposed automatic and sequential sampling and modeling algorithm for near-field scanning of printed circuit boards and/or integrated circuits is further optimized in two ways. The main goal of this optimization is to reduce the total measurement time needed to come to a complete model of the near-field distribution over the full scan area. The first optimization is that the new adaptive sampling algorithm in every step proposes a batch of N locations in which to measure the near-field while the original algorithm gave in every step exactly one location. The second optimization is that one tries to minimize the total distance that one needs to go through these N locations. These two optimization result in a significant reduction of the total measurements because the time needed to move the probe around is reduced and because much less models of the near-field distribution have to made. The performance and effectiveness of the proposed sampling algorithm is tested in practice on a simple printed circuit board. The time reduction that is achieved when increasing the batch size N is discussed. In the end a trade-off needs to be made. Choosing the bath size N larger means a reduction of the measurement time for the same total number of sampling points. However, a larger batch size N also means that for the same quality of the near-field scan, a larger total number of samples is needed.