Fast and Accurate Near-Field Measurement Method Using Sequential Spatial Adaptive Sampling (SSAS) Algorithm

Abstract

This paper presents an optimization of the Sequential Spatial Adaptive Sampling (SSAS) algorithm to accelerate nearfield scanning of printed circuit boards or integrated circuits. The first originality of this approach is to configure this algorithm from the spatial distribution characteristic of near field on a planar surface. The second originality is to propose three selection criteria to adjust the targeted accuracy with the objective to reduce the measurement time. The low time-consuming algorithm selects only measurement points, which carry the most information, in order to reduce significantly the number of captured points without increasing the measurement error. The analysis of measurement results, based on this algorithm applied on two case studies, validates its effectiveness compared to the classical full regular grid sampling.

Efficiency of Sequential Spatial Adaptive Sampling Algorithm to Accelerate Multifrequency Near-Field Scanning Measurement

“…This section gives the main principles of the SSAS algorithm used in this paper. A detailed description is provided in [1]. The method was developed and validated for NFS measurements done at only one frequency.…”

“…In order to limit the total measurement time, the purpose of this algorithm is to capture a limited number of positions in Ω allowing a complete reconstruction of the magnitude distribution of F in Ω. The selection criterion of SSAS algorithm selects these positions to ensure a reconstruction of F without any significant loss of information [1].…”

“…In the context of investigation activities, measuring only one field component (Hz) can be sufficient to identify the root cause of an EMC related issue. A complete characterization of the near-field emission above a complex printed circuit board (PCB) in a large frequency range may last several tens of hours [1]. Moreover, in the context of modeling activities, the minimization of the number of points required to S. Serpaud, F Coccetti are with the IRT Saint-Exupéry institute, Toulouse, France (e-mail: sebastien.serpaud@irt-saintexupery.com, Fabio.coccetti@irtsaintexupery.com) S. Serpaud, S. Ben-Dhia and A. Boyer are with LAAS, Univ.…”

“…However, the massive use of Kriging model and the random sampling lead to a significant increase of the time due to the algorithm computation and the probe movements. The Sequential Spatial Adaptive Sampling (SSAS), presented in [1] and [9], solves these limitations, which reduces by a factor 5 to 13 the total measurement time compared to the full sampling approach. Compared to other method, the SSAS algorithm is very low time consuming and is initialized from two set-up parameters (probe dimension and scans height).…”

“…EMC problems have often to be analyzed at a wide frequency range to be efficient. The purpose of this paper is to extend the SSAS algorithm presented in [1] to the multifrequency near-field scanning and determine the configuration that optimizes the measurement time reduction. Two strategies for applying the SSAS algorithm are discussed and completed by an evaluation of four frequency-segmentation approaches.…”