Numerical and experimental analysis of boundary fields in reverberation chambe

Abstract

Starting from the theory, this paper presents an analysis of the field properties in proximity of the reverberation chamber boundary. Three particular boundary geometries, a wall, a bend and a corner are analyzed, experimentally, measur- ing the three field components and calculating their statistics and variances, and numerically, with a previously assessed method using a summation of random plane waves and calculating fields and variances with FDTD formulas. I. INTRODUCTION Inside a well operating ideal Reverberation Chamber (RC) the electromagnetic field is uniform, isotropic and without a dominant polarization. These properties are well found in describing the chamber field by means of a plane wave integral representation (1). Inside a real chamber these characteristics are fulfilled inside a reduced space, called working volume, whose boundary is sufficiently far from the chamber walls, from the antennas, from the stirrers. This working volume is identified during the chamber calibration, according to stan- dardized procedures (2), and after defining the departure from the ideal behavior in terms of standard deviation. Therefore, the Device Under Test (DUT) must be placed inside this to carry out emission or immunity tests. In many applications, the DUT or the receiving antenna could be outside of this gold region, near the chamber walls. For example, transmission lines have been widely used to compare RC and Anechoic Chamber (AC) results, also to define the safety margin required to transfer the results from one test facility to another (3). Transmission lines were also used as a reference structure to compare different RCs (4) in order to facilitate the structure realization and its connection to the external instrumentation. These lines were realized directly placing a wire over the chamber floor, therefore in a boundary field condition. Another interesting application, where the chamber walls are the best candidates for the DUT insertion, is represented by the cable shielding effectiveness measurement. In this case, to allocate the Cable Under Test (CUT) inside the working volume, a high shielding performance cable must be used for the connection to the external instrument through the chamber wall (5): a direct connection of the CUT to the wall would facilitate the measurement, but it would be exposed to a non ideal field. Finally, consider the case of shielding ef- fectiveness measurement of material by the two reverberation chamber method (6): the sample under test is mounted in the wall that separates the two chambers and therefore exposed to a field that does not exhibit the ideal properties. The problem is partially overcame by using the nested reverberation chamber method (7), where a smaller RC is placed inside a greater one and the sample under test is mounted on one face of the smaller RC. When the size of the nested RC increases, the field uniformity in the outside chamber is more affected. The above mentioned practical situations justify the demand for a correct analysis of the field properties in proximity of the chamber boundary. In this sense, a paper is recently appeared that gives an analytical formulation for the boundary field properties in a reverberation chamber (8). The approach starts from the plane wave formulation of the chamber field and considers the field in proximity of the chamber walls, bends and corners. The deviation from the ideal behavior is predicted as function of the distance from the walls. Following this theoretical paper, the present paper experimentally analyzes the field in the boundary regions to give a confirmation to the theoretical results. In particular, a mechanical stirring mode chamber is analyzed, measuring the three field components and calculating their statistics and deviation. Moreover, also a numerical analysis of the field is carried out. It is based on a previously assessed method making use of a summation of random plane waves and numerically solved by a FDTD code (9). The method, initially designed to simulate an ideal chamber, is properly adapted to account for metallic walls in order to compute the field near walls, bends and corners. The statistics of the computed field are determined and compared to the experimental data.