Vojnotehnički Glasnik (Feb 2014)
Analysis and interpretation of the model of a Faraday cage for electromagnetic compatibility testing
Abstract
In order to improve the work of the Laboratory for Electromagnetic Compatibility Testing in the Technical Test Center (TTC), we investigated the influence of the Faraday cage on measurement results. The primary goal of this study is the simulation of the fields in the cage, especially around resonant frequencies, in order to be able to predict results of measurements of devices under test in the anechoic chamber or in any other environment. We developed simulation (computer) models of the cage step by step, by using the Wipl-D program and by comparing the numerical results with measurements as well as by resolving difficulties due to the complex structure and imperfections of the cage. The subject of this paper is to present these simulation models and the corresponding results of the computations and measurements. Construction of the cage The cage is made of steel plates with the dimensions 1.25 m x 2.5 m. The base of the cage is a square; the footprint interior dimensions are 3.76 m x 3.76 m, and the height is 2.5 m. The cage ceiling is lowered by plasticized aluminum strips. The strips are loosely attached to the carriers which are screwed to the ceiling. The cage has four ventilation openings (two on the ceiling and two on one wall), made of honeycomb waveguide holes. In one corner of the cage, there is a single door with springs made of beryllium bronze. For frequencies of a few tens of MHz, the skin effect is fully developed in the cage walls. By measuring the input impedance of the wire line parallel to a wall of the cage, we calculated the surface losses of the cage plates. In addition, we used a magnetic probe to detect shield discontinuities. We generated a strong current at a frequency of 106 kHz outside the cage and measured the magnetic field inside the cage at the places of cage shield discontinuities. In this paper, we showed the influence of these places on the measurement results, especially on the qualitative and quantitative changes of the cage resonant frequencies. Model of the cage On the basis of the testing, the initial simulation model of the cage in the Wipl-D program consists of plates with losses, including losses at plate junctions, wall ventilation openings, aluminum profiles, profile carriers, and screws. The analysis showed that this model is not satisfactory. To obtain a more accurate model of the cage, we need the additional and more precise model of the lowered ceiling. That would complicate the model and require a lot of additional testing. For obtaining a simpler, but more reliable simulation model, we dismounted the aluminum profiles from the carriers. Thereafter, the agreement between the simulation and the experimental results was better than for the initial model, but still is not good enough, due to the uncertainties caused by the strips carriers and the screws. Final model of the cage To avoid the uncertainty due to the strips carriers, we removed the carriers and the screws from the cage. A very good matching of the simulation and the experimental results was achieved. We showed that the relative difference of the resonant frequencies between simulation and measurement is less than 1 %. However, the relative difference of resonant frequency between simulation or measurement and theoretical results is also less than 1 %, except for TE101 and TE011 modes, for which the relative difference is under 5 %. We explained the exception by the influence of the places of the cage shield discontinuities on the measurement results. Conclusion The results of the simulations and the measurements for the final model show good agreement, both qualitatively and quantitatively. We showed that the resonant frequencies in the simulations and the measurement results have same positions and depths. Also, the relative differences of the resonant frequencies are less than 1 %. As future work, we will provide a technique for the identification of equivalent sources that represent the device under test. This would enable a characterization of the tested device in an arbitrary electromagnetic environment and allow a comparison of results obtained in different laboratories.
Keywords
