Modeling and Measurement of Shielding Enclosures

The shielding effectiveness (SE) of an enclosure has been simulated with CST MICROWAVE STUDIO® (CST MWS) and compared to measurement from a known reference case equipment. For the SE measurements and simulations the source was a 28-cm diameter loop (Fig. 1) fed from a shielded ACMOS digital gate rated at 8MHz clock. An essential feature of a source is that it covers an adequate bandwidth to excite the enclosure over the frequency range of interest. Furthermore, the source must be of sufficient size to adequately excite the enclosure and to have a variable polarisation and location to excite all possible cavity modes. Figure 1 also shows the simulation setup in CST MWS.

The radiating source was placed inside a shielded enclosure used for a switching telecommunication application (see Fig. 2). The structure was simulated taking only the main walls into account. The doors of the shielded rack were simulated with a perfect metal rectangular plate.  The narrow slots, which were created by introducing some sheet of paper between the horizontal door sides in order to cancel the gasket benefits, were simulated with rectangular slots of size: 2 mm x 50 cm and depth of 1.5 mm. The reflecting plane of the semi-anechoic chamber was accounted for using the condition of a metallic wall (Et=0).

Figure 3 compares measured and simulated values in the case of shielded enclosure with two narrow slots. The measurements were performed at two different periods in time on different enclosures of the same type and reported different values (continuous: old measurement and spotted lines: new measurement). The difference between the two measurements is probably due to the uncertainty of the EMC measurements and some slight differences in the mechanical structure such as joints, soldering, surface and gasket pressure.

Despite the fact that the shielded enclosure was simulated neglecting many details, the measured and calculated emission profiles shown in Figure 3 are in good agreement. Calculated values at some frequencies are a little higher than measured ones. The apertures were not simulated with the exact same size of a narrow slot caused by sheet of papers introduced between the horizontal gaskets of the doors. Furthermore, the simulation was done without the metallic losses which leads to a higher Q of the enclosure. Less than 4 hours were needed to simulate the shielded enclosure in the described setup on a Pentium II 233 MHz, 128 MB RAM.