SHIELDING EFFECTIVENESS MEASUREMENTS ON ENCLOSURES WITH VARIOUS APERTURES BY BOTH MODE-TUNED REVERBERATION CHAMBER AND GTEM CELL METHODOLOGIES

Transcription

1 Progress In Electromagnetics Research B, Vol. 2, , 2008 SHIELDING EFFECTIVENESS MEASUREMENTS ON ENCLOSURES WITH VARIOUS APERTURES BY BOTH MODE-TUNED REVERBERATION CHAMBER AND GTEM CELL METHODOLOGIES C. H. Fang, S. Q. Zheng, H. Tan, D. G. Xie, and Q. Zhang The National Key Laboratory of EMC China Ship Development & Design Center Wuhan , China Abstract The shielding effectiveness (SE) measurements on enclosures with various apertures by both reverberation chamber (RC) and GTEM cell (GC) methodologies are presented. Furthermore, computer simulation through commercial software (CST) is used to compare with the experimental data. Our results show that the SE of the rectangular enclosure with five different apertures from RC is usually larger than those from GC. The reason for this is related to the random polarization and various angles of incidence within the RC, where various polarization and incident angles are averagely and randomly coupled into the EUT. However, as for the GC, the normal incidence with vertical polarization usually corresponds to the worse case, which leads to the worse SE of the enclosure. Surprisingly, an abnormal phenomenon is found from the enclosure with nine periodic rectangles. According to Charles F Bunting s work, moreover, it is still reasonable because many rectangular apertures on the enclosure may lead to a fact that the coupled energy of oblique incidence may be larger than that of normal incidence. Finally, we obtain conclusions about the use of two experimental methodologies in SE measurement. 1. INTRODUCTION Shielding enclosures are employed to protect against radiation from both external fields and electromagnetic leakage from electronic devices, and hence meet the electromagnetic compatibility (EMC) requirements [1 7]. However, the integrity of these enclosures is often compromised by all kinds of apertures that interconnect various components, subassemblies, equipment, and subsystems. Such

2 104 Fang et al. openings allow exterior electric and magnetic fields to penetrate to the interior space, where they may couple onto printed circuit boards (PCBs) thus inducing currents and voltages on interior conductors. Therefore, it is very important and necessary to know the EM shielding effectiveness of shielding enclosures in the presence of these apertures. The SE can be obtained through numerical modeling or through measurements [8 12]. When the enclosures are with many complicated apertures, the ability of numerical models is not equal to their ambition at a certain extent and measurement techniques are used instead. The methodologies measuring the SE of the enclosures with various apertures have been the subject of a great deal of research and development for some time. The shielding industry in the area of enclosures has proposed a myriad of measurement techniques and procedures, hardware and software designs, literature descriptions, test standards and guidelines, and practices [13]. These are inclusive of an open area test site (OATS), anechoic (or semi anechoic) chamber, absorber-lined OATS, TEM cell, gigahertz TEM (GTEM) cell, and other shielding rooms, among which the mode-tuned reverberation chamber (RC) methodology and the GTEM cell ((GC) methodology are regarded as the most convenient and promising methods [14, 15] to measure the SE of enclosures over a wide frequency range. In this work, the shielding effectiveness of the enclosure with five kinds of apertures is measured by the mode-tuned RC and GTEM cell (GC) methodologies. Furthermore, we compare the two experimental results with the simulated results from CST (the two commercial software), and discuss the reason of the difference between them. 2. DEFINITIONS OF SHIELDING EFFECTIVENESS The common definitions of the electric and magnetic shielding effectiveness at an arbitrary point q within the shielded domain are given by [16] SE e = 20 log E unshield E shield db (1) SE m = 20 log H unshield H shield db (2) respectively. The numerators in (1) and (2) represent the amplitudes of the time-harmonic electric and magnetic field intensities, measured at q in the absence of the shield, while the denominators contain their values in the shielded case at the same locations. The advantage

3 Progress In Electromagnetics Research B, Vol. 2, of these definitions is that they are relatively easy to realize; in the meanwhile, they are intended mainly for low-frequency electric and magnetic fields, while the dimensions of the shield enclosures are less to the wave-length. 3. MEASUREMENT METHODOLOGIES AND RELATED SIMULATION In order to find the worst possible interference in a shielded chamber, it is necessary to rotate the EUT as the chamber contains a well-defined and inhomogeneous EM field. But this rotation is often impossible. Therefore a uniform environment must be reversely created within the chamber, which could decrease the dependence of the EUT on its location and orientation. Such a uniform EM environment can be created by rotating one or more conductive tuners (reflectors or stirrers) in the chamber. This kind of chamber is known as a modetuned (stepped rotating) or mode-stirred (continuous rotating) RC. The mode-tuned RC is per se an electrically large metallic cavity with high quality-factor, whose boundary conditions are continuously and randomly perturbed by the rotating conductive tuners. When sufficient modes are excited, the time-averaged fields (or power density) inside such a chamber are approximately equal in amplitude. The field distribution at each location in the working volume of the RC is assumed to be a composite of statistically isotropic, randomly polarized and uniformly homogenous plane waves. Measuring the SE by the use of the mode-tuned RC is a timeefficient and cost effective methodology over broad frequency spectra. The EUT may be placed at any convenient position or orientation within the working volume of the RC, where the field uniformity is guaranteed. A photograph of the mode-tuned RC for the SE measurements is shown in Figure 1. The mode-tuned RC, whose dimensions are m, is essentially a lossless shielded aluminum enclosure that includes an input antenna (log periodic antenna), a calibration antenna (isotropic electric-field probe), two orthogonally orientated conductive tuners, and controlling and testing accessories outside the chamber. The antennas are placed in the chamber such a way as to minimize direct coupling from the input antenna to both the calibration antenna and the EUT. To measure SE e, a monopole with a 1 mm diameter was mounted at the bottom of the enclosure. It was coupled via a cable link to the port of the frequency analyzer. Figure 2 shows the measurement set-up by the GC, where the EUT is horizontally put on an insulating table. A fiber is used to

4 106 Fang et al. Figure1. Photograph of a mode-tuned RC for the SE measurements of an enclosure. Figure2. Measurement set-up for SE tests of an enclosure using a GC. link the EUT and the fiber optic transmitter to make sure the EUT is fully saturated within the working volume of the GC, where an appropriate plane wave is excited. Figure 3shows the incident and polarized direction of incident wave in GC. Furthermore, the incident conditions of simulation are similar to that in GC, which usually mean the worst case.

5 Progress In Electromagnetics Research B, Vol. 2, Figure3. Measurement set-up for SE tests of an enclosure using a GC. The enclosure is constructed of aluminum plates. All of the side and top joints were soldered together and then covered on the outside with aluminum tape. Because more than one aperture configuration was intended, the front wall was constructed with interchangeable panels containing the apertures. This is illustrated in Figure 4. Besides, the enclosure size was taken as cm, the aperture size is varied, and the shielding effectiveness at the center of the enclosure is considered. Therefore, the first resonance of the enclosure is 606 MHz, and the higher resonance frequency of the enclosure is 958 MHz. Figure 4(a) presents the enclosures with an open side. Here, we consider the worst case of having the aperture to be as big as the wall on one side of the enclosure. Figures 4(b) and 4(c) presents the enclosures with fractal hole and fractal square apertures, respectively. As for the enclosures with fractal hole, there is a large hole within a radius of 3cm in the center of the front side while eight little holes within a radius of 1 cm are regularly around the large one. In the meanwhile, the distance between the center of the large hole and any little one is 5 cm. Similarly, as for the enclosures with fractal square apertures, there is a large square aperture within side length of 5 cm in the center of the front side while eight little squares within side length of 1 cm are regularly around the large one. In addition, the similar distances are 4.25 cm and 6 cm. Please note the size of every round hole is the same as that of the corresponding square aperture. Figure 4(d) presents the enclosure with eight cross slots in a side, while every cross slot is constitute vertically of two same narrow slots (5 0.3cm). The distances along x axis and z axis between the centers of two nearby cross slots are 7 cm. Figure 4(e) presents the enclosures with nine rectangular apertures in a side. The distances along x axis and z axis between the centers of two nearby rectangular apertures are 10 cm and 4 cm respectively. Figure 5 presents the modeling scenes of simulation in CST. Please note that the polarized direction of incident

6 108 Fang et al. (a) Opening at a side (b) Fractal hole (c) Fractal square (d) Periodic cross slot (e) Periodic rectangle Figure4. Photographs of five enclosures with various apertures. wave is vertical to the long side of enclosure. Furthermore, the grid size was made smaller from λ/20 (λ is the wavelength) to λ/25 in CST. 4. RESULTS AND ANALYSIS The SE results of the enclosure are derived by using Equation (1) within the frequency range of interest. The measured SE results of the enclosure under test are shown in Figure 6 to Figure 10, which are inclusive of five different apertures. The later four cases exhibit higher shielding effectiveness by about 30 db compared to that with the first case, since the later four apertures prevent the penetration of much coupling energy. The variation of the SE data in later four cases are

7 Progress In Electromagnetics Research B, Vol. 2, Figure5. around 20 db over the frequency range of GHz.It is also noted that the SE decreases significantly with the increase of frequency in the main. From Figure 6 to Figure 10, the results of simulation (CST) are in good agreement with the measured data from GC and have some difference with those from RC, except three frequency bands (0 200 MHz around 500 MHz, around 930 MHz). There may be two reasons about the disagreement. Firstly, as for SE at MHz, the field strength in the shielded case is too small, it is hard to be measured accurately due to the noise and accuracy limit of measurement equipment. Secondly, as for SE around 500MHz and 930 MHz, all results show that the enclosure resonates. Certainly, the resonance around 500 MHz is not enough clear, however, we can also see that there is a minimal value of SE around it. Moreover, according to the reference [17], there is a phenomenon of resonant frequency Figure6. The SE of the enclosure with opening at a side comparison by using two different measurement methodologies and simulation.

8 110 Fang et al. Figure7. The SE of the enclosure with fractal hole comparison by using two different measurement methodologies and simulation. Figure8. The SE of the enclosure with fractal square comparison by using two different measurement methodologies and simulation. offset in results due to the larger area of apertures. Therefore, the enclosure resonates around 500 MHz and 930 MHz, rather than the first and second frequencies (606 MHz and 958 MHz). By rule and line, from Figure 6 to Figure 9, some higher SE results are observed in the RC methodology compared to the GC methodology, which implies the worse scenario for the GC. The reason for this is related to the random polarization and various angles of incidence within the RC, where various polarization and angles incidences are averagely and randomly couple into the EUT. However, as for the GC, the normal incidence with vertical polarization usually corresponds to the worse case, which means the worse SE of the enclosure. Therefore it is reasonable that

9 Progress In Electromagnetics Research B, Vol. 2, Figure9. The SE of the enclosure with periodic cross slot comparison by using two different measurement methodologies and simulation. Figure10. The SE of the enclosure with periodic rectangle comparison by using two different measurement methodologies and simulation. the results of RC are usually larger than those of GC. Figure 7 and 8 show the results corresponding the fractal hole and fractal square apertures respectively. Obviously, both of them are very similar to each other in a whole frequency band. It is easy to understand that Turner have found that the shielding effectiveness of a round hole is approximately the same as that of a square aperture of the same size [18]. To our surprise, however, Figure 10 shows as though some abnormal phenomena. From GHz, the results of RC are less than those of GC. But if we consider the special configure of this

10 112 Fang et al. enclosure, the results is still reasonable. According to Charles F Bunting s work [19], the most common assumption of maximum field penetration for normal incidence is valid only for a single aperture on cavity, when number of apertures increases, the interaction between the radiation patterns of different apertures may result in larger enhancement of the fields for incidence angles other than the normal incidence. His work shows that when the rectangular enclosure has four rectangular apertures in a side, the SE corresponding to normal incidence is larger than those of oblique incidence in three frequencies, while the higher frequency may lead to the larger difference. As for the Figure 8 corresponds to the enclosure with nine rectangular apertures, this abnormal phenomenon will be more prominent which leads to the results after average in RC may be larger than those in GC. It is desirable to notice that the periodic cross slot has not appeared the similar effect. Actually, this compare shows that the presence of abnormal phenomena of shielding effectiveness of the enclosure greatly is affected by reducing the aperture size, as the electric field penetrate obliquely through the enclosure is preferably prohibited because of the smaller aperture width. 5. CONCLUSIONS The shielding effectiveness measurements for the rectangular enclosures with five different apertures were used by employing both the RC and GC methodologies. The CST was used to simulate the corresponding cases. Finally, the results indicated that the results of GC agree with the data of CST, which have some difference with those of RC. The SE results measured by the GC exhibited larger dynamic range compared to those measured by the reverberation chamber. Moreover, the SE results by the reverberation chamber were generally larger than those by the GC, which is mainly due to the random polarization and statistical average inside the RC. Surprisingly, the results of the enclosure with nine rectangular apertures from RC and GC show the abnormal phenomena. With Charles F Bunting s work, we can give a reasonable explaination about it. Because many apertures may lead to a fact that the coupled energy of oblique incidence is larger than that of normal incidence. However, the use of the mode-tuned RC for evaluating the shielding performance of the enclosure is still an important test methodology, one of whose advantages is the simulation of complicated electromagnetic environment. Therefore, in our view, as for simulating a complicated electromagnetic environment, using RC to evaluate the SE may be better; as for simulating a special incident case, especially

11 Progress In Electromagnetics Research B, Vol. 2, for the worst case, GC will be equal to. Of course, it should be note that the usual assumption about the normal incidence being the worst case scenario for SE may not be valid when there are more than one apertures in the cavity. The above conclusions will, therefore, be of use to testers or designers of shielded enclosures. ACKNOWLEDGMENT The authors would like to thank the reviewers and editors for their useful advice and help. REFERENCES 1. Robinson, M. P., T. M. Benson, C. Christopoulos, M. D. Ganley, A. C. Marvin, S. J. Porter, and D. W. P. Thomas, Analytical formulation for the shielding effectiveness of enclosures with apertures, IEEE Trans. Electromagn. Compat., Vol. 40, No. 3, , Aug Klinkenbusch, L., On the shielding effectiveness of enclosures, IEEE Trans. Electromagn. Compat., Vol. 47, No. 3, , Aug Ho, M., F.-S. Lai, S.-W. Tan, and P.-W. Chen, Nunerical simulation of the propagation of EM pulse through lossless ono-uniform dielectric slab using characteristic-based method, Progress In Electromagnetics Research, PIER 81, , Ansarizadeh, M. and A. Ghorbani, R. A. Abd-Alhameed, An approach to equivalent circuit modeling of rectangular microstrip antennas, Progress In Electromagnetics Research B, Vol. 8, 77 86, Robertson, J., E. A. Parker, B. Sanz-Izquierdo, and J. C. Batchelor, Electromagnetic coupling through arbitrary apertures in parallel conducting planes, Progress In Electromagnetics Research B, Vol. 8, 29 42, Hattori, H. T., Fractal-like square lattices of air holes, Progress In Electromagnetics Research Letters, Vol. 4, 9 16, Maab, H. and Q. A. Naqvi, Fractional rectangular cavity resonator, Progress In Electromagnetics Research B, Vol. 9, 69 82, Wang, Y. J., W. J. Koh, and C. K. Lee, Coupling cross section and shielding effectiveness measurements on a coaxial cable by both mode-tuned reverberation chamber and Gtem cell

12 114 Fang et al. methodologies, Progress In Electromagnetics Research, PIER 47, 61 73, Nie, X. C., N. Yuan, L. W. Li, and Y. B. Gan, Accurate modeling of monopole antennas in shielded enclosures with apertures, Progress In Electromagnetics Research, PIER 79, , Hussein, K. F. A., Effect of internal resonance on the radar cross section and shield effectiveness of a open spherical enclosures, Progress In Electromagnetics Research, PIER 70, , Lei, J.-Z., C.-H. Liang, and Y. Zhang., Study on shielding effectiveness of metallic cavities with apertures by combining parallel FDTD method with windowing technique, Progress In Electromagnetics Research, PIER 74, , Bahadorzadeh, M. and M. N. Moghaddasi, Improving the shielding effectiveness of a rectangular metallic enclosure with aperture by using extra shielding wall, Progress In Electromagnetics Research Letters, Vol. 1, 45 50, Tsaliovich, A., Electromagnetic Shielding Handbook for Wired and Wireless EMC Applications, Kluwer Academic Publishers, Jesch, R. L., Measurement of shielding effectiveness of cable and shielding configuration bymo de-stirred techniques, IEEE Trans. Electromagn. Compat., Vol. 30, No. 3, , Aug Crawford, M. L. and J. M. Ladbury, Mode-stirred chamber for measuring shielding effectiveness of cable and connectors an assessment of MIL-STD-1344A method 3008, Proc. of IEEE International Symposium on EMC, 30 36, MIL-STD-285, Method of attenuation measurements for enclosures, electromagnetic shielding, for electronic test purposes, Jun. 25, 1956 (replaced by IEEE-TD-299, Oct. 24, 1997). 17. Rajamani, V., C. F. Bunting, M. D. Deshpande, and Z. A. Khan, Validation of modal/mom in shielding effectiveness studies of rectangular enclosures with apertures, IEEE Trans. Electromagn. Compat., Vol. 48, No. 2, , May Turner, J. D., T. M. Benson, C. Christopoulos, D. W. P. Thomas, M. P. Robinson, J. F. Dawson, M. D. Ganley, A. C. Marvin, and S. J. Porter, Characterization of shielding effectiveness of equipment cabinets containing apertures, Int Symp. EMC, , Rome, Italy, Sept Khan, Z. A., C. F. Bunting, and M. D. Deshpande, Shielding effectiveness of metallic enclosures at oblique and arbitrary polarizations, IEEE Trans. Electromagn. Compat., Vol. 40, No. 3, , Aug