Product Focus
Last Updated: Apr 9th, 2008 - 15:00:00
|
See this article in our Digital Edition
Download a pdf of this article (right click to save the pdf)
In order to comply with the stringent radiated emission limits imposed by the standards, and taking into account the increasing clock speed and data rates of current high-speed digital electronics, it is often necessary to shield the enclosure where the printed circuit board (PCB) is located.
A thorough electromagnetic compatibility (EMC) or signal integrity (SI) assessment of a complete board or system is often a daunting task, due to the extreme complexity of modern electronic systems. But system designers would like to accurately evaluate the electromagnetic interference (EMI) produced by high-speed signals and, at the same time, to address the regulations due to the EMC community. This capability is useful in predicting and correcting interference problems at various stages in the design process.
There are several requirements for an accurate evaluation of these effects. First, the complicated multilayered board and package structures used in today’s designs (including signal traces, supply planes and vias) must be modeled in a way that takes into account full-wave effects. Secondly, the behavior of various shielding structures (such as metallic enclosures) must be taken into account as well. Finally, the issue of problems consisting of small structures embedded in large computational domains must be addressed. Performing such analysis is often a computational challenge.
A three dimensional (3D) solver is desirable for modeling the arbitrary shapes of the enclosures (which often includes slots, apertures [1]), but simulating a multi-layer board or package can be difficult and memory consuming. The principle problems result from the high complexity of modern boards and packages, and the 3D nature of the system enclosures surrounding them.
This problem can be solved by using either of the following approaches:
- A specialized 2D solver (transmission line (TL), cavity model [2]), 3D quasi-static (partial element equivalent circuit (PEEC) [3] or 2D FDTD [4]) can be employed to compute the current distribution on the power (PWR) plane of a complex PCB or the Nearfield of the entire board [4].
- A full-wave 3D EM solver (finite integration technique (FIT) [5] or transmission line matrix (TLM) [6]) can be instead used to analyze the shielding performance of a metallic enclosure due to the calculated current/near field distribution.
The described workflow is illustrated with the different steps in the block diagram of Figure 1.
Figure 1: Typical EMC-EMI workflow analysis
Model Validation
An important point is to check the reliability of the software used for both the numerical simulations. For example, Figure 2a illustrates a simple enclosure model (already studied in [7]), a rectangular box 30 x 12 x 30 cm in size, with a rectangular aperture of 10 x 0.5 cm in size, located at the center of the frontal wall (15, 6, and 0). The enclosure is illuminated by a normal incident plane wave (farfield source) at 0 degree polarization, and three probes are placed in the center position inside the enclosure in order to register the three components of the electric field (Ex, Ey, Ez) and to calculate the SE afterwards.
Figure 2b plots the shielding effectiveness (SE) results using one software tool (CST Microwave Studio®), other numerical techniques as well as measurements. A good agreement can be observed over the considered frequency range 0-1000MHz.
The use of 3D EM field solvers provides a visual idea of the field distribution and, therefore, allows the hot-spots of the design to be easily identified. For example, Figure 2c clearly shows how the field distribution remains confined outside the enclosure for the frequency value of 500MHz, but penetrates inside the enclosure thought the aperture located in the front panel of the box when the frequency is increased to 1GHz.
This type of consideration, which might seem easy and almost predictable, might be difficult to achieve in reality without the usage of a proper simulation tool.
Figure 2: Example of model validation by comparing results due to different numerical technique as well as measurements: a) 3D EM model, b) results, c) electric field plot at 500MHz (left side) and 1GHz (right side).
Example of EM Modeling and Simulation Strategy
A view of the metallic enclosure analyzed in this section is shown in Figure 3 [8-9]. It is a 370 x 90 x 296 mm box, with 2 rectangular apertures of dimensions 126 x 14 mm and 80 x 60mm in the front panel. On the top of the box a cover is mounted and inside there are two boards and a heat sink. The thickness of the metal walls (PEC) is t = 2 mm, and the dielectric material of the board has relative electric permittivity εr = 4.0.
Figure 3: View of a typical metallic enclosure (left side) and radiation source detail (right side)
The board to be analyzed is a typical 6 layers PCB with hundred of nets, vias and connections. The PWR plane is split in islands by means of gaps. Due the complexity of the board, the dynamic link with a PEEC based code [3] is used to evaluate the current distribution on the PWR plane. The calculated field can be then used within the selected software tool or [6] in order to perform the 3D numerical simulation of the metallic enclosure, according to the workflow design process summarized in Figure 1.
Electric field components are calculated in a specific location 3m distant from the front panel of the metallic box, and the SE is also analyzed. In the considered frequency range, the EM fields inside the enclosure are dominated by the first two waveguide modes, and the orientation of the slots located in the front part of the enclosure is such that the vertical electric field component can couple easily across the aperture.
The SE can be found from the ratio of the field strengths without and within the enclosure, as follows:
Due to the highly resonant behavior of the box, and in order to speed up the simulation time, a very small value of loss is distributed throughout the solution space by artificially assigning a conductivity (σ=0.002 S/m) to the free space cells of the calculation domain. In [10], it has been already demonstrated how this artifact has practically no influence of the far field calculated results.
Figure 4 illustrates a similar approach (as just described), invoking [4] to calculate the near field distribution due to a complex multilayer PCB, and imported as radiation source within [5] to determine the farfield data (left picture).
Figure
4: View of a metallic enclosure with the radiation source imported as
Nearfield (left side) and farfield results (right side)
Effect of Slots and Apertures
The effect of apertures’ shape and configuration on the value of the radiated electric field and the related shielding effectiveness (SE) is another important aspect to be analyzed. For instance, in Figure 5 it is shown that dividing a specified area into a combination of multiple apertures may reduce the value of the radiated emissions.
Figure 5: Example parametric analysis by dividing a slot located in the frontal part of the enclosure into multiple apertures.
The critical aspect of this kind of analysis is represented by the high number of details and high aspect ratio due to the possible presence of multiple slots, ventilation holes, honeycomb panels, etc. Some software tools prefer to adopt specific meshing techniques in order to be able to mesh the complex shape of enclosure, while other tools have a build-in library which allow for the easy characterization of slots and apertures without physically meshing them, therefore saving memory requirement.
One of the final goals for EMC-EMI analysis is represented by the matching of specific constrains determined by normative regulations; for this reason it is important to be able to perform this check within the same tool used to detect the possible EMI-EMC source problems.
Figure 6 illustrates the SE for a metallic enclosure in the frequency range of 0-1GHz and the comparison with FCC regulations (Class A and B) at 3 meters and 10 meters. In this specific case, the design corresponded to Figure 6a would not meet the regulations for the Class B.
Figure 6: SE results and comparison with FCC compliance (class A and class B) Honeycomb panels and treatment of multiple slots
It is straightforward to point out that the possibility of predicting this behavior a-priori (before any prototyping activity) will save time and money and speed up the design process, thereby also reducing the time-to market of the product.
The apertures located in the front panel of a metallic enclosure can be covered by honeycomb panels, as shown in Figure 7. For the left panel, circular holes of 2mm diameter and 1mm distant from each other are employed, while for the right panel circular holes of 4mm diameter and 2mm distant each other are modelled. The calculated results are reported in the same figure (left side) where the comparison with case of regular aperture (Figure 5, top picture, left side) is presented. A sensible improvement (of more than 100dB) can be observed in this case, which shows the high performance of honeycomb panels.
Figure 7: Model of enclosure with honeycomb panel and SE results
Often the system to be analyzed is more complex [11]. As an example, Figure 8 depicts a real view and the correspondent electromagnetic model of the loaded monopole antenna used to perform measurements inside a semi-anechoic chamber. The antenna consists of two disks, with a radius of r=48mm, a thickness of t=0.5mm, 1mm distant from each other, and separated by a dielectric material with relative electric permittivity of εr=8. The leg’s length is h=95mm, and the first resonance frequency is at 0.65GHz.
Figure 8: a): Model of a complex enclosure, b): measurement set-up inside semi-anechoic chamber
The results obtained by means of measurements are presented in Figure 9, where the vertical and horizontal (Ex and Ey) electric field components registered 3m distant from the metallic opened rack are compared with the corresponding results coming from the numerical simulation. Good agreement is also achieved in this case for both components.
Figure 9: Comparison between measured results and numerically simulated results: horizontal (Ex) and vertical component (Ey) of the electric field within the frequency range 30MHz-1GHz.
Integration of Analysis Tools
Another useful capability of EMC simulation software is the ability to operate in an integrated analysis environment, using the same model developed for mechanical/thermal analysis in the evaluation of EMC at the early design stages. This is valuable because thermal design often conflicts with EMC design, and fixes that are implemented to address thermal concerns often exacerbate or create EMC problems.
The most typical example is that thermal design requires large holes to enable adequate airflow, while EMC design requires small holes to reduce emissions. A hole will pass electromagnetic fields in and out of the enclosure if one or more of its dimensions are equal to or larger than the wavelength of the field.
As mechanical engineers begin to develop the physical design, they can drag the PCB model and drop it into a system-level thermal design. An integrated analysis environment not only ensures the transmission of accurate information to mechanical engineers, but also provides immediate notification of design changes.
The same model that is created for system level thermal analysis can also be used to address EMC issues far earlier than is normally possible. Figure 10 illustrates a complex metallic rack used for telecommunication equipment. The IGES mechanical file format can be easily imported into the simulation software tool and numerically simulated in order to obtain results in the Nearfield and/or farfield region.
Figure 10: Metallic enclosure used for telecommunication equipment: a) 3D EM model and b) E-field plot at 1.5 GHz.
Conclusions
This article has offered an overview of the different approaches which can be employed to address the challenging EMC-EMI analysis of complex PCBs located inside a metallic enclosure. It has also investigated possible workflow process, as well as the advantages, disadvantages and reliability of simulation software.
Multiple examples have been presented, like the analysis of the effect of different shapes and configuration of apertures on the SE of a metallic enclosure. The SE of some combinations of multiple apertures shows that dividing a fixed area into some smaller apertures will lead to more efficient shielding than implementing only one aperture. This is helpful when optimizing the shape of open area used to feed trough cables or heat dissipation. The SE when honeycomb panels are employed is also analyzed.
By combining a specialized board analysis tools based on PEEC method with a 3D full wave EM simulation tool, a consistent approach to compute the electromagnetic emissions of complex electronic systems can be demonstrated. The proposed workflow consists on three steps: 1) simulation of the complex PCB; 2) surface current distribution/Nearfield used to excite the enclosure; and 3) full wave simulation of the metallic box.
The reliability of 3D EM simulation is proven by comparing different technique as well as by providing measured data results. n
A. Ciccomancini Scogna is with CST of America, Inc., and can be reached at antonio.ciccomancini@cst.com.
References
- J. Parkes, J. Bracken, Z. Cendes, “EMC Simulation of Complex High Performance PCBs and Shielding Effectiveness”, in Proceedings of DesignCon 2006, Santa Clara, CA, USA.
- EZPP - www.ems-plus.com/
- PCBmod, Simlab – www.simlab.com
- Speed 2000 – www.Sigrity.com
- CST Studio Suite 2008TM – www.cst.com
- Microstripes – www.flomerics.com
- C. F. Bunting, Khan Z. A. and Deshpande, M.D, “Shielding effectiveness of metallic enclosures at oblique and arbitrary polarizations” IEEE Trans. Electromagn. Compat., vol. 47, no. 1, Feb. 2005, pp. 112–122.
- A. Ciccomancini Scogna, M.Schauer, “EMC Simulation of Complex PCB inside a Metallic Enclosure and Shielding Effectiveness Analysis”, on Proceeding of EMC Zurich 2007, Munich, September 2007.
- http://aces.ee.olemiss.edu
- G. Antonini, A.Ciccomancini Scogna, A.Orlandi, “Shielding analysis of a metallic enclosure by means of a statistical approach” in Proceedings of EMC Europe 04, Eindhoven, 6-10 September, 2004.
- Ciccomancini Scogna, “Shielding Performance of a Metallic Rack used for Telecommunication Equipments: FIT modeling and Measurements”, in Proceedings of IEEE Int. Symposium on EMC Zurich 06, Singapore, February 2006
© 2007 Conformity
Top of Page
|
