EMI Shielding: Design Guidelines

As is usual in engineering, the design stage is much more difficult than the analysis stage, and shielding does not represent an exception for this rule. The reason is manifold: on one hand, the solution of a shielding design problem is generally not unique, and therefore a number of choices has to be made involving very different aspects, often very far from electromagnetics. On the other hand, the concurrence of several coupling paths calls for some wise assumptions in order to simplify the procedure. The uncertainty as to the figures of merit to be selected and met to guarantee the correct operation of the shielding structure (i.e., the shielded devices and systems) is another delicate point that's often by-passed when compliance with standards is assumed to be sufficient for the achievement of EMC and EMI protection or information security objectives.

In general, any procedure aimed at designing a shielding structure consists of the following main steps:

1. Establishment of the shielding requirements, in terms of SE values or other figures of merit.

2. Assessment of type and number of functional discontinuities.

3. Assessment of dimensional constraints and non-electromagnetic characteristics of the materials.

4. Formulation of a hypothesis of shielding configuration with all the details defined.

5. Estimation of the shielding performance.

6. Check on specs fulfillment and , if necessary, modification of the characteristics of the most relevant coupling paths responsible for the failure in the design.

Of course, steps 4 to 6 should be repeated until the check guarantees that a reasonable probability of success is achieved by the shielding configuration, once it's realized. It is important to observe the core of the design procedure that's in steps 4 and 5: experience will be of fundamental guidance in the choice of the structure (e.g., single shield or multilevel), the components, and the materials. The first step is also critical because it may lead to an over- or underestimation of the required shielding levels.

The way in which the steps of the above-outlined procedure are dealt with can be very different in the two typical situations of susceptibility-oriented or emission oriented design. In the former, the frequency-spectrum of the EM field is assumed to be known, the immunity levels of the various components (in-band and out-of-band) are often known by manufacturers, and attention is paid to two key points: coupling paths and internal emissions (possibly adding to external threats). On the other hand, in emission-driven design procedures the EM field levels of the various subsystems and components are strongly dependent on assembly characteristics, while the emission level masks not to be exceeded are usually known and established by standards.

ESTABLISHMENT OF THE SHIELDING REQUIREMENTS

The frequency mask of the EM requirements is usually obtained by taking into account three terms that are then combined in different ways in case of emission or susceptibility problems:

• Levels of or limits for the environmental EM field, EL
• Margin, M
• Intrinsic and installation-dependent levels of EM susceptibility, S, and radiated emission of the involved components and systems, RE.

Each of the three terms calls for specific analysis and will be briefly considered in the following discussion by leaving to references a deeper treatment.

The margin M is often fixed between 10 and 30 dB, depending on the known accuracy of the adopted prediction methods and on the uncertainty existing on external field or radiated-emission levels and possible evolution in the system (system upgrade, time evolution of performance, etc.). However, a margin M 20 dB is often chosen.

When the susceptibility problem is considered, an upper bound has to be selected between the predetermined frequency masks and the possible application-oriented EM levels of external fields due to interfering sources. For instance, the mask may be that considered in MIL-Std. 461E [2] and the frequency spectrum that ensuing from an EM pulse [3]. Furthermore the presence of sources in the interior of the shielded volume should also be carefully considered; depending on the field levels and on the frequencies they may generate, this aspect will in fact be of guidance in the choice of the number of shield levels necessary to guarantee the correct operation of the system and the compliance with standards. In general, the most difficult task is that concerning the inventory of the susceptibility levels S of all the components or subsystems. A careful analysis must be conducted not only on the in-band and out of-band susceptibility of each component but also on their functional connection and on the influence that an external threat may have on non-exposed components through the exposed adjacent ones. In susceptibility-oriented shielding design, the required level of reduction is

SE = EL+M- S [eq. 1]

The design of shielding configurations aimed at limiting the unwanted emission from devices, apparatus, and systems requires a careful analysis of the characteristics of the sources, since the radiating elements and mechanisms are manifold, and moreover near-field problems are often encountered. Their identification may be not as trivial as it might be expected at a first glance. The listing of all the radiating components within the shielding structure will not circumvent the task; for each of them it's necessary to estimate the voltage and the current spectra, taking into account that, for most components, only the maximum time response is declared and guaranteed by manufacturers, leaving an important uncertainty on the maximum involved frequency. After this operation is completed, the geometry of the subsystem connections should be considered to account for the radiating components' additional radiation at prescribed distances where limits are ?xed. The shielding requirements are evaluated as

SE=RE+M-EL [2]

When both susceptibility and radiated emission are of concern, the selection between values ensuing from (1) and (2) is not automatically the largest. The source type involved in the two situations is generally different, and the a priori choice of the most critical SE is as difficult as the separate verification for the two cases. Thus the values ensuing from (1) and (2) must be compared with the shielding effectiveness (in a broad meaning) of the designed structure under both the test conditions.

ASSESSMENT OF THE NUMBER AND TYPES OF FUNCTIONAL DISCONTINUITIES

There are several types of functional discontinuities. The main and most frequently encountered discontinuities are junctions and seams, cable pass-throughs, visualization apertures, air vents, operational devices, and connectors.

As described previously, some functional discontinuities are dealt with by means of solutions enabling an adequate level of SE, whereas others can be rather critical. Shielded windows are probably the components with the lowest level of SE because they may achieve performance levels on the order of 40 to 50 dB (typical values for windows whose dimensions are up to few tens of centimeters in the maximum dimension), while levels up to and even above 100 dB are achieved with correctly designed air vents and shielded connectors. Windows of large dimensions can be expensive, so the use of multiple apertures for visualization is often preferred with respect to a unique large aperture. Very serious concern is generally directed to non-shielded operational devices (e.g., fuses, lamps, and switches) whose apertures offer an additional coupling path of large impact on the shielding performance of the overall system. Also the signal or power lines that connect the shielded region to the source region can be an efficient vehicle of coupling. However, the adoption of shielded operational devices and of proper grounding [4,5] of cable sheaths is usually sufficient to prevent any appreciable performance deterioration or to reduce to a minimum the undesired effects.

Junctions between panels are often sealed by gaskets, as described in Sect. 9.

When gaskets are correctly selected and maintained, they guarantee adequate levels of SE in almost all the applications. Much more critical situations occur in dealing with permanent or semi-permanent joints, typically welded, screwed, or bolted. At the design stage non-continuously sealed joints are worthy of a careful analysis. For instance, the improvement in the SE achieved by subdividing a large aperture, whose length and height are L_tot and h, respectively, into Na smaller apertures of length La is [1]

It should be noted that adding screws or other types of contact points will reduce the height of the apertures with respect to the original one, thus further improving the shielding performance.

ASSESSMENT OF DIMENSIONAL CONSTRAINTS AND NONELECTROMAGNETIC CHARACTERISTICS OF MATERIALS

Dimensional constraints and other non-electromagnetic considerations (corrosion, weight, cost, etc.) will influence the choice of materials and the shape of the shielding structure, which is closely related to the content's dimensions and to its emission and immunity characteristics. On the other hand, the shape and the geometrical dimensions of the shielded volume determine the frequency and the number of internal resonances. A preliminary and approximate analysis aimed at verifying whether or not the enclosure will work in an overmoded region of the frequency spectrum is useful to have as guidance on the importance that some aspects (e.g., the internal position of the most sensitive or radiating elements) can assume in the design.

Unfortunately, loading and apertures dimensions and positions can considerably affect the value of the resonant frequencies, so a careful numerical analysis is needed for a wise arrangement of components and subsystems in the shielded volume.

However, in most cases the arrangement of such components and subsystems is determined on the basis of non-electromagnetic considerations, such as those due to accessibility, ventilation, or visualization.

When compatible with costs and weight considerations, the use of partial internal shields is always recommended to protect the most susceptible components or to prevent the most radiating from causing both internal problems and emission levels larger than those admissible. Such a double level of shielding structure can also alleviate the requirements for the external shield with an overall advantage in terms of functionality and costs.

ESTIMATION OF SHIELDING PERFORMANCE

It should be recognized that shielding performance of actual configurations is very difficult to accurately predict. A rough estimation can be achieved by use of numerical simulations. These are generally accurate enough for prototyping purposes, provided that the constructive details are described with sufficient accuracy. However, depending on both system complexity and software/hardware characteristics, the input of details of an actual configuration, even in user-friendly commercial software, may require a lot of work and the output may be available after a long computational time. Therefore a preliminary approximate estimation can be very useful. Such an estimation of the SE can be obtained from the ''ideal'' SE achievable by means of a barrier of homogeneous material corrected by two terms: a worst-case term accounting for leakages and a rule-of-thumb term representing resonance deterioration. The consequent estimation of the SE is therefore SE SE_barrier DSE_tot leakage DSE_standing waves

The terms SE_barrier and DSE_leakage (due to one aperture) were presented in Sects 4, 6, and 7. Often, in a first attempt of shielding design, the deterioration effects due to internal resonances are taken into account by means of a 6 / 10 factor (in dB) [1].

If the leakage is due to several effects, their worst-case combination (i.e., the sum of in-phase field contributions arising from different coupling paths) is expressed as where Li dB is the leakage (in dB) of the ith coupling path. Thus the general expression for a first performance estimation could be

[equation omitted]

Next: Faraday Screens

Prev: Questions/Answers about Switching Power-Supply Topology

top of page   Home