Getting the Most out of an EMI Control Plan

William D. Kimmel and Daryl D. Gerke

Savvy contractors in military development programs can turn a paperwork requirement into an effective project tool.

Anyone working on a military development project has probably encountered the EMI control plan, a document formally known as "Electromagnetic Interference Control Procedures (EMICP)" and described in DI-EMCS-80199B.1 It is one of three EMI-related developmental documents typically required for a military program, the other two being the "EMI Test Plan" and the "EMI Test Report," which are not discussed here.

Military programs may seem overburdened with paperwork, especially to people who are new to defense projects, but the documentation is necessary if the complex systems are to perform as intended. Any program may have multiple layers of suppliers. Below the prime contractor will be several echelons of subcontractors. The content of each contractor's plan will depend on where in the supply chain the company falls.

The purpose of the EMI control plan is to ensure that EMI gets its rightful share of attention throughout the program. It provides design guidance to each contractor's design staff and describes to company management–that of both the contractor and its customer–how EMI requirements are going to be met. The plan is created early in the design stage, perhaps even when the contractor submits a proposal. No one is expected to have detailed knowledge of all aspects of EMI at the beginning of the program. The plan is intended to be a living document. Though initiated early in the project, it is updated as necessary throughout the program until the time for testing.
This article examines EMI control plan requirements, intending to supply insight into what each requirement means and why it is there. The discussion is organized as the plan is, with sections covering management, design, analysis, and developmental testing.

Management

Assessing Unknowns. The first order of business is to get a handle on the extent of the EMI problem, including identification of the problem areas. In the early part of the program questions are many and answers few. Some contractors, on some projects, will have a pretty good idea of what needs to be done. In other cases, they start with little information and thus have to map out a method of getting the information they need.

The next thing that must be done is to decide on an overall design approach. This will encompass box shielding, cable shielding, signal and power filtering, and transient protection, as well as internal design, printed circuit board (PCB) design, cables, and grounding. A complex system, as mentioned, will involve many suppliers, ranging from the prime contractor down through several layers of subcontractors. A project participant using subcontractors will need to apportion the responsibility for designing to control EMI.

If a contractor is allocating responsibility to a supplier, the contractor has to establish test requirements. Sometimes the requirement may be flowed down to the supplier, but in other cases the contractor will need to take on that responsibility. A good example of the latter situation is if the contractor is responsible for the enclosure but has suppliers providing circuit boards. Radiated-EMI requirements cannot realistically be levied on the circuit boards; those will need to be handled at the box level. On the other hand, conducted requirements may well need to be handled by the suppliers. They certainly will if the components interface directly with the outside world.

The contractor may also need to look at government-furnished equipment (GFE), which may include other equipment or existing-installation constraints, and which is usually unmodifiable. If the contractor uncovers a problem involving GFE, it is probably going to have to work around it. The test requirements have to be specified. This may be a simple matter of reproducing the requirements that have been levied on the contractor by the customer. If there is disagreement between customer and contractor–for example, should the contractor believe the suggested test levels are inappropriate–the contractor should present rationale for the requirements. If the contractor is levying requirements onto the supplier, then test levels should be specified in detail.

Unless the contractor has expert understanding of the test levels, it may be best to start with conservative requirements. Stringent requirements allow room for error or other contingencies. Also, it is really tough to tighten up requirements late in the program. The customer may well have built some margins into the contractor's requirements, too; thus, if the contractor runs into a problem in a later stage of the program, it may be able to get relief on some of its own requirements.

Getting Answers. Once the EMI unknowns have been assessed, the next step in devising the initial EMI control plan is figuring out how the answers to questions are going to be acquired. The original design requirements may be established by the analysis that is to be included in the plan (and which is discussed below), or later on as a result of development testing. Simple back-of-the-envelope worst-case analysis can be performed relatively easily, but it tends to be quite conservative. Nevertheless, it is worth a first cut, as it does highlight both problem areas and areas that can be eliminated from consideration as problems. Detailed computer analysis may be carried out, although it is not particularly accurate either.

Better answers can be obtained through developmental testing. However, that requires preparation of a suitable test setup with the appropriate test fixtures, which may or may not be feasible when the plan is being drafted. Sadly, radiated testing is not feasible until the hardware can be mounted in a prototype enclosure. That is often later than the contractor would like. Anyway, the contractor should describe its intentions in the plan.

A final alternative is to adopt the conservative design approach, that is, designing for a near worst case, perhaps as estimated by that back-of-the-envelope analysis. In some cases, this is an easy way out. However, this approach often collides with other constraints, such as weight, size, and, increasingly, cost.

This last approach might be moderated by means of the fallback method, in which the contractor begins by assessing what would be a near-worst-case need and then backs up to see where some corners might be cut. For example, there may be a question of whether gasketing is needed or not. The solution is to make provision for gasketing, just in case, but to design in the hope that it won't be necessary. EMI engineers find that the best way to mitigate the effect of Murphy's Law is to have a good fallback position ready to implement.

Clearly, the better the data available early in the design phase, the fewer the contingencies that need to be allowed for.

Involving Others. The plan will need to incorporate the normal management information, as well. Because it may be assumed that the plan will be read by people not steeped in EMI lore, it should include an overview of the system. Also, the plan should be composed such that, should the author be run down by the metaphorical train at a critical point in the program, someone else will be able to pick it up and run with it.

The organization chart should be available from the program manager. It is important that the EMI contact and reporting path be shown. This information assures the customer that when a concerned engineer talks, someone will listen. It also gives the management team the message that an avenue is available for the resolution of problems or disputes.

The contractor's plan should identify the lead EMI person in the company, along with the key customer and supplier contacts–this again just in case someone else needs to pick up the project and continue it.

The final component of this section of the plan is the project schedule, usually furnished by the program management office. It should be checked to make sure it contains the EMI schedule, including dates for documentation and testing.

Design

After establishing the ground rules in the management section of the EMI control plan, the contractor must turn to design specifics. Basic design concepts have to be fleshed in enough detail to provide guidance to the in-house design team or to the suppliers' designers. There are a lot of misconceptions about EMI that must be countered by careful specification in the design section of the plan. Things needing to be defined are the enclosure shielding, cable shielding and filtering, internal cabling and grounding, and the PCB and its grounding.

Enclosure Shielding. The enclosure construction and materials need to be specified. In most cases, any metal will be satisfactory from a shielding standpoint; however, corrosion treatment must be considered, especially at the mating surfaces.

If an effort is being made to get away without gasketing, seams and other openings in the enclosure will have to be specified, as will spacing of the fasteners. Unless engineers are quite confident that gasketing will not be needed, the designer should be urged to make fallback provisions for gasketing. Guidance regarding acceptable gasket materials should be included in the plan.

Something to keep in mind is that mechanical designers may have only a dim notion of what it takes to make a Faraday cage. They tend to think in terms of line of sight for shielding, and may have little understanding of the significance of slots and other openings.

Cable Shielding and Filtering. A decision must be made with respect to which signal and power lines are to be shielded and which are to be filtered.

For lines that are going to be shielded, the plan must specify the method of shield termination at both the cable connector and the bulkhead.

For lines that are going to be filtered, how the filtering will be accomplished must be specified. The most effective method–which also is the most expensive and involves the longest lead times–is to use filter connectors. A second choice is to put the filters immediately behind the connector. This is less certain but a lot cheaper. Power lines would probably have high-frequency filtering mounted in a doghouse immediately behind the connector. The plan also needs to specify whether transient protection is intended to be used and, if so, what kind of devices and how they will be mounted.

The contractor should be alert for cases where filtered and shielded lines are in the same connector. Although military-style connectors will accommodate this design practice, the isolation is not optimal. Such a design configuration should be avoided if possible. Another problem with it is that the shielded lines will inevitably have at least a short pigtail termination, which degrades high-frequency shielding effectiveness.

Internal Grounding and Cabling. Next to be specified is the internal grounding and cabling. High-frequency grounds will have to be explained to the mechanical designers; to them, a ground is a ground, and any metallic connection will do. Ground straps should be specified if high frequencies are needed. Also, the plan needs to specify bond impedance and how to achieve it.

The specification of internal cable layout should include the presence of a shield or ground layer for ribbon or flex (microstrip line) or, if there is no ground plane, the positioning of ground traces. Assuredly, if the ground returns are not specified, the designer will come up with one or two perhaps located in the worst possible location. When a microstrip line or shielded cable is part of the design, whether the ground layer is to be connected to the enclosure ground or circuit ground should be noted.

The cable routing also must be specified. Cables often carry large amounts of high-frequency energy and will couple to portions of the enclosure. If the cable is routed close to a cover plate or a seam, those elements will be energized, causing, at the least, a radiated-emission problem. This is especially important if the contractor hopes to avoid gasketing.

PCB Design. Military electronics designers don't pay much attention to circuit board design, mostly because stiff military EMC requirements call for shielding and careful filtering that make PCB design less critical. But in these highly cost-conscious days, shielding may be less than generous, and thus also marginally effective. Anything that can be done to enhance the design of the PCB may prove to be the difference between passing or failing developmental tests.

So the plan should define some basic good PCB design techniques. The board design does not require extraordinary effort, just attention paid to the basics. Rise-time control and filtering of critical signals, cross-talk-minimizing component placement and trace routing, maintenance of signal path return continuity, plane isolation, and decoupling techniques should all be employed. Good board design minimizes the time spent running down gremlins, as well.

Also requiring specification is the manner in which the board is going to be grounded, both within the board (as when maintaining analog-digital isolation) and to the enclosure. If this is left unstated, the design could include undesirable ground connections in one place and missing grounds in another. If single-point ground is necessary, it may be advisable to specify frequent decoupling between the signal and enclosure ground.

If the PCB has input/output signals, it may well be necessary to have some filtering on board in addition to that in the connector area.
 
Analysis

Everyone would love to be able to do a good job of analysis, but it is difficult to achieve in practice. The contractor can and should produce a reasonable estimate of conducted interference, but analyzing radiated interference is not feasible. Anything beyond basic worst-case analysis will require the aid of analytical software. Considering the present state of the art, even that is not going to ensure much accuracy. What is needed in this section is basic analysis grounded in realistic assumptions and real-world experience.

Even the best analytical software and all the inputs an engineer can supply will provide at best one-order-of-magnitude prediction of radiated emissions. Susceptibility can't even begin to be calculated.
 
Developmental Testing

Because it is the only way to acquire reliable data early in the design stage, developmental testing is encouraged. However, it takes some ingenuity to come up with a test that is both useful and feasible.

For example, a test configuration might involve mocking up an enclosure with representative slots and openings, inserting a radio-frequency (RF) source, and measuring emissions. That would give an indication of shielding effectiveness. Still, it will not tell how much shielding is needed. Such knowledge at the outset comes only from experience gained from previous projects. By the time enough of the system electronics is built to provide a valid emission source, the test stage is near anyway.
 
Preliminary Design Review

The contractor's EMI control plan will be submitted for approval prior to preliminary design review (PDR). At this time, everyone with a project interest will have a shot at criticizing it. The object is to uncover any potential problems in a timely manner. If the contractor lacks ready answers, a reference should be inserted in the plan to flag the issue. This will prevent errors of omission, which are often overlooked.
 
Conclusion

When writing an EMICP, the contractor should keep firmly in mind its intent, which is to avoid as many problems as possible while providing a road map for identifying and resolving any problems that may materialize. Realistically, plenty of problems can be expected. However, if the plan writing is done right, these will be minimized, and there will be no disasters.

The purpose of the plan is not always fulfilled, unfortunately. The idea is to start with the plan and then fill in the blanks as the project progresses. All too often, however, the plan is written but never referred to after PDR. This is a big mistake. The EMI control plan is a valuable tool.
 
Reference

1. DI-EMCS-80199B, "Electromagnetic Interference Control-Procedures (EMCP)," 1990. Available from Internet: http:// www.jsc.mil/jsce3/emcslsa/stdlib.asp.
 
William D. Kimmel, PE, and Daryl D. Gerke, PE, are partners in Kimmel Gerke Associates Ltd., an EMC consulting firm with offices in St. Paul, MN, and Mesa, AZ, that specializes in EMC design, troubleshooting, and seminars. Together, they have more than 70 years of EMC experience. The NARTE-certified EMC and ESD engineers may be reached by phone at 888-EMI-GURU or via their Web site at www.emiguru.com.