How to Select EMI Shielding for PCB Assemblies

A shield that looks effective on a bench can fail once the enclosure is closed, cables are connected and the product enters its real electromagnetic environment. That is why knowing how to select EMI shielding is not simply a material choice. It is a system-level engineering decision involving the noise source, coupling path, sensitive circuitry, mechanical construction and production process.

For compact electronics, AI hardware, robotics and high-speed PCB assemblies, late-stage EMI fixes are expensive. The right approach is to define the interference problem early, then select shielding that performs at the frequencies that matter without compromising heat dissipation, serviceability, weight or manufacturability.

How to Select EMI Shielding: Start With the Failure Mode

EMI shielding should be selected against a defined problem, not a generic requirement for “more shielding”. First establish whether the design is emitting excessive energy, receiving interference from nearby equipment, or both. An emissions issue may originate from a switching regulator, clock line, processor, display interface or cable. A susceptibility issue may affect low-level analogue signals, wireless radios, sensors or high-speed data channels.

The source-path-victim model is useful here. Identify the noise source, determine how energy is coupling through radiation, conduction, capacitive coupling or inductive coupling, then identify the circuit or subsystem being affected. Shielding is most effective when it interrupts the actual coupling path. A can placed over a noisy IC may offer little benefit if the main antenna is an unshielded flex cable or a poorly bonded enclosure seam.

Frequency defines much of the solution. Low-frequency magnetic fields are difficult to attenuate using thin conductive foils, while higher-frequency electric fields can often be controlled with conductive coatings, shielding cans and well-grounded enclosures. Broadband noise may require a combination of shielding, filtering, controlled return paths and improved PCB layout.

Before choosing a material, capture the operating bands, clock harmonics, switching frequencies and relevant compliance limits. Include expected cable lengths and enclosure apertures. A small slot can become an efficient radiator when its dimensions become significant relative to the wavelength of the interference.

Define the Required Shielding Effectiveness

Shielding effectiveness is commonly expressed in decibels. It indicates how much unwanted electromagnetic energy is reduced between one side of a shield and the other. A higher value is not automatically better if it adds cost, mass or assembly complexity without addressing the dominant leakage point.

As a practical starting point, define the attenuation needed across the relevant frequency range, then assess the complete shielded structure. Material data sheets can show impressive attenuation figures under controlled conditions, but final performance depends on installation. Gaps, apertures, poor seams and inadequate grounding can reduce real-world effectiveness far more than the difference between two otherwise suitable materials.

The enclosure and its interfaces deserve the same attention as the shield itself. Consider:

• openings for connectors, displays, ventilation and controls
• joints between enclosure sections and removable covers
• cable shields and their termination method
• grounding points, return-current paths and bonding resistance
• tolerances that may create inconsistent contact in production

A conductive enclosure with an unbonded lid is not a continuous Faraday cage. Equally, a shielded cable with a long pigtail termination can lose much of its high-frequency benefit. At higher frequencies, short, wide, low-inductance connections are generally preferable.

Choose the Right EMI Shielding Material

Material selection should balance conductivity, magnetic performance, mechanical form, corrosion resistance and ease of manufacture.

Copper offers excellent electrical conductivity and is highly effective for electric-field shielding. It is commonly used in foil tapes, conductive fabrics, flex circuits and grounding features. Its limitations are cost, oxidation risk and relatively low stiffness unless supported by another structure.

Aluminium is lightweight, readily available and well suited to machined or formed enclosures. It provides good electric-field shielding but requires deliberate surface treatment and bonding design. Anodised surfaces, for example, are electrically insulating unless contact areas are masked or otherwise prepared.

Nickel-plated materials and conductive coatings can provide durable contact surfaces and useful corrosion resistance. Conductive fabrics, foams and elastomer gaskets are valuable where two parts must maintain electrical continuity despite dimensional movement, vibration or repeated opening. Their performance depends on compression, contact geometry and environmental durability.

For low-frequency magnetic fields, high-permeability alloys may be required. These materials guide magnetic flux rather than primarily reflecting it. They can be effective around transformers, inductors and sensitive magnetic sensors, but they are more specialised, can be costly and may lose performance if formed or handled incorrectly.

In many products, no single material is sufficient. A metal enclosure may contain the main radiated field, a conductive gasket may close the seam, and a local PCB shield can may isolate a radio, switching supply or processor section. The correct architecture depends on where the energy is generated and how it escapes.

Treat Grounding as Part of the Shield

A shield without a controlled reference can behave unpredictably. For electric-field shielding, grounding gives intercepted current a defined return path. The goal is not to connect every metal part indiscriminately, but to create low-impedance bonds appropriate to the frequencies involved.

At low frequencies, DC resistance is a useful indicator. At high frequencies, inductance becomes critical, so bond length and geometry matter. A broad chassis connection can outperform a narrow wire even where both measure near-zero resistance with a multimeter.

Separate functional ground, protective earth and chassis ground according to the system architecture and applicable safety requirements. Then review where cable shields terminate. A 360-degree termination at the point a cable enters the enclosure is often far more effective than routing a drain wire across the PCB to a distant ground point.

For PCB-level shields, provide multiple low-inductance ground vias around the perimeter. Via spacing, solder continuity and the quality of the underlying ground plane all influence performance. Avoid treating a shield can as an isolated mechanical part. It should be designed into the RF and return-current strategy from the outset.

Account for Flex Cables, Connectors and Apertures

Interconnects often become the weakest part of an otherwise well-shielded product. A flex cable can carry high-speed signals, switching noise and common-mode current between compartments. If it crosses an enclosure boundary or runs close to a sensitive analogue section, its construction and routing need deliberate consideration.

A grounded layer in a flex design can reduce coupling and provide a controlled return path, but it also changes flexibility, thickness and bend life. Stiffeners, shielding films and conductive adhesive systems should be selected around the required bend radius and movement profile, not added after the mechanical design is fixed. Connector shell bonding and pin assignments matter as much as the cable construction itself.

Apertures require the same discipline. Ventilation may be essential for thermal performance, yet large openings can compromise shielding. Honeycomb vents, conductive meshes or smaller distributed apertures can offer a better balance where airflow is needed. The right answer depends on frequency, cooling demand and the permitted enclosure volume.

Validate the Design Before Compliance Testing

Pre-compliance testing is the fastest way to confirm whether the selected shielding strategy is addressing the problem. Use near-field probes to locate local sources, current probes to investigate cable-borne noise and spectrum analysis to track changes across relevant bands. Test representative mechanical assemblies, not only bare PCBs.

Change one variable at a time where possible. If adding a shielding can improves one frequency band but worsens another, investigate grounding, resonances and altered return paths before adding further material. EMI control is often iterative, particularly in dense systems with multiple radios, high-speed interfaces and switch-mode power supplies.

Manufacturing validation is equally important. Check gasket compression, plating quality, lid fit, conductive adhesive cure, screw torque and ground-via soldering across production tolerances. A design that passes with a hand-built prototype must still perform when assembled repeatedly at volume.

For bespoke interconnect and PCB projects, bringing shielding requirements into the design review early allows cable stack-up, grounding features and enclosure interfaces to be engineered together. Cocom supports this approach through custom flexi and PCB engineering built around the practical constraints of next-generation electronics.

The strongest shielding solution is rarely the thickest or most expensive one. It is the one that closes the real coupling path, survives production variation and leaves the product free to perform exactly as intended.