Board Level Shielding Design With Cans and Fences

A board level shield is a metal enclosure soldered over a circuit to keep electromagnetic energy in or out. It is a mechanical part, a thermal part and an electrical part at the same time, and the design has to satisfy all three. A shield that is electrically correct may block airflow to the components inside, and one that is easy to assemble may be impossible to rework. This article covers the design decisions.

What a Shield Does

The shield reflects and absorbs electromagnetic energy. For emissions, it keeps the field generated by the circuit inside, and for immunity it keeps external fields out. The effectiveness depends on the material, the thickness and, above all, on the continuity of the enclosure and its connection to the ground reference.

A shield is not a sealed box. Airflow, test access and the need to rework components all require openings, and every opening is a path for the field. The design question is not whether to have openings but whether the openings are small enough and placed so that the frequencies of interest cannot pass.

Can and Fence Construction

A one piece can is a formed metal cover that is soldered directly to the board. It is simple, has no removable parts and gives the best electrical continuity, but it must be removed destructively to reach the components inside. A fence and lid construction separates the two functions: a frame is soldered to the board and a removable lid clips or screws onto it, which allows access without desoldering.

The trade is between electrical performance and serviceability. A removable lid introduces a seam, and the contact between the lid and the fence must be maintained over the life of the product. Where the frequency or the sensitivity demands the best performance, a soldered can is preferred, and where the product needs field service, the fence construction is worth the small performance penalty.

Aperture Rules

An opening in a shield behaves as a slot antenna, and its effect depends on the longest dimension of the opening relative to the wavelength. A long narrow slot is much worse than a small round hole of the same area, so the design should avoid long slots and distribute the required open area as many small openings instead.

Where a slot is unavoidable, it should be oriented so that the current it interrupts flows across the short dimension rather than along the long one. Ventilation patterns should be designed with this in mind rather than chosen for appearance, and the pattern should be reviewed against the frequencies the product must control.

Grounding and Attachment

The shield is only as good as its connection to ground. The attachment points should be distributed around the perimeter so that the current can flow around the enclosure without a long detour, and they should be close enough together that the gap between them is short relative to the wavelength. A shield soldered at four corners with a long gap between them behaves as a slot at the frequencies of interest.

The ground reference matters as much as the attachment. The shield connects to the board ground, so the return current has to be able to flow in the plane underneath. A plane that is fragmented under the shield, or a ground that is connected through a long narrow trace, defeats the connection. The grounding approach used for mixed signal boards is described in mixed signal design guidelines.

Thermal Implications

A shield that covers dissipating components traps heat. The temperature inside the enclosure rises above the board ambient, and the components inside run hotter than the thermal calculation assumed if it was based on an open board. The effect can be large enough to change a design that was comfortable in free air.

Thermal design therefore has to be part of the shield design. Ventilation openings, a thermal path through the board via a copper area under the components and a shield material with good conductivity all help. Where the shield is also used as a heatsink, the contact between the component and the shield has to be defined, and the interface material considerations are the same as those described in thermal interface material application.

Assembly, Rework and Inspection

The shield is attached after the components are placed and reflowed, either by a separate soldering operation, by a conductive adhesive or by a mechanical clip. The attachment method determines the rework path: a soldered can has to be removed before a component inside can be replaced, and the removal can damage the board if it is not done with proper heating.

Inspection has to confirm both the attachment and the internal condition. The solder fillet or the clip engagement is visible, while the components inside are not. Where the shield covers a critical circuit, the components should be inspected before the shield is attached, and the process should include a check that the shield is the correct part for the board rather than a visually similar variant.

Materials and Surface Treatment

The shield material and its surface treatment affect both the electrical performance and the solderability. A conductive material with a stable surface gives a low resistance contact to the ground pad, while an oxidised or poorly plated surface adds contact resistance that changes over time. Where the shield is expected to be soldered, the finish on the shield has to be compatible with the alloy and the flux used on the board.

Sheet metal thickness is a mechanical decision with an electrical consequence. A thicker material is more rigid and easier to handle but heavier and harder to form, while a thin material is light and forms easily but may not maintain contact pressure against a fence. Where a removable lid is used, the material has to provide the spring force that keeps the lid engaged over repeated removals.

Design Checklist

The shield footprint should be on the board outline, with the attachment pads, the ground plane and the keep out areas defined. The aperture pattern should be reviewed against the frequencies of interest, the attachment spacing should be set by the electrical requirement and the thermal effect should be estimated. The rework path should be decided at the same time, because it determines the attachment method.

The gopcb engineering team reviews shield designs together with the EMC and thermal requirements, since the three interact and a change to one affects the others. The acceptance criteria for the shield attachment follow the same class based logic as any soldered joint, and the record of the design decisions is kept with the product so that a later change does not silently degrade the performance.

Testing Shield Effectiveness

The shield has to be verified as installed rather than as a material. Measuring the field around the product with and without the shield, using a near field probe or a chamber measurement at the frequencies of interest, shows the actual attenuation. The difference between the two measurements is the only direct evidence that the design achieves what it was intended to.

The test should be repeated after any change to the attachment, the aperture pattern or the ground connection, because a small change can move the performance significantly. Where the measurement is taken in a chamber, the mounting and the cable routing must be identical between the two configurations, otherwise the difference reflects the setup rather than the shield. The same measurement discipline used for the checks in EMI immunity design applies to the shield evaluation.

FAQ

Does a shield have to be soldered? No. Clips and conductive adhesive are used where rework access matters more than the best possible continuity.

Can ventilation holes be added to a shield? Yes, as many small openings rather than a few long slots, and the pattern should be checked against the frequencies involved.

Why does a shielded circuit run hotter? Because the enclosure restricts convection. The thermal design has to account for the trapped air and provide a path for the heat.