Board Level Shielding Effectiveness and Can Mounting

A metal can over a circuit is a simple idea with a complicated set of consequences. It blocks the coupling between the circuit and the outside world, but it also adds a resonant cavity, changes the thermal path and introduces a soldering and rework step. Used well, it solves an electromagnetic compatibility problem cleanly. Used as a late fix, it creates new problems at the same frequency it was meant to suppress. Board level shielding works best when the requirement is defined before the layout rather than after the first failed test.

What a Shield Can Does

A conductive enclosure around a circuit attenuates the electric and magnetic fields that would otherwise cross it. The attenuation depends on the conductivity and the thickness of the material, on the frequency and on the integrity of the enclosure, and the last of those is usually the limiting factor.

The can is also a mechanical part. It protects the components from handling, it defines a controlled environment and it can carry a heatsink or a thermal pad. Those secondary functions often decide the mounting method as much as the shielding requirement does, and a can that has to be removed for rework cannot be soldered on every edge.

Shielding Effectiveness and Frequency

Shielding effectiveness is quoted in decibels as the ratio of the field without the shield to the field with it. It falls as the frequency rises for a given material thickness, and it is dominated by the openings rather than by the metal once the openings are large compared with the wavelength.

A can that is specified at a single figure is therefore an oversimplification. The requirement should be stated across the frequency band that matters, with the most demanding point identified, because the design is set by the worst case rather than the average. Board level shielding specifications should therefore name the band and the limiting frequency.

Apertures and Their Effect

An aperture in a shield behaves like a slot antenna, and it couples energy efficiently when its longest dimension approaches a half wavelength. Several small openings are therefore much better than one large opening of the same total area, which is why ventilation and inspection features should be broken into an array of small holes.

The seams are the other weak point. A joint between two parts of the enclosure, or a gap where the can meets the board, can radiate as effectively as a deliberate opening. Overlapping seams and a close contact pitch are what keep the shield behaving as a continuous surface. A long narrow gap between the can and the board behaves like a slot radiator.

Cavity Resonance

The space enclosed by the can and the board is a cavity, and it resonates at frequencies set by its dimensions. At resonance the field inside can be much larger than outside, which can increase the coupling to the enclosed circuit rather than reduce it. The cavity is therefore a design feature to be managed rather than a side effect.

The practical answer is to place the resonance where it does not matter, to break the cavity into smaller volumes with internal walls or dividers, or to absorb energy with a lossy material. Our EMI immunity notes describe how the same considerations apply to the immunity side of the design.

Grounding the Can

The can has to be connected to the reference ground at multiple points, and the spacing between those points sets the frequency above which the connection stops behaving as a short. A single connection at one corner leaves the rest of the perimeter floating and turns the can into a coupled element rather than a shield.

The ground pads also have to be connected to the reference plane with vias, not merely to a local copper area. A can grounded to an isolated copper island provides no shielding at all, which is a failure that is invisible until the product is measured. Every ground pad should have vias to the reference plane directly beneath it.

Mounting Options: Clips, Solder and Adhesive

A soldered can is connected electrically along its whole perimeter, which gives the best shielding and the most robust mechanical result. It is also the hardest to remove, so it is unsuitable where rework is expected.

A clip mounted can uses spring fingers that press on ground pads, which allows removal and rework at the cost of a contact that can oxidise or lose its spring force. A conductive adhesive or gasket is a third option, used where soldering is not possible, and it depends on the surface preparation and on the compression the mechanical design provides. The mounting choice is therefore a reliability decision as much as an electromagnetic one. Our ferrite bead notes describe how the coupling that remains can be reduced with components when the mechanical shielding is limited.

Solder Process Considerations

A soldered can is a large thermal mass on a thin board, and it changes the reflow profile locally. The perimeter joints need enough heat to wet the ground pads while the rest of the board stays inside its own window, which often means a longer soak and a careful distribution of the ground pads.

The can also shields the enclosed circuit from the reflow heat, which is an advantage for a temperature sensitive part and a disadvantage for a joint underneath it that needs to reach liquidus. That conflict should be resolved at design time rather than at the oven.

Thermal Effects of a Can

A sealed metal can traps heat, and a power device under it will run hotter than the same device on an open board. The thermal path has to be provided deliberately, usually with a thermal pad or a gap filler to the inside of the lid, and the lid then has to be coupled to the outside. A can with no thermal path turns a marginal design into a failing one.

The pad also has to be considered from the shielding point of view, because a conductive pad can create a coupling path between the enclosed circuit and the lid. Our solder defect notes cover the joint failures that appear when a large metal part is soldered to a thin board.

Verification and Inspection

Shielding effectiveness is verified by measurement in a chamber or in a near field setup, and the measurement should be made on the assembled product rather than on the bare board. Near field scanning can locate the leak that matters and show whether it is a seam, an aperture or a ground connection.

Production inspection concentrates on what can be seen: the presence and the wetting of the perimeter joints, the planarity of the can and the absence of a short to a neighbouring feature. Our quality guide describes how a can mounting defect is classified at gopcb.

FAQ

Is a thicker can always better? Thickness helps at low frequency, but above the point where the openings dominate it makes little difference. The aperture and seam design usually matters more than the material thickness.

Can a clip mounted can match a soldered one? It can come close if the contact pitch is tight and the pads are clean, but the contact degrades with time and with contamination. For a strict requirement, a soldered perimeter is the safer choice.

Why does the shield sometimes make the problem worse? Because of cavity resonance or a poor ground connection. The enclosed volume can amplify the field at a particular frequency, and a can grounded at only one point behaves as a coupled element rather than a shield.