RF Shield Can Design And Attachment

A shield can is a metal cover placed over a circuit that would otherwise radiate or receive energy. It is also a mechanical part: it has to be attached to the board, to survive the assembly process, and to allow the components underneath to be inspected and, when necessary, replaced. The shielding it provides depends less on the metal than on the way the metal is connected to the ground of the board.

This article explains how a shield works, how the can types differ, how the attachment and grounding are designed, and what has to be checked after assembly.

How A Shield Works

Shielding is mostly reflection with some absorption. The metal presents an impedance discontinuity to an incoming wave and reflects most of the energy, and the small part that penetrates is attenuated as it passes through the material. A solid metal can of a reasonable thickness is an excellent shield on its own, and the performance of a real enclosure is decided almost entirely by its openings.

An aperture in a shield radiates or admits energy efficiently when its longest dimension approaches a half wavelength of the frequency of interest, so a single long slot is far worse than a row of small holes with the same total area. That is why a shield often looks like a perforated box rather than a solid one, and why the joint between the can and the board is the critical feature: a gap in the attachment is a slot, and a long, thin gap behaves like an efficient antenna.

The frequency range that has to be shielded sets the physical size of every feature. A slot of a centimetre is a small fraction of a wavelength at a few hundred megahertz and a significant one at several gigahertz, so a shield that is acceptable for a power converter is not necessarily adequate for a radio. The useful exercise is to take the highest frequency that carries significant energy, convert it to a wavelength, and dimension the apertures, the via spacing, and the can height from that figure rather than from a general impression of what looks solid.

Can Types

A one piece stamped can is soldered directly to pads on the board. It is the simplest and the cheapest, it has the fewest apertures, and it is the most difficult to remove, which makes rework of a component underneath a destructive operation. A two piece design separates the frame from the lid: the frame is soldered or otherwise attached to the board and the lid clips or screws onto it, which allows access for inspection and rework at the cost of a second part and of a joint that has to make good contact.

A clip on can uses spring fingers that press against ground pads, which makes it removable without solder, and a shield with a conductive gasket or foam is used where the can also has to seal against a housing. Each option trades shielding performance, cost, height, and reworkability, and the choice follows from how likely an intervention under the can will be.

Attachment And Grounding

The can is grounded along its perimeter, and the quality of that grounding is what sets the shielding. A can that is tacked at two or three points and open along the rest of its edge has a long slot between the metal and the board, and that slot radiates at a frequency set by its length. The correct arrangement is a continuous ground ring under the can wall, or a dense row of pads that the solder bridges into an effectively continuous connection.

The ground ring has a second function, which is to provide a path from the can to the ground plane inside the board. The ring alone sits on the surface, separated from the inner plane by the dielectric, and the connection is made with vias. The via stitching should be placed at intervals along the ring, close enough that the distance between them is a small fraction of the wavelength of the highest frequency of concern, because the inductance of a long path between vias is what allows the shield to leak. The design principles behind that arrangement are described under EMI suppression design principles, and the switching regulator case under radiated EMI from a switching regulator.

Layout Rules Under The Can

The components that will be covered are placed inside the ring with a clearance to the wall, because the wall has to be soldered and because a component that touches the can creates a short. The area also has to be reachable by the placement nozzle, so the wall clearance has to accommodate the pick and place head as well as the solder fillet. Test points under a can are of no use unless the can is removable, and a design that puts the only test access inside a soldered can has removed the ability to test the product.

The height of the can is a mechanical constraint on the assembly. It has to clear the tallest component inside it, it adds to the total height of the product, and it affects the thermal behaviour of the components it covers. A can prevents convection from carrying heat away from the parts beneath it, so a temperature measurement inside a shielded area is worth taking during development.

Assembly And Rework

A can that is soldered in the reflow oven has to be placed like a component. Its mass and its shape affect the placement and the reflow, and it can shift during the process if the paste does not hold it, so the footprint and the paste volume are designed for the part rather than for a generic pad. Where the can is large, the assembly may be placed after the main reflow and attached with a separate soldering step, which avoids the risk of moving it and allows the interior to be inspected first.

Cleaning is the other process consequence. The flux under a soldered can cannot be washed out afterwards, so a no clean process has to be qualified for the residue that will remain inside the enclosure, and a product that needs a cleaned assembly has to be cleaned before the can is attached. The general assembly constraints are described under manufacturable design guidelines.

Verification

Three checks close the loop. The electrical one is a continuity measurement from the can to the ground plane at several points around the perimeter, which finds a leg that was not soldered. The functional one is a shielding effectiveness measurement, taken on a representative assembly, which confirms that the apertures and the joints perform as designed. The mechanical one is a visual inspection of the fillet and a check that the can does not touch a component or a trace on the board.

FAQ

Does a shield can have to be soldered all the way around? It has to be grounded all the way around. Solder is the usual method, but a spring finger or a gasket can serve the same purpose if it makes continuous contact with the ground ring.

Why are holes better than slots in a shield? Because an aperture couples efficiently when its longest dimension is a significant fraction of a wavelength. Many small holes have the same open area as one slot and couple far less energy.

Can the flux under a soldered can be cleaned? Not after the can is attached. The process has to be designed so that the residue inside the enclosure is acceptable, which usually means a qualified no clean process.