PCB Edge Plating: Grounding, Shielding and Card Edges

Edge plating is copper that wraps around the outside of a circuit board, connecting the surface layers to the edge of the board. It is used to ground a board to a chassis, to close the seam between a shield and an enclosure, to make a card edge that plugs into a socket, and to carry current around the perimeter of a power board. It is also one of the more demanding processes to specify, because a plated edge is defined as much by the routing that follows the plating as by the plating itself.

What Edge Plating Is

A copper track that runs to the board outline would normally be cut away when the board is routed out of the panel, exposing bare laminate and leaving the copper flush with the edge or slightly recessed. Edge plating changes that by depositing metal on the cut edge itself, so the copper on the top surface, the metal on the edge and the copper on the bottom surface become one continuous conductor.

The result is a conductive rim around part or all of the board. It can be narrow, covering only a small strip where a contact is needed, or continuous around the whole perimeter where the board has to close an opening in a metal enclosure.

Three other features are often mentioned alongside edge plating, and the differences matter.

Castellation is a row of half holes on the edge, used to solder a module onto a carrier. The holes are plated before they are cut, so the castellation is really a cross section of a plated barrel rather than a plated edge.

A card edge connector is a continuous strip of plated copper on the edge, used for insertion into a socket. It is edge plating applied for a specific purpose, with its own thickness and wear requirements.

Half holes and edge castellation are related to castellation and are covered in our article on castellated holes, which sets out how the cut through a plated hole behaves.

Why It Is Used

Four reasons account for most edge plating.

Grounding to a chassis. A board that mounts into a metal case needs a low impedance connection between its ground and the case. A plated edge pressed against a machined ledge or a spring finger gives a continuous connection along the whole side rather than a connection through one screw.

Shielding seams. Where a shield can or a metal cover meets the board, the seam between the cover and the enclosure has to be closed. An edge plated board that fits closely into a metal housing turns the housing into part of the enclosure, which is more effective than a can alone. Our article on EMI shield cans covers the related geometry.

Current carrying. On a power board, an edge plated strip can carry current around the perimeter where a wide trace would not fit, and it can also spread heat from the perimeter into a heatsink or a chassis wall.

Contact and connection. Card edge contacts and the solderable edge of a module both rely on plating that continues around the corner, so that the contact surface is not interrupted by the laminate edge.

How It Is Made

Plating a board edge is not the same as plating a hole. The hole is already surrounded by copper and the chemistry has a defined path; the edge is a cut through laminate, and the plating has to be applied to a surface that was not designed to be conductive.

The standard approach is to plate before the final routing. The panel is processed with copper running all the way to the board outline, and the copper connects to a plating bus on the panel frame. After plating, the edge carries metal along its full height, and the final routing operation trims the board to size, leaving the plated edge in place on the finished outline.

The alternative is to route first and then plate, which requires the boards to be held in a frame and re-contacted individually. It is used where the geometry makes the first approach impossible, and it costs more.

Two process details decide the quality of the result. The first is the plating bus and how it is removed: the copper that fed the plating current has to be cut away without tearing the edge plating off the board. That is done by routing a small step, and the tolerance on that step is what determines whether the finished edge is clean or ragged.

The second is the plated thickness and uniformity on the edge. The edge is a narrow surface at the panel periphery, which is where plating current distribution is least even, so the achievable uniformity is worse than on the board surface. A specification that demands a tight thickness tolerance on an edge plated surface should be checked with the fabricator before it is written into the drawing.

Design Rules

• Copper must reach the board outline wherever plating is required, because plating cannot bridge a gap in the copper.
• Define the extent of the plating on a mechanical layer: which edges, how much of each edge, and whether it is continuous.
• Keep the copper land on the surface wide enough to be handled and inspected, typically 0.5 mm or more.
• Keep the plating away from areas where a burr or a raised edge would interfere with assembly or with a mating part.
• Specify the plating thickness and the surface finish for the edge, which may differ from the finish on the faces.
• Where the edge is a contact, define the wear surface: hard gold is usually required for a card edge that will be inserted repeatedly.
• Allow for the routing tolerance on the board outline, and add it to the clearance between the plating and any nearby feature.
• Confirm the plating bus removal method with the fabricator, and ask what the resulting edge tolerance will be.

Two of these are repeatedly ignored. The copper-to-edge requirement is absolute, so a design that has a mask-defined clearance around the outline will not plate. And the routing tolerance has to be added to the land width, because the router will remove some of the copper on the surface as it trims the board.

Choosing Between the Edge Features

The four edge features solve different problems, and picking the wrong one produces a board that is more expensive than it needs to be.

• Edge plating (a continuous strip): for grounding to a chassis, closing a shielding seam, or carrying current around the perimeter.
• Castellation (a row of half holes): for soldering a module to a carrier board, where each half hole is one connection.
• Card edge contacts: for repeated insertion into a socket, with the finish and thickness chosen for wear.
• Press fit: for a connector pressed into plated holes, where the hole and barrel do the work rather than the edge.

A design can use more than one on different edges, and this is common on a power module that has castellations for its signals and an edge plated strip for a chassis ground. They should be specified separately on the drawing, because each has its own process and its own tolerance.

Assembly and Mechanical Fit

Most edge plating failures in the field are mechanical rather than electrical, and they come from the interface with the housing.

Where the plating presses against a machined surface, the contact force has to be defined. A board held by screws with no compliance can lose contact as the assembly ages, and a board held by spring fingers can wear through the plating if the surfaces move against each other. The usual solution is a defined compression through a spring element or a conductive gasket, with the plating finish chosen for wear rather than for solderability.

Where the plating forms part of a shielding seam, the gap between the plated edge and the housing has to be small compared with the wavelength of interest, exactly as it does for a shield can. A gap of a millimetre is a slot, and a slot radiates.

The third mechanical factor is flatness and burrs. A board edge that has been routed can carry a small burr, and a burr on a plated edge is both a handling hazard and a contact problem. The specification should say whether the edge has to be deburred, and the fabricator should be asked what they can achieve.

Finally, the finish matters for the assembly process: an edge that will be soldered needs a solderable finish, and an edge that will be pressed against a gasket needs one that resists wear and corrosion. The two requirements rarely coincide, which is a good reason to decide which edges are doing which job before the stackup is released.

Cost and Tolerance

Edge plating is a process step with a tooling implication rather than a routine part of the flow, so it carries a premium that depends on how much of the perimeter is involved.

A single plated strip on one edge is the cheapest case, since it can be handled with the existing plating bus arrangement. Plating on all four edges, or a plated edge with a tight tolerance, costs more because the routing step has to be controlled more closely and the yield on the edge is lower. Where the edge is also a wear surface, the hard gold plating adds a separate operation.

The tolerance to expect is looser than on the board surface. The plating thickness on the edge is less uniform, and the routing that defines the edge has its own positional tolerance, so a specification of plus or minus 0.1 mm on the edge position is realistic while a much tighter figure should be justified. Asking the fabricator for the achievable tolerance before writing the drawing avoids a specification that cannot be inspected.

FAQ

• Does edge plating require copper to the board edge? Yes. The plating will only form where the copper reaches the outline, which is why the copper design is part of the requirement.
• Is edge plating the same as castellation? No. Castellation is a row of half holes for soldering; edge plating is a continuous strip of metal on the cut edge.
• Can edge plating be applied to part of an edge only? Yes, and partial plating is common where only a contact area needs metal.
• Does edge plating improve EMC? It does when the plated edge closes a seam against a metal housing or a shield. Used without a mating surface it has little effect.
• Can the edge be soldered? Yes, if the finish is solderable, which is a separate specification from the finish used on the board faces.

Summary

Edge plating wraps the circuit onto the board outline. It exists to ground a board to a chassis, to close a shielding seam, to carry current around the perimeter and to make a contact surface that continues around the corner, and each of those uses has its own finish and tolerance requirements.

It is a fabrication feature that depends on the design: copper has to reach the edge, the extent of the plating has to be drawn, and the routing tolerance has to be allowed for. Where that is done, the result is a board that fits into a metal housing as part of the mechanical and electrical design rather than as an afterthought. Confirming the achievable tolerance and finish with the fabricator during the layout stage is the step that prevents an expensive surprise, and the capabilities review is where the mechanical fit with the enclosure should be checked.