Gold Finger PCB: Plating and Fabrication Guide

A gold finger is the row of exposed contact strips along the edge of a board that plugs into a slot. The name comes from the plating, which is usually hard gold over nickel, and the reason for that plating is the same reason a connector contact is plated: the surface has to survive thousands of insertion cycles without wearing through or building up an insulating film. A card edge is a connector that is part of the board itself, and designing one means designing a connector.

Why Hard Gold Rather Than a Solder Finish

A solderable finish is designed to be wetted by molten solder once. A card edge needs something quite different: a hard, wear resistant surface with low and stable contact resistance, which does not oxidise in storage and which does not cold weld to the mating contact. Gold provides all of that, and the hardness comes from alloying it with a small amount of cobalt or nickel rather than from the gold itself.

The plating is applied over a nickel barrier. The nickel stops the copper from diffusing into the gold and provides the mechanical support that keeps the thin gold layer from being pushed aside by the mating contact. The gold thickness is a wear specification, not a conductivity specification, and it is chosen from the number of insertion cycles the product will see. A few hundred cycles needs much less than a memory module that is swapped repeatedly.

Edge Geometry and Beveling

The edge of the board is chamfered so that it enters the mating connector without catching. The chamfer angle and depth are specified on the fabrication drawing, and they are what allow the board to be inserted by feel rather than by sight. A board with a square edge will either damage the connector contacts or fail to enter at all.

The finger geometry matters as much as the plating. Fingers are spaced to match the mating connector, and the tolerances on the board outline, the finger width and the finger position all have to be taken from the connector’s own specification rather than chosen. A small error in the board outline position shifts every finger relative to the connector, and the result is intermittent contact on the outermost fingers. Our component tolerance and reliability notes describe how that tolerance stack is assessed.

Fabrication Sequence

Gold fingers are formed by the same imaging and etching process as any other copper feature, but the plating is handled differently. The fingers are connected to a common rail on the panel so that they form a continuous conductive path for electroplating, because electroplating requires current to flow to every feature being plated. That rail is removed at the end, either by routing or by a final etching step.

The plating itself is applied before the solder mask in most processes, and the area is masked during the subsequent operations so that the gold is not contaminated. This is why a gold finger board has a different process flow from an ordinary board, and why the specification has to state the plating thickness explicitly rather than relying on the standard finish.

Panel Design and Routing

Gold fingers sit on the board edge, which means the panel has to be arranged so that the fingers can be plated and then separated cleanly. A board placed at the edge of a panel cannot have fingers along that edge, so the panel layout is driven by the requirement that each board has its finger edge free. That usually means fewer boards per panel and a higher unit cost, which is one of the reasons card edge products are more expensive than their circuitry would suggest.

Where the fingers run along one edge, the copper routing on the board has to approach them from the side away from the chamfer, and the spacing between the fingers limits how many traces can escape into the board. A dense finger pattern usually forces the first routing onto an inner layer almost immediately, which in turn drives the layer count up.

Design Rules That Keep Contacts Reliable

Keep the gold area free of solder mask and of any exposed copper other than the fingers themselves. Leave the fingers unconnected in the schematic sense until the plating rail is removed, and confirm with the fabricator which end of the board the rail will be removed from so that the copper around it is planned. Keep the fingers short relative to their width, because a long narrow finger is more likely to be damaged during handling.

Protect them during assembly as well. A gold finger board should be handled by its edges and stored in a sleeve, and if the assembly process moves the board through equipment that touches the fingers, the plating will be worn before the product is ever used. Our design release checklist covers the handling requirements that should be documented.

Testing and Inspection

Gold thickness is measured by x-ray fluorescence on each panel, and the hardness of the deposit is a process parameter that a supplier should be able to report. Visual inspection looks for nicks, scratches and plating voids along the finger row, and for gold that has bled under the solder mask during processing, which indicates a mask adhesion problem.

Electrically, the fingers are part of the normal test network, and the continuity through each finger is verified in the same pass as everything else. Where the product is critical, a contact resistance measurement on a sample confirms that the plating was applied correctly, because a finger that is electrically continuous can still have a contact resistance high enough to affect the signal.

Bevel, Chamfer and Insertion Mechanics

The bevel on the leading edge of a card is what allows it to enter the connector at all. The angle is specified on the fabrication drawing, and it has to be matched to the connector: too shallow and the board will not guide itself into the slot, too steep and the bevel reduces the usable length of the finger. Because the bevel is cut after plating in most processes, it must not extend into the plated area, or the gold will be cut away at the point where the contact first touches.

Insertion force rises with the number of fingers and with the plating thickness, and the mechanical design of the enclosure has to provide something for the user or the assembly operator to push against. A card that can only be inserted by pressing on the components will eventually be inserted by pressing on the components, and the parts nearest the edge will be the ones that fail first. Our thermal management notes describe how the heat from the card is carried into the chassis through the same connector region.

Storage, Handling and Field Wear

Gold fingers are damaged more often before assembly than during service. Boards sliding against one another in a box, handling with bare hands and repeated test insertions at the factory all wear the plating, and the wear is invisible until the contact resistance rises. Storage in a sleeve, handling by the non-plated edges and a limit on the number of test insertions are the practical controls.

In service the enemy is contamination. Dust and skin oils on the finger row form an insulating film that the wiping action of insertion may not remove, and the symptom is an intermittent fault that appears only after the product has been in use for some time. Where the environment is dirty, a connector with a wiping contact is preferred to a simple edge socket.

FAQ

Why is hard gold used instead of soft gold? Because the contact is a sliding, wearing interface. Pure gold is soft and deforms readily; alloying it with cobalt or nickel produces a hard deposit that resists wear over many insertion cycles.

What is the nickel layer for? It forms a diffusion barrier between the copper and the gold and provides mechanical support, so the thin gold layer is not pushed through or worn away by the mating contact.

How many insertion cycles can a gold finger survive? It depends on the gold thickness and the mating contact. A few hundred cycles is normal for a fixed installation, and thousands are achievable when the plating is specified for it.