PCB Short Circuits: Tracing the Cause in Fabrication

A short circuit found after fabrication is expensive, but the conditions that produce one are usually visible in the process long before the boards reach electrical test. Most shorts originate in the wet process, where a thin residue of metal survives the etching step and bridges two conductors that were meant to be separate. Understanding where that happens makes it possible to prevent rather than to sort out.

How a Short Survives the Etching Step

Etching removes the copper that is not protected by the resist pattern. If a thin layer of another metal remains on the surface, that layer can act as a mask and prevent the copper beneath it from being removed. The result is a narrow bridge between two conductors that looks intact and passes a visual check.

The mechanism is a residue problem rather than an etching problem. Tin from the plating or stripping operation is the usual culprit: it deposits on the copper surface, resists the etchant, and leaves copper in place that should have been etched away. Once the residual metal is understood as the cause, the controls that prevent it become obvious.

Stripping Chemistry and Handling

The stripping solution removes the plating resist after the pattern has been plated. Three parameters control whether the removal is complete: the concentration of the solution, its temperature and the time the board spends in it. A concentration that is too high, or an immersion that is too long, attacks the resist before it has been fully removed, which leaves fragments behind. Those fragments, or the metal they hold, can then interfere with the following steps.

Handling during the same stage matters as much as the chemistry. Boards stacked together in the stripping tank trap solution and debris between them, so the faces that are in contact are not exposed properly. Racks that hold the boards separately, with no face-to-face contact, remove the problem at its source. Our etching process notes cover the sequence around this step.

Wet Boards and Transferred Metal

A related cause appears when boards are stacked while still wet. Solution retained between two wet surfaces dissolves metal from one and redeposits it on the other, and the transferred metal then protects the copper underneath it during etching. Preventing it requires drying the boards before they are stacked, and stacking them only when they are dry and separated.

Both of these mechanisms produce the same defect signature, which is what makes the diagnostic step useful: a short that appears on many boards from a batch, with a characteristic appearance, points to a chemical or handling control problem rather than to a design issue.

Incomplete Etching From Uneven Plating

The second family of shorts comes from plating thickness variation. If the plated copper is much thicker in one area than another, the etching step may remove the thinner regions completely while leaving copper in the thicker ones. The leftover copper appears as a short or as a partial feature, and the defect is often accompanied by traces that are narrower than specified.

The remedy lies in the plating process: an automated line with consistent current density, current adjusted according to the plated area, equal plating times for every panel, and shielding arranged so that the potential difference across the panel is small. Where plating is done manually, the practical measures are to keep the current density uniform, to use double clamps on large panels, and to make the timing explicit rather than estimated. Our plating thickness notes cover the thickness requirement, and the throwing power material explains why the distribution is never perfectly even.

Alkaline Etching Parameters

Where an alkaline etchant is used, the parameters are narrow enough that they should be controlled continuously rather than checked occasionally. The pH is held in a band around eight to nine; a lower value thickens the solution and slows the etch, which promotes undercutting and incomplete removal. The chloride content is held within a defined range using an etching salt and a replenisher. The specific gravity, which reflects the dissolved copper content, is maintained within a band and measured at short intervals. The temperature is held in the upper forties to low fifties Celsius, both because the etch rate depends on it and because the tank material has a thermal limit that a runaway temperature will exceed.

Etch rate also has to be matched to the copper thickness being processed. A speed that is right for a thin foil will under-etch a thick one, and the copper left behind is a short. Where several foil weights are processed on the same line, the parameters should be recorded per product rather than adjusted by feel.

Diagnosing a Short After the Fact

When a short is found, the first question is whether it is a bridging defect from fabrication or a genuine design error. A short that exists on every board from a batch, at a location where the artwork clearly separates the conductors, is a fabrication defect. A short that appears at a location where the artwork is marginal, such as a spacing below the process capability, is a design issue.

The second question is whether the defect is a residue-type short or an etching-type short, because the corrective action differs. A residue short disappears when the stripping and drying controls are corrected; an etching short requires the plating uniformity to be addressed. Our short circuit inspection notes describe how to localise the fault, and the design release checklist covers the spacing checks that prevent the marginal cases.

Design Contributions to the Problem

Not every short is a process failure. A spacing that is below the capability of the process will produce bridges intermittently even when every parameter is in range, and the defect will be attributed to the line rather than to the artwork. The same applies to a copper feature that is isolated from the rest of the pattern, which can lift and fold during processing, and to an annular ring that is at the registration limit. Checking the design against the fabrication capability before release removes those contributors and leaves the process controls to deal with the rest.

FAQ

Can a short be repaired? It depends on the cause and the location. A localised bridge on a conductor that is not part of a fine-pitch pattern can often be removed, and the board can then be retested. A short caused by residual metal inside a fine-pitch area is difficult to clear without damaging adjacent features, and the board is usually scrapped. Repair decisions should be based on the cause, not only on the location.

Why do shorts appear on only some boards of a batch? Because the mechanisms are process excursions rather than systematic errors. A stripping solution that is slightly out of specification, a stack of boards that stayed wet a little longer, or a plating current that drifted during one panel will affect some boards and not others. That is also why the defect rate tends to rise gradually before anyone notices, which makes the parameter records the most useful diagnostic tool.

What does gopcb check to prevent shorts? We control the stripping concentration, temperature and dwell time, dry the boards before stacking, and monitor the etching parameters within their bands so that the etch is complete for the copper weight being processed. Because a short found at electrical test costs the whole panel, those controls are treated as process discipline rather than as inspection.