PCB Failure Analysis: From Symptom to Root Cause

Failure analysis is the discipline of finding out why a board failed rather than guessing. It matters because the alternative, replacing a component or adjusting a process until the symptom disappears, leaves the cause in place and usually brings the problem back in a different form. A structured approach moves from what was observed to what actually happened, and it produces evidence that can be acted on.

Start With the Symptom, Not the Suspect

The first task is to describe the failure precisely: when it appears, how often, under what conditions, and which boards are affected. A failure that appears only at temperature, only under load or only on one panel position carries a great deal of information. A failure described only as not working carries almost none.

From that description, build a list of candidate mechanisms and rank them by how well they explain the pattern. This is the stage where premature conclusions do the most damage, because a plausible suspect that is investigated first can consume the whole budget while the real cause sits a few steps away.

Non Destructive Examination First

Before anything is cut, examine what is already visible. Optical microscopy reveals cracks, contamination, discolouration and damage from handling. X-ray imaging shows hidden joints, voids and barrel defects without disturbing the assembly, and a thermal camera can localise a hot spot to a small area of the board.

Electrical measurements follow, and they are most useful when they are localised: resistance measurements between specific nets, a curve trace on a suspect pin, or a voltage measurement compared with a known good board. The comparison board is one of the most valuable tools available and the most frequently neglected.

Destructive Analysis and Cross Sections

When the external examination has narrowed the possibilities, a cross section answers questions that no surface technique can. Mounting a section of the board in resin, grinding and polishing it, and examining the result under a microscope shows the internal geometry of a joint, the plating thickness in a barrel, the presence of voids, and the interface between materials.

Sectioning is destructive and slow, so it should be targeted. Choose the location from the evidence rather than randomly, and prepare more than one section if the mechanism is expected to vary across the panel. Our notes on copper plating defects and prevention describe the defect morphologies that a section is used to confirm.

Chemistry, Contamination and Materials

Not every failure is mechanical. Residues, corrosion products and contamination are analysed by techniques such as energy dispersive X-ray spectroscopy or ion chromatography, which can identify the elements and ions present at a failure site. That identification often distinguishes between a process problem and an environmental one.

Material analysis follows when the failure suggests a bulk property: a laminate that delaminated, an adhesive that released or a solder alloy whose composition has drifted. Comparing a failed sample with a reference sample is what makes these results interpretable. The electrochemical mechanisms that produce failures long after assembly are described in our notes on HDI failure mechanisms.

Reproducing the Failure

A failure that can be reproduced under controlled conditions is one that can be fixed with confidence. Reproduction may mean a thermal cycle, a humidity chamber, a specific mechanical stress or an electrical condition that pushes a marginal design over the edge. Where the failure appears only in the field, reproducing it may require understanding the service environment rather than the bench.

Where reproduction is not possible, the analysis has to rely on the physical evidence alone, and the conclusion should be stated with the corresponding degree of uncertainty. Claiming a root cause without a reproducible mechanism is how the same problem appears again six months later.

From Cause to Corrective Action

A root cause is only useful if it leads to an action. The action should address the mechanism rather than the symptom, should be testable, and should include a way to verify that it worked. Where the cause is a design issue, the fix belongs in the design; where it is a process issue, it belongs in the process control plan.

Verification closes the loop. Run the affected product through the process, apply the same test that reproduced the failure, and record the result. That record is what allows the organisation to know, rather than assume, that the problem has been resolved.

Recording What Was Learned

Failure analysis produces knowledge that is worth keeping. A short report describing the symptom, the investigation, the evidence and the conclusion turns one investigation into a reference for every similar product that follows. Without it, the same analysis is repeated by a different engineer with a different board.

The record also feeds back into design reviews. A checklist item that says, for example, to verify the thermal pad paste pattern after a past failure is worth more than a general instruction to be careful, because it names the specific mistake that experience has shown to be likely.

Process Control and Verification

Reviewing the design before the data is released is far cheaper than correcting it after the panel is in the tank, because every step downstream inherits the decision made at the front end. Where a value sits close to a process limit, the drawing should say so, since the shop can then open the process window rather than working to a nominal figure that carries no tolerance.

The measurements that matter are the repeatable ones: conductor width and spacing, annular ring, finished hole size, plating thickness and surface finish are all verifiable on a coupon that travels with the panel. Documenting the assumption is part of the design work, and a short note on the drawing prevents a question that would otherwise arrive a day later and cost a day of schedule.

Where a value sits close to a process limit, the drawing should say so, since the shop can then open the process window rather than working to a nominal figure that carries no tolerance. Running a first article through the same checks as the production panel confirms that the two agree, and that comparison is the cheapest form of process control available at prototype stage.

Related reading: our fabrication notes, board quality and design release notes cover the same ground.

FAQ

How long should a failure analysis take? The investigation should be proportional to the impact. A single field return with a clear pattern justifies a full analysis, while an isolated unit with no safety implication may only warrant recording. What matters is that the decision is deliberate rather than a function of how busy the team happens to be.

Should I always cross section a failed joint? No. Sectioning is justified when the external evidence points at an internal mechanism and when knowing the exact geometry changes the corrective action. Where the cause is already clear from other evidence, the section adds cost without adding information.

What is the most common mistake in failure analysis? Fixing the first plausible cause. A board that fails at temperature has several possible explanations, and stopping at the first one that seems reasonable often replaces a temporary symptom with a permanent misunderstanding. Enumerating the candidates and testing them in order of likelihood is slower at first and faster overall.