PCBA Failure Analysis: Root Cause Investigation, Verification and CAPA

A failed printed circuit board assembly does not automatically reveal why it failed. A product that will not power on may have a damaged component, an open solder joint, incorrect firmware, an unstable power rail, a test-fixture problem, or a defect introduced during handling. Effective PCBA failure analysis is a structured investigation used to move from a reported symptom to a verified root cause. It combines failure reproduction, production-history review, visual inspection, electrical troubleshooting, nondestructive analysis, destructive analysis where necessary, and controlled verification.

For US OEMs, the quality of the investigation matters as much as the speed of the response. Replacing a component may restore one board, but it does not prove why the failure occurred or whether other units are at risk.

A complete investigation should answer five questions:

What failed?
How did it fail?
Why did it fail?
Which products or lots may be affected?
What action will prevent recurrence?

Main content

What Is PCBA Failure Analysis?

PCBA failure analysis is the systematic process of identifying the physical, electrical, material, design, manufacturing, or operational cause of a circuit-board assembly failure.

The process normally includes:

Confirming the reported failure
Preserving the original sample condition
Collecting product and production records
Reproducing the symptom
Localizing the faulty circuit or component
Identifying the failure mechanism
Developing possible root-cause hypotheses
Testing those hypotheses
Defining containment and corrective action
Verifying that the action is effective

A structured investigation may use bench electrical testing, visual inspection, X-ray, computed tomography, cross-section analysis, microscopy, material analysis, and manufacturing-data review. An IPC technical paper on a structured approach to failure analysis and root-cause determination demonstrates how several methods can be combined to narrow a complex PCB problem to a defensible cause.

Failure analysis is different from routine production inspection. Inspection determines whether a unit meets defined acceptance criteria. Failure analysis investigates why a unit did not meet those criteria or stopped performing as intended.

Failure Symptom, Defect, Failure Mechanism and Root Cause

A strong report must distinguish between the observed symptom and the underlying cause.

Term	Meaning	Example
Failure symptom	What the user or test system observes	The product does not power on
Defect	The physical or electrical nonconformity found	An open solder joint is present
Failure mechanism	The process through which performance was lost	A solder crack created an intermittent connection
Root cause	The underlying condition that produced the mechanism	Excessive board flex during depanelization damaged the joint
Contributing factor	A condition that increased the likelihood or severity	The component was located close to the panel breakaway tab

Finding a cracked solder joint is not always the end of the investigation. The crack may have resulted from:

Inadequate solder formation
Thermal cycling
Mechanical shock
PCB bending
Component-package interaction
Poor pad design
Rework damage
Inappropriate depanelization
Excessive operating temperature

The investigation should continue until the evidence explains both the physical defect and the conditions that created it.

When Is PCBA Failure Analysis Required?

Formal analysis may be needed when:

A first article does not meet requirements
Production yield falls unexpectedly
ICT or functional failures repeat at the same circuit location
An intermittent problem cannot be reproduced consistently
A customer returns a failed unit
Several failures occur within the same production lot
A product fails under temperature, vibration, humidity, or electrical load
A previously stable product develops a new failure pattern
A component supplier, process, design, or firmware revision changes
The potential consequence of failure is significant

Not every isolated production defect requires a laboratory-level investigation. However, repeated, safety-related, field, or unexplained failures generally require a documented process beyond routine repair.

A review of the existing PCBA testing process is often the first step because test coverage, fixture condition, test limits, and historical results can help determine whether the problem originated in the product or in the verification system.

The PCBA Failure Analysis Process

The exact sequence depends on the failure, but a disciplined investigation usually follows the stages below.

Step 1: Preserve the Sample and Collect Failure Information

The failed board should be preserved before anyone cleans, repairs, reflows, or replaces components.

Early handling can destroy evidence such as:

Flux or contamination residue
Burn marks
Corrosion
Cracked solder
Foreign material
Loose hardware
Connector damage
Intermittent contact
Signs of previous rework

Photograph both sides of the assembly and any relevant connectors, cables, enclosures, labels, or damage. Record who received the sample, when it arrived, and what actions were performed.

Product identification

Collect:

Product name and model
PCB part number
Assembly revision
Serial number
Production date
Production lot
Bare PCB lot
Component lots where available
Firmware version
Test-program revision
Rework history

Failure information

Document:

Reported symptom
Operating condition
Failure frequency
Whether the failure is permanent or intermittent
Environmental conditions
Applied input and load
Time in service
Customer troubleshooting already performed
Whether the board was dropped, exposed to moisture, or modified
Whether the failure changes with temperature, pressure, movement, or time

The initial problem statement should describe observable behavior rather than assume a cause.

For example:

“The output voltage falls below the specified limit after 20 minutes at rated load.”

This is more useful than:

“The power component is defective.”

The second statement introduces an unverified conclusion before testing begins.

Step 2: Verify and Reproduce the Failure

A reported failure must be confirmed under controlled conditions.

The investigation should determine:

Whether the symptom can be reproduced
Which inputs or operating modes trigger it
How long it takes to appear
Whether the failure is temperature-dependent
Whether movement or vibration changes the result
Whether the condition disappears after a power cycle
Whether firmware or configuration affects it

Verify the test system first

Before concluding that the PCBA is defective, check:

Power supply
Test fixture
Probes and connectors
Cables
Software
Test-program revision
Calibration status
Limit settings
Golden sample
Operator setup

A worn fixture contact or incorrect program limit can create a false production failure.

The difference between electrical verification methods is explained further in ICT versus functional testing.

Compare a failed board with a known-good board

A good-board comparison can reveal:

Different current consumption
Missing power rails
Abnormal waveforms
Temperature differences
Incorrect resistance
Unexpected firmware behavior
Component-marking differences
Variation in startup sequence

The comparison board must be the same approved design and configuration. Comparing different revisions may introduce misleading differences.

Step 3: Conduct Visual and Microscopic Inspection

Visual inspection is usually one of the first nondestructive steps.

Inspect the board under suitable lighting and magnification for:

Missing or shifted components
Incorrect polarity
Solder bridges
Insufficient solder
Cracked joints
Lifted leads
Lifted pads
Damaged traces
Burned areas
Corrosion
Contamination
Bent connector pins
Foreign objects
Mechanical impact
Previous rework
Damaged coatings or labels

A detailed understanding of visible production issues can be supported by reviewing common PCB assembly defects and common SMT and wave-soldering defects.

Inspect the entire assembly

The obvious damaged area may not be the origin of the failure.

For example:

A burned component may be the result of an upstream overvoltage.
A blown fuse may have protected the circuit from another short.
A cracked connector may have been caused by enclosure misalignment.
Corrosion may extend beyond the visibly affected area.

Inspection should include mechanical interfaces, cables, mounting points, heat sinks, shields, fasteners, and any enclosure features that can apply stress to the PCB.

Step 4: Perform Electrical Troubleshooting

Electrical troubleshooting narrows the failure from a system symptom to a circuit, net, component, or interface.

Common checks include:

Continuity
Resistance to ground
Power-rail voltage
Current consumption
Leakage current
Signal waveforms
Clock presence
Reset behavior
Communication activity
Component input and output
Short-circuit localization

Use the schematic and expected sequence

The engineer should understand:

How power enters the board
Which rails start first
Which enable signals are required
How reset is released
Which communication paths are active
What feedback controls the circuit
What expected values should appear at each stage

Replacing components randomly can introduce new damage without identifying the faulty circuit.

Use controlled isolation

Useful methods may include:

Disconnecting loads
Isolating circuit branches
Comparing resistance with a good board
Injecting a controlled voltage
Monitoring thermal response
Swapping a socketed or removable module
Replacing a suspected component only after documenting evidence

Component replacement may support a hypothesis, but it is not sufficient by itself to establish root cause.

If replacing one integrated circuit restores function, the investigation must still determine whether the component failed because of:

An internal device defect
Electrical overstress
Excess temperature
Incorrect surrounding components
Poor soldering
Handling damage
Firmware behavior
An external load

Step 5: Use AOI, X-Ray and Advanced Analysis Where Appropriate

The selected method should match the suspected failure mechanism.

Automated Optical Inspection

Historical AOI images and results may help determine whether a visible defect was present at production.

AOI may provide evidence for:

Missing components
Wrong components
Polarity errors
Placement shift
Visible solder bridges
Some solder-joint abnormalities

However, a current AOI inspection cannot always prove when a defect occurred. Review machine records, program revision, verification results, and repair history where available.

More information is available in SHDC’s guide to Automated Optical Inspection in PCB assembly.

X-ray inspection

X-ray can reveal conditions that are hidden beneath components, including:

BGA solder connections
QFN and bottom-terminated joints
Voids
Bridges
Missing solder
Solder-ball alignment
Some internal cracks
Connector or package features not visible optically

Standard two-dimensional X-ray may not reveal every crack orientation. CT or other methods may be required when the suspected feature overlaps with surrounding structures.

Cross-section analysis

Cross-sectioning can expose:

Solder-joint structure
Intermetallic layers
Cracks
Voids
Pad separation
Plated-through-hole conditions
Via defects
Delamination
Internal laminate damage

Because cross-sectioning destroys the sample, it should be planned after nondestructive evidence has identified the correct location.

Dye-and-pry testing

Dye-and-pry analysis may be useful for identifying cracks beneath BGA or similar packages. Dye penetrates existing fractures before the package is mechanically separated, helping show where joints were cracked before removal.

The process is destructive and requires careful interpretation. Damage created during package removal must not be mistaken for the original failure.

Material and contamination analysis

Depending on the hypothesis, testing may include:

Ionic contamination analysis
Fourier-transform infrared spectroscopy
Scanning electron microscopy
Energy-dispersive X-ray spectroscopy
Metallographic analysis
Surface-residue analysis
Component decapsulation

IPC test methods provide defined procedures for evaluating specific printed-board failure mechanisms. For example, an IPC-TM-650 test method for conductive anodic filament growth addresses electrochemical migration risks associated with materials, design, and manufacturing conditions.

Not every supplier maintains every advanced analytical capability internally. Complex analysis may require an independent laboratory with appropriate equipment and expertise.

Step 6: Identify the Failure Mechanism

The failure mechanism describes the physical or electrical process through which the assembly lost functionality.

Common PCBA failure mechanisms include:

Open circuit
Short circuit
Intermittent connection
Solder fatigue
Brittle solder fracture
Pad cratering
Plated-through-hole cracking
Via failure
Delamination
Corrosion
Electrochemical migration
Electrical overstress
Electrostatic discharge
Thermal damage
Mechanical cracking
Component degradation
Insulation breakdown
Connector wear
Firmware corruption

The same symptom may result from different mechanisms.

For example, intermittent communication may be caused by:

A cracked solder joint
A marginal connector
A damaged cable
A power-rail disturbance
Clock instability
Firmware timing
Electromagnetic interference

The investigation should select the mechanism supported by physical and electrical evidence.

Step 7: Determine the Root Cause

After identifying the mechanism, determine why it occurred.

Root-cause analysis may include:

5 Why analysis
Fishbone diagram
Fault-tree analysis
Process-data review
Design review
Material-lot comparison
Good-versus-failed sample comparison
Timeline analysis
Controlled experiments
Statistical analysis
Supplier investigation

Review the complete production history

Potential evidence includes:

Incoming inspection results
Solder-paste lot
SPI data
Placement records
Reflow profile
AOI results
ICT and FCT results
Repair records
Operator and line
Machine-program revision
Equipment maintenance
Approved deviations
Engineering changes

A board may have passed production testing but failed later because the test did not expose the relevant condition. This is why design for testability should be considered during development rather than after failures begin.

Consider design, process and use conditions together

Root causes may originate from:

Circuit design
PCB layout
Footprint design
Material selection
Component quality
Soldering process
Assembly handling
Rework
Mechanical integration
Test coverage
Packaging
Transportation
Customer use
Environmental exposure

Avoid assuming the cause belongs to the department that first detected the issue.

Step 8: Verify the Root Cause

A root cause should be demonstrated, not merely described.

Verification methods may include:

Reproducing the failure under the suspected condition
Removing the suspected condition and confirming the failure disappears
Repeating the experiment across multiple samples
Comparing affected and unaffected production lots
Rebuilding boards using the corrected process
Testing an updated design
Reviewing performance after corrective action
Conducting environmental or accelerated testing where appropriate

For solder-joint fatigue, IPC-9701B defines a thermal-cycling method used to characterize surface-mount solder-attachment fatigue life. Such qualification testing is different from routine troubleshooting, but it may help validate a design or process correction when thermal-mechanical reliability is involved.

Example of root-cause verification

Suppose returned products contain cracked ceramic capacitors close to the PCB edge.

An initial hypothesis may be excessive bending during depanelization.

Verification could include:

Documenting crack location and orientation.
Comparing failures with tab or V-score positions.
Measuring board strain during the current separation process.
Separating new samples using the existing method.
Inspecting the capacitors.
Moving the tab, changing the tool, or increasing component clearance.
Repeating the experiment.
Confirming the crack no longer occurs under the revised condition.

This evidence is stronger than stating, “The capacitor was damaged during production.”

Determine the Affected Scope Through Traceability

Failure analysis must identify which other units may share the same risk.

Relevant traceability data may include:

Product serial number
PCB lot
Component lot and date code
Material supplier
Production date
SMT line
Machine program
Reflow profile
Test program
Firmware version
Rework history
Shipment lot

IPC-1782B establishes risk-based minimum requirements for manufacturing and supply-chain traceability for electronic products. The appropriate traceability level should be agreed between the user and supplier according to product risk.

A detailed discussion is available in SHDC’s guide to traceability in electronics manufacturing.

Traceability helps the team avoid two opposite errors:

Containing too few products and allowing additional failures to escape.
Containing every product unnecessarily when only one material or process lot was affected.

Containment, Correction and CAPA

These actions should not be treated as interchangeable.

Action	Purpose	Example
Containment	Protect customers while investigation continues	Hold the affected lot and inspect warehouse stock
Correction	Address the current nonconforming unit	Replace the cracked connector
Corrective action	Eliminate the verified cause	Redesign the fixture that applies excessive force
Systemic preventive action	Apply lessons to related risks	Review similar fixtures and products
Effectiveness validation	Confirm the action works over time	Monitor subsequent lots and confirm no recurrence

Immediate containment

Containment may include:

Stopping shipment
Holding work in process
Identifying affected serial numbers
Inspecting warehouse inventory
Sorting a production lot
Notifying the OEM
Reviewing units already shipped
Implementing temporary additional testing

Containment should begin before the investigation is complete when customer risk justifies action.

Corrective action

Corrective action may involve:

Updating a process parameter
Changing tooling
Improving a work instruction
Revising a PCB layout
Changing an approved material
Adding a test point
Updating the test program
Revising the control plan
Retraining personnel
Strengthening incoming inspection
Improving supplier controls

Changes to released design or manufacturing documentation should follow a controlled Engineering Change Order process.

Effectiveness validation

Do not close CAPA immediately after implementation.

Effectiveness should be evaluated through:

Subsequent production lots
Yield and defect trends
Repeat testing
Customer-return data
Process audits
Reliability testing
Verification that revised documents are in use
Confirmation that the problem has not moved elsewhere

A corrective action is not effective simply because the first rebuilt board passes.

What Should a PCBA Failure Analysis Report Include?

A customer-facing report should make the evidence and reasoning easy to follow.

1. Product identification

Include:

Product and part number
Revision
Serial number
Production lot
Firmware
Sample quantity
Receipt date

2. Failure description

State:

Reported symptom
Verification result
Test conditions
Failure frequency
Environmental conditions
Pass/fail criteria

3. Sample condition

Document:

Physical condition on receipt
Packaging
Signs of rework
Damage
Contamination
Missing parts
Photographs

4. Investigation methods

List:

Visual inspection
Microscopy
Electrical measurements
AOI review
X-ray
Cross-section
Material analysis
Environmental testing
Production-record review

5. Findings and evidence

Present:

Measurements
Images
Waveforms
Test data
Lot comparison
Process records
Physical-analysis results

Clearly separate observations from interpretations.

6. Failure mechanism

Explain how the identified defect caused the reported symptom.

7. Root cause

State the verified underlying cause and supporting evidence.

8. Affected scope

Define:

Affected lots
Potentially affected serial numbers
Inventory status
Shipped quantity
Basis for the containment boundary

9. Containment and corrective action

List:

Immediate controls
Permanent actions
Responsible owner
Due date
Document revisions
Customer approvals required

10. Effectiveness validation

Define:

Validation method
Sample size or production period
Acceptance criteria
Results
CAPA closure decision

Common PCBA Failure Analysis Mistakes

Confusing a symptom with a root cause

“Product does not start” is a symptom, not a cause.

Replacing a component and ending the investigation

A failed component may have been damaged by another condition.

Altering the sample before documenting it

Cleaning, reflowing, probing, or removing parts may destroy evidence.

Failing to check the test equipment

A fixture, program, or cable may create false failures.

Comparing different product revisions

The observed differences may be intentional design changes.

Selecting a favorite hypothesis too early

Investigators may search only for evidence that supports their first assumption.

Using destructive analysis too soon

Cross-sectioning the wrong location can destroy the only available sample without finding the defect.

Ignoring production and traceability records

Physical analysis alone may not reveal when or where the problem was introduced.

Failing to verify root cause

A plausible explanation is not the same as demonstrated causation.

Closing CAPA before validating effectiveness

The problem may return in later lots.

PCBA Failure Analysis Checklist for US OEMs

When reviewing a supplier report, ask:

Was the original symptom reproduced?
Was the test equipment verified?
Was the failed sample preserved?
Was a known-good board used for comparison?
Are observations separated from assumptions?
Was the failure mechanism identified?
Does the evidence support the claimed root cause?
Was the affected production scope defined?
Which lots and serial numbers were contained?
Were material and production records reviewed?
Was the hypothesis experimentally verified?
Does the corrective action eliminate the cause or only repair the board?
Were relevant procedures, control plans, or drawings updated?
Was customer approval obtained for product changes?
How will effectiveness be monitored?
What evidence is required before CAPA closure?

US OEMs evaluating a new supplier should also review its diagnostic, testing, traceability, and corrective-action capabilities during supplier qualification. SHDC’s PCB Assembly Vietnam guide provides additional questions to consider before sending an RFQ.

How Better Design and Production Planning Reduce Failure-Analysis Risk

Failure analysis is necessary when problems occur, but early planning can make failures easier to detect, isolate, and prevent.

Useful practices include:

Accessible test points
Clear power-domain separation
Diagnostic firmware
Controlled production files
Traceable component lots
Defined test limits
Stored test data
First article approval
Production validation
Documented change control

The purpose of Production Validation Testing is to confirm that the product and manufacturing process can meet requirements before full production scale-up.

Likewise, reviewing the full PCBA manufacturing process helps OEM teams identify where quality records should be created and retained for future investigation.

PCBA Failure Investigation Support at SHDC

SHDC provides EMS activities from component soldering and assembly through testing and final packaging. Its current factory scope lists four high-speed SMT lines, three DIP lines, one assembly line, one test line, and one packaging line.

The documented production flow includes IQC, screen printing, component placement, reflow, AOI, repair, ICT, visual inspection, FCT, OQC, and packaging. Listed production equipment includes 3D SPI, 3D AOI, ICT, wave soldering, nitrogen reflow, and functional-testing equipment.

These production checkpoints can provide important first-line evidence for a failure investigation, including inspection results, electrical-test outcomes, process history, and repair status. Advanced methods such as CT, SEM, cross-sectioning, or specialized material analysis may require appropriate external laboratory support depending on the failure.

US companies evaluating an electronics manufacturing partner can review SHDC’s broader PCBA Vietnam capabilities.

Conclusion

Effective PCBA failure analysis does more than locate a defective component. It connects the reported symptom to a physical failure mechanism, verifies the root cause with evidence, identifies the affected scope, and establishes corrective action that prevents recurrence.

The strongest investigations preserve the sample, reproduce the failure, verify the test system, compare good and failed assemblies, review traceability records, and select analytical methods according to a clear hypothesis.

For US OEMs, a supplier’s ability to produce a structured failure-analysis report is an important indicator of engineering maturity. A useful report should explain not only what was found, but also why it happened, which products are at risk, what changed, and how the supplier confirmed that the solution works.

Frequently Asked Questions

What is PCBA failure analysis?

PCBA failure analysis is a structured investigation used to identify the failure symptom, physical or electrical mechanism, verified root cause, affected production scope, and required corrective action.

What is the difference between a defect and a root cause?

A defect is the observed nonconformity, such as a cracked solder joint. The root cause is the underlying condition that created it, such as excessive PCB bending during depanelization.

What methods are used for PCB failure analysis?

Methods may include visual inspection, microscopy, electrical troubleshooting, AOI record review, X-ray, CT, cross-sectioning, dye-and-pry testing, SEM, contamination analysis, and environmental testing.

Should a failed PCBA be repaired before analysis?

Normally, the board should first be photographed, documented, and tested in its received condition. Premature repair or cleaning can destroy evidence.

Is replacing a failed component enough to prove root cause?

No. Replacement can confirm that the component is involved, but the investigation must still determine why the component or connection failed.

How is an intermittent PCBA failure investigated?

The team may vary temperature, load, supply conditions, movement, time, firmware, and operating mode while monitoring signals and comparing the failed board with a known-good assembly.

What should a PCBA failure analysis report contain?

It should include product identification, failure description, sample condition, investigation methods, evidence, failure mechanism, root cause, affected scope, containment, corrective action, and effectiveness validation.

What is the difference between containment and corrective action?

Containment protects customers while the investigation continues. Corrective action permanently eliminates the verified cause of the failure.

How does traceability support failure analysis?

Traceability links the failed assembly to material lots, production records, inspection results, test data, firmware, rework history, and shipment information.

When can a CAPA be closed?

CAPA should be closed only after the corrective action has been implemented and objective evidence confirms that it is effective and the failure has not recurred.