8.3 Root cause analysis & CAPA

A useful principle in manufacturing is that addressing the same production issue repeatedly indicates that the CAPA (Corrective and Preventive Action) system needs improvement. Root Cause Analysis (RCA) distinguishes between merely “fixing” an immediate symptom and permanently “solving” a systemic problem. Effective organizations investigate beyond convenient explanations, such as “Operator error,” preferring to drill down to the underlying systemic failure. This module defines the methodology for permanently eliminating a defect.

The RCA mindset

A primary rule of RCA is that quality cannot be achieved simply by replacing personnel. If the final RCA conclusion attributes the fault solely to the operator (“Man”), the engineering investigation is likely incomplete. Human performance naturally varies; manufacturing processes must be mechanically and digitally robust enough to handle that variation safely.

The Investigation Hierarchy:

1. Physics First: It must be checked if the machine, the fixture, or the raw material failed to operate as intended.
2. Process Second: It must be reviewed whether the documented method allowed for uncontrollable variation.
3. People Last: It must be determined if the operator deviated from a clear, possible, and verified instruction.

The 5 whys (drilling for oil)

The first answer in an investigation is often just a local symptom. Peeling back multiple layers (often five, though the number varies) usually reveals the true, actionable systemic root cause.

Example: A Solder Short on U12

• Why 1: Solder bridged two pins on the BGA. (The Symptom)
• Why 2: Too much solder paste was deposited on the pads. (The Direct Cause)
• Why 3: The stainless steel stencil aperture was cut too large. (The Process Cause)
• Why 4: The CAD stencil design followed the copper pad 1:1 without the required area reduction. (The Design Cause)
• Why 5: The Engineering DFM (Design for Manufacturing) Guideline for 0.4mm pitch components was outdated and unaudited. (The Systemic Root Cause)

The Action: Formally update the DFM Guideline in the document control system. Resolving the immediate short with a soldering iron (Correction) guarantees the problem will reappear on the next batch.

The fishbone (Ishikawa) logic

Systematically using the “6Ms” structures the engineering brainstorming process, ensuring that critical interaction variables are evaluated.

• Man: Training gaps, physical fatigue, visual acuity under microscope lighting.
• Machine: Calibration drift, mechanical wear, unverified software settings, missed preventative maintenance.
• Material: Undocumented vendor BOM changes, expired shelf life, excessive moisture content.
• Method: SOP clarity, tool sequencing, lack of Poka-Yoke (mistake-proofing) tooling.
• Measurement: Poor Gauge R&R (Repeatability & Reproducibility), incorrect lighting, parallax visual error.
• Mother Nature: Uncontrolled ambient humidity (affecting ESD/MSL), radical temperature drift on the SMT floor.

Pro-Tip: When a Root Cause is listed as “Operator Training,” a comprehensive review should still examine the “Method.” Retraining cannot fully compensate for a process that demands an unrealistic level of attention.

CAPA: Correction vs. Corrective Action

These terms are distinct concepts during customer audits and everyday operations.

Correction (the temporary action):

• Definition: Immediate action to fix a specific, localized non-conformance.
• Example: Manually reworking a specific solder bridge on U12 with an iron.
• Result: The single product is functional, but the systemic factory risk remains active.

Corrective action (the permanent solution):

• Definition: Systemic engineering action to eliminate the underlying cause of the non-conformance entirely.
• Example: Redesigning and replacing the stencil aperture, and updating the global DFM rules in the CAD system.
• Result: The specific defect will not reliably reoccur on this product line.

Preventive action (the global update):

• Definition: Strategic action to proactively eliminate a potential cause across all other product lines before they experience failure.
• Example: Applying the new 0.4mm pitch aperture design rule laterally to all future PCB layouts across the company portfolio.

Final Checkout: Root cause analysis (RCA) & CAPA

Control Point	Engineering Requirement	Risk Avoided
Investigation Depth	Focus on the System/Process level, avoiding “Human Error” as a root cause.	Endlessly Recurrent Failures.
Scientific Evidence	The root cause should ideally be demonstrated (e.g. via physical simulation).	Guesswork and superficial fixing.
Action Assignment	A CAPA should establish a physical or digital process change, rather than merely suggesting retraining.	Ineffective fixes.
Follow-up Verification	The Quality team should formally audit the effectiveness of the fix 30-60 days after closure.	Defect recurrence.
Lateral Scope	”Preventive Action” should be applied laterally across similar product families.	Recurrent Cross-Product issues.