8.3 Root Cause Analysis (RCA) & CAPA

If you are fighting the same fire this week that you fought last month, your CAPA system is broken. Root Cause Analysis (RCA) is the difference between "fixing" a problem and "solving" it. Most organizations stop at the symptom ("Operator error"); true engineering discipline drills down to the systemic failure ("Why did the system allow the error?"). This chapter defines the logic required to permanently kill a defect.

The RCA Mindset: Blame vs. Physics

The managementfirst rule of Non-ConformingRCA Materialis: (NCM)You andcannot fire your way to quality. If you blame the Corrective and Preventive Actionoperator (CAPA)"Man"), systemyou formhave failed the "immuneinvestigation. system"Humans ofare variable; processes must be robust.

The Investigation Hierarchy:

1. Physics First: Did the factory.machine or material fail?
2. Process Second: Did the method allow variation?
3. People Last: Did the operator willfully violate a clear, physically possible instruction?

The 5 Whys (Drilling for Oil)

Do not stop at the first "Why." The first answer is usually a symptom. The fifth answer is the root cause.

Example: Solder Short on U12

• Why 1: Solder bridged two pins. (Symptom)
• Why 2: Too much solder paste deposited. (Direct Cause)
• Why 3: The stencil aperture was too large. (Process Cause)
• Why 4: The stencil design followed the pad 1:1 without reduction. (Design Cause)
• Why 5: The DFM Guideline for 0.4mm pitch components was outdated. (Systemic Root Cause)

Action: Update the DFM Guideline. Just cleaning the board (Correction) ensures the problem will happen again tomorrow.

The Fishbone (Ishikawa) DiagramLogic

When a defect occurs,Use the investigation"6Ms" mustto structure your brainstorming. If you don't look at sixall distinct6, categoriesyou ("Thewill 6 Ms") to avoid blamingmiss the operator.interaction variables.

• Man (People):Man: Training, fatigue.fatigue, visual acuity.
• Machine: Calibration, wear, settings.settings, maintenance.
• Material: ComponentVendor vendors,changes, shelf life.life, moisture content.
• Method: WorkSOP instructions,clarity, routing.sequence, tooling.
• Measurement: Gauge error,R&R, lighting.lighting, parallax error.
• Mother Nature (Environment):Nature: Humidity,Humidity temperature.(ESD/MSD), temperature drift.

Pro-Tip: If your Root Cause is "Operator Training," I demand to see the "Method" analysis. Training cannot fix a process that requires superhuman attention span.

TheCAPA: 5Correction Whysvs. Corrective Action

Drill down from the symptom to the systemic root cause.

• Problem: Solder bridge.
• Why? Paste volume too high.
• Why? Stencil aperture too large.
• Why? Designed 1:1 with pad.
• Why (Root)? DFM Guidelines didDo not specifyconfuse aperturethese reductionterms. forThey thisare pitch.

legally

distinct in an audit.

CAPACorrection (CorrectiveThe & Preventive Action)Band-Aid):

• Correction:Definition: FixImmediate theaction immediateto problem (rework).
• Corrective Action: Eliminatefix the rootnon-conformance.
• Example: Reworking the solder bridge.
• Result: The product is good, but the risk remains.

Corrective Action (The Cure):

• Definition: Action to eliminate the cause to prevent recurrence (fixof the stencil).non-conformance.
• Verification of Effectiveness:Example: AuditRedesigning the processstencil 30-60aperture.
• Result: daysThe later.defect Ifcan never physically happen again.

Preventive Action (The Vaccine):

• Definition: Action to eliminate a potential cause in other products.
• Example: Applying the problemnew returned,aperture thedesign CAPArule failed.to all future PCB layouts.

Final Checklist

StepControl Point	Critical Requirement	Risk Avoided
Depth	Drill down to System/Process level, not "Human Error."	Recurrent Failures
Evidence	Root cause must be able to turn the problem On and Off (simulation).	Guesswork
Action	CAPA must include Ownerprocess change
Containment	Segregate, &not Redjust Tag"Retrain Operator."	ProductionIneffective / QA
Root Cause	Fishbone / 5 Whys	Process Engineer
Correction	Implement Poka-Yoke	Engineering"Fixes"
Verification	Audit effectiveness 30-Day60 Effectivenessdays Auditafter close.	QualityZombie ManagerDefects
Scope	Apply "Preventive Action" to similar product families.	Cross-Product Contamination