9.2 RMA Processing & Field Failure Analysis

Field returns (RMA) are the only source of truth in manufacturing. While internal yield metrics measure your ability to build the product, RMA data measures the product's ability to survive reality. A robust Field Failure Analysis (FFA) process is not a "returns department" function; it is a forensic engineering discipline. The objective is not to issue a credit note, but to extract the failure mode, replicate it, and feed the corrective action back into the Design (NPI) or Manufacturing process to prevent recurrence.

Intake & Safety Protocols (The Air Gap)

Returned units are unknown variables. They may have been installed in hazardous environments (hospitals, oil rigs) or subjected to bio-contamination.

Decision Logic for Intake:

• IF Unit arrives from Medical/Industrial field:
  • THEN Quarantine immediately. Do not open without Decontamination Certificate.
  • Risk: Exposure to pathogens or chemicals.
• IF Li-Ion Battery is present and swollen/damaged:
  • THEN Classify as HAZMAT. Store in fireproof cabinet. Do not charge.
• Action: Record serial number and photograph external condition before any testing. This distinguishes "shipping damage" from "customer abuse."

The Verification Gate (Confirming the Failure)

The most expensive outcome in RMA is NTF (No Trouble Found). NTF is not a customer error; it is often a gap in your test coverage.

Testing Hierarchy:

1. Visual Inspection: Look for burn marks (EOS), corrosion (liquid ingress), or impact damage (cracked housing).
   • Decision: If physical damage is evident, stop electrical testing. The root cause is likely "Misuse."
2. Functional Verification: Attempt to replicate the failure using the customer's reported scenario.
   • Pro-Tip: Customers often report "Dead Unit" when the issue is actually a specific sleep-mode bug. Do not just run the standard "Pass/Fail" script; try to break the unit.
3. NTF Protocol: If the unit passes all tests:
   • Action: Perform thermal cycling (0˚C to 50˚C) or vibration. Intermittent failures (cold solder joints) often hide at room temperature.

Root Cause Analysis (The "Smoking Gun")

When a defect is confirmed, categorize it into one of three buckets to assign ownership.

Bucket A: Electrical Overstress (EOS)

• Signs: Burnt components, carbonized PCB, melted traces.
• Physics: Massive energy surge (Lightning, wrong power supply, short circuit).
• Owner: Typically Customer (Application) or Design (Inadequate Protection).

Bucket B: Electrostatic Discharge (ESD)

• Signs: SEM analysis shows gate oxide puncture. No external burns.
• Physics: Latent defect from assembly or improper handling.
• Owner: Manufacturing Process (EPA Violation).

Bucket C: Workmanship / Component

• Signs: Cold solder, missing component, wrong value, dead silicon.
• Owner: Manufacturing or Supplier.

The Feedback Loop (CAPA)

RMA data is useless if it sits in a spreadsheet. It must trigger the Corrective Action (CAPA) system.

Trigger Thresholds:

• IF Safety Incident (Fire/Shock): THEN Immediate Stop Ship + Recall Analysis.
• IF New Failure Mode detected: THEN Issue CAR (Corrective Action Request) to Engineering.
• IF Repeat Failure Rate > 1%: THEN Initiate Process Audit.

Final Checklist

Control Step	Parameter	Critical Limit / Rule
Safety	Decontamination	Zero Handling without PPE/Cert
Verification	Replicate Failure	Use Customer Scenario, not Factory Test
NTF Rate	"No Trouble Found"	Target < 10%. (If higher, review Test Specs)
Analysis	EOS vs. ESD	Differentiate Energy Source (External vs. Internal)
Disposition	Scrapping	Demolish (Do not allow gray market resale)
Data	Feedback	RMA Metrics must drive FMEA updates