High Defect Rate Process

PPM is high — what is the fastest path from containment to a defensible root cause without shipping suspect product?

• Typical PPM: 10,000+ PPM

1. Activate containment: 100 % sort with independent inspection, marked and traceable.
2. Bracket the suspect production window using SPC, downtime, and setup logs.
3. Quarantine WIP and finished goods within the window; recall from FGI if required.
4. Verify the gauge and the defect definition — many "spikes" are inspection changes.
5. For variable characteristics, run a fresh capability study on the stable segment.
6. For attribute defects, Pareto the failure modes and target the top mode first.
7. Correlate with tool change, material lot, shift, operator, program revision.
8. Implement corrective action via 8D D5; validate with a confirmation run.
9. Update Control Plan reaction rules, PFMEA severity/occurrence, and inspection frequency.
10. Close containment only after D6 verification confirms sustained performance.

• Sort operators trained inconsistently; the sort itself leaks defects.
• Defect definition changed between shifts (calibration of visual criteria).
• Attribute gauge (go/no-go) drift after mishandling.
• Mixed material lots inside the sort window mask the true onset.
• Reaction plan not executed because SPC limits were too wide from a pre-improvement era.
• PM overdue on the station driving the defect mode.

• High PPM with acceptable Cpk almost always means one of: non-normality, mixed streams, an unstable process, or a redefined defect.
• Containment is not corrective action. Auditors and customers treat conflation of the two as a systemic finding.
• Defect spikes that appear at shift boundaries are process-handover issues; defect spikes at tool-change boundaries are setup issues. The pattern is the diagnosis.
• A Pareto that "looks flat" usually means the defect coding is too coarse — split the top category and re-Pareto.

• Do not compute Cpk to justify the sorted stream — sorting is not capability.
• Do not use capability at all on attribute data — quantify by defect rate, DPMO, or Poisson methods.
• Do not close 8D on the basis of "no defects since Monday" — validate under representative volume.

• Cpk vs PPM: A 1.5σ shift assumption is common but must be disclosed; PPM without disclosure is not comparable.
• Cpk vs DPMO: DPMO includes defect opportunities and is used across characteristics; do not equate to single-CTQ PPM.
• Cpk vs FTY / RTY: Rolled throughput yield exposes multi-station defect stacking that Cpk on one CTQ hides.

1. Contain (100% sort).
2. Verify MSA and spec.
3. Capability study to quantify Cpk.
4. 8D root cause.
5. Use the free Process Capability Calculator to compute Cp, Cpk, Pp, Ppk, sigma level, and predicted PPM from your measurement data — no sign-up required.

• Never let containment lapse before corrective action is validated on a full production window.
• Never ship a suspect lot on the strength of a sample re-inspection alone.
• Never close an 8D without updating the PFMEA and Control Plan.

• risk-based validation and evidence — High defect risk can require validation evidence.

• ✓Containment active with independent verification of the sort
• ✓Suspect window bracketed with production and setup logs
• ✓Customer notified per CSR when applicable
• ✓Root cause identified and validated in 8D D4
• ✓Corrective action confirmed under representative volume in D6
• ✓Control Plan reaction rules and PFMEA scores updated
• ✓Containment lifted only after sustained defect rate confirmed

• Process condition: statistical stability is required.
• Distribution assumption: use a distribution model justified for the data.
• Confirm process stability and measurement-system adequacy before interpreting a capability index.
• Use a justified distribution model or non-normal method when the normal model is unsuitable.