Choosing the right sampling plan is critical for effective quality control, balancing inspection costs with the risk of accepting defective products or rejecting good ones. Here’s a step-by-step guide to help you select the optimal plan for your product:

• Product Criticality:
  • High-risk products (e.g., medical devices, aerospace parts) require stricter plans (e.g., lower AQL, tighter sampling).
  • Low-risk items (e.g., non-critical plastic parts) may use relaxed plans.
• Type of Defects:
  • Critical Defects: Safety or non-compliance issues → Use zero acceptance (e.g., 100% inspection or plan with c=0).
  • Major Defects: Functionality issues → Tight AQL (e.g., 0.65%).
  • Minor Defects: Aesthetic issues → Higher AQL (e.g., 2.5%).
• Lot Size:

  Larger lots typically require larger sample sizes (per standards like ISO 2859).

• Supplier Reliability:

  Reliable suppliers may use reduced sampling; new/suspicious suppliers need tightened plans.

• Cost vs. Risk Trade-off:
  • High inspection costs? → Use variables sampling or reduced plans.
  • High failure costs? → Use tightened/100% inspection.

Choose Sampling Standards

• Attribute Sampling (Pass/Fail):
  • ISO 2859-1 (AQL-based): Most common for general manufacturing.
  • MIL-STD-105E (Historically used in defense; similar to ISO 2859).
  • ANSI/ASQ Z1.4 (US version of ISO 2859).
• Variables Sampling (Measurable data):
  • ISO 3951: Used when defects are measurable (e.g., dimensions). More efficient for high-quality lots.
• Special Plans:
  • c=0 Plans: Zero acceptance (e.g., NISSAN Zero Sampling).
  • Dodge-Romig Plans: Focus on limiting outgoing defects (LTPD) or average outgoing quality (AOQL).

Define Acceptance Quality Limit (AQL)

• AQL = Maximum defect rate you consider acceptable.
  • Common AQLs:
    • Critical: 0.01%–0.1%
    • Major: 0.65%–1.5%
    • Minor: 2.5%–4.0%
• How to Set AQL:
  • Regulatory requirements (e.g., FDA, ISO 13485).
  • Historical defect data.
  • Customer specifications.

Select Sampling Type

• Single Sampling:
  • Draw one sample → Accept/reject lot based on defects found.
  • Best for: Quick decisions, stable processes.
• Double Sampling:
  • First sample → If inconclusive, draw second sample.
  • Best for: Cost savings when defect rates are low.
• Multiple Sampling:
  • Up to 7+ samples.
  • Best for: Minimizing sample size but increases complexity.
• Sequential Sampling:
  • Inspect items one-by-one until accept/reject decision is made.
  • Best for: High-value items or destructive testing.

Determine Sample Size & Acceptance Criteria

• Use standard tables (e.g., ISO 2859):
  • Lot Size → Code Letter (e.g., lot size 5,000 → Code Letter L).
  • Code Letter + AQL → Sample Size (n) & Acceptance Number (c).
  • Example:
    • Lot size: 5,000, AQL 1.0% → Code L → n=200, c=5.
    • Accept if ≤5 defects; reject if ≥6.
• Switching Rules:
  • Normal → Tightened: If 2+ lots rejected in 5.
  • Tightened → Normal: If 5 consecutive lots accepted.
  • Reduced Sampling: For stable suppliers (e.g., 20% smaller samples).

Validate & Monitor

• OC Curve Analysis:

  Plot probability of acceptance vs. defect rate. Ensure the plan meets your risk tolerance (e.g., ≤5% chance of accepting a 3% defective lot).

• Continuous Improvement:

  Review sampling data quarterly. Adjust AQL/sample size if defect patterns change.

• Documentation:

  Record sampling plan, results, and decisions per ISO 9001.

Decision Flowchart

graph TD
  A[Product Criticality?] -->|High| B[Use c=0 or Tightened Plan]
  A -->|Low| C[Standard AQL Plan]
  B --> D[Define Critical AQL <0.1%]
  C --> E[Set AQL: Major 0.65-1.5%, Minor 2.5-4.0%]
  D --> F[Choose MIL-STD-105E/ISO 2859]
  E --> F
  F --> G[Lot Size → Code Letter]
  G --> H[Sample Size (n) & Acceptance Number (c)]
  H --> I[Implement Switching Rules]
  I --> J[Monitor OC Curve & Adjust]

Key Considerations

• Regulatory Compliance: Medical/automotive may require specific standards (e.g., IATF 16949).
• Destructive Testing: Use variables sampling to minimize waste.
• Supplier Audits: Combine sampling with on-site process audits.
• Cost Impact: For high-volume items, even small sample sizes can be costly.

Final Tip: Start with ISO 2859-1 for attribute sampling and ISO 3951 for variables. Use AQL calculators or software (e.g., Minitab) to simulate plans. Always align with customer requirements and industry best practices!