ONE ENGINEERING QUESTION AT A TIME

How Should AQL Sampling Be Used?

AQL is an indexing parameter for acceptance-sampling plans over a continuing series of lots.

Purchasing & Quality5 min readEngineering data requires project validation
DIRECT ANSWER

AQL is an indexing parameter for acceptance-sampling plans over a continuing series of lots. It is not “the percentage defective allowed in this lot” or permission to ship defects. ISO 2859-1:2026 uses lot size, inspection level, AQL and normal/tightened/reduced status to determine sample size and acceptance/rejection numbers.

TECHNICAL USE NOTE

Values in this article support engineering screening during design review, RFQ and prototyping. They do not replace formal calculations, material data sheets, qualified procedures or contractual standards.

01

CORRECT WORKFLOW

1. Classify critical, major and minor defects with objective criteria. Critical defects commonly use zero acceptance, though Ac=0 still does not mean zero risk. 2. Define a lot as completed, traceable product made under common conditions; never split a problem lot to obtain a more favorable plan. 3. Use the contractual standard edition: lot size and inspection level give a code letter; the AQL table then gives n, Ac and Re. Do not copy a fixed “inspect 80” rule online. 4. Sample randomly across boxes, pallets, times and machines; top-layer convenience sampling is not random. 5. Handle a rejected lot under predefined return, 100% screening, rework or concession rules; resubmission after screening must be traceable.

02

STATE THE RISKS

Acceptance sampling has producer’s risk α (rejecting a good lot) and consumer’s risk β (accepting a bad lot). The plan’s OC curve shows acceptance probability at different true defect levels; AQL alone does not show the full risk. For isolated lots consider ISO 2859-2:2020. Where measurable continuous data meet assumptions, ISO 3951-1:2022 variables sampling retains more information than reducing dimensions to pass/fail.

03

PRACTICAL BOUNDARIES

AQL supports shipment acceptance; it does not replace SPC, process audits, mistake-proofing or 100% automated checks for critical safety features. A systemic error—wrong material, revision or missing operation—should trigger quarantine and investigation even below Re, because the random-defect model no longer applies. ======================================================================

Submit Drawings for Project Review

SHARED FAQ

Before Using Data from These Articles

FAQ

Engineering Use Boundaries

Can the numerical rules in these articles be copied directly onto drawings?

Not without validation. Use them for initial DFM, validate with the actual grade/temper, equipment, tooling and samples, then place the approved range in the drawing or control plan.

Does passing salt spray, adhesion or Cpk prove part reliability?

No. They address a test environment, coating-separation behavior or statistical stability and form only part of the evidence. Reliability also depends on service load/environment, design details, process audit and durability tests.

Why do the articles emphasize standard editions?

Standards are revised or withdrawn. State number and year contractually and ensure customer, manufacturer and inspector use the same edition. For example, ISO 2859-1 has a 2026 edition; an old table does not automatically govern a new contract.

TURN INSIGHT INTO A MANUFACTURABLE PLAN

Submit Drawings for Project Review

Send the 2D drawing, 3D model if available, material, quantity, critical tolerances, finish and service environment. We will provide specific feedback on DFM, process route, risk points, inspection, cost and lead time.

Submit Drawings for Project Review