AQL Sampling Plan Calculator

Why sample instead of inspecting every unit

Inspecting every unit in a lot is often impractical — too slow, too costly, or (for destructive tests) impossible. Acceptance sampling inspects a smaller, defined sample and decides whether to accept or reject the whole lot based on how many defects turn up in that sample.

ISO 2859-1 formalizes this as a lookup: lot size and an inspection level together select a sample-size code letter, and that code letter combined with the target Acceptable Quality Limit (AQL) selects a sample size and a pair of accept/reject numbers.

The AQL is not a defect rate you expect to see — it is the boundary quality level the plan is designed around. A lot at exactly the AQL defect rate still has a meaningful chance of being rejected by chance; the plan only guarantees that lots at or better than the AQL are accepted most of the time, and progressively worse lots are rejected with rising probability.

Inspection level controls how much discriminating power the plan has for a given lot size: Level II is the general-purpose default, Level I uses a smaller sample for less discrimination, Level III uses a larger sample for more discrimination, and the S-1 through S-4 special levels trade discrimination for a much smaller sample when inspection is destructive or expensive.

When to Use This

• Setting up an incoming-inspection plan for purchased parts against a supplier quality agreement.
• Defining an outgoing-inspection sample size before a lot ships to a customer.
• Checking what sample size a specified AQL and inspection level actually requires, before committing inspection labor to a lot.

Common Mistakes

• Treating the AQL as an expected defect rate rather than a plan design parameter — a plan built around AQL 1.0% will still reject some lots that are genuinely at 1.0% defective; that is how the statistics are supposed to work.
• Reusing the same inspection level across very different products without considering that Level I trades discrimination for smaller samples — appropriate for low-risk items, not for safety-critical ones.
• Forgetting that sample size is capped at the lot size — a very small lot may require inspecting every unit even though the table nominally calls for a larger sample.

FAQ

What does the accept/reject number pair mean?

If the count of defective units found in the sample is at or below the accept number, the lot is accepted. If it is at or above the reject number, the lot is rejected. For a single sampling plan these two numbers are consecutive integers (e.g. 2 and 3), so every possible defect count leads to one outcome or the other.

Why did the AQL I entered get replaced with a different value?

ISO 2859-1 defines AQL as specific table columns (0.065, 0.1, 0.25, 0.4, 0.65, 1.0, 1.5, 2.5, 4.0, 6.5, and others). If a requested AQL is not one of those columns, the calculator uses the nearest column value and reports it separately from what was entered — the sampling plan itself is only ever pulled from an actual table column.

Does a smaller sampling percentage mean a lot is less thoroughly checked?

Sampling percentage falls as lot size grows because sample size does not scale linearly with lot size in the ISO 2859-1 tables — larger lots get proportionally smaller samples relative to their size while still delivering the discrimination the inspection level specifies. It reflects sampling economics built into the table, not a decision made per lot.