What AQL means
In acceptance sampling, AQL stands for acceptance quality limit. ISO 2859-1 describes it as the worst tolerable quality level for a continuing series of lots submitted for sampling. The older name, acceptable quality level, is still widely used. The US equivalent is ANSI/ASQ Z1.4; both standards descend from MIL-STD-105 and their tables closely match.
In practice, an AQL tells an inspector how many units to pull from a lot and how many defects can be found before the lot fails. “AQL 2.5 for major defects” does not mean the shipment contains at most 2.5% defective units. Plans are built to accept lots at the AQL most of the time, and a lot somewhat worse than the AQL can still pass.
How a sampling plan is chosen
Under ISO 2859-1 and Z1.4, a single sampling plan comes from three inputs:
- Lot size: the number of units in the batch being inspected.
- Inspection level: decides how lot size translates into a sample size code letter.
- AQL: chosen separately for each defect class.
The code letter gives the sample size, and the AQL column gives two numbers: the acceptance number (Ac) and the rejection number (Re). If the defects found in the sample are at or below Ac, the lot is accepted; at or above Re, it is rejected. In some cells the table points to a different sample size instead; follow the plan it points to.
Worked example. For a lot of 5,000 units at general inspection level II under normal inspection, the code letter is L and the sample size is 200. At AQL 2.5 the lot is accepted with up to 10 defective units and rejected at 11. At AQL 4.0 it is accepted with up to 14 and rejected at 15. Always check plans against the edition of the standard your agreement names.
Inspection levels
- General levels I, II and III. Level II is the default. Level I uses smaller samples and separates good lots from bad less reliably; level III uses larger samples and separates them better.
- Special levels S-1 to S-4. Much smaller samples, for tests that are destructive, slow or expensive, such as drop tests, teardown checks or long functional tests.
The standards also define switching rules between normal, tightened and reduced inspection based on recent lot history. Many pre-shipment inspections apply normal inspection to each lot on its own and never switch. That is simpler, but it ignores the supplier’s track record in both directions.
Defect classes
AQLs are set per defect class, and the classes need written examples in your inspection specification:
| Class | Meaning | Common convention for consumer products |
|---|---|---|
| Critical | Could cause harm, or breaks a legal or safety requirement | Zero accepted; any finding rejects the lot |
| Major | Likely to cause failure or a return, or to reduce usability | Often AQL 2.5, or tighter |
| Minor | Unlikely to affect use, such as small cosmetic flaws | Often AQL 4.0 |
These values are a convention, not a rule. Products with higher risk or higher price often use tighter AQLs for major defects.
What AQL does not do
AQL sampling detects bad lots after they are made. It does not prevent defects, measure process capability or make a poor process acceptable. A supplier running at a steady 2% defective will pass most AQL 2.5 inspections and still send you a steady stream of returns. Use lot sampling alongside in-process control (IQC, IPQC and OQC) and process data (Cpk, first pass yield).
Common mistakes
- Reading the AQL as the maximum defect rate in the lot.
- No written defect list, so the inspector and the factory classify the same flaw differently.
- Pulling samples from the most convenient cartons instead of at random across the lot.
- Using level II for destructive or long tests, then cutting the sample informally.
- Judging cosmetics without a golden sample or limit samples.
- Treating a passed inspection as proof that the process is healthy.
Where Zefinity helps
Many AQL disputes are about classification rather than arithmetic: whether a flaw is major or minor, and whether it sits on a surface that matters. Through QA and corrective action follow-up, Zefinity helps teams write inspection criteria that the factory and inspectors apply the same way, and traces recurring defects back to their cause instead of relying on lot rejection. The guide to building a practical hardware quality plan shows where lot inspection fits among the other controls.
Frequently asked questions
What does AQL 2.5 mean?
It is the quality level used to select the sampling plan. Together with the lot size and inspection level, it sets the sample size and the number of defects allowed. It does not mean that up to 2.5% of shipped units may be defective, and a lot somewhat worse than that level can still pass. Treat it as a lot acceptance rule, not a quality target.
What is the difference between ISO 2859-1 and ANSI/ASQ Z1.4?
Both descend from the US military standard MIL-STD-105 and use closely matching tables for sampling by attributes. ISO 2859-1 is the international standard; ANSI/ASQ Z1.4 is the US national standard. Name one of them, and its edition, in your inspection agreement so the supplier and the inspector use the same plans.
Which inspection level should I use?
General inspection level II is the usual default for visual and basic functional checks. Use a special level, such as S-2 or S-3, for tests that destroy the unit or take a long time. Moving to level I or III changes how well the plan separates good lots from bad ones, so make that choice deliberately rather than to save inspection time.