Acceptance sampling makes a decision about a lot by inspecting a randomly selected sample. 100% inspection checks every part against the characteristics defined in the scope. The choice depends on risk, process performance, and applicable requirements, not just lot size. Neither method alone guarantees zero nonconforming parts.
In automotive, before using an AQL table, the control plan and the customer requirements must be reviewed. When applying IATF 16949:2016 paragraph 8.6.6, attribute sampling must respect the zero-acceptance requirement. Therefore, an example of a table with a higher than zero acceptance is not an authorization to release material under that requirement.
What Is AQL?
AQL stands for Acceptance Quality Limit. It is a parameter of sampling schemes associated with the worst tolerable average quality level for a continuous series of lots.
It does not represent a guaranteed maximum percentage of nonconforming parts in each lot, nor does it authorize shipping parts already known to be out of specification. Sampling is used to make an acceptance decision under defined risks, not to prove that every part conforms.
Among the references for attribute sampling are ISO 2859-1:2026 and ANSI/ASQ Z1.4-2003 (R2018). The required standard and edition for the operation must be identified; they should not be treated as interchangeable without verifying their requirements.
How Is an AQL Table Used?
Before consulting the table, the characteristic, the classification of non-conformities, the acceptance criteria and the applicable standard must be defined.
For a single sampling plan, normal inspection:
- The lot is identified and its size is defined.
- The level of inspection is selected according to the agreed procedure. General level II is an option of the system, not a universal selection for any risk.
- The lot size and inspection level give you the code letter.
- With the letter and the AQL, the sample size and the acceptance (Ac) and rejection (Re) numbers are calculated, following the instructions and arrows in the table.
- A random sample is selected and inspected according to the established method.
- The result is compared with Ac and Re and the decision is documented. In a series of lots, the scheme’s rules for switching between normal, tightened, and reduced inspection must also be applied.
Counting nonconforming parts must be distinguished from counting nonconformities. They are different evaluation bases and should not be mixed.
Table Reading Example: 2,000 Parts
This example corresponds to ANSI/ASQ Z1.4-2003 (R2018), single sampling, normal inspection, and evaluation by percent nonconforming units.
| Parameter | Value |
|---|---|
| Lot Size | 2,000 parts |
| General Inspection Level | II |
| Code Letter | K |
| Sample Size | 125 parts |
| AQL | 1.0 |
| Acceptance Number (Ac) | 3 |
| Rejection Number (Re) | 4 |
Under that plan, the result meets the acceptance criterion with up to three nonconforming units in the sample and fails with four or more. Nonconforming units detected must be identified and controlled, even if the lot is accepted.
The example is instructive. It should not be used as a criterion for release when applicable requirements require zero acceptance. It should also not be modified from Ac 3 to 0 while keeping the rest and stating that the plan remains the same: the sample size and statistical protection must correspond to the selected zero-acceptance plan.
Limitations of Sampling in the Automotive Industry
Accepting a sample does not prove that the rest of the lot is free of nonconforming parts. Even a plan with c = 0, which does not accept nonconforming findings in the sample, maintains a risk of accepting a lot that contains them outside of it.
In the event of a claim, alert, or process change, the risk and control plan must be reassessed. Containment, increased frequency of verification, or 100% inspection may be required, depending on applicable requirements; not all changes require the same treatment automatically.
A rejected lot requires controlled disposition. Drawing successive samples until one allows acceptance is not appropriate. Sorting, rework, and subsequent verification must follow the authorized procedure.
When Is 100% Inspection Needed?
It may be required, among other situations, when there is suspect material and parts must be separated for a defined failure mode; when the customer requires 100% containment; or when releasing the material requires individual verification according to the approved scope.
In GP12 (Early Production Containment), the plan sets the characteristics, additional controls, frequencies, and exit criteria. The program name does not, by itself, mean that every characteristic is inspected 100%.
In Controlled Shipping CS1/CS2, the customer's specific requirements must be respected, including the scope of the inspection and the responsibilities for execution and release.
For critical or safety characteristics, increasing the number of parts inspected is not enough: a method capable of detecting the condition being evaluated and controlling the risk must be used. A visual check does not replace a functional or dimensional test when one is needed. If the test is destructive, inspecting every part may not be a viable way to release the lot.
Best Practices for 100% Inspection
- Approved technical criteria: characteristics, limits, method, and representative OK/NOK references. Marketing illustrations do not replace validated technical standards or evidence.
- Appropriate methods and personnel: instruments and verification of the measurement system when appropriate, lighting conditions, training and proficiency testing.
- Traceability and segregation: identification of inspected, pending, and nonconforming material; prevention of mix-ups and accidental reintroduction.
- Records and verification: part number, lot, shift, quantities inspected, and results by failure mode; periodic effectiveness checks and documented material disposition.
The inspection complements the prevention and correction of process causes. It does not replace them nor is it infallible in reviewing all the parts.
How RIDYS Group Supports
RIDYS Group performs sorting, inspection, and rework of automotive parts at the customer’s plant, in San Luis Potosí and the Bajío region. It holds ISO 9001 certification and REPSE registration.
Learn about our services part sorting, quality inspection, GP12 y Controlled Shipping CS1/CS2.
Need to inspect or contain a lot at your plant? Share the part number, quantity, location, and the customer’s criteria or requirement through our contact page, to review the scope of the service.
Frequently Asked Questions
Does AQL Mean Defective Parts May Be Shipped? No. Define a parameter of the sampling scheme; it does not eliminate the product specifications or authorize the sending of known parts as non-conforming.
Which Standard Defines the AQL Tables? ISO 2859-1 and ANSI/ASQ Z1.4 are reference standards for attribute sampling. The applicable standard and edition of the contract and procedure must be used, without combining tables or editing rules that are not validated.
Does Zero Acceptance Mean the Entire Lot Has Zero Defects? No. It means that the criteria do not allow for non-conforming findings in the sample. It does not demonstrate that all the uninspected parts meet the requirements.
Does 100% Inspection Guarantee Zero Defects? No. Its effectiveness depends on the method, the characteristics inspected, the measurement system, and the execution. Its goal is to control risk within a defined scope, not to promise infallibility.
Technical References
- ISO 2859-1:2026, Sampling procedures for inspection by attributes — Part 1.
- ANSI/ASQ Z1.4-2003 (R2018), Sampling Procedures and Tables for Inspection by Attributes.
- IATF 16949:2016, section 8.6.6, acceptance criteria.
- NIST/SEMATECH, e-Handbook of Statistical Methods, sections 6.2.1 and 6.2.2.
