Inspection planning for automotive production

Choosing Production Samples for an AOI Trial

Choose an AOI trial sample set that represents production variation, includes known outcomes and supports a meaningful review of the results.

A gloved worker completing a clipboard record
Illustrative photo.

A trial sample set should reflect the decision the machine will face in production. Sending only a few obvious defects and unusually clean good parts may demonstrate a concept without answering the difficult application questions.

Record the source and condition

Identify each sample and its expected result under the agreed criteria. Note the part variant and relevant production context. Avoid changing its appearance through unnecessary handling or labelling over the inspection area.

Include realistic variation

Discuss colour, finish, orientation and other variation with the quality team. Use evidence from the actual process to decide what belongs in the set. There is no universal sample count that establishes suitability for every automotive application.

Distinguish setup material from evaluation material

Agree which examples the supplier may use to develop the inspection and how the later evaluation will be conducted. If criteria or settings change during the session, record the change and the results that apply to it.

Track what comes back

Keep sample identities matched to the report and arrange appropriate handling and return. The production AOI evaluation guide connects the sample set to the wider acceptance decision.

Frequently asked questions

How many samples are enough for an AOI evaluation?

Have the responsible quality team determine a plan based on the requirements, variation and risk. A fixed number does not suit every application.

Why identify each trial sample separately?

It allows the reported result to be compared with the agreed reference judgement and investigated if the two differ.

Should setup results and later evaluation results be mixed together?

Keep their conditions clear. Record which data was obtained during tuning and which came from the agreed evaluation procedure.