An instruction to inspect a few boards is incomplete. The result cannot support a clear lot decision until the team defines which population the sample represents, what counts as a nonconforming unit, and what action follows the observations. Establish those details before material is selected.
NIST’s single-sampling-plan discussion identifies sample size and acceptance number as key plan parameters. These parameters belong to a defined statistical and operational plan; choosing them informally after seeing the sample defeats the purpose of an agreed decision rule.
Describe the population being judged
Identify the board revision, supplier lot, process route, and relevant production interval. Determine whether a shipment contains several materially different populations. Combining them under one delivery number may hide a localized problem or make the sample unrepresentative of part of the shipment.
State the unit of inspection. A panel, an individual PCB, a solder joint, and an assembled product are different units. If one board contains several inspected features, define whether the decision counts failed boards, individual defects, or both.
Complete the plan before inspection
| Plan item | Required clarity |
|---|---|
| Lot identity | Which material the decision covers |
| Inspection method | What evidence is collected and how |
| Sample selection | How units enter the sample without convenience bias |
| Acceptance and disposition | What observations trigger acceptance or further action |
Use the sampling standard or agreement applicable to the project and retain its identity and relevant settings. Do not present an AQL designation as a guarantee that every accepted lot contains fewer defects than that number. A sampling plan has probabilities of accepting and rejecting different quality levels.
Preserve the actual execution
- Record the selected unit identities or locations, including panel position where repeated neighboring units could otherwise dominate the selected inspection sample unintentionally.
- Retain observations before calculating the summary.
- Document inaccessible or substituted sample units.
- Apply the predefined decision rule.
- Record the disposition of both sampled and remaining material.
Do not replace a failed sample with a passing unit and continue as though the first observation never occurred. Additional inspection may be justified, but it needs a defined purpose and disposition route rather than an improvised search for a passing subset.
Close the inspection with the plan, observations, and lot decision together. If the lot definition or selection method was not followed, state the limitation and resolve it before claiming the planned statistical interpretation. The defect-count guide helps keep the numerical summary consistent with the selected inspection unit.
Move from reading to a build-ready project
Use the fabrication and assembly checklists to prepare drawings, production files and test requirements.
Open the DFM checklist