The short answer
Accept a first article against a written specification and a pre-agreed test plan, then retain an identifiable reference sample with its records. Sample approval establishes a bounded baseline; it does not prove the conformity of every future production lot.
Establish what the sample represents
Ask whether the sample came from normal production, a pilot run or a hand-prepared prototype. Record the manufacturer, production location, part number, specification revision and lot if available. A prototype can be useful for development, but it should not be silently presented as evidence of routine process consistency. Require the supplier to identify any difference between the sample and the intended production construction before testing starts.
Write acceptance criteria before seeing results
Separate identity and workmanship checks from functional testing. The first group can include agreed material documentation, ring and seam construction, cleanliness, label accuracy and visible damage. The second must address the application: filtrate quality, clean-start behavior and operation under representative approved conditions. State the method, units, acceptance limit, number of samples and decision authority for each check. A box marked sample approved without the tested scope leaves too much open.
Keep the retention requirement linked to a test method, not just a micron number. Eaton's PROGAF data illustrates how one supplier states conditions for its own performance measurements. It does not validate another sample or provide a universal acceptance threshold. Resolve method differences before comparing results from different laboratories.
Build a sample-disposition matrix
Agree the treatment of each result before the sample arrives. A concession should identify a specific deviation and limited use; it should not become an undocumented permanent specification change.
| Check or outcome | Evidence to retain | Disposition |
|---|---|---|
| Identity and documents match | Part, revision, lot and document references | Proceed to the agreed next test |
| Visible defect or wrong construction | Photographs and inspection result | Hold and request investigation |
| Functional test meets defined criteria | Raw results and operating conditions | Approve only the tested scope |
| A required test remains incomplete | Named missing evidence and owner | Keep approval conditional or on hold |
| Sample differs from production proposal | Written difference assessment | Require relevant requalification before routine supply |
Retain a reference without destroying its meaning
Where practical, retain an unused representative unit separately from a used trial bag. Seal and label it with sample identity, date, approval status and a link to the report. Record storage conditions appropriate to the materials and an agreed review or retention period; do not invent a universal shelf life. A used bag may be useful for investigating deposits, but it is not an unchanged reference for original cleanliness or material condition.
Identify who holds the reference, who can open it and how access is recorded. If buyer and supplier each retain a sample, document whether they came from the same lot. Photographs help describe appearance but do not replace a physical specimen for every future question.
Connect first-article approval to incoming control
Translate accepted characteristics into the routine purchase specification and incoming inspection plan. NIST explains that lot acceptance sampling supports a decision on a particular lot; it is not the same as proving long-term quality. Select any sampling plan according to risk and the applicable agreement, rather than copying an arbitrary sample count. Revisit qualification when an approved material, manufacturing method or test basis changes. Track later deliveries against the baseline and investigate drift instead of repeatedly approving isolated attractive samples.
Your review checklist
- Sample origin and production representativeness declared.
- Part, revision and lot or prototype status recorded.
- Acceptance methods and limits agreed before testing.
- Raw results and deviations retained with the decision.
- Unused reference and used trial specimens distinguished.
- Routine lot control and requalification triggers defined.
Limits of this guidance
General first-article planning, not a JINLV sample approval or engineer signoff. Engineering and quality personnel must approve application tests, operating limits, deviations and any regulated-use requirements.
Questions buyers ask
Is a retained sample the same as a guaranteed standard?
No. It is a reference linked to an approved specification and test record. Aging, storage and handling can change it, so written requirements remain essential.
How many samples should we request?
There is no universal number. Consider destructive tests, variability, failure consequences and the decision being made. Agree the sample plan with quality personnel before the supplier selects the units.
Technical references
References support general principles or the named supplier's own products; they do not certify or validate a JINLV configuration.
AI-assisted technical synthesis of the references below, with editorial checks. This is general procurement guidance, not an independently signed engineering review or a product validation report.
English is the indexing version. Chinese and Indonesian are reading translations of the same article.