Published:
Start with the inspection agreement
Before opening cartons, collect the purchase order, product specification, defect list, and buyer sampling requirements. A calculator cannot decide which quality risks the buyer will accept.
- Lot size: the number of units in the lot presented for inspection.
- Inspection level: the agreed general or special level. Do not select a smaller level to save time.
- Inspection severity: normal, tightened, or reduced, as required by the applicable switching rules.
- AQL: the agreed value for each defect class.
- Counting basis: nonconforming units or nonconformities per 100 units. Record which basis applies.
ASQ describes Z1.4 as an attributes sampling system with switching rules for a continuing series of lots. A single calculator result does not implement the whole system. Check the supported mode before using the result.
Enter the lot size and read the complete result
Enter the number of finished units in the defined lot. Do not enter the number of cartons unless cartons are the inspection unit in the agreement. Confirm that different orders or product variants may be grouped before treating them as one lot.
Record the sample-size code letter, actual sample size, acceptance number (Ac), and rejection number (Re). Follow any table arrows or plan instructions. Keep a copy of the inputs and result with the report.
Example: count units against Ac and Re
This is an illustrative counting example, not a Z1.4 table lookup. Suppose the agreed plan calls for 80 units, Ac 2, and Re 3 for major nonconforming units. Two such units meet that criterion. Three reach the rejection number. Other checks can still prevent release.
If one unit has two major defects, record both observations. Under a nonconforming-unit plan, that unit counts once within the major class. Under a nonconformity-count plan, the counting rule differs. Do not mix the two methods in one report.
AQL 2.5 does not mean you can multiply the lot size by 2.5% and ship that many bad units. It is an index used to select the sampling plan. Sample acceptance is not proof that every untested unit conforms.
Select the sample before judging the product
Choose units across the defined lot with a random selection method. Avoid a sample made only from the nearest open cartons. Record the carton or pallet positions so another inspector can identify where the units came from.
- Check the lot identity and available quantity.
- Select the required sample without factory staff choosing the best pieces.
- Inspect against the current specification and approved defect examples.
- Count each class using the agreed rule.
- Record the sampling result separately from the final release decision.
Handle critical defects and failed samples
A zero-acceptance rule means no critical findings are accepted in the inspected sample. It does not prove that the full lot has zero critical defects. Do not treat “AQL 0” as a standard table column without checking the agreed procedure.
When a sample fails, identify and hold the affected lot. Record the defect, evidence, owner, and next action. Do not draw repeated samples until one passes. Sorting, rework, and reinspection must follow the buyer-approved process.
Keep the original failed report linked to any later inspection. This gives the office clear proof of what changed and prevents a new spreadsheet from hiding the first result.
Key takeaways
- Agree the sampling inputs before inspection.
- Read Ac and Re from the selected plan, not from a lot-size percentage.
- Keep defect observations separate from the agreed counting basis.
- Sampling supports a lot decision; it cannot guarantee zero defects.
Sources and related guides
Put this into practice with KaizenQ
KaizenQ gives your team a structured digital inspection workflow: defects logged with photos on the floor, a live report ready to share as soon as inspection is complete. No manual compilation. No end-of-day rework.