Reading a sampling plan correctly means keeping three ideas separate: the level sets how many pieces to pull, the AQL values set how many defects each class may contain, and the selection method decides whether the sample truly represents the lot. For injection molded parts, where defects cluster by cavity and shift, the third idea often matters most. Agree all three in writing, switch levels only on defined history, and record enough detail that the verdict can be reconstructed months later.
Key takeawaysSample size comes from lot size plus inspection level via a code letter, while the pass or fail verdict comes from applying AQL values to defects found in that sample. · Documented quality management principles, including process control and evidence-based decisions described in the ISO 9000 family foundations[1], support consistent sampling discipline across repeat molded-parts orders. · Measurement credibility behind dimensional checks depends on calibration chains covered by mutual recognition between accreditation bodies[2]. · Tightened inspection should follow failures and good history should be required before any move to reduced sampling, with mold or material changes resetting the clock. · Carton selection spread across pallets, shifts, and cavities matters as much as sample size for molded parts with clustered defects.
The first part of this guide explained what sampling levels control and when each level fits injection molded parts. This second part goes further into the working detail: how sample sizes are actually read from the tables, when to tighten or relax a plan, and which documentation habits keep results credible when a shipment is disputed.
Molded parts deserve this extra attention because their defects are rarely random. A worn cavity flashes on every twentieth shot, a cold joint appears when the barrel temperature drifted for an hour, a color shift starts when regrind ratio was adjusted mid-run. Sampling mathematics assumes the sample represents the lot, and with molded parts that assumption has to be earned through careful carton selection and cavity-aware checking.
What follows is practical rather than theoretical: worked logic you can follow without owning the standard, switching rules that respond to history, and a documentation checklist that protects both buyer and supplier. The goal is a sampling plan that survives contact with a real factory floor.
From lot size to sample size without guesswork
The standard route from lot size to sample size runs through a code letter. Each lot-size range has a row, each inspection level has a column, and their intersection gives the letter. The letter then points to a fixed sample size in the master table. A buyer does not strictly need to memorize the tables to use sampling well, but should understand this two-step logic so that discussions with the factory stay concrete. injection molded parts inspection
Consider a molded enclosure order of several thousand pieces. At Level I the code letter sits lower in the alphabet and the sample is comparatively small; at Level II the letter moves up and the sample grows; at Level III it grows again. The same lot therefore faces visibly different scrutiny depending on the level, which is why the level belongs in the purchase order rather than in a verbal understanding. Professional injection molded parts inspection applies this mapping without shortcuts, so the sample size on the report always traces back to the agreed inputs.
Two refinements matter for plastics. First, keep colors and mold cavities as separate lots where volumes allow, because blending them averages away cavity-specific problems. Second, when the lot spans multiple production days, draw cartons from each day in proportion to output. A sample pulled from a single day's pallets cannot represent a week of molding, whatever the code letter says.
Tightened, normal, and reduced in plain terms
| Regime | What changes | Trigger for molded parts | Return path |
|---|---|---|---|
| Normal | Standard sample and acceptance numbers | Default at order start and after stable history | Baseline |
| Tightened | Larger sample or stricter acceptance | Failed lot, repeated major defects, new mold teething problems | Several consecutive passes under tightened rules |
| Reduced | Smaller sample with same AQL values | Long run of clean passes on identical mold and resin | Any failure or process change ends reduced status |
| Discontinued pending action | Sampling pauses while process is fixed | Critical defects or evidence of sorting without disclosure | Root cause action and fresh first-article approval |
Why molded parts punish lazy carton selection
Attribute sampling mathematics works best when every piece in the lot had a roughly equal chance of landing in the sample. Molded production violates that ideal whenever output is stacked pallet by pallet and the inspector samples only the nearest ones. Cavity wear, a hot-runner imbalance, or a regrind adjustment can concentrate defects in a slice of the lot that a convenience sample never touches.
Systematic selection beats convenience selection. Number the finished cartons, generate spread intervals across the full quantity, and pull from front, middle, and rear pallets. Where the factory tracks cavity numbers or shift marks, include pieces from each source in the sample rather than trusting a single stream. Record the carton numbers selected so the selection can be audited later.
This discipline also protects the supplier. When a lot fails on defects concentrated in one identifiable batch, the factory can contain and sort that batch instead of disputing the whole verdict. Records showing broad, even selection make that conversation factual. Quality management thinking emphasizes exactly this kind of process-based, evidence-driven control rather than inspection as a lottery[1].
Steps that make the verdict defensible
- Fix the lot before the visitWrite the lot as order line, part and mold number, resin batch, colors, and production dates. Note any split between first-article, bulk, and top-up quantities so the sample can be allocated fairly.
- Confirm size and acceptance numbersRead the code letter from lot size and level, then the sample size from the letter. Confirm the AQL acceptance numbers for critical, major, and minor defects at that sample size before opening a single carton.
- Select cartons systematically and inspectDraw cartons across pallets, days, and cavities using pre-set intervals. Verify counts, check carton condition and labeling, then examine each sampled piece for appearance, color, dimensions, and fit.
- Log defects with evidence and instrumentsLog every defect with its class, location, and a photo, and record gauge identities with their calibration status. Confidence in dimensions rests on calibration chains recognized through international accreditation cooperation[2].
- Write the verdict and the next actionState the verdict per defect class, reference the agreed AQL values, and spell out the consequence: release, sort with re-inspection, or reject. Circulate the report to buyer and supplier together with the photos.
Paperwork buyers actually reuse
The most useful sampling record fits on a few pages but answers every later question: what was the lot, which level and AQL values applied, how many pieces were drawn from which cartons, what was found, which instruments were used, and what was decided. Attach the defect photos and the gauge calibration notes rather than describing them from memory. Future orders reuse this file as the starting template, which is how sampling discipline compounds over time.
Keep a small history table per part number: date, lot size, level, AQL values, result, and dominant defect type. After several consecutive clean lots on an unchanged mold and resin, a buyer may discuss reduced sampling; after a failure, tightened rules apply until stability returns. The history table turns those transitions from opinions into entries. Anyone reviewing the file a year later can see why the plan changed and when.
Sources
- Wikipedia: ISO 9000 —— Overview of the ISO 9000 family of quality management standards.
- ILAC: International Laboratory Accreditation Cooperation —— ILAC homepage describing the international cooperation on laboratory and inspection body accreditation.
Frequently asked questions
How do I read the sample size from the level and lot size?
Start from the agreed lot size and read across the row for the chosen level to find the code letter, then read down to the sample size for that letter. Apply the agreed AQL values per defect class to the defects found in that sample. The verdict only means something when the lot, the level, and the AQL values were all fixed before the inspection.
What are tightened and reduced inspection?
Tightened inspection means larger samples or stricter acceptance after quality problems, while reduced inspection means smaller samples after a sustained run of good lots. Both switches should follow written rules agreed in advance, not ad hoc decisions on the day. For molded parts, a cavity change, resin change, or new operator team usually resets any move toward reduced inspection.
Why can a passed lot still contain a pocket of defects?
It means the sample was drawn from only part of the production, often the most accessible cartons, so clustered defects in the unsampled portion were never given a chance to appear. Cavity-specific flash, color drift after a mid-run material top-up, and warpage in early-shift output are classic cases. Spreading carton selection across the whole lot is the main defense.
What belongs in the sampling record?
At minimum the lot definition, the level, the sample size actually drawn, the AQL values applied, the defect log with photos, the instruments used, and the pass or fail verdict with its reasoning. Calibration status of gauges matters because the credibility of measurement results depends on recognized accreditation behind the laboratory or calibration chain. A report missing any of these elements is hard to defend in a claim discussion.