Quick answer
To approve plastic samples before mass production, first freeze the acceptance basis. Next, verify sample traceability and test representative parts from every active cavity under production-intent conditions. Approve dimensions, appearance, assembly, function, material and compliance evidence, and pack-out separately. Finish with a signed disposition and change triggers.
A handpicked sample can look perfect while hiding cavity imbalance, unstable processing, sorting, rework, or a substitute material. A passing first article answers, “Can this setup make a conforming part?” It does not automatically answer, “Can the normal process keep making acceptable parts at the required rate?”
The release rule: no mass production until every required check has a pass, an approved temporary deviation, or a clearly assigned corrective action that blocks release.
Six questions a sound approval must answer
Know which kind of sample you are approving
Terms such as T0, T1, prototype, first-off, first article, pilot sample, golden sample, and limit sample are not used consistently by every buyer and factory. Do not release an order because a label sounds advanced. Define the sample’s source, required evidence, and allowed next step.
| Sample stage | What it can prove | What it does not prove |
|---|---|---|
| Prototype | Concept, fit, capacity, handling, or selected tests. | Production material, finish, shrinkage, rate, or capability. |
| Tool-trial sample | Tool operation, fill, ejection, early dimensions, and defects. | Final process, all-cavity balance, stable output, or final pack-out. |
| First article | A representative production item meets defined requirements. | Long-run capability, delivery rate, or every future lot. |
| Production-intent pilot | The planned process, rate, yield, controls, and pack flow work together. | Unchanged performance after an unreviewed change. |
| Golden or master sample | Approved appearance or assembly feel when attributes are named. | All drawing, test, legal, or process requirements. |
| Limit sample | The agreed boundary for a named subjective defect. | Permission for worse defects or unrelated nonconformance. |
Gate 1 · freeze the target
Freeze the approval basis before the sample run
A supplier cannot prove conformance to a moving target. Release one controlled package with the part number, CAD and drawing revision, critical dimensions, general tolerances, material specification, color and texture references, artwork, mating-part revisions, performance tests, appearance zones, packaging, labeling, and applicable market requirements.
For each requirement, name the method, sample condition, acceptance limit, and owner. If a lid must survive filling, sealing, freezing, shipping, and consumer opening, a room-temperature fit check alone is not the acceptance plan.
Also define which differences require buyer authorization. A sample that matches an obsolete drawing or an unapproved resin is not approvable, even if it looks good.
Demand sample genealogy, not a bag of anonymous parts
Genealogy is the recorded origin of each sample. It connects the physical part to the conditions that produced it. Label each group so a result can be traced to:
Record actual conditions. A form that says “production material” or “standard settings” is weak evidence. If a failed test appears later, the team needs enough detail to reproduce and investigate the sample.
Gate 2 · represent production
Use a production-intent run and expose the full output
A production-intent run uses the planned production tool, resin, press, auxiliaries, cycle, process window, secondary work, inspection, operators, and packaging as closely as practical. Record every departure. A slow, heavily adjusted trial on a special machine may help improve a tool, but it should not silently authorize normal production.
After a justified stabilization period, collect traceable parts from every active cavity at the beginning, middle, and end of the run. Add material lots, shifts, or restarts when risk requires them. Keep cavity data separate; pooled results can hide one weak cavity.
Ask for total output, accepted quantity, scrap, rework, sorting, downtime, alarms, setting changes, blocked cavities, and manual intervention. The best ten pieces tell less than the honest run history.
Quarantine and identify samples before testing
When samples arrive, keep them on hold. Compare labels, count, cavity marks, revision, material records, process report, and requested deliverables. Photograph the received condition. Separate groups without breaking their identity. If parts are missing labels or mixed, stop and resolve the genealogy before measurements consume time and create doubtful data.
Inspect for shipping damage, contamination, incomplete conditioning, or distortion from tight packaging. Record when the part was molded and when it was measured. Some plastics continue changing after molding or absorb moisture; the timing and environment can affect a close result.
Never let sample urgency erase control. An unlabeled sample approved from a phone photo becomes an unreliable reference, especially after the team, tool, or supplier changes.
Gate 3 · verify the part
Approve dimensions with a usable measurement plan
Start with a ballooned drawing or numbered characteristic list. Link every measured result to its requirement, sample ID, cavity, method, instrument, operator, and date. Review critical-to-quality features first, but do not ignore an unmeasured requirement by calling the layout “complete.”
Agree the datum setup, fixture support, clamp force, free or assembled state, instrument program, conditioning time, temperature and humidity, and decision rule near a limit. ISO 291:2008 provides standard-atmosphere concepts for conditioning and testing plastics. ISO 14253-1:2017 addresses conformity decisions that consider measurement uncertainty.
If buyer and supplier gauges disagree near a limit, quarantine the result and run the agreed correlation or referee method. Do not convert a measurement-system disagreement directly into a tool change.
Conformance is not capability
One passing part cannot prove repeatable production
A first article verifies that a representative item meets defined requirements. Process capability compares a stable process’s variation and centering with specification limits. It needs representative observations over time, trustworthy measurement, cavity identity, and an appropriate statistical method.
IL NIST process-capability guidance explains that capability estimates depend on stability and distribution assumptions. Do not demand a universal sample count or capability index without defining the study. The number should follow feature risk, production volume, customer rules, cavities, process behavior, and the cost of failure.
Practical distinction: use the first article to verify product definition. Use the pilot run, time-sequenced data, yield, and capability evidence to verify the production system. Keep both decisions visible in the approval record.
Verify the exact material and the compliance file
A resin family such as PP, ABS, PC, or PA does not identify a production material. Verify the manufacturer, exact grade, additives, reinforcement, color masterbatch, recycled or regrind rule, and supplier source. Review the certificate of analysis or conformity when required, but make sure the document matches the actual lot and specification.
For regulated or safety-related products, sample approval and compliance approval are separate gates. The test report must match the submitted material, design, manufacturing source, intended market, and conditions of use. A past report for a similar material does not automatically cover the production article.
Food-contact example: the U.S. FDA explains that the status of food-contact material depends on the substances expected to migrate and their authorization, specifications, limitations, and intended conditions. Some Food Contact Notifications are manufacturer- and use-specific. Verify the actual resin, additives, food type, time, temperature, and repeat-use conditions; a good-looking or dimensionally conforming sample does not prove full U.S. food-contact compliance. See the FDA food-contact status guidance.
Gate 4 · appearance and workmanship
Use golden and limit samples for named attributes
A golden or master sample shows the approved target. A limit or boundary sample shows the edge of acceptance for a defined defect. They are most useful for color, gloss, texture, printing position, weld lines, flow marks, specks, sink, flash, mismatch, trimming, and assembly feel that written words alone cannot describe well.
Define viewing distance, time, angle, light source, background, appearance zones, and whether the reference is an ideal target or an acceptance boundary. Identify every physical reference with the part, revision, material, cavity, date, controlled attribute, and signatures. Store protected duplicate sets with buyer and supplier, and review them for fading, wear, contamination, and deformation.
A master sample does not cancel the drawing, functional tests, or law. If it controls color but has an out-of-tolerance dimension, write “color reference only.”
Test the assembled function and real use condition
A part can pass every listed dimension and still fail in use. Assemble it with controlled production-intent mating parts and check the functions that matter: insertion and removal force, snap retention, hinge movement, seal or leak performance, torque, load, drop, stack, nesting, filling-line handling, closure opening, visual gap, or electrical and thermal behavior.
Use both sides of an important mating tolerance where practical. One “friendly” mating sample may hide a stack-up problem. Record the fixture, equipment, speed, orientation, number of cycles, environment, acceptance limit, raw results, and failure mode.
Gate 5 · pilot and delivery
Make the pilot test the whole production flow
A useful pilot does more than keep the molding press running. It challenges resin handling, changeover, startup approval, all active cavities, part removal, trimming, printing or labeling, inspection, assembly, counting, packing, and material movement at the planned rate.
Verify that downstream stations are not the real bottleneck. Record good output per hour, scrap by reason, rework, unplanned stops, labor, cavity restrictions, and the setting window. If the supplier selects and repacks only good-looking pieces before the buyer sees them, the pilot has not demonstrated the release system.
Approve pack-out with the production count, nesting direction, separators, bags, cartons, pallet pattern, labels, storage, and transport conditions. Inspect scuffing, deformation, contamination, mixed parts, count accuracy, barcode readability, and damage after the required distribution test.
Choose sample quantity by risk—not a universal number
Five parts may be excessive for a destructive concept test and far too few for a multi-cavity capability study. Decide quantity after asking what the test must prove, how many cavities and material lots are involved, whether results vary over time, whether the test destroys the part, and how severe a missed failure would be.
As a robust general rule, every active cavity should be represented and traceable for the relevant checks unless a documented risk-based plan justifies another approach. Time-sequenced samples are needed when repeatability or drift matters. A critical safety characteristic may need a special validation or control plan. Document the rationale instead of copying a generic number from another project.
AQL is not sample approval. ISO 2859-1:2026 provides AQL-indexed sampling plans for lot-by-lot attribute inspection in a continuing series. It does not approve the design, prove process capability, promise zero defects, or mean every accepted lot contains the AQL percentage defective. Define lot, defect classes, sampling plan, acceptance and rejection rules, and lot disposition separately.
Gate 6 · the release decision
Write a disposition that cannot be misread
Avoid “looks good,” “OK to proceed,” or an unqualified signature on a sample bag. State the approved part and revision, attributes, tool and cavities, material, site and process, quantity or order scope, open issues, containment, expiry, and required resubmission.
Importante: artwork approval is not dimensional approval. Color approval is not functional approval. Tool-trial approval is not mass-production approval. Use separate boxes or explicit wording so one signature cannot be stretched beyond its evidence.
What belongs in the signed approval pack?
The pack can be simple for a low-risk product and more formal for a regulated or high-consequence part. It should still preserve one evidence chain:
Utilizzo PPAP when the automotive customer, contract, or applicable system requires it. IAQG 9102 is a first-article framework for aviation, space, and defense and may be invoked elsewhere. For ordinary custom plastic products, build a scaled approval dossier with the necessary evidence without falsely calling every sample pack PPAP or aerospace FAI.
When a sample fails, verify the cause before changing the tool
First confirm sample identity, drawing revision, test method, fixture, conditioning, cavity, material, process record, and mating parts. Re-measure a retained sample with the agreed method when correlation is doubtful. Then separate design, tool, material, process, measurement, handling, and packaging causes.
Ask the supplier to describe the nonconformance, affected cavities and time range, suspected cause, containment, proposed correction, risk to other features, and resubmission plan. If the tool changes, identify the modified area and recheck features that may be affected. A correction to one dimension can move another, alter wall thickness, appearance, venting, ejection, or assembly.
Keep failed and corrected evidence linked. Replacing the report with a clean final version hides learning. A controlled history helps future teams understand why the tool, limit sample, or process window is defined as it is.
Control the first production lot and future changes
Sample approval starts a controlled production baseline; it does not end quality planning. Apply enhanced checks to the first mass-production lot: identity, all cavities, critical dimensions, appearance, functional tests, packaging, quantity, traceability, and comparison with the approved references. Review early complaints and process trends quickly.
Your quality agreement should require notice before changes to the drawing, resin grade or manufacturer, additives, colorant, regrind, tool repair, insert, cavity, gate, cooling, press, auxiliary equipment, automation, process window, manufacturing site, subcontractor, secondary operation, test laboratory, artwork, or packaging. Also define review after a long shutdown, tool damage, transfer, repeated escape, or damaged reference sample.
Not every change needs a full repeat of every test. Use the change’s risk and customer or sector requirements to choose partial or full resubmission. Document the rationale and never allow the supplier to decide silently that a change is harmless.
Fast buyer check
Before you sign: a ten-minute approval review
Frequently asked questions
Can I approve mass production from a 3D-printed prototype?
Usually not. A prototype can approve selected design decisions such as shape, fit, capacity, or ergonomics. It normally does not reproduce the production resin, mold, shrinkage, surface, cycle, cavities, process variation, or pack flow. State the attributes it approves and require production-intent evidence later.
How many samples should I request from each cavity?
There is no universal number. Include every active cavity, then add a time sequence and enough observations for the decision being made. Dimensional conformance, capability, destructive reliability, color approval, and lot acceptance need different plans. Write the sample rationale into the approval plan.
Is a golden sample enough to control quality?
No. It is useful for named visual or tactile attributes, but it cannot replace measurable dimensions, functional tests, material identity, legal requirements, process controls, or packaging specifications. Mark the attributes it controls and keep protected, identified duplicates.
Only if the agreed release plan says so and the other gates also pass. Confirm sample genealogy, production-intent conditions, appearance, assembly, function, material and compliance evidence, pilot performance, and packaging. A report can also hide an unmeasured requirement, mixed cavities, or a nonrepresentative run.
What if I must start production with one open issue?
Use a written conditional approval or deviation. Define the nonconformance, risk review, containment, allowed quantity or order, affected cavities or attributes, inspection, segregation, owner, expiry, and resubmission. Never let a temporary commercial need become a permanent unwritten tolerance.
Should I use PPAP for a non-automotive plastic product?
Use it when a contract, customer, or relevant quality system requires it. Otherwise, scale the approval package to product risk. You can still use the useful principles—released requirements, production-intent samples, traceability, measured evidence, capability where needed, and controlled sign-off—without labeling the package PPAP.
Supplier evidence, not marketing shorthand
Plan sample approval before the mold trial
Quello di Honokage IML solutions page states that its workflow can include design, development, prototyping, in-house moldmaking, production, inspection planning, and logistics. Its testing page shows supplier-listed steps from incoming-material checks and mold setup through first-sample and finished-product inspection. These are supplier-provided capability statements, not project approval evidence by themselves.
For a real quotation, ask for the project-specific sample plan, tool and cavity identity, material traceability, measurement methods, pilot conditions, acceptance criteria, records, and change controls. Honokage’s FAQ page also describes a supplier-side sequence from product design and manual sample confirmation to mold and trial-product confirmation; buyers should convert that sequence into signed gates.
Need a project-specific sample and pilot plan?
Send the drawing revision, critical requirements, exact resin and use conditions, mating parts, tests, color or artwork standard, annual volume, production rate, packaging, and target market. Ask the team to return proposed sample stages, quantities, evidence, timing, and release gates.
Technical sources and scope
Standards status and technical links were reviewed on August 13, 2026. Project contracts, customer-specific requirements, laws, and current editions take priority.
AIAG PPAP · IAQG 9102 First Article Inspection · NIST Process Capability
ISO 291:2008 · ISO 14253-1:2017 · ISO 2859-1:2026 · ISO 10007:2017 configuration management
This buyer guide is general information, not a substitute for product-specific engineering, regulatory, statistical, or legal review. Define acceptance and release requirements with the responsible customer, manufacturer, laboratory, and qualified specialists.