Every plastics manufacturer knows the scenario: a batch of parts passes the go/no-go check, ships to the customer, and comes back rejected. Or the toolroom adjusts a mould based on CMM data, and the problem persists. The instrument wasn't the issue — the inspection method was.
Optical metrology — profile projectors, quick measuring machines (QMMs), and vision measuring machines (VMMs) — changes what's measurable in plastics inspection. Not by magic, but because it's genuinely better suited to the geometry and material behaviour of injection-moulded parts than contact-based methods.
This article covers the six most common dimensional failure modes in injection-moulded parts, why optical metrology detects them more reliably, and how to build an inspection protocol that actually feeds useful data back to your toolroom.
Why Injection-Moulded Parts Are Dimensionally Difficult
Plastics behave differently from metal. The same feature, measured on the same part, will give different readings at 30 minutes post-ejection versus 24 hours post-ejection. That's not instrument error — it's physics.
Three material-level factors make injection-moulded parts inherently harder to inspect than machined metal:
- Anisotropic shrinkage: Parts shrink at different rates in flow direction versus transverse direction. A part that's "in tolerance" on one axis may be significantly out on another.
- Residual stress relaxation: Warpage continues for 24–72 hours after ejection in unfilled materials. Inspecting too early produces data that doesn't represent the part's final geometry.
- Surface compliance: Soft or thin-walled plastic deforms under contact probe force. A 0.5N stylus on a 0.8mm wall is not measuring geometry — it's measuring material response.
CMMs handle none of these issues gracefully. Optical metrology handles all three.
The Six Dimensional Failure Modes Optical Metrology Catches
Warpage / Bow
Deviation of flat surfaces or reference planes from nominal. Particularly common in large flat parts, lids, and panels. A VMM measures flatness deviation across a surface in a single scan.
Sink Marks / Shrinkage
Localised volumetric shrinkage at thick sections. Detectable optically by surface profile deviation. Correlates with wall thickness ratios and cooling channel placement.
Parting Line Offset
Mismatch between mould halves creates a step at the parting line. Profile projectors measure this directly from the silhouette — no fixturing on the feature required.
Flash at Parting Line or Gate
Thin membrane of material where polymer has flowed into the parting line gap. Optical inspection detects flash as small as 0.05 mm — below go/no-go gauge resolution.
Gate Vestige Height
Residual material at the gate point after degating. Controls assembly clearance in snap-fit and mating surface designs. Measured directly on a profile projector overlay.
Draft Angle Deviation
Incorrect draft causes ejection marks and surface drag. VMMs measure angular deviation from nominal across the full surface, not just at a single point.
Which Instrument for Which Defect?
Not all optical instruments are equal for plastics inspection. The right choice depends on part complexity, production volume, and what specifically you're inspecting.
| Defect / Feature | Profile Projector | QMM | VMM |
|---|---|---|---|
| 2D contour profile | Excellent | Excellent | Excellent |
| Flash dimension | Excellent | Good | Excellent |
| Gate vestige | Excellent | Good | Excellent |
| Parting line offset | Excellent | Limited | Good |
| Warpage / flatness | Limited (2D) | Good | Excellent |
| Multiple features / GD&T | Manual / limited | Excellent (CNC) | Excellent |
| High-volume (100+ ppm) | Slow (manual) | Excellent (<30 s/part) | Moderate |
| Toolroom R&D | Good | Moderate | Excellent |
Most plastics manufacturers need two instruments: a QMM on the shop floor for production inspection (high-speed, multi-feature, CNC-driven), and a VMM in the toolroom for mould qualification and first-article inspection. Using a CMM for either is a compromise — it works, but it's slow, contact-dependent, and operator-sensitive.
Inspecting Injection-Moulded Parts?
Optomech's applications team can review your current inspection method and recommend the right optical system — including sample part trials before purchase.
What Most QA Teams Get Wrong About Plastics Inspection
Three mistakes appear repeatedly in plastics quality systems — and each one compounds the others.
1. Inspecting Parts Too Soon After Ejection
Polypropylene parts shrink approximately 1.5–2.0% over the first 4 hours post-ejection and stabilise near 24 hours. Measuring immediately gives dimensional data that's accurate — for that moment. But the part that ships to your customer has already changed.
Industry practice: allow parts to condition at 23°C ± 2°C for a minimum of 16 hours (ISO 294-4 for moulding shrinkage) before dimensional measurement. This is non-negotiable for critical features.
2. Measuring Only What the Drawing Asks For
Most customer drawings specify 8–12 critical dimensions. But a mould that's producing parts with 0.3 mm of warpage will fail at assembly even if all the specified dimensions are in tolerance — because the part doesn't sit flat.
A proper mould qualification protocol should include flatness, perpendicularity of datum features, and parting line offset — regardless of whether these appear on the customer drawing. These are the features that predict assembly failure.
3. Using the Same Fixture for Inspection and Moulding
A common toolroom practice: clamp a warped part in a fixture to "flatten" it, then measure. The part is now in tolerance. The part that ships is still warped. This produces perfect inspection records for non-conforming parts.
Optical measurement — particularly on a VMM with a non-contact sensor — measures the part as it sits in free-body state, which is the actual state the customer will receive.
Building an Optical Inspection Protocol for Injection Mouldings
-
Define the inspection tier Separate first-article inspection (FAI) from production inspection. FAI covers every GD&T callout on the drawing plus toolroom diagnostics (warpage, shrinkage deviation, draft). Production inspection covers critical functional features only, measured at throughput speed.
-
Condition parts before measurement Set a minimum conditioning time (typically 16–24 hours for unfilled materials, 4–8 hours for highly filled grades). Record ambient temperature and document it in the inspection record. Humidity matters for hygroscopic materials like nylon and ABS.
-
Establish datum reference frames Define a 3-2-1 datum scheme consistent with how the part will be assembled. Measure datums first — every subsequent measurement is only as good as the datum registration. Use purpose-made fixtures, not gravity or hand-holding.
-
Capture a full warpage map at FAI Use a VMM to measure Z-height at a grid of points across the primary flat datum. Flatness deviation, bow, and twist should be reported for every new mould or after any tooling modification. This catches mould errors before they become delivery problems.
-
Pre-load part programs for production On a QMM, pre-programmed CNC part programs eliminate operator-to-operator variation in production. Once the program is validated (Gauge R&R <10%), the measurement result is as repeatable as the instrument — not the operator.
-
Feed data back to the toolroom Export dimensional data in SPC-compatible formats. Trend analysis on shrinkage deviation and parting line offset over tool life is a leading indicator of tool wear — not trailing confirmation that parts are bad.
Optical Metrology and Mould Qualification: The Toolroom Benefit
The biggest underused application of optical metrology in plastics is mould qualification — and it's where the return on investment is most obvious.
A new mould goes through typically 3–6 trial shots before it's approved for production. Each trial involves adjusting process parameters, measuring the output, and iterating. With manual gauging or CMM inspection, each measurement cycle takes 2–4 hours. A VMM reduces this to 20–40 minutes per part, with a complete GD&T report that tells the toolmaker exactly which features are out and by how much.
The practical result: mould qualification that took 2 weeks now takes 3–4 days. That's not a marginal improvement — it's a toolroom capacity multiplier.
If your toolroom is spending more than one day per trial shot on dimensional inspection, you need a VMM — not a faster CMM. The bottleneck isn't instrument speed; it's the combination of part fixturing, manual measurement, and data compilation that CMMs impose. VMMs eliminate all three.
What Optical Metrology Cannot Replace
It's worth being direct about the limits. Optical metrology for injection mouldings works exceptionally well for 2D and 3D feature geometry. It has genuine limitations:
- Deep internal features: Profile projectors and VMMs measure what the camera or optic can see. Blind holes, undercut features, and internal thread geometry require tactile measurement or CT scanning.
- Very large parts: Standard QMMs and VMMs have measuring ranges up to ~600 mm × 500 mm. Automotive bumpers, large panels, or complex assemblies may need CMM capacity.
- Translucent or reflective surfaces: Standard optical systems can struggle with highly transparent or mirror-finish parts. Frosted glass, some clear polycarbonate, and chrome-plated surfaces may require special lighting configurations or contact methods.
For the majority of injection-moulded components in automotive, consumer goods, medical device, and packaging applications — parts in the 10 mm to 300 mm range with standard finishes — optical metrology is not just adequate. It's the better method.
Real-World Performance Benchmarks
Based on Optomech installations at Indian plastics manufacturers:
- Inspection time per part reduced from 3–6 minutes (manual gauging) to under 45 seconds (QMM) for multi-feature inspection of auto plastic components
- Gauge R&R improved from 18–30% (manual) to 6–9% (QMM CNC program) on critical diameter and profile features
- Mould trial cycles reduced by 40–60% in toolrooms using VMMs for first-article analysis
- Parting line and flash defects caught at source, before 100% rework cycles were triggered at the customer end
These are not exceptional results — they're what you should expect from a properly implemented optical inspection system. The caveat is always in the implementation: fixturing, conditioning time, datum definition, and a validated part program.