Quality control with 3D scanning has become a major tool for printed, machined, moulded, thermoformed and assembled parts. Unlike point measurements with calipers or a CMM, 3D scanning provides a global view of the geometry. It reveals deformation, shrinkage, warping and form errors across the whole surface. But this rich data must be structured to avoid misleading conclusions.
1. Define the inspection objective
Before scanning, define what you need to validate. Are you checking dimensional conformity, comparing several batches, analyzing deformation after 3D printing, validating tooling or documenting a supplier non-conformity? Each objective influences the alignment method and final report.
2. Prepare the part and CAD model
The part must be clean, stable and accessible. Glossy surfaces may be matted. The CAD file must match the correct revision, unit system and orientation. A CAD version error is a common cause of false deviations. For assemblies, decide whether to inspect each component or the assembled unit.
3. Scan acquisition
The scan must cover the inspected areas with sufficient overlap. Markers can stabilize tracking on large or low-texture parts. It is better to capture functional areas from several angles to reduce holes and noise. The chosen resolution must remain coherent: scanning a ±0.5 mm tolerance part at very high density does not improve the decision, but it does make processing heavier.
4. Choose the right alignment
Alignment is the core of inspection. A global best-fit minimizes the average deviation between scan and CAD. It is useful for visualizing overall deformation, but it can artificially move a part and hide a local defect. Functional-reference alignment uses the drawing datums: bearing planes, axes, bores or mounting surfaces. This is often the most relevant choice for deciding conformity.
In some cases, local alignment on a stable area is used to analyze the displacement of another area. For example, aligning a flange on its three bearing points and then measuring the drift of a boss.
5. Read a colour deviation map
A colour map indicates deviations between the real surface and the CAD model. Red may mean excess material and blue missing material, depending on the software convention. The colour scale must be set to the real tolerance. A scale that is too tight dramatizes insignificant deviations; a scale that is too wide hides defects.
Stray points must also be filtered: thin edges, poorly scanned areas, dust, reflections or filled holes. The report should distinguish reliable measured deviations from low-confidence areas.
6. Dimensional measurements and GD&T;
The deviation map does not replace critical dimensions. For complete inspection, extract distances, diameters, centre distances, flatness, coaxiality or surface profiles according to the drawing. 3D metrology software can generate reports with annotations, sections, histograms and deviation statistics.
7. Define an acceptance rule
A real part always contains noise and small variations. It is therefore useful to define a rule before measurement: tolerance by area, acceptable surface percentage, exclusion of non-functional zones, maximum threshold on critical interfaces. For example: the outer surface can accept ±0.30 mm, but mounting surfaces must remain within ±0.08 mm.
8. Report and traceability
A quality report should include the part reference, CAD version, scanner, date, resolution, alignment method, tolerances, screenshots and conclusions. This traceability makes it possible to compare batches and defend a decision with a supplier or customer.
Conclusion
Comparing a real part to CAD by 3D scanning is extremely powerful, provided the inspection is treated as a complete measurement method. Acquisition, alignment, tolerance, deviation interpretation and reporting must be coherent. That is how 3D scanning becomes a true quality control tool, not just an attractive colour image.
Machines-3D resources to go further
