The accuracy figure on a scanner’s spec sheet comes from a calibrated test object measured under controlled conditions. Your part will be larger, shinier, or harder to hold still, and each of those adds error. 3D scanner accuracy is how close the measured points land to the true surface. Published figures run from about 0.02 mm on metrology-grade handheld systems to about 0.1 mm on general-purpose handhelds, while phone-based LiDAR (light detection and ranging) lands around a centimeter.
The number worth asking for is the accuracy your tightest tolerance needs, since anything beyond that adds cost without changing the finished part. It also helps to keep accuracy apart from resolution, because a scan can be rich in detail and still measure wrong. If the technology itself is new to you, start with our overview of what 3D scanning is.
How Accurate Are 3D Scanners?
Published figures are given in millimeters or microns, and they vary widely by class of equipment. Manufacturers measure them under good conditions, so read each one as the best case:
- Metrology-grade handheld systems: about 0.02 to 0.03 mm, verified against a published acceptance test. They are portable, so the scanner goes to the part.
- Professional handheld structured light scanners: about 0.05 to 0.1 mm. Models built for small parts sit at the fine end, and large-volume models sit at the coarse end.
- Photogrammetry: the model is solved from many overlapping photographs, so accuracy depends on the camera, the number and layout of the shots, and the size of the subject. Its strength is holding accuracy across big objects, which is why it anchors scans of large structures.
- Phone-based LiDAR: about plus or minus 1 cm on objects larger than 10 cm in published testing. Fine for visual reference, too coarse for parts that have to fit.
We scan with a Creaform Go!SCAN SPARK, a handheld white light scanner rated by its maker to 0.050 mm. Read the volumetric accuracy as well as that headline number. The SPARK’s volumetric figure is 0.050 mm plus 0.100 mm per meter, so the length term equals the base figure at half a meter, and a one-meter part can carry up to 0.15 mm. For anything longer than about half a meter, we would compare scanners on the volumetric figure, since that is the one describing your part.
3D Scanner Resolution vs Accuracy: What Is the Difference?
Resolution is how much surface detail a scanner records, usually given as point spacing: the distance between neighboring points in the point cloud a scan produces. At 0.1 mm spacing, engraving, sharp edges, and small fillets stay crisp; at 1 mm they blur into the surface.
Accuracy is about position, and resolution is about density. Picture darts on a board: resolution is how many you throw, and accuracy is how close they land to the target. A scanner can place dense points slightly off, giving a detailed scan that is dimensionally wrong, or place sparse points exactly right, giving a true scan that smooths away fine features.
Match each one to the part. Fine texture on a decorative casting calls for resolution, while a bolt pattern on a bracket calls for accuracy. Paying for the extreme end of whichever one the part doesn’t need adds cost and nothing else.
What Changes 3D Scanner Accuracy on a Real Job?
Most of a good scan is controlling the error sources a spec sheet leaves out. These four are the ones to plan for:
- Size: error grows with length. On the SPARK’s volumetric spec, a two-meter part can carry up to 0.25 mm, five times the headline figure.
- Registration: most objects need several passes, and aligning them into one dataset lets small errors build up across the part. A network of reference targets spread over the whole object keeps that drift in check.
- Surface: shiny, dark, and clear surfaces return noisy or missing data. A vanishing scanning spray solves most of this. One coat adds roughly 0.008 to 0.015 mm, small next to a 0.050 mm spec but worth allowing for on the tightest features.
- Temperature and stability: parts deflect with handling, vibration, and heat, and a part that shifts on its stand mid-scan will not align cleanly.
We recommend settling how the part will be held and targeted before anyone picks up the scanner. A part that shifts, or a join with too few targets, can leave a scan that looks smooth up close while far-apart features sit in the wrong place.
Checking the result takes minutes. We would caliper two or three well-defined features on the part and measure the same features in the scan data, because agreement on features the alignment never used is an independent check that the passes came together correctly. A mismatch caught there costs a rescan instead of a part cut to the wrong shape.
How Much Accuracy Does Your Project Need?
A common rule of thumb in inspection, the rule of tens, asks for a measuring method about ten times finer than the tolerance band it checks. Where that is impractical, four-to-one is the usual floor, the ratio long written into U.S. calibration standards.
Applied to a 0.050 mm scanner, the arithmetic is short. The rule of tens covers tolerance bands of 0.5 mm and wider, roughly plus or minus 0.25 mm. The four-to-one floor reaches bands of 0.2 mm, roughly plus or minus 0.1 mm. Features tighter than that need a finer method, such as a metrology-grade system.
That split shapes first article inspection, where a part from the first production run is checked against its drawing. Before quoting that kind of job, we would sort the drawing’s tolerances into the ones a scan can check and the ones it cannot, so the report never leans on a number the scanner can’t support.
On a subwoofer enclosure for a Ferrari F355 GTS, the scan and the fit check worked as a pair. We scanned the rear deck behind the seats at 0.050 mm surface accuracy, modeled the enclosure against that scan, and checked the fit digitally in CAD (computer-aided design) before the nine-part body was cut from MDF (medium-density fiberboard) on a CNC (computer numerical control) router. The finished enclosure was then test-fit in the car itself.
Some measurements don’t need a scanner at all. For a handful of hole diameters or wall thicknesses, we would reach for calipers or a micrometer first, as our comparison of 3D scanning vs measuring tools explains. If you are weighing a DIY (do-it-yourself) scanner against hiring a service, our guide to DIY versus professional 3D scanning shows where equipment accuracy starts to matter.
The right method follows from the tolerance, so you pay only for the accuracy the part needs. Call 718-557-9578, or send the drawing and a note on what the scan data is for through our contact form. We will reply with the dimensions a scan can check and the ones that need another tool.
