Metrology, explained straight
3D scanning vs a CMM: they answer different questions.
A coordinate measuring machine and a 3D scanner are both dimensional inspection tools, and they do not compete so much as cover different ground. One touches a few precise points, the other captures the whole surface at once. The honest answer to which is better is the one that names your part, your tolerance, and your feature. This page draws that line clearly, so you can tell where a scan is the right tool and where a CMM still wins.
What is the real difference between the two?
A CMM uses a touch probe to capture a few dozen or a few hundred discrete points per feature. Each point is placed with high single-point accuracy, and the contact method is largely immune to a part's color, finish, or the light in the room. That precision is why a CMM is the reference for tight callouts.
A 3D scanner captures the whole surface as millions of points in seconds, a full field instead of a handful of touches. That coverage catches things point probing misses, a warp between two probed points, a dent, a surface that drifted off nominal across a whole panel. The tradeoff is that optical accuracy is more sensitive to surface and environment than a probe.
What do the accuracy standards actually say?
Both technologies are held to published performance standards, and the specs are worth reading before anyone quotes you a single accuracy number. A CMM is verified under ISO 10360, which defines the acceptance and reverification tests that prove its length-measuring performance. Optical scanners are evaluated under VDI/VDE 2634 and the newer ISO 10360-13, using calibrated spheres and length artifacts to measure probing error, sphere-spacing error, and flatness across the working volume.
The number that matters is the one measured the way you will use the tool. A structured-light scanner holding about 0.02 mm on a granite table in a metrology lab can drift to 0.15 mm or worse once it moves to a warm, vibrating production floor. That is not a defect, it is the environment rewriting the result, and it is exactly why a lab-bench spec and a shop-floor spec are different conversations.
Which one wins, feature by feature?
| What you need | CMM | 3D scanning |
|---|---|---|
| Tight GD&T on a bore or edge | Wins | Supporting |
| Deep holes and internal features | Wins | Line of sight only |
| Free-form surface deviation, whole part | Slow, point by point | Wins |
| Speed across many features | Slower | Wins |
| Reverse engineering an existing shape | Impractical | Wins |
| Large part, no lab, on site | Hard to move | Wins |
The two are often used together. A scanner captures the whole surface fast, and a CMM confirms the handful of tight callouts that carry the closest tolerance. On many parts that pairing is the right answer, not one or the other.
Where a scan is the right tool.
When the question is the whole surface rather than a few tight features, scanning wins on coverage and speed. Free-form panels, castings, and complex shapes get a full deviation map instead of a scatter of probed points. Reverse engineering a part that has no drawing needs the entire form captured, which is a scan by definition. And a large part that cannot come to a lab gets measured where it sits.
When the deciding feature is a tight-tolerance bore, a sharp edge, or a deep internal geometry, a CMM still holds the reference. Knowing which of those describes your part is the whole decision, and we will tell you honestly which way it falls.
Tell us about the part and the tolerance.
Send us the part, the features that matter, and the tolerance you are holding to. We will tell you whether a scan, a CMM, or the two together is the right inspection for it, and quote the work.
Get a quoteQuestions about a measurement? Reach us at info@scanningandmodeling.com or (772) 210-4584.