Quick Answer
Mesh validity and manufacturability are different properties. A valid STL is watertight, free of self-intersections and consistently oriented; these are attributes of the triangle mesh. Millability additionally requires that the described geometry can be produced in a specific material on a specific machine: sufficient wall thickness, no unreachable undercuts, stable edges and an intact margin. Generic repair tools ensure the first. They cannot ensure the second.
“The file checks out fine.” Every CAM technician has heard this sentence from a lab, and every CAM technician has milled the counterexample.
Both sides are right, and that is the problem. The lab ran a mesh check, and the mesh is genuinely flawless: watertight, no intersections, clean normals. The milling center then loses a blank to a wall that was thinner than the material allows. The file was valid. The part was impossible.
This distinction, mesh validity versus manufacturability, explains most of the friction between labs and milling centers over “good” files that fail. It is worth getting precise about, because the tools most people use to check STL files only ever test one half of it.
Two words, two different tests
A valid STL is a statement about the triangle mesh. Every edge belongs to exactly two triangles, no surface passes through itself, every normal points outward, the expected number of shells is present. Validity can be tested by any competent mesh tool, with no knowledge of dentistry, materials or machines.
A millable STL is a statement about production. The geometry the mesh describes can be manufactured on the intended machine, in the intended material, and the result seats on the preparation. Millability cannot be tested without knowing the material, the indication and the machining setup.
Validity is necessary for millability. It is nowhere near sufficient. The full pre-CAM sequence is covered in How to Check Whether a Dental STL File Is Millable Before CAM.
Five files that pass every mesh check and still fail
1. The crown with a paper-thin fissure. Watertight, clean, beautiful. The occlusal fissure drops below the minimum wall thickness for the chosen ceramic. It survives milling and cracks at try-in, or it does not even survive milling. No mesh checker will ever flag it, because thin is not invalid.
2. The bridge with an unreachable connector zone. The geometry is closed and consistent, but the connector region cannot be reached from the tool directions available in a 3-axis setup. On a 5-axis machine the same file might pass. Millability depends on the machine; validity does not.
3. The coping with an undercut to the insertion path. Every triangle is correct. The restoration still will not seat, because a region of the inner surface undercuts the path of insertion. This failure is invisible to any tool that does not know what an insertion direction is.
4. The inlay with a ragged margin. The margin line wanders, breaks and runs out into segments thinner than any cutter can reproduce. The mesh around it is perfectly valid. The fit is not.
5. The case exported at the wrong scale. A flawless mesh at one tenth of the intended size. Validity checks pass; the “crown” is four millimeters wide. Only a plausibility rule that knows what a crown is will catch it.
Why generic STL repair tools stop halfway
Repair tools do exactly what their name says: they take a defective mesh and return a closed, consistent one. Holes get filled, fragments removed, normals unified. For mesh defects, automatic repair is genuinely useful, and a production workflow should include it.
But “repaired” describes the mesh, not the part. A repair tool has no concept of zirconia versus PMMA, of insertion directions, of margin lines, of what a premolar crown should measure. It will happily return a perfectly valid mesh of an unmanufacturable restoration, and it will do so with a green checkmark.
Repair answers: is this file well-formed? Production needs the answer to: can I clamp a blank for this and bill the result?
What a production-grade validation adds
The difference is domain rules on top of mesh checks. A production-grade validation knows the material of the actual blank and its minimum wall thickness. It knows the machining setup and computes reachability, not just geometry. It finds the margin line and judges its quality. It applies indication logic: how large a crown may be, what a bridge connector must carry. And it separates what can be fixed automatically from what has to go back to the designer with a marked-up report.
This is the layer SimQ DENTAL Mill Check adds on top of automatic mesh repair, per file, in seconds, in the browser. Every check is documented on What We Check.
FAQ
Is a watertight STL automatically millable?
No. Watertightness is a mesh property. Millability additionally depends on wall thickness, tool access, margin quality and material rules, none of which are visible to a mesh check.
My repair tool says the file is fine. Why did the case fail?
Because the tool tested validity, not manufacturability. It confirmed the mesh is well-formed. It never looked at wall thickness, undercuts, the margin or the material.
Who is responsible when a valid file is unmillable?
Practically: whoever detects it last pays the most. The economical answer is to detect it at file intake, automatically, and return unproducible cases to the designer with a visual report before any machine time is spent.
Can millability be checked without knowing the machine?
Only partially. Mesh integrity, dimensions, wall thickness and margin quality can be tested from file and material alone. Tool reachability depends on the setup, which is why serious validation is configured per machine strategy rather than one-size-fits-all.
Design with proof. Produce with confidence. See what a production-grade check finds in your files: start the free browser test. No installation, no credit card.