Scan to CAD: what you actually get back from the service, and what it costs

A scan-to-CAD service sounds like one product. In practice it is two jobs joined together, and a quote only makes sense once you know which half you are buying. The first is capturing the part as a 3D scan. The second, and usually the larger, is turning that scan into a clean, editable solid model that a designer, a machinist or a printer can actually work from. Most disappointment in this area comes from people expecting the second job and receiving the first.

This is the buyer's view of what comes back from our scanning and reverse engineering service: what the deliverables are, how to tell a proper rebuild from a shortcut, and what moves the price.

Mesh versus solid: the difference that decides everything

A scan produces a mesh: a very large number of tiny triangles that record the surface exactly as it was, including wear, casting flaws and the dent where someone dropped it. Meshes (STL, OBJ, PLY) are perfectly good for printing a copy, keeping a record or comparing a shape against a reference. What a mesh does not have is design intent. It does not know that a hole is a hole, that two faces are meant to be parallel, or that a wall should be three millimetres thick. You cannot change a bore diameter, add a fillet, or hand it to a machine shop and expect clean toolpaths.

Scan-to-CAD is the work of replacing that mesh with a parametric solid model. Someone sits down with the scan as a reference and rebuilds the part feature by feature in CAD software, which is why it sits with our custom CAD modelling work rather than the scanner. The result is normally delivered as a STEP file, which opens in any mainstream CAD package, and in a native format if you tell us which software you use. A simple rule of thumb: if the goal is a printed copy, a mesh is often enough and is the cheaper route. If the goal is to change the part, make it by another process, or keep an editable master, you want the solid.

What a proper rebuild should contain

A good rebuild is not the mesh wrapped in a skin. When you open the file you should find real geometry: planes, cylinders and cones fitted to the scan data, so that a bore is genuinely round and a mounting face is genuinely flat, rather than a lumpy approximation of both. You should find a feature tree you can edit, so that changing a hole diameter changes the hole and nothing else. Damage, wear and moulding flash should be dealt with on purpose, not copied across by accident. And the dimensions that matter to you, the ones you told us about, should be held and checked.

We also compare the finished model back against the scan and can show you where the two differ, as a colour map. That is a useful sanity check and it is the reason you can trust the rebuild. It is not, however, a certified inspection report, and we do not describe it as one.

Copy the part as it is, or as it was meant to be?

This is the decision most buyers have not thought about until we ask. A worn part scanned faithfully gives you a model of the worn part. If the thing it mates with is also worn, that may be exactly what you want, because the new part will fit the old assembly. If you are replacing a part on a good machine, you probably want the geometry as designed: sharp edges restored, worn bearing surfaces returned to nominal, a bent flange straightened. Both are legitimate. They are different jobs, they take different amounts of modelling time, and it is far cheaper to settle the question before the work starts than after.

A worked example (representative)

This is a representative job, not a specific customer. A small manufacturer has an injection-moulded plastic bezel from a control unit that is no longer made. The tool has gone and they need forty replacements. We scan one good bezel, then rebuild it as a STEP model with the snap-fit clips, the screw bosses and the display recess modelled as proper editable features. Because the original is slightly warped from years in service, we agree up front that the rebuild should follow the intended flat faces, not the warp. We print a test fit in resin to check the clips engage, adjust the clip thickness by a fraction of a millimetre, and only then print the batch. The same loop of scan, rebuild, test print and adjust is what got the part right in our ribber dial case study, where a small error in the model was enough to stop a hundred-year-old mechanism meshing.

Notice what the customer ends up with: not just a part, but a model they own. If the bezel needs a design change next year, nobody has to start from a scan again.

What drives the cost

We quote scan-to-CAD per job, because two parts of the same size can differ by a factor of ten in modelling time. The capture half, covered by our 3D scanning service, is usually the smaller cost. The hours sit in the rebuild, and they are driven by a handful of things: how many features the part has, whether it is mostly prismatic (flats, bores, ribs) or mostly sculpted, how tight the tolerances are on the dimensions that matter, how much repair is needed, and how many of the part's surfaces genuinely need to be editable. That last point is where you can save money. A hybrid approach models only the critical features parametrically and leaves the rest as scan data, which is often enough for a printed replacement. For the full workflow behind this, see our reverse engineering page.

We give a written estimate before any modelling starts, saying what is included and in which file formats, so you can see what you are paying for and choose a lighter option if a full rebuild is more than the job needs. Photos, rough dimensions and a note of what you intend to do with the result are enough to get a first view.

If you have a part you need as an editable model, send photos, the dimensions you care about and what you plan to do with the file. We will tell you whether a full rebuild, a hybrid or just a mesh is the right fit, and give you a written estimate.