2026-09-24 · 7 min · By Alcott Dube
How to prepare a cad file for 3d printing
I prepare printable models by checking solid geometry, choosing realistic wall thicknesses, planning orientation and supports, then inspecting the exported file in a slicer.
To prepare a cad file for 3d printing, make it a closed solid, size walls and clearances for the chosen process, and choose an orientation that balances strength, finish and supports. Export at the correct scale with enough mesh detail, then inspect the sliced layers before printing.
Choose the printing process before changing the model
I start with the process, material and purpose of the part. A visual enclosure, a loaded bracket and a casting pattern need different decisions. For filament printing, I also need the nozzle diameter and intended layer height. Without those inputs, a minimum wall thickness is only a guess.
Protolabs publishes separate design guidance for stereolithography and selective laser sintering because their constraints differ. Stereolithography uses supports, while surrounding powder supports parts during selective laser sintering. That changes which cavities, overhangs and surfaces are practical. I wouldn't transfer a thickness recommendation from one process to another without checking the material-specific guidance.
For the working example, I'd use a small enclosure printed in pla with a 0.4 mm nozzle and 0.2 mm layers. Those are starting conditions, not a universal recipe. I'd identify the mating edges, screw locations and visible face before editing anything, because those features decide where compromises can go.
Check that the cad model is a closed solid
I check the native model before exporting it. Each intended solid should enclose a volume, without missing faces, accidental internal surfaces or edges shared by more than two faces. Multiple bodies are fine when they represent separate parts. Intersecting bodies that should form one printed object usually need a boolean union.
Thin surface geometry is a common trap. A surface can look complete on screen while having no printable thickness. I'd thicken it deliberately in cad, then inspect sections through ribs, corners and junctions. Automatic mesh repair can close a hole, but it can't reliably infer the engineering intent behind that hole.
I also check the bounding dimensions and assembly state. An enclosure exported with its lid fitted may become two trapped or overlapping shells. I'd export mating components separately unless the design deliberately uses a print-in-place joint. Repair tools belong after these checks, not instead of them.
Set wall thickness for the nozzle and material
For filament printing, I relate thin walls to the intended extrusion width, not just the nozzle diameter. With a 0.45 mm line width, a 1.35 mm wall can accommodate three lines in a straightforward section. Modern slicers can vary line width, so exact multiples aren't mandatory, but they remain a useful starting point.
That example is a toolpath decision, not a strength guarantee. A tall unsupported wall may flex at that thickness. A screw boss may split despite being much thicker. I'd add material around the actual load path, use fillets at abrupt junctions, and inspect how ribs connect to the outer shell. More infill won't rescue a badly placed thin neck.
For the enclosure, I'd trial walls around 1.35 to 1.8 mm and inspect the sliced result before committing. Small lettering, sharp tips and narrow ribs need their own check because they can disappear or become single-line features. For resin or powder printing, I'd replace these nozzle-based assumptions with the supplier's limits for the selected material and feature type.
Choose print orientation for strength and surface finish
I choose orientation before polishing cosmetic details. In filament printing, loading that pulls layers apart can expose weaker interlayer bonding. I'd rotate a loaded bracket so its main tensile load runs within the layers where practical, rather than relying on bonds between stacked layers. Material, temperature and print settings still affect the result.
Next, I rank the surfaces. A face against the build plate takes on its texture and may show first-layer expansion. A supported underside can carry contact marks. A shallow curved top may show visible stepping. I'd put the enclosure's most visible face somewhere that avoids support contact, unless that creates a worse mechanical compromise.
Height matters too: a tall, narrow part can be less stable and take longer than a lower orientation. I compare two or three plausible positions in the slicer and record estimated time, support material and critical surface placement. I don't choose the shortest print automatically if it weakens a mounting tab.
Reduce supports without hiding inaccessible geometry
I treat supports as temporary tooling that needs an exit route. Before generating them, I check whether pliers or another removal tool can reach every contact area. A supported ceiling inside a closed enclosure can turn a printable model into an unusable part. Soluble supports change the options, but still need suitable access for removal.
For ordinary filament printing, I use a surface angled about 45 degrees from vertical as an initial overhang test, not a guaranteed limit. Cooling, material, layer height and geometry all matter. Bridges are different: a short span anchored at both ends may print without support even when a comparable cantilever would fail.
I'd replace an unnecessary horizontal ledge with a chamfer, rotate a hole, or split the enclosure into two components before accepting extensive supports. Splitting adds alignment and fastening work, so it isn't free. I compare those costs with support removal, damaged surfaces and the risk of breaking a small feature during cleanup.
Export a 3mf or stl file at the correct scale
I keep the native cad file as the editable master. For transfer to a compatible slicer, I prefer 3mf because it can carry units and multiple objects. An stl file remains widely accepted, but it doesn't define units. I export in millimetres and verify a known dimension after import rather than trusting the apparent size.
Mesh resolution needs a deliberate choice. Too coarse, and a round boss becomes visibly faceted. Excessively fine, and the file becomes slow without improving the print. For a small mechanical part, I'd start with a maximum chordal deviation around 0.02 to 0.05 mm, then inspect the curves. That's a starting setting, not a supplier specification.
Cad exporters use different controls, sometimes combining surface deviation with angular tolerance. I check both where available and inspect small holes as well as the outside silhouette. A finely tessellated mesh won't correct an undersized hole or a weak wall. If a print service accepts native or step geometry, I'd follow its upload guidance instead of making a mesh unnecessarily.
Inspect sliced layers and print a fit test
The export isn't my final check. I open it in the target slicer with the actual printer, nozzle and material profile. UltiMaker Cura provides a layer preview; I use it to inspect the generated paths, not just the model view. A feature visible in the model can still disappear when sliced.
I scroll through thin walls, hole roofs, the first unsupported layers and the start of small islands. I look for missing paths, unexpected gaps, supports inside inaccessible spaces and sparse material beneath loaded features. I also check the first layer for adequate contact. A plausible time estimate says nothing about whether those details will succeed.
For mating parts, I'd print a small coupon before the full enclosure. A trial with 0.2, 0.3 and 0.4 mm clearance per mating side gives me a practical comparison, not a guaranteed fit range. I measure the result, adjust the cad dimensions and save the successful slicer settings separately. That keeps geometry corrections distinct from process adjustments when the next revision arrives.
Questions people ask
Can I send a cad file straight to a 3d printer?
Usually, I need to pass the model through a slicer first. Most filament printers require machine instructions generated for their configuration, while resin printers use their own prepared job formats. A print service may accept the native cad file and handle preparation.
What is the minimum wall thickness for 3d printing?
I wouldn't specify one minimum across every process and material. For filament printing, I start with extrusion width and the number of wall lines, then account for height and loading. For resin or powder processes, I check the provider's feature-specific guidance.
Is 3mf better than stl for 3d printing?
I prefer 3mf when the receiving software supports it because it records units and can contain multiple objects. Stl is useful for broad compatibility. Neither format makes unsuitable geometry printable.
Why are holes too small in my 3d print?
I check mesh resolution, extrusion settings, orientation and material behaviour before changing the hole diameter. First-layer expansion can also constrict openings near the bed. For a precise fit, I'd print a hole coupon or leave allowance for drilling.