Post-Machining 3D Printed Parts on a CNC Router: Adding Precision Features to FDM Prints

3D printing is excellent at complex geometry and terrible at the handful of features that actually need to be precise — a bearing bore that has to be within a couple thousandths, a flat mating face that can't have visible layer lines, a hole pattern that has to match a real fastener spacing exactly. Rather than fighting a printer for tolerances it was never designed to hit, this guide covers the increasingly common hybrid approach: print the complex bulk shape on an FDM printer, then finish the features that actually matter for fit and function on a CNC router. It's a workflow this site's CNC and 3D printing content each touch on separately but never as its own combined process.

Why Combine the Two Processes at All

FDM printing and CNC routing fail in almost exactly opposite ways. A printer can produce a organic, hollow, internally-complex shape that would be impossible or absurdly expensive to machine from solid stock — but its dimensional accuracy on any single feature is limited by layer height, hole compensation, thermal shrinkage, and the general fact that a hole "printed round" is really a many-sided polygon approximation. A CNC router, by contrast, struggles with the same complex internal geometry a printer handles easily, but can hold a bore, a flat face, or a hole pattern to real machinist tolerances all day long. Printing the bulk shape and reserving CNC time only for the two or three features that need to be precise gets you both strengths without paying for either weakness — full-part CNC machining of a complex shape in plastic wastes huge amounts of stock and machine time, while trying to print-to-tolerance on features a printer just isn't built for wastes prints and PLA.

Designing the Part for a Print-Then-Machine Workflow

The workflow only works if the part is designed for it from the start, not decided after the fact:

Leave stock material on features that will be machined. Any surface or bore that's going to see a cutting tool afterward needs to be printed slightly oversized (typically 0.3–1mm depending on tolerance needed and how much your printer's dimensional accuracy varies) so the CNC pass removes only printed material, not the surrounding structure — printing a hole to exact final size and then trying to "clean it up" with a machining pass leaves you removing material unevenly from an already-undersized feature.
Orient the print so the machined faces are accessible. A feature that needs a 3-axis CNC router to reach it in a single setup has to face up, or be on a face the part can be safely fixtured against — design the print orientation around the machining operation, not just around print quality.
Use a print pattern that gives the machined area solid backing. Standard sparse infill under a surface that's about to be faced or bored leaves the cutting tool breaking through into voids partway through the cut, which chatters and can tear the surface. Increase infill locally (a modifier mesh in most slicers, or simply increasing overall infill percentage for smaller parts) anywhere machining will happen.
Add fixturing features on purpose. Printed tabs, a flat reference face, or a registration boss that exists only to locate and clamp the part in the CNC's vise or fixture (and gets machined off or trimmed away afterward) makes workholding dramatically easier than trying to clamp an organic printed shape directly.

Workholding: The Actual Hard Part

Machinists take rigid, uniform stock for granted; a 3D print is neither. Printed parts are comparatively soft, can deform under standard vise clamping pressure, and rarely have a flat, parallel reference surface unless you deliberately printed one. The workholding techniques covered in this site's CNC workholding guide apply, with printed parts favoring a few specific approaches: double-sided tape or CA glue directly to a sacrificial spoilboard works well for parts with a flat printed face and light cutting forces, low-pressure soft-jaw or 3D-printed custom fixture clamps (yes, printing a fixture to hold another print) distribute clamping force over more surface area than a standard vise, and vacuum fixturing works if the part geometry gives enough flat sealing surface. Whatever method you use, keep clamping pressure well below what you'd use on wood or metal — it's easy to crush or distort a printed wall that a vise wouldn't even notice on aluminum.

Feeds, Speeds, and Material Behavior

MaterialMachining notes PLAMachines cleanly and predictably at modest speeds — behaves similarly to a soft, brittle plastic; low melting point means light cuts and adequate chip clearance matter more than raw feed rate to avoid local melting/regrinding of chips into the surface PETGGummier than PLA and prone to melting/re-welding onto the cutter at high speed or with dull tooling — sharp single-flute or two-flute upcut bits and slower spindle speeds than PLA reduce this significantly ABS/ASAMachines well and produces manageable chips, similar cutting behavior to machining cast acrylic in some respects, but generates more fine dust — treat it with the same ventilation caution as machining any thermoplastic NylonTough and somewhat gummy like PETG, benefits from sharp tooling and light passes; poor heat dissipation in the material means repeated light passes beat one aggressive pass

Across all of these, treat printed plastic more like machining a soft engineering plastic than like wood or metal: sharp bits matter more than raw horsepower, and lower spindle speeds than you'd use on the same material in a solid billet reduce melting and re-deposition significantly, since a printed part's internal structure (infill, layer boundaries) dissipates heat differently than solid stock.

Setting Work Zero on an Irregular Printed Part

Locating a precise work zero on a part that doesn't have a machined reference surface yet is the other real challenge of this workflow. The most reliable approach is designing a small flat reference feature into the print specifically for probing — even a 5x5mm flat boss in a known, documented location relative to the feature you're about to machine — and touching off from that with a standard touch probe or edge finder as covered in this site's CNC touch probe guide, rather than trying to eyeball zero off an organic printed surface. For features that need to align to something already on the printed part (a hole that has to be concentric with an already-printed boss, for instance), a dial indicator sweep off the existing printed feature gets you close enough for most FDM-tolerance applications, even though it will never match indicating off a truly machined reference.

When This Workflow Is (and Isn't) Worth It

This hybrid approach earns its extra setup time specifically for parts where one or two features need real precision inside an otherwise complex or large shape that would be impractical to machine from solid stock — a robotics chassis with precision bearing bores, an enclosure with a machined flat sealing face and printed complex internal structure, or a jig body with a few precision-bored locating holes. It's not worth the extra complexity for parts where the printer's native tolerance is already good enough, or where the whole part genuinely needs to be machined (at which point it should probably just be CNC-machined from stock or cast, rather than printed and then reworked).

Printing the bulk shape and reserving the CNC for only the features that actually demand precision plays to the real strength of each process instead of asking either one to do a job it wasn't built for — the extra design discipline of leaving machining stock and building in fixturing features pays for itself the first time a bearing bore actually fits without needing to be reamed, sanded, or epoxy-shimmed into submission.

Related Guides

Loading_