CNC Turning and 3D Printing: How Each Process Forms a Part

• 1
  How each process forms a part
• 2
  Where the tolerance and surface limits sit
• 3
  Geometry decides the process before cost does
• 4
  What drives cost and lead time in each route
• 5
  When each process is the wrong answer
• 6
  A five-step way to choose between them
• 7
  CNC turning and 3D printing compared on the shop floor

CNC turning holds a bar or casting in a chuck and spins it. A single-point tool travels along the Z and X axes and peels off material. The part is round by definition, because the workpiece rotates. A mill-turn center adds a milling spindle and a Ø400 mm rotary table, so the same setup can cut flats, cross holes and slots without re-fixturing.

3D printing goes the other direction. A slicer converts the CAD body into layers, then the machine deposits or cures material a layer at a time. Fused deposition modeling pushes a thermoplastic filament through a heated nozzle. Resin processes cure a liquid photopolymer with a laser or a projector. Metal powder processes weld or sinter each layer.

The practical split follows from that difference. Turning needs a continuous chip to form, so the material must be machinable, and the tool must reach the surface. Printing needs a support strategy, so overhangs, internal channels and trapped volumes drive the setup more than the material does.

Both processes start from the same 3D model. The difference is what the machine does with it: turning reads a toolpath, printing reads a stack of 2D contours.

• 1
  Turning is subtractiveStock is removed by a defined cutting edge.
• 2
  Printing is additiveMaterial is added layer on layer from a digital model.
• 3
  Turning needs reachIf the tool cannot enter, the feature cannot be cut.
• 4
  Printing needs supportOverhangs and internal voids set the setup plan.

Turning holds ±0.005 mm (±0.0002 in) on diameters and faces when the setup is rigid and the thermal drift is controlled. That number comes from a stable machine, a sharp insert and a short tool overhang. Long boring bars, thin walls and deep pockets push the achievable band wider, sometimes to ±0.05 mm.

Surface finish moves with the same variables. A fine turning pass reaches Ra 0.2–0.8 μm. A normal production pass lands at Ra 0.8–1.6 μm, and a roughing cut sits at Ra 1.6–3.2 μm. Feeds and speeds set the scallop height between the tool marks, so finish is a choice you make in the program, not a fixed property of the machine.

Printed parts behave differently. The layer thickness sets the vertical resolution: 0.1 mm layers read smoother than 0.3 mm layers on the same machine. But the as-printed surface still shows layer lines, and the XY accuracy depends on the extrusion width or the laser spot, not on a single-point tool.

A printed bore rarely holds a press fit straight off the bed. The usual path is to print oversize and ream or bore the critical diameter on a lathe. That hybrid route keeps the printed geometry and puts the tolerance where it matters.

Holes are the clearest example. A turned and reamed hole can hold a sliding fit for a shaft. A printed hole usually needs 0.2–0.4 mm of clearance, or a drilled finish pass.

Ask one question first. Can a cutting tool reach every surface? If the answer is yes, turning usually wins on tolerance, finish and material choice. Turned shafts, bushings, fittings, valve bodies and connector shells all fall into that group, and the round form is an advantage rather than a constraint.

If the answer is no, printing earns its place. Conformal cooling channels, lattice infill, hollow shells with internal ribs and organically shaped brackets are hard or impossible to cut. A printed version exists in days without a fixture, and the internal geometry costs nothing extra in setup.

Sharp internal corners are a turning problem, not a printing problem. A tool nose radius sets a minimum fillet, so a square internal corner needs an EDM or a broach. Printing produces that corner directly, which matters for flow paths and for parts that seat against a mating face.

Thin walls split the two processes as well. Printed walls below 1 mm print reliably on resin and FDM. Turned walls below 0.5 mm deflect under cutting force, so they need light passes, a support mandrel or a change of process.

Threads follow the same logic. Cut threads hold class fits and load. Printed threads work for low-load covers, and a heat-set insert is the usual answer when the joint must survive repeated assembly.

• 1
  Round and reachableTurn it, then finish the secondary features on the mill-turn.
• 2
  Internal and organicPrint it, then machine only the sealing or bearing surfaces.
• 3
  Square internal cornerPrint, or plan for EDM on a turned part.
• 4
  Thin wall under loadPrinting avoids cutting deflection but adds its own stiffness limits.

Turning cost is mostly machine time plus setup. On a bar-fed lathe, a simple part runs in minutes, and the program is written once. A complex mill-turn part with four setups costs more in fixture time than in cutting time. That is why turning gets cheaper per part as volume rises but stays efficient at quantity one as well.

Printing inverts part of that curve. There is no fixture and no program in the machining sense, so the first article costs about the same as the hundredth. Machine time, however, is long and does not shrink much with volume. A printed part that takes nine hours on the bed takes nine hours whether you order one or fifty.

That gives a simple crossover. Below a few hundred units, printing often wins on total cost for complex geometry. Above that, turning or molding spreads the setup over more parts and wins. The exact crossover depends on cycle time, material price and post-processing, so it should be quoted rather than assumed.

Material price matters more than people expect. A turned aluminium part wastes chips, but aluminium stock is cheap and recyclable. A printed PEEK part uses little material by volume, yet the filament itself is expensive, so the saving in scrap does not always show up in the invoice.

Post-processing closes the gap. Anodizing, plating and polishing are standard on turned parts. Printed parts usually need support removal, sanding or a machined finish pass on critical faces, and those hours are real cost.

Turning is the wrong answer when the part has deep internal channels, a closed hollow section or a lattice. No tool reaches those surfaces, and no amount of fixturing fixes it. It is also the wrong answer when the wall thickness cannot survive cutting force, or when the geometry changes every revision and a fixture would be scrapped each time.

Printing is the wrong answer when the part carries a structural load across a layer boundary. Layer adhesion is the weak direction, so a printed bracket loaded in Z fails earlier than the same bracket loaded in XY. It is also the wrong answer for a sealing face, a bearing bore or a thread that must hold torque without an insert.

Printing is also the wrong answer when you need a documented material certificate matched to a specific alloy heat. Turned stock arrives with a mill certificate, and the part traceability follows the bar. Printed metal powder is certified as powder, not as a finished part.

The honest position is that the two processes solve different problems. Turning gives dimensional control on round, reachable geometry. Printing gives shape freedom where tool access fails. Most production parts that use both stages do exactly that: print the shape, then turn the interfaces.