CNC Processing Plastics: How Material Behavior Drives the Cut

• 1
  What CNC Processing Plastics Actually Does to the Material
• 2
  Which Plastics Machine Well, and Which Ones Fight Back
• 3
  Feeds, Speeds and Heat: Where Plastic Jobs Go Wrong
• 4
  Reading Tolerances and Finishes on a Plastic Print
• 5
  When CNC Processing Plastics Beats Injection Molding
• 6
  Finishing, Inspection and Documentation
• 7
  Plastic Machining Comparison by Behavior
• 8
  Machining vs Molding: Pick by Volume and Geometry

A metal cutting tool removes material by shearing a crystal lattice. Plastics do not have one. The cutter instead pushes polymer chains ahead of the edge, then breaks them. Some chains spring back, some smear, some fracture along the grain of the material. That is why a plastic job that looks simple on a print can machine badly: the same feed and speed that works in 6061 aluminum will melt a polycarbonate wall or chip a PMMA edge.

Three numbers govern most of the outcome. Cutting temperature, because thermoplastics soften and then gum up on the tool. Chip clearance, because a plastic chip is light and tends to recirculate instead of falling away. Tool sharpness, because a dull edge rubs the surface and raises local heat instead of cutting. Control those three and the cut is clean. Ignore one and you get a burr, a melted edge or a dimension that moves after the part cools.

Plastic also moves after machining. Absorbed moisture leaves the surface, internal stress from extrusion or molding relaxes, and the part grows or shrinks a few tens of microns over hours or days. For a bracket that is fine. For a 300 mm seal plate with a ±0.05 mm flatness callout, it is the whole problem. Design the print so the tolerance is measured after a defined rest period, not immediately off the machine.

Hardness and filler content set the ceiling. Unfilled ABS and PP cut fast and forgive a lot. Glass-filled PA or PEEK will hold a tighter tolerance but will wear a carbide edge in a way aluminum never does. Carbon-fibre sheet behaves differently again: it is abrasive in-plane and can delaminate if the tool lifts the top ply. We treat each of these as a separate process, not as one plastic setting.

ABS, POM, PC, PMMA, PA and PP are the workhorses. ABS machines clean and takes a fine bead-blasted finish; it is the default for enclosures and covers where stiffness is moderate. POM (acetal) is the best all-round machining plastic for tight tolerance parts: it is dimensionally stable, cuts with a clean chip, and holds ±0.02 mm on a 50 mm feature without drama. It is the material we reach for on gears, bushings and sliding blocks.

PC and PMMA both cut optically clear if you slow down and use sharp, polished tooling. PMMA is more brittle and chips at the exit edge, so climb milling and a support backing plate matter. Polycarbonate is tougher but gummier; it needs a generous coolant or air blast and a sharp edge, otherwise it smears. Both are common in light guides, guards and display covers.

PA (nylon) absorbs moisture and changes size with humidity. Machining is easy; holding a tight tolerance over weeks is not. If a PA part must fit a metal housing, call out the tolerance after conditioning, or switch to POM. PP is soft, chemically resistant and cheap, but it is flexible and deflects under cutting force, so thin walls need light passes and good support.

PEEK and carbon-fibre grades are the high end. PEEK machines well but the stock is expensive, so a wrong first article costs real money. Carbon-fibre composite is abrasive and should be cut with diamond-coated tooling where possible. Neither belongs in a job where ABS would pass the functional test and cost a tenth as much.

The instinct carried over from aluminum is to run fast and flood with coolant. That fails on plastics. The correct starting point is a high surface speed with a light chip load, a sharp two- or three-flute cutter with a large helix, and enough air or mist to clear chips without thermal shock. Water-based flood coolant can stain or stress-craze PC and PMMA, so we often run dry with compressed air and a vacuum shoe.

Heat is the enemy, and it shows up as a colour change, a glossy smear on the cut face, or a stringy chip welded to the tool. When you see any of those, the answer is rarely more speed. Reduce the chip load slightly, sharpen or change the tool, and increase chip evacuation. On deep pockets in POM or PC, a peck cycle with full retract clears the swarf that would otherwise be re-cut and melted.

Fixturing decides the dimension. Plastics are roughly ten times more flexible than steel, so a vise tightened the way you would clamp a steel block will bow the part and cut a concave face. Use soft jaws machined to the part profile, vacuum chucks for flat panels, or a sacrificial backing plate for thin stock. Where a wall is thin, support it from behind rather than pushing harder from the front.

Deburring is part of the process, not an afterthought. A cleanly machined plastic edge can still carry a fine fuzz, especially on PP and PA. A light hand scrape, a tumbling cycle, or a controlled bead blast removes it without rounding a critical edge. For a sealing face, specify the deburr method on the print; a chamfer callout alone does not tell the shop how to reach the edge.

A general tolerance block copied from a metal drawing is usually wrong for plastics. Machined plastics can hold ±0.005 mm on a small, well-supported feature, and we do it regularly. But that number must be attached to a feature that can physically support it: a short bore, a face on a thick section, a position between two machined datums. Across a thin free wall, or over a 400 mm span, the material will move more than that no matter how good the machine is.

Practical bands look like this. Small features on POM or PC: ±0.01 to ±0.02 mm is routine. Medium features on ABS or PA: ±0.05 mm is comfortable. Long spans or thin walls: hold what the material allows and specify a functional fit instead of a blanket number. If a part mates with a metal component, tolerance the plastic side looser and put the control on the metal side, which is stable.

Surface finish is specified in Ra and behaves differently than on metal. As-machined plastic typically lands around Ra 1.6–3.2 μm. Careful work with sharp tooling reaches Ra 0.8–1.6 μm on POM, ABS and PC. Below that, polishing is a secondary operation and results depend on the polymer: PMMA polishes to a clear optical surface, while PP and PA stay slightly matte. Do not call out Ra 0.2 μm on a plastic part unless the function truly needs it.

Wall thickness and corner radii matter as much as the tolerance block. A 1 mm wall in PC will flex under cutting load; a sharp internal corner concentrates stress and can crack during assembly. Machine a 0.5 mm minimum internal radius where the design allows, and keep walls uniform so cooling and stress relaxation stay even. Those two choices prevent more scrap than any tolerance change.

CNC processing plastics wins when the part count is low, the geometry is complex, or the design is still moving. There is no tooling to cut, so a revision costs a new program rather than a new mold. We machine from one prototype to 10,000+ part runs with no minimum order quantity, and plastic prototypes commonly ship in 3–5 days. For a bridge build ahead of a mold, or a fixture that will never be made twice, machining is the cheaper path.

Injection molding wins when the geometry is stable and the annual volume is high. A molded part carries a lower unit cost at volume and can hold features, like molded-in bosses, that a cutter cannot reach. The trade is time and tooling: a mold must be designed, cut, tested and sometimes revised before the first good part exists. If your design freeze is months away, machining the plastic now removes that wait.

Some geometries only machining can make. Deep pockets with square internal corners, tight-tolerance bores, threaded plastic ports, and one-off manifolds all fall into that group. Transparent PMMA and PC parts are also a natural fit, because a machined and polished surface can match an optical requirement without a polished mold cavity.

There is a middle route worth knowing. If a plastic part will eventually be molded, machine a small batch first, test the fit, and let that feedback drive the mold design. The machined parts absorb the risk, and the mold is cut once. That sequence has saved more than one program from a second tooling round.

Machined plastic takes most cosmetic and functional finishes, though the list is shorter than for metal. Bead blasting gives a matte, uniform surface and hides tool marks. Tumbling softens edges on small parts in bulk. Brushing and polishing work well on PMMA and PC. Laser marking handles part numbers, lot codes and traceability marks; the minimum character height we mark is 1.5 mm, so keep labels readable rather than tiny.

Metal-style finishes such as anodizing, plating and powder coating are for the metal components in an assembly, not for the plastic. Clear-coat and painting can be applied to plastics but change the surface chemistry, so specify them only when the function requires it. For a plastic part, the finish that matters most is usually the as-machined surface and a controlled deburr.

Inspection is where a plastic job earns its tolerance. We inspect 100% of parts before shipment, with a raw material check, in-process monitoring and a final dimensional inspection. Reports are available on request. For a tight feature on POM or PC, measure after the part has rested, since the reading taken hot off the machine will differ from the reading a day later.

Documentation keeps a plastic project predictable. First article inspection reports, material certificates and dimensional reports can be supplied when the print requires them. If your part feeds a regulated program, tell us at quote stage, because the inspection plan and the paperwork should be agreed before the first chip is cut, not after.