Optimize plastic parts with CNC processing

• 1
  Why plastic behaves differently at the spindle
• 2
  Which plastics machine well, and which fight back
• 3
  Tool geometry, feeds, and speeds for plastic
• 4
  Fixturing and clamping without deforming the part
• 5
  Measuring and finishing plastic parts
• 6
  Material and process selection for plastic parts

Plastic removal is a heat problem before it is a cutting problem. The thermal conductivity of unfilled POM or ABS sits in the range of 0.2–0.4 W/m·K, roughly two orders of magnitude below aluminum. Heat generated at the shear zone has nowhere to go, so it stays in the chip and in the workpiece. Melting is not the usual failure; softening is. A part that reads 60 °C at the surface can be 20 °C warmer in the core after a heavy pass.

That heat does three things. It drops the elastic modulus, so the material deflects away from the tool instead of shearing cleanly. It raises thermal expansion effects, where a 40 °C rise on a 100 mm POM part moves the dimension about 0.1 mm. And it leaves a smeared surface that looks polished but fails a profilometer check.

Elastic recovery is the second mechanism. Plastics spring back after the edge passes, so the finished wall can measure 0.03–0.08 mm larger than the cutter path. On a 10 mm bore with a 0.05 mm target, that springback alone eats the whole tolerance band. Rigid metals do not do this.

Chip evacuation is the third. Long stringy chips wrap the tool, re-cut, and drag heat back into the slot. Plastics that machine well produce short, powdery chips; if you see ribbons, the feed per tooth is too low or the tool geometry is wrong.

Amorphous plastics such as ABS, PC, and PMMA soften gradually over a wide temperature band. They cut with a continuous chip, hold a decent edge, and tolerate moderate cutting speeds. They also scratch easily, so secondary handling matters as much as the program.

Semi-crystalline grades behave differently. POM, PA, and PEEK have a sharp melting point and low thermal conductivity. They produce clean chips and good surface finish when the feed is high enough, but the window is narrow. Too slow and the material rubs; too fast and the chip welds to the flute.

Filled grades change the rules again. Glass-filled PA or carbon fibre reinforced stock is abrasive. Carbide tool life drops to a fraction of what it is on unfilled stock, and the fibers leave a fuzzy edge unless the tool is sharp and the last pass is light.

Soft, low-modulus materials like HDPE, PP, and soft PVC are the hardest to hold. They deflect under clamping pressure, so the part springs back when you unload it and the dimension you measured in the vise is not the dimension the customer receives.

Two-flute and three-flute carbide end mills with polished flutes are the default. The open flute form gives chip room, which matters more than edge strength on plastics. For finishing passes on transparent PC or PMMA, a single-flute cutter with a 0° helix leaves the fewest witness marks.

Rake angle should be positive and generous, usually 15–20° for acrylics and 10–15° for POM. A neutral or negative rake rubs the material instead of shearing it, and the surface turns cloudy.

Surface speed is the parameter most people get wrong. For POM and PA we run 200–400 m/min with a chipload of 0.05–0.15 mm per tooth. For acrylics we drop to 100–200 m/min to keep the edge from melting. For glass-filled grades, 80–150 m/min with coated carbide.

Cooling is not optional. Compressed air at 4–6 bar clears chips and pulls heat away without thermal shock. Flood coolant on unfilled plastics can cause stress crazing in acrylic and PC, so we use it only on filled grades and PEEK.

Vacuum chucks work well on flat plastic plates because the load spreads across the whole face. We use a fixture plate with a sealed O-ring groove and 0.6–0.8 bar of vacuum. That grips a 300 × 300 mm PC sheet firmly enough to take a 2 mm depth of cut without marking the back face.

For parts that cannot be vacuumed, soft jaws machined to the part profile are the next option. Aluminum jaws with a urethane liner spread the clamp load. Torque values stay low: 2–4 N·m on an M6 clamp screw is usually enough, and it is easy to overtighten a thin-walled plastic part past the point where it recovers.

Thin ribs and tall walls need support, not more clamp pressure. We leave sacrificial tabs or machine a support web that gets cut away on the last operation. The tab thickness is typically 0.5–1.0 mm for small parts.

When the part is very flexible, we rough it in a stressed state, stress-relieve it, then finish. That means removing the part, letting it sit, and re-clamping for the finishing pass. It costs an extra setup and it is the only reliable way to hold parallelism on a 0.8 mm wall.

Caliper readings on plastic are misleading if you take them immediately. The part is still warm from machining and it has not finished relaxing. We wait 30–60 minutes at room temperature before final dimensional checks on tight-tolerance features.

For bores and pockets, a bore gauge or a coordinate measuring machine gives a truer number than a caliper, because the contact pressure of a caliper jaw can compress a soft wall by a few hundredths of a millimeter.

Surface finish targets on plastic run from Ra 1.6–3.2 μm as machined, down to Ra 0.2–0.8 μm after polishing. Bead blasting hides tool marks on opaque grades and gives a uniform matte look. On transparent parts, flame polishing or vapor polishing is the only way to get optical clarity, and both need a clean, scratch-free surface first.

Stress cracking is the failure mode that shows up last. Annealing PEEK or PC parts at 120–150 °C for 2–4 hours before the finishing operation relieves internal stress and reduces the chance of a crack forming weeks later in service.