Precision Plastic Finish: How Surface Quality Is Actually Achieved

• 1
  Why Machined Plastics Never Come Off the Tool Clean
• 2
  How Material Choice Changes the Achievable Finish
• 3
  Machining Parameters That Set the Starting Surface
• 4
  Where a Precision Plastic Finish Stops Making Sense
• 5
  How to Specify and Verify the Finish on a Drawing
• 6
  Finishing Processes and Where Each One Fits

Metals cut by shearing. Plastics do not. A polymer responds to a passing edge with a mix of cutting, smearing, elastic recovery and localized melting. That is why a surface that looks acceptable on a steel part can look scratched and cloudy on ABS or PC from the same machine. The tool is doing the same thing. The material is not.

Three forces set the as-machined texture. First, the cutting edge radius: if the edge is blunter than the intended chip load, the tool rubs instead of cuts, and the surface tears. Second, heat. Thermoplastics have low thermal conductivity, so friction heat stays at the surface instead of dissipating into the bulk. Above the glass transition temperature, the surface smears and drags. Third, elastic recovery: the polymer springs back behind the edge and rubs against the flank face, leaving a burnished line.

The practical result is that a precision plastic finish is mostly a machining problem, not a polishing problem. If the as-machined surface is torn, no amount of buffing will restore dimensional accuracy after you remove the damaged layer. The finishing step should remove microns, not fix a bad cut.

This matters most on thin walls and on parts with tight tolerances. A polish that removes 0.03 mm from a 1.5 mm wall changes stiffness. A polish that removes 0.03 mm from a bearing bore changes fit. Plan the finish allowance into the model from the start.

• 1
  Rubbing, not cuttingAn edge radius larger than the chip load burnishes instead of shearing.
• 2
  Heat stays localLow conductivity traps friction heat at the surface and softens it.
• 3
  Elastic recoveryThe polymer springs back and scratches itself against the flank face.

Amorphous plastics such as ABS, PC and PMMA behave differently from semi-crystalline ones such as POM, PA and PP. Amorphous polymers soften gradually, so they smear and smear cleanly. Semi-crystalline polymers hold a sharper edge but can chip at the crystal boundaries, which shows up as a dull, slightly frosty surface even after fine machining.

Filled and reinforced grades are the hardest case. Carbon fibre and glass fibre compounds are abrasive. A carbide cutter that holds an edge for a full day in unfilled POM may dull in 40 minutes in a 30% glass-filled PA. Dull edges raise cutting temperature, and the resin matrix around the fibres starts to pull out. The result is a finish with visible fibre pull-out, which no bead blasting will hide.

PEEK is the opposite problem. It machines cleanly and takes a very fine finish, but it is expensive enough that the finishing allowance should be kept small. A part that starts at 20 mm and finishes at 19.7 mm wastes material cost that the customer pays for.

So the first question in any finishing discussion is not which process. It is which polymer, and whether the drawing allows enough stock for the process being proposed. For a precision plastic finish, the material datasheet and the tolerance stack have to agree before the toolpath is written.

• 1
  Amorphous (ABS, PC, PMMA)Smears rather than chips; responds well to polishing and vapor processes.
• 2
  Semi-crystalline (POM, PA, PP)Holds an edge but can chip; usually lands at a matte or satin look.
• 3
  Fibre-filledAbrasive and prone to pull-out; keep expectations at a functional finish.

If the finishing operation is going to remove only a few microns, the starting surface has to be good. That comes from three decisions on the machine: cutter geometry, chip load and cooling.

Use two-flute or single-flute cutters with a polished flute for plastics. High helix angles help clear the soft chip before it rubs. Keep the cutting edge sharp and replace it on a count, not on feel. For finishing passes on unfilled thermoplastics, a stepover around 0.05–0.1 mm and a depth of cut under 0.2 mm keeps radial force low and reduces deflection on thin walls.

Cooling matters more than most shops admit. Compressed air is usually the right choice for plastics because it clears chips and removes heat without a thermal shock. Flood coolant can work on POM and PA but may cause stress crazing in PC if the fluid is not compatible. Mist cooling sits in between and is common on 5-axis work where the nozzle can follow the tool.

Spindle speed should be high enough to keep the chip thin, but not so high that friction heat builds at the edge. When the surface starts to look glossy or waxy rather than cut, the speed is too high or the feed is too low. That gloss is melted polymer, and it will not hold tolerance after it cools.

• 1
  Stepover 0.05–0.1 mmKeeps radial load low on finishing passes.
• 2
  Depth of cut under 0.2 mmReduces wall deflection on thin features.
• 3
  Air over floodClears chips and avoids stress crazing in PC.

Finishing is not free, and it is not always useful. There are four situations where we tell customers to skip it or reduce the specification.

The first is a non-visible internal surface. If a bracket sits inside an enclosure and only carries load, an as-machined finish at Ra 1.6–3.2 μm does the job. Spending on polishing adds cost and lead time for no functional gain.

The second is a part where the finish interacts badly with the function. A textured or bead-blasted bore holds lubricant, which is good for a sliding fit but bad for a pneumatic seal that needs a smooth land. A polished face looks better but reflects light into an optical sensor. Match the finish to the function, not to the drawing default.

The third is a part with a very tight tolerance on a thin wall. Any mechanical polishing removes material unevenly. If the tolerance band is already at ±0.005 mm, hand polishing is likely to push the part out of spec. In that case, achieve the finish through the machining pass and leave the surface alone.

The fourth is a prototype where the design is still moving. If the geometry may change next week, invest in the finish after the design freezes. A precision plastic finish is worth paying for once, on the right revision.

• 1
  Hidden surfacesAs-machined finish is enough when nothing sees or touches it.
• 2
  Function conflictTexture helps lubricant retention but hurts seal lands.
• 3
  Tight toleranceHand polishing can consume the whole tolerance band.

Ra alone is a weak specification. It is a two-dimensional average that says nothing about the direction of the marks or the presence of a few deep scratches. Two parts can share the same Ra and look and behave very differently.

On a drawing, state the finish with three pieces of information: the numerical Ra limit, the process (for example, bead blast or fine machined), and the measurement direction relative to the lay. If the surface is cosmetic, add a visual standard or a photograph. If the surface is functional, add the contact or sealing requirement that the finish supports.

For inspection, a portable stylus profilometer is the usual tool. Measure at least three locations on the functional face and record the cutoff length. On curved or small features, a replica tape can be used, though accuracy drops. For transparent parts, a light transmission check is often more meaningful than an Ra number.

We inspect 100% of parts before shipment, including raw material verification, in-process checks and final inspection, with reports available on request. If a finish is critical to the function, say so at the quote stage so it can be built into the inspection plan rather than added at the end.

• 1
  State the processBead blast and polished can share an Ra value but not a look.
• 2
  State the directionLay matters on sealing and sliding surfaces.
• 3
  Measure in more than one spotOne reading on one face does not describe a part.