What Can CNC Milling Do?

• 1
  The Cutting Mechanics Behind Every Milled Part
• 2
  Tolerance and Surface Finish: What You Can Realistically Hold
• 3
  3-Axis, 4-Axis, and 5-Axis: Picking the Right Setup
• 4
  Materials That Mill Well and Materials That Fight Back
• 5
  From One Prototype to 10,000 Parts on the Same Process
• 6
  Matching the Milling Setup to the Part

A milling machine holds the workpiece still and spins the tool. The spindle turns an end mill, face mill, or ball nose cutter at a set surface speed, while the X, Y, and Z axes feed the part past the flutes. Each flute peels off a chip. Chip size is set by feed per tooth and spindle speed, not by how hard the operator pushes.

That distinction matters when you read a drawing. Milling removes material as a shearing action, so the cut produces heat in the chip rather than in the part. With the right coolant and speeds, a 6061 aluminum block can leave the machine warm to the touch. Titanium and Inconel behave differently: they hold heat at the cutting edge, so surface speed drops and tool life shortens.

The tool geometry sets what shapes are possible. A flat end mill leaves a square internal corner with a radius equal to the cutter radius. A ball nose cutter can trace a curved surface but leaves scallops that must be sanded or polished out. A dovetail or T-slot cutter reaches undercuts that a standard end mill cannot. None of these tools can cut a true sharp internal corner.

For any part, the machinist reads three numbers from the CAD file: the smallest internal radius, the deepest pocket, and the tightest tolerance. Those three numbers decide whether the part is straightforward, tricky, or impossible on a given machine. Ignore them at the design stage and the quote comes back with caveats.

• 1
  Chip load is the real controlFeed per tooth sets chip thickness; too thin and the tool rubs instead of cutting.
• 2
  Heat follows the chipGood chip evacuation keeps the workpiece cool and holds tolerance.
• 3
  Tool radius rules cornersEvery internal corner carries the cutter radius unless you specify EDM or a relieved design.

Milling tolerances depend on the machine, the material, and the feature being cut. On a rigid 5-axis center, GreatLight holds ±0.005 mm (±0.0002 in) on critical dimensions when the setup is stable. That number is not free. It requires a qualified machine, a checked fixture, and in-process probing on tight features.

As-machined surfaces land around Ra 1.6–3.2 μm. A finishing pass with a sharp cutter and light radial engagement gets you to Ra 0.8–1.6 μm. Below that, you are usually looking at a separate process: bead blasting, tumbling, polishing, or a coated insert run. The right finish callout on the drawing saves a secondary operation.

Deep pockets and thin walls are where tolerance claims break down. A wall 0.8 mm thick will deflect under cutting force even if the machine is perfect. The fix is usually design-side: add a rib, split the part, or accept a looser tolerance on the non-functional face. Machinists can compensate for tool deflection, but not for a wall that springs back after the cut.

Inspection closes the loop. GreatLight checks raw material on receipt, monitors in-process, and inspects 100% before shipment, with reports available on request. If your drawing calls a true position or profile tolerance, say so up front. It changes the fixturing plan and the inspection method.

• 1
  ±0.005 mm is a controlled claimIt applies to specified features under stable setup, not to every dimension on the print.
• 2
  Finish is a separate passMoving from Ra 1.6–3.2 μm to Ra 0.2–0.8 μm adds time and often a second operation.
• 3
  Thin walls deflectMachining force moves material; stiffness comes from geometry, not from the spindle.

A 3-axis mill moves the tool in X, Y, and Z only. The part sits in one orientation. This covers the majority of plates, brackets, housings, and manifolds. It is the fastest and cheapest route when the part can be reached from one or two directions. GreatLight runs 27 three-axis machines for exactly this kind of work.

A 4-axis mill adds rotation around one axis, usually the A axis. That lets the tool cut around a cylindrical part or index to several faces without re-clamping. Twelve four-axis mills handle shafts, rotary components, and parts with features on multiple sides. Indexing is the key benefit: fewer setups means fewer datum shifts and better feature-to-feature alignment.

A 5-axis mill adds two rotary axes, so the tool can approach the part from nearly any angle. The payoff is real for contoured surfaces, deep cavities, and parts with features on five faces. Machining in a single setup removes the error stack that builds up when you re-clamp a part three times. It also lets a short, stiff tool reach deep pockets that a long 3-axis tool cannot.

Five-axis is not automatically better. Programming takes longer, the machine hour rate is higher, and simple prismatic parts gain nothing from the extra motion. The rule we use: if the part can be made in two 3-axis setups with a stable fixture, do that. Reserve five-axis for geometry that genuinely needs the access.

• 1
  3-axis: fast and economicalBest for plates, brackets, and parts reachable from one or two directions.
• 2
  4-axis: fewer setupsAdds rotation for cylindrical parts and multi-face indexing.
• 3
  5-axis: access and accuracySingle-setup machining of contoured and deep-cavity parts.

Aluminum is the default for milled parts. Grades 6061 and 6061-T6 machine fast, hold tolerance, and take anodizing well. 7075 offers higher strength for aerospace and stressed components. 2024, 5052, 5083, 6063, 6082, and ADC12 cover most other aluminum needs. If you are unsure, 6061-T6 is the safe starting point.

Stainless and steel need more attention. Grades 303 and 304 cut cleanly; 316 and 316L resist corrosion but work-harden if the feed is too light. 17-4PH (SUS630) is common for medical and high-strength parts. Tool steel and 4140, 4340, 4130 appear in molds and structural work. The trade-off is always the same: harder material, slower cut, shorter tool life, higher cost.

Titanium and nickel alloys sit at the difficult end. Ti-6Al-4V (TC4) and TA1, TA2 hold strength at temperature but conduct heat poorly. Inconel is worse. These materials demand low surface speeds, rigid setups, and plenty of coolant. They can be milled to tight tolerance, but the cycle time and tooling cost are several times an aluminum part of the same size.

Plastics behave differently again. ABS, PC, PMMA, POM, PA, PEEK, PP, and HDPE all mill, but each has a preferred chip load. POM cuts clean; PEEK is abrasive and expensive; PMMA chips and can crack at sharp corners. Carbon fibre is abrasive and needs diamond tooling. Machining plastic is not a soft version of machining metal; it is its own process.

• 1
  Aluminum is the benchmark6061-T6 balances machinability, strength, and finish.
• 2
  Stainless work-hardensLight feeds and dwelling dull the tool; keep the chip load up.
• 3
  Titanium and Inconel are slowHeat stays at the edge, so speeds drop and costs rise.
• 4
  Plastics are not soft metalsEach grade has its own chip load, clamping, and cooling needs.

Milling scales without changing the basic method. The first part comes off the same kind of machine as the ten-thousandth. What changes is the fixture, the tooling plan, and the inspection frequency. That continuity is useful: a design validated on a prototype will behave the same way in production, provided the fixture is stable.

At the prototype stage, the goal is form, fit, and function. One or two parts, quick turnaround, and a willingness to adjust the design between runs. GreatLight has no minimum order quantity, so a single part is a normal job. Free DFM analysis comes back with the quote within 12 hours, and production can start within 24 hours.

At the production stage, the goal shifts to repeatability. Fixtures get more robust. Tool changes are scheduled. In-process probing holds critical dimensions. The historical late-delivery probability is below 2%, and parts ship in 3–5 days on standard work. The process is the same; the discipline around it is different.

The mistake to avoid is designing for one part and assuming it will scale. A feature that needs a hand-finished blend on a prototype will need a repeatable method in production. Flag those features early. A small drawing change at the prototype stage costs nothing; the same change after tooling is built costs a lot.

• 1
  Same process, different disciplinePrototype and production use the same cutting method; fixtures and inspection differ.
• 2
  No minimum orderOne prototype or a 10,000+ part run both fit the same workflow.
• 3
  Design for repeatabilityHand-blended features on a prototype need a production method.