CNC machining Coventry: how the process actually behaves

• 1
  What CNC machining Coventry actually does to metal
• 2
  Three-axis, four-axis and five-axis setups: what changes
• 3
  Where ±0.005 mm is realistic and where it is not
• 4
  Material behaviour: aluminium, stainless, titanium and plastics
• 5
  When CNC machining is the wrong route
• 6
  Choosing a machining route by part feature

CNC machining removes material with a rotating cutter that follows a toolpath generated from your CAD model. The machine does not know what the part is for. It only knows coordinates, feed rates and spindle speeds. Everything that matters in the result comes from three things: how rigid the setup is, how sharp and how long the tool is, and how well the toolpath matches the geometry.

Take a 12 mm carbide end mill cutting 6061 aluminium at 8,000 rpm and 2,400 mm/min feed. The cutter sings through the part and leaves a finish near Ra 0.8–1.6 μm. Swap that for 17-4PH stainless and the same tool needs roughly one-third the surface speed, more coolant, and a lighter radial engagement. The machine is identical. The physics is not.

This is why quoting a part from a drawing alone is risky. Two parts with the same envelope can behave completely differently depending on wall thickness, corner radii, thread depth and how the part is held. A good shop asks about the setup before it talks about price. If nobody asks, the first article will tell you the truth.

The cutting edge shears material. It does not push it neatly aside. Below a certain chip thickness the tool rubs instead of cuts, work-hardening the surface and dulling the edge fast. That boundary, not the machine's resolution, is what usually sets the real floor on feature size and finish.

• 1
  Rigidity beats spindle speedA short, thick tool in a solid fixture will out-cut a long tool on a faster spindle.
• 2
  Chip thickness mattersToo light a cut rubs the surface instead of shearing it.
• 3
  Setup drives accuracyMost dimensional drift comes from how the part is held, not the control.

A three-axis machine moves the table in X, Y and Z. The tool always approaches from above. Simple plates, brackets, pockets and drilled holes are cheap and fast on this platform. The part may need several re-clamps to reach features on different faces, and every re-clamp adds error and labour. GreatLight runs 27 three-axis machines for exactly this kind of work, with travels from 500 × 310 × 200 mm up to 750 × 1,150 × 550 mm.

A four-axis machine adds a rotary table, usually Ø400 mm, so the part can spin under the cutter. This cuts setups on shafts, flanges, impellers and parts with radial features. The tool still comes from one general direction, so undercuts and steep side walls remain awkward. GreatLight keeps 12 four-axis mills and 16 mill-turn centers for parts that need turning and milling in one setup.

Five-axis machining adds two rotary axes and tilts the tool relative to the part. The real gain is not speed. It is access. Deep pockets, compound angles, ported manifolds and impeller blades can be cut with a short, stiff tool because the table or head brings the surface to the cutter. GreatLight runs 16 simultaneous five-axis machining centers, with a maximum processing size of 4,000 mm.

Five-axis is not automatically better. A flat bracket with ten holes is cheaper on a three-axis machine, and the tolerance will be just as good. Five-axis earns its cost when the alternative is four or five separate fixtures, or when a feature simply cannot be reached from above. Ask which features drive the axis count, then decide.

• 1
  Three-axisPlates, pockets, holes. Lowest cost per part, most setups.
• 2
  Four-axisShafts and radial features. Fewer setups than three-axis.
• 3
  Five-axisCompound angles and deep pockets reached with a short tool.

A tolerance is a promise about a dimension, not a mood. On a rigid, well-supported feature in aluminium or brass, ±0.005 mm (±0.0002 in) is achievable and repeatable. That is the tightest band GreatLight holds across a production run. On a thin wall, a long overhang, or a deep bore in titanium, the same number may be impossible without extra operations and extra cost.

Temperature is the quiet variable. Aluminium expands about 23 μm per metre per °C. A 300 mm part that sits 5 °C warmer than the inspection room has already moved 0.035 mm before you measure it. Shops that hold tight tolerances control the room, let parts stabilise, and measure at a known temperature. If your drawing demands ±0.005 mm on a 500 mm aluminium part, ask how it will be measured.

Surface finish and tolerance are linked but separate. Ra 0.2–0.8 μm is a fine finish, usually reached with a finishing pass, a small step-over and a sharp tool. Ra 0.8–1.6 μm is a normal high-quality machined surface. Ra 1.6–3.2 μm is as-machined and perfectly fine for most brackets. Specifying a fine finish on a face that never touches anything only adds cycle time.

The practical rule: put tight tolerances only where the function needs them. A bearing bore, a dowel hole or a sealing face earns ±0.005 mm. A clearance hole for an M6 bolt does not. Datum choice matters just as much. Pick datums that a machinist can actually hold and measure, and the whole tolerance stack gets easier to meet.

• 1
  Tight where it mattersBores, dowels and sealing faces justify ±0.005 mm.
• 2
  Watch temperatureA 300 mm aluminium part moves 0.035 mm per 5 °C shift.
• 3
  Finish costs timeRa 0.2–0.8 μm needs a dedicated finishing pass.

Aluminium is the default for prototypes and most enclosures. Grades 6061 and 6061-T6 cut fast, hold a good finish and anodise cleanly. 7075 is stronger but gummier and more prone to distortion after machining. 2024 and 5052 behave differently again. GreatLight stocks 6061, 2024, 5052, 5083, 6063, 6082, 7075 and ADC12, so grade substitution is a real conversation, not a guess.

Stainless steel is where feeds and speeds separate good shops from cheap ones. Grades 303 and 304 are common, 316 and 316L resist corrosion, and 17-4PH (SUS630) can be heat treated after machining. Stainless work-hardens if the tool rubs, so the cut must stay under the hardened layer. That means heavier feeds, sharp tools and no dwelling. A light finishing pass on already-hardened 304 will polish the surface, not remove it.

Titanium moves the problem again. TC4 (Ti-6Al-4V) has low thermal conductivity, so heat stays in the cutting edge. Tools wear fast and the part can spring back as the cutter passes. Light radial engagement, high pressure coolant and a rigid setup are not optional. Inconel is harder still, and small features in it are usually a sign that another process, or a redesign, should be considered.

Plastics are not soft metals. POM and PEEK cut cleanly but hold heat and can warp. ABS and PC melt and smear if the tool lingers. Carbon fibre and glass-filled grades are abrasive and will wear a carbide edge in minutes. For these, toolpath strategy and sharp, single-purpose tooling matter more than machine power. Vacuum holding often beats clamping for thin plastic plates.

• 1
  AluminiumFast, stable, anodises well. 6061 is the safe default.
• 2
  StainlessWork-hardens if the tool rubs. Keep the cut under the hard layer.
• 3
  TitaniumHeat stays at the edge. Light radial cuts and high-pressure coolant.
• 4
  PlasticsHeat and abrasion are the limits, not spindle power.

CNC machining wins on accuracy, material choice and geometry freedom. It loses on unit cost at high volume and on hollow, thin-walled shapes. If a part needs 50,000 identical copies of a simple bracket, die casting or stamping will beat machining on price once tooling is amortised. Machining stays competitive when the part is complex, the volume is low to medium, or the design is still changing.

Thin walls are a boundary condition, not a preference. A 0.5 mm wall in aluminium can be machined, but it will deflect under cutting force and may need support material, light passes and extra handling. A 0.5 mm wall in titanium is a different problem. If the design allows a 1.5 mm wall, the part gets cheaper, straighter and faster. Material is often the cheapest thing on the drawing.

Deep, narrow features are the other limit. A pocket deeper than about four times its cutter diameter needs a long tool, which flexes and chatters. A hole deeper than ten times its diameter in stainless needs peck drilling and careful coolant. These are not impossible. They just move the part from a three-axis job to a slower, more expensive one. Knowing this early saves a redesign later.

Finally, consider the whole part. If it needs a complex internal channel, a machined part may be better split into two pieces and joined than hogged from solid. If it needs a smooth organic shell, 3D printing or vacuum casting may be the right first step. Machining is a strong tool, not the only one. The question is always whether the geometry and the volume suit it.

• 1
  High volume, simple partCasting or stamping usually wins on unit cost.
• 2
  Thin wallsBelow about 1 mm, deflection drives cost and risk.
• 3
  Deep pocketsOver 4× cutter diameter, the tool gets long and flexible.
• 4
  Organic shapesAdditive or casting may beat hogging from solid.