CNC shell processing: where the efficiency boundary actually sits

• 1
  What counts as a shell in CNC shell processing
• 2
  Why thin walls move: force, heat and residual stress
• 3
  Workholding decisions that keep the wall still
• 4
  Tool path and parameters for thin-wall milling
• 5
  When 3-axis is enough and when 5-axis pays
• 6
  Which process route fits the shell

A shell is a part where the wall thickness is small compared with the overall envelope. A 180 mm gearbox cover with a 2 mm wall is a shell. The same cover with an 8 mm wall is just a milled part. The distinction matters because everything that governs the cut changes once the wall is thin: stiffness, clamping force, heat path and measurement.

In CNC shell processing the removed volume is usually large and the remaining wall is small. Roughing a pocket from solid stock is not the hard part. The hard part is the last 0.3 mm of finishing pass, when the wall has already lost most of the material that was holding it rigid.

Two numbers tell you whether a geometry behaves as a shell. The first is the wall-to-span ratio. Below roughly 1:20 the wall starts to deflect under its own clamping load. The second is the unsupported height. A 1.5 mm wall that is 20 mm tall behaves very differently from the same wall that is 100 mm tall, even though the thickness is identical.

Shells show up in engine housings, drone frames, battery enclosures, pump bodies, RF shields and medical instrument casings. They share one trait: the function needs a closed or semi-closed surface, and the mass budget does not allow that surface to be thick.

• 1
  Wall-to-span below 1:20Treat as a shell and plan the fixture around deflection.
• 2
  Unsupported height over 5× wallSupport or step the cut; the wall will chatter otherwise.
• 3
  Closed or semi-closed surfaceTypical of housings, covers and enclosures.

Three mechanisms move a wall during cutting, and they arrive in a fixed order. Cutting force bends the wall while the tool is engaged. Heat from the cut expands the local material, and the wall grows into the tool. Then, after the part cools and the fixture is released, residual stress from the parent stock redistributes and the wall springs to a new shape.

Radial cutting force on a thin wall scales with depth of cut and drops fast when you reduce it, but the wall stiffness falls with the cube of thickness. Halve the wall from 4 mm to 2 mm and you lose roughly 87 percent of its bending stiffness. The tool path that worked on the 4 mm wall will sing on the 2 mm wall.

Heat is the quieter problem. Aluminium conducts heat away quickly, so a 6061 wall may only grow 8–12 μm during a finishing pass. Titanium and stainless hold heat at the edge, and a 0.5 mm wall in Ti-6Al-4V can move 30 μm or more between the roughing and finishing passes. That is six times the ±0.005 mm tolerance we hold on the finished part.

Residual stress is the mechanism that surprises people. Rolled plate and extruded bar carry internal stress from the mill. Machine one side and the balance tips. The part bows. Rough, stress-relieve, then finish is the standard answer, and it is the reason a shell often needs two setups even when the geometry looks like a single-op job.

• 1
  ForceBends the wall while the tool is in the cut.
• 2
  HeatExpands the wall into the tool path; worst in titanium and stainless.
• 3
  Residual stressReleases after unclamping and bows the part.

Clamping a shell in a vise is the fastest way to scrap it. A 2 mm wall will take the vise load as a bending moment and yield before the first cut. The fix is to clamp on a thick boss, a flange or a sacrificial tab, and let the thin wall float free with no external load on it.

Where the part has no thick feature to grip, cut soft jaws to the finished outside profile. The jaw contact becomes a form fit rather than a point load, and clamping pressure spreads over the whole arc. For a 1.5 mm wall in aluminium, keep jaw pressure low and use a torque wrench on the vise screw so the setup is repeatable across the batch.

Vacuum fixturing suits flat shells and covers. A 0.8 mm aluminium panel holds well on a grooved vacuum plate because the load is spread over the full face. The limit is part size and porosity. A casting with surface porosity will leak at the seal and lose hold.

For tall shells, add a temporary bridge. Leave 3–4 mm of material across the open end, machine everything else, then cut the bridge in a final light pass. The bridge raises stiffness during the heavy cuts and costs one extra operation. On a 300 mm tall housing, it is usually cheaper than a custom fixture.

• 1
  Clamp the boss, not the wallThin walls yield under vise pressure.
• 2
  Form-fit soft jawsSpreads load over the profile instead of a point.
• 3
  Vacuum for flat panelsFails on porous castings that leak at the seal.
• 4
  Leave a bridge3–4 mm across the open end, removed last.

The single biggest lever is axial depth of cut. Reduce it and the radial force drops almost linearly. On a 2 mm aluminium wall, a 0.5 mm axial depth with a 6 mm three-flute carbide end mill at 12,000 rpm and 900 mm/min keeps the wall quiet. Push the axial depth to 3 mm and the same wall will chatter within 20 mm of travel.

Use climb milling on the finishing pass. The tooth enters at maximum chip thickness and exits at zero, which pulls the wall toward the tool instead of pushing it away. Conventional milling on a thin wall tends to lift the part off the fixture and leaves a tapered surface.

A high-helix variable-flute cutter is worth the extra cost on shells. The variable pitch breaks the regular tooth impact that excites wall vibration, and the high helix pulls chips up and out of a deep pocket. On titanium shells, a 38–45° helix with through-spindle coolant keeps the edge cooler and reduces the thermal drift between passes.

Leave 0.2–0.3 mm of radial stock for the finishing pass, never less. A 0.05 mm finishing allowance lets the tool rub instead of cut, which raises temperature and work-hardens stainless and titanium surfaces. Rubbing also blunts the edge, and a dull edge on a thin wall is a chatter source.

• 1
  Low axial depth0.5 mm on a 2 mm aluminium wall keeps force down.
• 2
  Climb mill the finishPulls the wall toward the tool, cleaner surface.
• 3
  Variable-flute cuttersBreak the regular tooth impact that excites vibration.
• 4
  0.2–0.3 mm finish stockLess than that rubs and work-hardens the surface.

Most shells with a single open face and no undercuts run fine on a 3-axis mill. If every feature is reachable from one direction, a 3-axis machine with good workholding will match a 5-axis machine on cycle time and beat it on setup cost. Do not buy simultaneous motion you will not use.

The case for 5-axis appears when the shell has features on multiple faces or a curved outer skin that must be machined in one continuous pass. Repositioning a thin shell between setups is where the tolerance disappears. Each unclamp and reclamp cycle on a 1.5 mm wall can shift the datum by 20–40 μm, and that error stacks.

Simultaneous 5-axis also lets you keep the tool normal to a curved wall. On a spherical or cylindrical shell, a normal approach distributes the cutting force along the wall instead of peeling it sideways. That is the difference between a stable cut and a wall that deflects away from the cutter.

For shells up to 4,000 mm in one axis, we run large-travel machines with a Ø400 mm rotary table so the part is machined in one setup. The trade-off is fixture mass. A heavy tombstone stabilizes the cut but adds thermal inertia, so the part should be allowed to reach shop temperature before the finishing pass.

• 1
  3-axisSingle open face, no undercuts; lowest setup cost.
• 2
  5-axis indexingMultiple faces, one setup, datum stays fixed.
• 3
  5-axis simultaneousKeeps the tool normal to curved walls.
• 4
  Large travelUp to 4,000 mm with a Ø400 mm rotary table.