Multi-Axe Cuttings CNC Programming for Complex Laser Optical Paths

• 1
  What multi-axe cuttings CNC programming actually controls
• 2
  Why the post-processor decides whether the part is right
• 3
  Focus, kerf and angle on complex laser optical paths
• 4
  Planning toolpaths and collision checks before the first cut
• 5
  When multi-axis cutting is the wrong choice
• 6
  3-axis, 3+2 and continuous five-axis compared

A multi-axe machine moves the cutting point along a path that has more than three degrees of freedom. Two rotary axes tilt the head or the table, so the beam or the tool arrives at the surface at an angle instead of straight down. The programmer's job is to keep the focal point or the tool tip on the intended curve while the rotary axes swing underneath it.

That is the whole idea. On a 3-axis machine the nozzle stays vertical, so a hole drilled into a sloped wall comes out elliptical and a cut along a curved surface changes focus depth as the part rises and falls. Adding rotary axes lets the head follow the surface normal. The cut width stays even and the kerf does not taper.

The cost is not the machine. It is the programming. Every extra axis adds a coordinate that has to be solved, a kinematic chain that has to be modeled, and a post-processor that has to output the right G-code for that specific machine. Get any of the three wrong and the part is scrap, not slightly off.

For laser work the problem is sharper than for milling. The beam has a focal length, a Rayleigh range and a kerf that all depend on the angle between the nozzle and the surface. A 15° tilt changes the effective spot size on some optics. Programming has to account for that, not just for the tool center point.

• 1
  Three coordinates plus two rotationsX, Y, Z with A and B or C tilting the head or table.
• 2
  Tool center point controlThe controller keeps the tip on path while the rotary axes move.
• 3
  Surface normal followingThe head stays perpendicular to the local surface, not to the machine bed.

CAM software works in a clean mathematical world. The post-processor translates that into the real machine, and the real machine has offsets, pivot distances and axis limits that the CAM kernel does not know about. A head-and-table machine and a gantry machine with two head rotations behave differently even for the same toolpath.

The pivot distance is the usual source of trouble. If the distance from the C-axis centerline to the tool tip is entered wrong by 0.05 mm, every tilted move is off by that amount in the direction of the tilt. On a flat cut nothing shows. On a five-axis contoured cut the error appears as a step where the tilt changes.

We verify post-processor output before cutting metal or optics. A simple test block with a known sphere and a known cylinder is cut first, then measured. If the sphere is out of round or the cylinder wall thickness varies, the kinematics are wrong and no amount of tweaking the CAM file will fix it.

Rotary axis limits matter too. Many machines cannot rotate A or C continuously, and a toolpath that needs a 400° swing will hit a soft limit, alarm out, or worse, unwind mid-cut and leave a mark. The programmer has to plan the winding direction of the rotary axes, not just the path of the tip.

• 1
  Check pivot offsetsA wrong pivot distance tilts every rotary move by a constant error.
• 2
  Respect rotary travelPlan winding so the C axis never has to unwind mid-cut.
• 3
  Cut a test sphere firstRoundness and wall thickness reveal kinematic errors fast.

A laser cut is not a milling cut. The material is removed by melt and vaporization, and the kerf width depends on the spot size at the surface and on the assist gas flow. When the head tilts, the gas jet no longer blows straight into the kerf, so dross can form on the lower edge even if the beam itself is perfectly focused.

Focus position is the second variable. On a flat sheet the focal point sits a fixed distance below the surface. On a curved optical path or a contoured tube the surface height changes along the cut, and the controller has to move Z to keep the focus at the same depth. That is a programming problem, not an operator problem.

For thin material, up to about 3 mm, the process is fairly forgiving. A small tilt change does not move the focus much and the kerf stays stable. Above 6 mm the depth of focus becomes short relative to the wall thickness, and the tilt has to be controlled within a few degrees or the lower edge will not cut through cleanly.

The same logic applies when the laser is cutting a curved optical path in a tube. The cut follows a 3D curve on a cylinder, so the head rotates to stay normal to the surface. If the tube is not centered on the rotary axis, the surface moves in and out as it spins and the focus drifts. Programming assumes a true center; setup has to deliver one.

• 1
  Tilt affects gas flowDross on the lower edge often comes from a deflected assist jet.
• 2
  Walls above 6 mm are less forgivingKeep tilt within a few degrees or the lower edge will not cut.
• 3
  Center the tube on the rotary axisRunout shows up as focus drift and an uneven kerf.

Once the kinematics are correct, the toolpath still has to be planned. Continuous five-axis motion is slower than 3+2 positioning, so the first decision is whether the part really needs the rotary axes to move during the cut. Many contoured features can be cut in a series of fixed tilts, which is faster and easier to verify.

When continuous motion is needed, the programmer has to keep the rotary axes from making large swings between adjacent points. A sudden 30° flip will show up as a gouge or a witness mark, because the machine cannot follow it at feed rate. Smoothing the tilt over several millimeters of path keeps the actual motion close to the commanded one.

Collision checking is not optional. A tilting head plus a fixture plus a tube can interfere in ways that a 3-axis simulation never shows. The check has to include the nozzle body, the fixture clamps and the rotary table itself, not just the tool tip.

A short checklist before releasing a program: verify the post-processor with a test sphere, confirm the rotary winding, check the focus compensation table, and run a full collision simulation with the real fixture model. Four checks, and they catch most of the failures we see.

• 1
  Prefer 3+2 where possibleFixed tilt positions are faster and simpler to prove out.
• 2
  Smooth the tilt changesSpread a 30° swing over several millimeters of path.
• 3
  Simulate the fixtureNozzle body and clamps collide where a tool-tip check says clear.

Multi-axis programming is worth its cost when the part has features a 3-axis machine cannot reach, or when the surface angle changes enough that a fixed head would produce an unacceptable cut. Tubes, cones, curved optical paths and contoured panels are the classic cases.

It is not worth it for simple prismatic parts with holes and slots on one face. A 3-axis mill with a good fixture will be faster, cheaper and easier to inspect. Programming time for a five-axis contoured cut can run several times the cutting time on a small batch, and that cost does not go away with volume unless the setup is amortized over thousands of parts.

There is also a material boundary. Aluminium and mild steel cut and machine predictably with tilted axes. Titanium and Inconel generate more heat at the contact point, and a tilted tool or beam changes the heat distribution. On those materials the tilt range has to be narrower, or the process has to be re-qualified.

The honest rule: if the part is flat or has features on one side, use 3 axes. If it has a true 3D contour that a fixed head cannot follow, multi-axis is the only way. If it sits in between, run the numbers on both before committing the fixture.

• 1
  Good fitTubes, cones, contoured panels, curved optical paths.
• 2
  Poor fitFlat plates and single-face prismatic parts.
• 3
  Material limitTitanium and Inconel need a narrower tilt range.