CNC Tour Analysis and Programming: How a Turned Part Gets Planned

• 1
  What CNC tour analysis actually decides
• 2
  Reading the part: features that drive the program
• 3
  Turning parameters that hold CNC tour tolerance
• 4
  Workholding and setup choices that change the plan
• 5
  Checking the program before the first cut
• 6
  Turning route vs. part characteristics

A CNC tour is the turning operation on a lathe, where the workpiece rotates and the tool feeds along the Z axis. Before programming, the shop runs a tour analysis: which surfaces must be concentric, which faces are the datums, and how much stock each pass removes. Get that wrong and the G-code is fine but the part is scrap.

The analysis starts from the drawing, not from the geometry model. A model shows shape. A drawing shows what matters: the Ø tolerance band, the concentricity callout, the surface finish on a sealing face. Two bores may look identical in CAD and carry very different requirements in the title block.

The output of the analysis is a short process sheet. It lists the datum, the operation sequence, the tool for each feature, and the inspection points. A programmer who skips the sheet still ends up making those decisions, just later and less carefully.

For simple turned parts the whole analysis takes minutes. For a housing with a bore, a thread, two grooves and a milled flat, it can take an hour. That hour is cheaper than a rework loop.

• 1
  Datum firstChoose the surface that carries the most tolerances, usually a bore or a turned face.
• 2
  Sequence secondRough, finish, then any feature that must stay concentric with the datum.
• 3
  Tools thirdFewer tools means fewer setups and less accumulated error.

Turned features fall into a few families. Each one has a preferred tool and a preferred approach. External diameters are the easy case: one or two passes with a turning insert, then a finish pass. Internal bores are harder because tool rigidity drops with depth.

A bore at 4× diameter depth is routine. At 8× diameter the boring bar starts to deflect and chatter. The usual fix is a smaller depth of cut and a slower feed, or a step-bore approach that removes material in stages. Neither fix is free.

Threads, grooves and undercuts each need their own tool and their own entry path. A relief groove at the end of a thread looks like a detail in the model. In the program it decides whether the threading tool can exit cleanly or has to be backed out.

Faces that must run true are the real constraint. If two bores share a concentricity band of 0.02 mm, the program has to finish them in the same setup. Separate setups add fixture error that the tolerance band cannot absorb.

• 1
  Bore depth ratioKeep boring bars under 6× diameter where the tolerance allows.
• 2
  Thread reliefConfirm the groove exists before programming the thread exit.
• 3
  Thin wallsBelow 1 mm wall thickness, plan light finishing passes to limit deflection.

Cutting parameters on a lathe come down to surface speed, feed per revolution and depth of cut. For aluminum 6061, roughing at 200 to 300 m/min with 0.2 to 0.3 mm/rev is normal. Stainless 316 runs slower, often 120 to 180 m/min, because it work-hardens at the cut.

Feed per revolution matters more than most programmers expect. It sets the theoretical surface finish on a turning insert with a given nose radius. When a drawing calls for Ra 0.8–1.6 μm, the feed has to drop to roughly 0.1 mm/rev or the finish pass will not meet it.

Depth of cut controls tool load and heat. On a rigid setup with a 4,000 mm maximum processing size machine, a roughing pass can take several millimeters. On a slender shaft, the same depth will push the part away from the tool and cut a taper instead of a cylinder.

Compensation is the last variable. Tool nose radius compensation changes the programmed path so the insert edge, not the tool center, follows the profile. Turning it off saves a few lines of code and produces a wrong chamfer.

• 1
  Aluminum 6061200–300 m/min, 0.2–0.3 mm/rev roughing.
• 2
  Stainless 316120–180 m/min, keep the tool moving to avoid work hardening.
• 3
  Finish pass0.1 mm/rev or finer to reach Ra 0.8–1.6 μm.

A three-jaw chuck is the default and the least accurate option. Jaw runout of 0.03 to 0.05 mm is common. For parts with tight concentricity, a collet or a soft bored jaw brings that to under 0.01 mm without extra cost.

Between centers is the right answer for long shafts. A steady rest adds support at mid-length and controls deflection. Both approaches limit access to the part, which changes the tool path and sometimes forces the operation to be split.

A mill-turn center removes a setup by doing turning and milling in one machine. Sixteen mill-turn centers in our shop handle parts that would otherwise need two fixtures and two datum transfers. The trade-off is programming time, since the post-processor has to handle both modes.

Bar feeders suit high-volume runs of small parts. They are a poor fit for a one-off prototype because setup time dominates. The process plan should say which case applies before the programmer starts.

• 1
  Collet or soft jawsUse when runout must stay under 0.01 mm.
• 2
  Steady restLong shafts above 8× diameter usually need one.
• 3
  Mill-turnBest when milling and turning share a datum.

Simulation catches most collisions and rapid-move errors. It does not catch a wrong datum or a missing chamfer. Those show up only when someone compares the simulated profile against the drawing.

A dry run with the tool offset set high runs the full path without cutting material. It confirms the tool reaches every feature and clears the chuck. On a part with an internal groove, this step often reveals that the bar cannot reach the back of the bore.

First-article inspection closes the loop. Measure the datum feature first, then the features that depend on it. If the first diameter is out, measuring the rest is wasted effort.

Our inspection runs 100% before shipment, with raw material check, in-process monitoring and final inspection. Reports are available on request. Qualification rate on turned work runs at 99.99%.

• 1
  SimulateCheck collisions and rapid moves against the model.
• 2
  Dry runVerify reach and clearance before cutting metal.
• 3
  First articleMeasure the datum before any dependent feature.