Lathe insert pressed into shaper duty for deburring

Oct 17, 2026 Last reply: 1 day ago 4 Replies
Jump to replies (4)

When this setup runs, a CNC lathe behaves like a shaper: a clamped indexable insert sweeps along the edge of an aluminum part and takes off the burr. The workpiece sits in a rotating chuck while a metal nozzle pours coolant over the cut, and the maker reports the insert cuts the burr clean off rather than leaving anything behind.



The hardware side is straightforward. The insert clamps into a standard lathe tool holder, and the part grips in a HAINBUCH SPANNTOP chuck with Kombi Axfix markings, rated up to 6000 rpm. The machine runs a Siemens 840d control, which relies mostly on built-in cycles instead of long stretches of direct G code. On the programming side, the path was laid out in Fusion as a 2D contour on an open chain, going the right way, with the lead-in radius set to move in x only. The insert was programmed as a 1/16 ball end mill, the operation targets a .010 inch chamfer, and each pass takes only a couple thou. Everything runs in the C axis.



Two details stand out to me. The maker stripped the live spindle M code out by hand to save wear, which reads as a tidy bit of housekeeping. And the shallow depth of cut seems like a sensible way to treat an insert doing a job it was not shaped for, though that is just my read.



The clip shows the insert engaging the outer diameter and peeling off fine shavings, with large curly swarf piling up inside the enclosure. One open question remains: the insert's real lifespan was never stated, only the maker's guess that it should hold up a while on aluminum at that shallow depth.


Curious about one thing: did the maker have to hand write the toolpath for this, or was it all generated in CAM? Getting a contour to run on a lathe in C axis sounds like it could go either way.

Deleting the live spindle M code by hand to save wear is a nice touch in my view. It would have been easy to leave the code as generated and let the spindle spin for nothing. Small edit, but it shows attention to what the machine is actually being asked to do on every cycle.

Since only a guess exists on insert life, a simple next step would be to log the part count per edge and inspect the cutting edge now and then. That would turn the maker's hunch about holding up a while on aluminum into a real number, and it costs nothing but a little attention between runs.

The maker has said the toolpath was programmed in Fusion, set up as a 2D contour on an open chain going the right way, with the lead-in radius made to go in x only. So it was not hand written from scratch; the CAM side handled the path generation.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required