Application of GIBBSCAM Turning and Frail

A two-axis lathe will accept usable code from almost any CAM package. The toolpath is short, the geometry is simple, and the operator can adjust offsets at the control. GIBBSCAM starts to pay off when the part leaves that comfort zone: live tooling, a sub-spindle, or a B-axis head. At that point the programmer is no longer writing a single operation. They are sequencing a set of operations across two spindles and several axes, and the order matters more than any individual pass.

The feature that changes the workflow most is the coordinate system. Define a new CS on the face of a cross-hole or a Y-axis groove, and GIBBSCAM keeps the tool vectors relative to that feature instead of making the programmer recalculate every position by hand. On a part with eight radial holes at two angles, that is the difference between an hour of careful work and an afternoon of trigonometry. Full machine simulation is the second reason shops keep the software: collision checking between the turret, the sub-spindle, and the part is cheaper than a crashed B-axis head.

Turning output is only as good as the post. A generic post will run, but it will not use the machine well. Feed per revolution, constant surface speed, spindle sync for pick-off, and bar-puller macros are all post-level decisions. When we bring a new turn-mill center online, the post gets validated on scrap before any real job touches it.

Frail is not a formal category. It is a practical label for parts where the workpiece moves more than the slide does. A thin-wall tube, a long slender shaft, a thin disc, or a part with an interrupted cut all qualify. The failure mode is usually the same: cutting force deflects the part, the tool leaves a mark, and the next pass starts from a surface that is no longer where the program thinks it is.

Wall thickness is the first number to check. Below roughly 1.5 mm on aluminum, or 1 mm on stainless, a three-jaw chuck will distort the bore even at moderate clamping pressure. The part springs back after unclamping, and the diameter measures wrong on the CMM. Boring soft jaws to the finished diameter, or switching to a collet or expanding mandrel, usually fixes it. On very thin rings, we machine with the part supported from the inside and take the final wall in two light passes.

Length-to-diameter ratio drives the second problem. A shaft over about 8:1 will chatter without a steady rest or tailstock support. Above 12:1, the deflection is often larger than the tolerance, and no amount of feed adjustment saves the job. The honest answer is a different setup, not a different speed.

A C-axis alone indexes the spindle; the tool still sits on center. Add a Y-axis and the live tool can move off center, which is what makes face grooves, slots, and hole patterns possible without a second setup. In GIBBSCAM this is where the coordinate system work shows up most clearly. Each feature gets its own CS, the tool axis follows the feature normal, and the posted code carries the right C rotation and Y offset.

Spindle orientation accuracy matters here. The C-axis repeats to a fraction of a degree, but the part has to be clamped so that the rotation is repeatable across loadings. On a part with a tight angular tolerance between two groove sets, a small clamping error shows up as a large positional error at the outer diameter. A stop pin or a serrated jaw face is often the cheapest fix, and it costs nothing at the programming stage.

Not every shop needs a Y-axis. For a handful of radial holes at one angle, a C-axis lathe with an axial live tool does the job, and the GIBBSCAM program is short. The Y-axis earns its cost when the part has off-center features that would otherwise require a mill fixture and a second op. That second op is where position error and handling damage usually creep in.

A thin-wall housing with a cross-hole and a face groove is a good example. The blank is turned in soft jaws to the finished OD and bore, then the part is moved to a mill-turn center for the cross-hole and groove while it stays on the same spindle. Keeping the part on one machine avoids re-clamping a wall that is now 1.2 mm thick. On a part like this, re-clamping is often the largest single source of scrap.

Cutting parameters are set conservatively for the finishing passes. Light radial engagement, higher spindle speed where the tool allows, and a sharp uncoated or PVD insert on aluminum. We inspect with a micrometer at the wall, not just at the ends, and we check roundness at three angular positions. A part that measures round at one spot can still be triangular after clamping release.

Everything ships after 100% inspection. Raw material certificates, in-process checks, and final inspection reports are available on request. Our tolerance capability is ±0.005 mm and surface finish reaches Ra 0.2–0.8 μm on finishing passes where the geometry allows it. Those numbers assume the setup is right; a frail part in the wrong chuck will miss them by a wide margin.