Are Large Arc Surfaces Treated With Four-Axis Machine Tools?
Large arc surfaces are treated on four-axis mills more often than people expect. The hard part is not the arc itself. It is the offset between the arc center and the rotary table center. This page covers the geometry, the macro logic, the pass layout and the cases where a four-axis machine is the wrong choice.

What this page covers
A four-axis horizontal mill can cut a large radius surface with a face mill, but only if the arc center, the table center and the part datum are all tied together in one coordinate model.
Why a large arc becomes a four-axis problem at all
A three-axis mill has three linear axes and no way to index the part. Cutting a 745 mm radius with a ball nose on a three-axis machine means a long, shallow scallop, a small stepover and a lot of air cutting. Cycle times climb and the tool wears on the same band of the insert.
Add a rotary A axis and the picture changes. The part indexes around the arc center, and a face mill or a bull nose cutter runs along the bus bar. One pass at a fixed tool angle removes a strip of the arc instead of a single scallop line. Large arc surfaces treated this way need fewer passes and hold a more even surface finish.
The catch is alignment. The rotary axis turns about the table center. The arc turns about its own center. Those two centers rarely coincide on a real part. When they do not, the tool path has to compensate for the offset at every index position, and that is where the programming work sits.
For reference, a part like the one in the classic layout is 280 mm × 210 mm × 114 mm with an outer arc radius of 745 mm. Parts at that scale fit our four-axis envelope without trouble: 4,000 × 400 × 150 mm travel and a Ø400 mm rotary table.
Setting the coordinate model when the centers do not match
Put the X and Z zero at the rotary table center. That gives the control a fixed physical reference that does not move when the part is rechucked. Then measure where the arc center sits relative to that table center in X and Z. Call those offsets a and b. They are the whole problem in two numbers.
The arc can sit in any quadrant relative to the table center. Sign errors here are the most common reason a first article comes out mirrored or shifted. Work out the quadrant from the measured signs before you write a single line of macro, not after the first crash.
The tool reference also has to be defined. On a horizontal four-axis machine the tool cuts along the bus bar, so the effective radius at the cut is the distance from the arc center to the tool tip, not the corner radius of the insert. Set the tool offset against a known diameter and verify it on the machine.
Keep the arc center axis perpendicular to the work surface during tightening and alignment. If the part is tilted in the chuck, every calculated angle is wrong by the tilt amount, and the error grows with the radius.
- 1X and Z zero at table centerFixed reference that survives a rechuck.
- 2Record a and bArc center offset from table center, in X and Z.
- 3Confirm the quadrantSigns decide the rotation direction and the angle formula.
- 4Set the tool offset on a known diameterEffective cutting radius, not the insert corner radius.
The macro logic, step by step
A macro is worth writing here because the pass count changes with the arc length. Hard-coding every index position works once and breaks the next time the radius changes. A parameterized macro takes the arc start angle, the end angle, the step angle and the offsets a and b, then generates the positions itself.
The sequence is short. Assign the parameter values, decide the quadrant from the arc center coordinates, compute the angle between the arc center and the current tool position, index the A axis to that angle, then feed along the bus bar. Repeat until the end angle is reached.
The feed has to be set for the worst case in the pass, which is usually the point where the tool contact is widest. A constant feed across the whole arc will either burn the insert at one end or waste time at the other. On a 745 mm radius the difference between the two ends is measurable in the surface finish.
Step angle is a trade-off. A 3° step on a 745 mm radius leaves a scallop that a Ra 1.6–3.2 μm as-machined finish hides. Drop to a 1° step and the finish improves, but the pass count roughly triples. Pick the step from the print tolerance, not from habit.
Write the macro so the operator can enter the radius, the start angle, the end angle and the step at the top of the program. That turns a one-off job into something the next operator can run in ten minutes.
- 1Read the offsets firsta and b define every downstream angle.
- 2Compute, then indexAngle from the arc center, not from the table center.
- 3Feed for the widest contactConstant feed across a large radius wastes or burns.
- 4Expose the variablesRadius, start, end and step at the top of the program.
Four-axis or something else: a quick read
Match the case to the machine before you spend time on a macro.
| Case | Better choice | Why |
|---|---|---|
| Arc radius above 300 mm, single bus bar | Four-axis horizontal mill | One index per strip, face mill along the bus bar |
| Arc center close to table center | Four-axis, simple program | Offsets a and b near zero, direct indexing |
| Arc center far off the table center | Four-axis with macro | Angle must be recomputed at every index |
| Arc plus compound angles on the same face | Five-axis simultaneous | Tool axis has to tilt while it feeds |
| Free-form blend into a flat, tight tolerance | Five-axis with ball nose | Stepover control beats index-and-feed |
| One or two parts, loose finish | Three-axis, long stepover | Macro setup costs more than the cut |
Where the four-axis method stops working
If the arc surface also carries a taper, a twist or a blend into a second curved face, index-and-feed leaves witness lines at the step boundaries. A five-axis machine tilts the tool axis through the move and removes them. That is the cleanest reason to move the job off a four-axis machine.
Very small radii are the opposite problem. Below roughly 50 mm, the arc is short enough that a ball nose on a three-axis machine with a fine stepover is faster than setting up a macro. The setup time dominates.
Hardened material changes the arithmetic too. In 17-4PH or Inconel, a face mill along the bus bar loads the insert heavily at the widest contact point. Reduce the step angle, reduce the feed, or switch to a smaller-diameter cutter and accept more passes.
Deep arcs that need a long reach will chatter before they cut cleanly. When the tool overhang passes about four times the cutter diameter, the four-axis method loses its surface finish advantage. Shorten the overhang, or move to a machine with a shorter reach to the work.
For parts inside our four-axis envelope we run 12 four-axis mills and 16 simultaneous five-axis centers. The choice is made on the geometry, not on machine availability.
Questions engineers ask before quoting
Can a four-axis machine hold ±0.005 mm on a 745 mm radius arc?
The tolerance is set by the setup and the machine, not by the radius. A four-axis horizontal mill with a verified table center, a measured arc center offset and a tool offset set on a known diameter can hold ±0.005 mm on the arc profile.
The risk is thermal and alignment drift over a long cycle. On a long arc the pass count is high, so we check the first article and re-verify the table center if the cycle runs long.
Do you need a macro, or can CAM post the passes?
CAM can post the passes if the arc center offset is fixed for the batch. A macro earns its place when the radius, start angle or step angle changes between parts.
In practice we do both. CAM for the first article, then a parameterized macro for the repeat runs so the operator can adjust the step without a new post.
What step angle should be used on a large arc?
Start from the print. If the print allows Ra 1.6–3.2 μm as machined, a 3° step on a 745 mm radius usually passes. Tighter finish or a visible cosmetic surface needs a smaller step.
Step angle and feed work together. Halving the step without adjusting the feed can leave a worse finish, because the insert rubs instead of cutting.
When is five-axis the better option for the same part?
When the arc is not a simple bus bar surface. A compound curve, a taper along the arc, or a blend into a second face needs the tool axis to tilt while it feeds.
It is also the better call when the witness lines from index-and-feed fall inside a sealing or mating surface.
Which materials are practical for this method?
Aluminium grades such as 6061, 6061-T6, 7075 and 6082 cut well and hold the finish. Mild and alloy steels like 1018, 1045 and 4140 also run well with a reduced step.
Stainless 304, 316 and 17-4PH, plus titanium TC4 and Inconel, need lower feed and a smaller step angle. Budget more cycle time for those.
How is the arc inspected before shipment?
The arc profile is checked against the model, and the arc center position is verified from the same datum used in the setup. That catches a mirrored or shifted arc before it reaches the customer.
Every part gets inspected before shipment, with raw material checks, in-process monitoring and a final inspection. Reports are available on request.
Send us the arc and we will tell you which machine it belongs on
Upload the model and we return a quotation with a free DFM analysis within 12 hours. If a macro is the right answer, we say so. If a five-axis move is cleaner, we say that too.
12-hour quote100% inspectionNDA on request