CNC Turning vs Milling: How the Part Decides

Two kinematics, and everything that follows

Both processes remove metal with a cutting tool, and the difference between them is which element moves.

Turning rotates the workpiece and feeds a single-point tool along and into it. The tool is stationary in the sense that it does not spin; the work does. Because the work turns on an axis, the process naturally produces features that are symmetric about that axis: diameters, faces, tapers, grooves, threads, chamfers and radii — a body of revolution.

Milling holds the workpiece still and spins a multi-tooth cutter, moving it through the material on multiple axes. Because the cutter can be driven in any direction and the part can be repositioned, milling produces shapes that have no axis of symmetry: flat faces, pockets, slots, steps, contours, and hole patterns.

	Turning	Milling
What rotates	The workpiece	The cutting tool
Tool	Single-point, stationary	Multi-tooth, rotating
Natural output	Bodies of revolution	Prismatic shapes
Stock	Round bar	Plate, block
Typical parts	Shafts, pins, bushings, spacers, rings, threaded parts	Brackets, plates, housings, manifolds, fixtures
Features	Diameters, faces, tapers, threads, grooves	Pockets, slots, flats, contours, hole patterns
Concentricity	Natural — one axis, one setup	Setup dependent
Round features	Inherent to the process	Produced by interpolation
Material utilisation on round parts	Excellent	Poor — a lot of the block becomes chips
Material utilisation on prismatic parts	Not applicable	Good
Tolerances	±0.01 mm on precision features	±0.01 mm on precision features

Two rows in that table do most of the practical work: what each process naturally produces, and how much material each wastes getting there.

The decision rule: rotational symmetry

If there is one question that settles the process, it is this: is the part mostly a body of revolution?

Mostly round — turn it. A shaft with steps, a pin with a groove, a bushing with a bore, a spacer, a ring, a threaded stud, a flanged sleeve. Turning these from bar is fast, accurate and economical, and the material is already close to the finished form.

Mostly not round — mill it. A bracket with pockets and a mounting hole pattern, a housing with a cavity and a machined face, a manifold with ports on several sides, a fixture plate, a cover. These shapes have no axis of symmetry to turn about, and a lathe simply cannot reach the features.

The rule gets interesting at the boundary, where a part has both. A shaft with a milled flat, a keyway or a cross-drilled hole is a turned part with a small amount of milling on it. A housing with a large bored bore is a milled part with a turning operation on it. Those are the normal cases, not the exceptions — which leads to the point that matters most for cost.

Setup count is the real cost driver

On any machined part, the number of setups — the number of times the part has to be fixtured in a new orientation — drives the price far more than the volume of metal removed.

Each setup costs time in fixturing and in proving the position. More importantly, each setup introduces a new source of positional error. When a feature is machined in a different setup from the feature it has to relate to, the tolerance between them is the stack of both setups’ errors, plus the fixture’s. That is why a part with tight relationships between faces on different sides is expensive: the money is going into holding the relationship, not into cutting.

The practical consequences:

• Features that can be machined in one setup should be. A turned part that keeps everything on the lathe avoids stacking errors between diameters and faces.
• A feature that is toleranced to another feature belongs in the same setup where possible. This is the instruction that most improves a drawing, and it is worth stating explicitly rather than assuming.
• A part that needs five faces is a different proposition from a part that needs one. That is where five-axis machining earns its cost — not by cutting faster, but by reaching more faces in one setup and protecting the relationships between them. See 5-axis CNC machining for where that line sits.
• Reducing the number of machined faces is the strongest cost lever available. A face that does not need to be machined does not need a setup.

This is also why the process choice sometimes follows the fixturing rather than the geometry. Where a round part needs one small milled feature, whether that feature is done on a lathe with live tooling or on a mill as a second operation can change the price by more than the geometry suggests.

Material utilisation: bar against block

The second structural difference is how much of the bought material becomes a part and how much becomes chips.

Turning starts from round bar. For a round part, the bar is already close to the finished diameter, so the removed material is a thin layer. On a long shaft, the waste is a small fraction of the bought weight.

Milling starts from plate or block. For a prismatic part, the block is reasonably close to the finished envelope, and the waste is the pockets and the profile cut away — acceptable. But milling a round part from a block is a different matter: everything outside the finished cylinder becomes chips. On a long round part, the volume removed can exceed the volume of the part itself.

That matters most on expensive materials. On a titanium or 17-4PH part, the difference between starting from a bar close to size and starting from a block that is mostly removed can be a larger cost item than the machining time — and it is decided by which process the part is routed through. Where the geometry allows, the material-efficient route is usually worth choosing even if the cutting is slightly slower.

The same logic runs through the choice of stock form, which is covered in the comparison of hot rolled and cold drawn bar.

Concentricity, roundness and the nature of the accuracy

Both processes hold ±0.01 mm on precision features with ISO 2768-m as the general block, so the headline tolerance is not the differentiator. What differs is where the accuracy comes from.

In turning, the part turns on its own axis, so a diameter is produced by a single-point tool held at a fixed distance from that axis. Roundness comes from the spindle, and concentricity between two diameters comes from cutting them in the same setup without removing the part. That is why turning is the natural process for anything where two diameters, a face and a bore have to share an axis — shafts, bearing seats, spigots, sealing faces.

In milling, a round feature is produced by interpolating the cutter around a circle, and its position depends on where the part is fixtured. Roundness comes from the machine’s interpolation and the cutter, and position comes from the setup. Both are good, and both are governed by different things — which is why a bore on a mill is usually reamed or bored with a single-point tool when roundness and size matter, rather than interpolated.

The practical instruction that follows: where a round feature has to be concentric with another round feature, or with a turned diameter, say so. It is the single clearest signal that the operation belongs on a lathe, and it is often the difference between a part that assembles and one that does not.

Threads, holes and the operations that overlap

Both processes can produce threads and holes, and the choice follows the same logic.

Threads on an axis of revolution — a threaded shaft end, a threaded spigot — are normally turned, because the thread has to be concentric with the diameters it sits between and keeping it in the same setup protects that. At volume, threads on round parts are often rolled or cut on a screw machine rather than single-point turned, which is its own comparison.

Threads in a hole in a prismatic part — a tapped hole in a bracket or a housing — are drilled and tapped on a mill. There is no axis of revolution to keep them concentric to, and the operation belongs where the hole is.

Holes on a bolt circle in a round part can go either way. On a mill, the circle is interpolated or the part is indexed and drilled; on a lathe with live tooling, the C-axis positions each hole. Which is cheaper depends on the number of holes and the tolerance on the pattern, and it is a genuine judgement call rather than a rule.

Cross-holes in a shaft are the classic case of a secondary operation. They are usually drilled on a mill after turning, unless the lathe has live tooling and the position can be held in the same setup.

Most real parts go through both

It is worth stating plainly, because the framing of “turning versus milling” implies a choice that often does not exist.

A typical turned part with one milled feature goes through a lathe and then a mill, or through a lathe with live tooling that does both. A typical milled part with a precision bore goes through a mill and then a boring operation, sometimes on a lathe. A hydraulic manifold is milled, drilled and tapped, then bored and faced. A valve body may be turned, milled and bored.

The order matters, and it is chosen deliberately:

• Turn first, then mill where the round features define the part’s axis — the milling is then positioned relative to a datum created on the lathe.
• Mill first, then turn where a prismatic body has a bore that has to be true to a machined face.
• Keep related features together wherever the tolerance demands it, even if that means a less obvious process order.

Where a part needs heavy machining on several faces and tight relationships between them, the setup-reduction argument for a multi-axis machine or a mill-turn becomes strong, and it is usually made on accuracy grounds rather than speed. Our Swiss screw machining capacity covers the small-diameter, high-volume end of the turned-part range, where the part is often finished in a single cycle.

Where each process lives

Turning shows up as:

• Shafts, spindles, pins and axles — see custom shafts
• Bushings and sleeves
• Spacers, standoffs and rings
• Threaded studs and turned fasteners
• Flanges, spigots and sealing faces
• Any part whose function depends on diameters sharing an axis

Milling shows up as:

• Brackets, plates and mounting hardware — see custom brackets
• Enclosures and housings — see enclosures and housings
• Manifolds, valve bodies and ported components
• Fixture plates and tooling
• Any part with pockets, slots or hole patterns
• Complex 3D surfaces

Cost

Neither process is inherently cheaper. The cost of a part is driven by the material, the number of setups, the tolerance, the finish and the cycle time — and the process choice acts on all of those through the geometry.

What can be said with confidence:

• A round part belongs on a lathe. Turning it is faster, more accurate where it matters, and wastes far less material.
• A prismatic part belongs on a mill. There is no cheaper way to make a shape with no axis.
• A part with features on many faces gets expensive in setups, and the answer is often to reduce the number of machined faces or to use a multi-axis machine, rather than to change the process.
• Material utilisation can outweigh cutting time on expensive alloys, and it is decided at the stock and process choice.

The most useful thing a drawing can do is to make the functional relationships explicit — which faces have to be true to which, which bore has to be concentric with which diameter, which surfaces are datums. Those statements let the shop route the part in the fewest setups that still meet the requirement, which is where the cost actually comes from.

The callout that belongs on the drawing

Instead of	Write
Machined part	State whether the part is predominantly turned or milled, if known
Round part	Turned from bar; state the concentricity requirement between diameters
Bore, tight tolerance	State whether the bore must be concentric with a turned diameter or a milled face
Flat, tight position	State which face is the datum and what the flat has to relate to
Threaded part	Say whether the thread is on an axis of revolution or in a hole in a prismatic body
Five faces machined	Say so — it changes the machine and the setup count

The single most valuable line is the datum and the relationship. It tells the shop what has to be held and lets the process route be planned around it, rather than discovered afterwards.

Send the drawing with the material, the quantity and the relationships that matter, and the quote comes back within 24 hours with DFM feedback — including where the geometry is adding setups and what could be changed to remove them. Orders start at 50 pieces per style and material, samples from a single piece in 5–10 days, and every batch ships with a material certificate (MTR) tracing the grade to its mill heat plus pre-shipment photos.