Basic Knowledge of CNC Side Milling
This cut forms a vertical wall with the periphery of the cutter, not the tip. The guide below covers cutter geometry, depth settings, workholding and the cases where another strategy is the better call. Written for engineers and buyers who need to judge a part before it goes on the machine.

What this page covers
A working explanation of the process: how the cut is made, which numbers matter, and where it stops being economical.
What the cut actually is
Material is removed by the cylindrical periphery of the cutter while the tool axis stays parallel to the wall being produced. The cutting edges on the side of the tool do the work, and the corner radius or tip only touches the floor of the feature. That is the difference from face milling, where the bottom edge of the tool sweeps a flat surface and the side edges barely engage.
Because the wall is generated by the side of the tool, the finished surface carries the tool marks of the flute geometry. A four-flute cutter leaves a different pattern than a two-flute or a variable-pitch tool. On aluminium a high helix pushes chips up and out; on stainless a lower helix and a tougher substrate keep the edge from chipping.
Any feature with a straight or contoured vertical face can be cut this way: pockets, slots, shoulders, bosses, open profiles and the outside of a plate. The same motion also does the roughing pass that clears material before a finishing pass runs along the wall at a smaller radial stepover.
- 1Peripheral contactThe side of the cutter generates the wall; the tip generates the floor.
- 2Axial depth (ap)How deep the tool engages along its axis, up to the flute length.
- 3Radial depth (ae)How much of the cutter diameter bites into the material.
Matching the cutter to the wall
Cutter diameter sets the minimum internal radius you can cut. A Ø10 mm end mill cannot produce a corner tighter than about 5 mm, so a 3 mm radius in a pocket corner forces a smaller tool or a separate operation with a smaller cutter and a longer reach. Engineers should check corner radii against the drawing before releasing the part, because that single number often decides how many setups are needed.
Flute count trades chip room against rigidity. Two and three flutes clear large chips and suit aluminium, plastics and roughing cuts where the chip is bulky. Four and five flutes give a stiffer core and a smoother wall, which helps on steel, titanium and finishing passes. Variable-pitch geometry breaks up chatter harmonics and is worth the cost on thin walls.
Helix angle controls the direction the cutting force pushes the chip and the work. A 45° helix is a common general-purpose choice. Higher helix angles, 50° to 60°, pull chips up and reduce recutting in deep pockets, but they also pull the part upward, so workholding has to resist that lift. On a thin floor, that lift is the failure mode you watch.
- 1Corner radius ruleSmallest internal radius equals half the cutter diameter.
- 2Reach-to-diameterKeep the gauge length under about 4×D to limit deflection.
- 3CoatingAlTiN for steel and stainless; uncoated or DLC for aluminium.
Depth of cut, stepover and feed
Two numbers define the load on the tool: axial depth and radial depth. Traditional practice uses a small axial depth and a large radial engagement, often 50% of the diameter or more. High-efficiency milling flips that, using the full flute length axially and a radial stepover of 5% to 10% of the diameter. The second approach spreads wear along more of the edge and cuts the cycle time on deep walls.
Feed per tooth is the number that matters, not table feed alone. Surface speed drops as the tool gets smaller, and feed per tooth drops with it. A Ø12 mm carbide cutter in 6061 aluminium may run at 300–500 m/min surface speed and 0.05–0.10 mm per tooth; the same cutter in 316 stainless runs far slower, closer to 60–90 m/min, with a lighter chip load.
Climb milling is the default on a CNC with ball screws. The cutter engages the material at maximum chip thickness and exits at zero, which pushes the tool away from the finished wall and leaves a cleaner surface. Conventional milling does the opposite and is mostly reserved for rough castings with a hard skin or for older machines with backlash.
- 1Light radial, full axialGood for deep walls and long reach tools.
- 2Heavy radial, light axialGood for shallow pockets and rigid setups.
- 3Climb millingDefault on modern CNC; better finish and tool life.
Typical starting points by material
Values are starting points for carbide tooling with good coolant and a rigid setup. Adjust to the actual machine and holder.
| Material | Surface speed (m/min) | Feed per tooth (mm) | Notes |
|---|---|---|---|
| 6061 aluminium | 300–500 | 0.05–0.10 | High helix, uncoated or DLC |
| 7075 aluminium | 200–350 | 0.04–0.08 | Lower speed, sharp edges |
| 303 / 304 stainless | 60–90 | 0.02–0.05 | AlTiN coating, flood coolant |
| 17-4PH stainless | 40–70 | 0.02–0.04 | Rigid holder, light stepover |
| 1018 / 1045 steel | 120–200 | 0.03–0.07 | AlTiN or TiAlN coating |
| Ti-6Al-4V | 30–60 | 0.02–0.05 | High pressure coolant, sharp edge |
| POM / PEEK | 200–400 | 0.05–0.12 | Sharp edge, air blast |
Holding a straight wall and a tight tolerance
Tool deflection is the main enemy of wall accuracy. A long, thin cutter pushed sideways bends away from the material, so the top of the wall ends up oversize and the bottom undersize. The fix is a shorter gauge length, a larger diameter where the geometry allows it, and a lighter radial stepover. A finishing pass of 0.2–0.5 mm radial depth removes the deflected surface left by the roughing pass.
Thermal growth matters on long runs. The spindle, the ball screws and the workpiece all move as the machine warms up. On parts with a ±0.005 mm tolerance, that drift alone can consume the whole band. Warm-up cycles and in-process probing keep the tool on the nominal path instead of chasing the error after the fact.
Thin walls add a second problem: the wall moves as the material around it is removed. Rought out both sides, leave a uniform allowance, and finish in a sequence that keeps the wall supported. Sometimes the answer is a fixture that backs the wall, or a change in the order of operations so the stiff side is cut last.
- 1Finish allowanceLeave 0.2–0.5 mm radial for the finishing pass.
- 2Spring passesA zero-stepover pass cleans up deflection marks.
- 3ProbingUse in-process probing on tolerances under ±0.01 mm.
When this strategy is the wrong call
A wall shorter than the cutter diameter is usually cut faster with a face mill or a shoulder cutter. Those tools engage more edge at once and clear the floor in fewer passes. Reserving the side strategy for walls deep enough to need the flute length keeps the cycle time honest.
Internal corners with a radius under about 1 mm are a different problem. The tool that can reach the corner is small enough to deflect or break, and the wall beside it suffers. It is often cheaper to relax that corner to a standard radius, or to cut the sharp detail on a sinker EDM after milling the rest of the cavity.
Very deep, narrow slots push the reach-to-diameter ratio past what a solid carbide cutter can hold. Above roughly 6×D, chatter and taper show up even with reduced feeds. A larger cutter with a relieved shank, or a different process such as wire EDM for a through slot, usually beats fighting the deflection.
- 1Shallow wallsFace milling or shoulder milling is faster below about 1×D.
- 2Sharp cornersRelax the radius or plan an EDM operation.
- 3Deep narrow slotsAbove 6×D reach, expect taper and chatter.
Questions engineers ask
What is the difference between side milling and face milling?
Face milling cuts with the bottom edge of the tool to produce a flat surface; the tool axis is perpendicular to the surface. Side milling cuts with the periphery to produce a vertical wall; the tool axis is parallel to that wall.
Many parts need both. A pocket floor is face milled and its walls are side milled, often in the same toolpath with the same cutter.
How deep can one pass go?
It depends on the axial depth the tool can take and the rigidity of the setup. A stub-length carbide cutter in aluminium can take its full flute length in a high-efficiency pass with a 5% to 10% radial stepover.
In steel or titanium, start at one-third to one-half of the diameter axially and increase only when the spindle load and the finish allow it.
Why does my wall taper from top to bottom?
Taper almost always comes from tool deflection. The cutting force pushes the bottom of the tool away while the top stays on the nominal path.
Shorten the gauge length, reduce the radial stepover, or run a finishing pass at 0.2–0.5 mm radial depth. Check the holder too; a worn collet adds runout that shows up as taper.
Climb or conventional milling for a finishing pass?
Use climb milling on any machine with ball screws and low backlash. The edge enters at maximum chip thickness and exits at zero, so it tends to pull the tool away from the finished surface.
Conventional milling is the safer choice on an older machine with visible backlash, or when cutting through a hard cast skin that would chip a climb-milling edge.
Which materials can be side milled?
Aluminium, stainless steel, carbon steel, copper and brass, titanium alloys, magnesium and most engineering plastics can all be cut this way. The tool geometry, coating and surface speed change with the material, not the strategy.
Hardened tool steel above roughly 45 HRC is usually cut on a dedicated hard-milling machine with CBN or ceramic tooling, and the feeds drop sharply.
What tolerance and finish can a side milling operation hold?
On a rigid machine with a short, stiff cutter, wall position can hold ±0.005 mm and surface finish can reach Ra 0.2–0.8 μm with a finishing pass.
A general as-machined finish sits around Ra 1.6–3.2 μm. Tighter numbers need a controlled setup, temperature stability and in-process inspection.
Send the drawing, get a process plan
Upload your part and we will review wall thickness, corner radii and tolerances, then quote the operation and the finishing pass needed to hold it.
12-hour quote100% inspectionNo minimum order quantity