Two knobs, one load
Every milling cut is described by two engagement numbers alongside the speed and feed. Axial depth of cut is how far the tool reaches down into the material; radial width of cut is how much of the tool's diameter is buried sideways as it passes. Their product, multiplied through by chip load and flute count, is the mechanical load picture the tool experiences - which is why engagement is where the real strategy lives. Two programs can carry the same spindle speed and feed rate and be entirely different jobs, depending on whether the tool is buried one diameter deep or one twentieth of a diameter wide. The speed and feed formulas on the speeds and feeds reference stay the same; the engagement decides what those numbers cost the tool.
Slotting: full width, humble depth
A slot is the heaviest engagement a mill can see: the full diameter buried, chips confined between two walls, evacuation possible only up the flute. Everything the chip-packing failures teach applies at full strength - chip room rules, depth comes in steps, air or flood earns its place. In aluminum, generous flutes and high-helix geometry keep the stringy chips moving; in steels, the depth steps are humbler and the coating does hotter work. Slotting is where the flute count decision bites hardest: the two- and three-flute tools that leave room for chips are the ones that survive deep slots. The general frame holds - depth is built in passes, each one adjusted to how the last cut sounded and looked, never programmed to a chart's courage.
Side milling: light radial, deep axial
Turn the engagement sideways and the tool's strengths invert. In side milling - profiling a wall, taking a finish pass - the radial width is a thin slice while the axial depth can run most of the tool's length of cut. This plays to carbide's nature: the cutting pressure arrives mostly along the tool's axis, where a solid carbide column is stiffest, and the slender direction is barely loaded. Walls come out straighter, deflection worries shrink, and the tool's whole flute length works instead of just its tip. Finishing passes live here almost entirely: a light radial slice, climb direction, the full depth of the wall in one pass where the setup allows. The direction choice interacts with wall forces - the climb vs conventional note covers which way the cutter pushes on a thin wall.
Light-radial roughing: the modern default
The strategy that reshaped roughing over the last generation combines slotting's axial ambition with a modest radial width: the tool runs at or near full depth of cut but takes a light radial slice, advancing fast. Each edge meets a thin, well-formed chip; the heat rides out on the chip instead of soaking into the tool; and the engagement stays short enough in time per tooth that the load remains steady. The feed per tooth runs higher than the light radial alone would suggest, because at thin widths the chip thins - the edge takes less than the programmed feed implies - a mechanism explained in the chip load note. Tool paths built around this idea are the high-efficiency milling strategies covered separately. The honest framing: the strategy trades spindle time for tool consistency, and where it fits, it fits well; where the machine or the fixture cannot hold the pace, conventional stepped passes remain entirely respectable.
An order of decisions that works
| Strategy | Axial engagement | Radial engagement | Where it fits |
|---|---|---|---|
| Slotting | Built in steps, modest per pass | Full diameter | Keyways, grooves, pockets without pre-drilled starts |
| Side milling / finishing | Deep - most of the length of cut | Thin slice | Walls, profiles, finish passes |
| Light-radial roughing | Deep, often full depth of cut | Light, with thinned-chip feeds | Pocket and contour roughing on rigid machines |
| Stepped pocketing | Modest steps | Wide, then narrowing | Pockets on less rigid machines, tough materials |
| Plunge roughing | Deep, axial | Stepover between drops | Deep cavities, long overhangs - see the plunge note |
Run the decisions in an order that respects dependency. Rigidity first: overhang, holder, fixture, part - these set the ceiling on everything downstream. Engagement second: choose the strategy the part and the machine can hold. Chip load third: set within the engagement, adjusted per the calculated baseline and the machine's response. Speed last: placed by the material family's band, then trimmed by what the cut reports. Shops that start with speed and retrofit engagement spend their lives chasing chatter and broken corners; the order is the strategy.
Roughing and finishing as a pair
Most real parts use more than one of the rows above, and the planning principle is simple: roughing removes the bulk and deliberately leaves a thin, even layer for finishing to take. That allowance matters because a finish pass that varies - thin here, thick there - deflects unevenly and telegraphs the variation into the wall. The roughing tools are chosen for removal: serrated periphery on the roughing family breaks chips into short segments at high removal rates, at the cost of a scalloped wall that finishing exists to erase. Corners get their own care, since a square corner in the pocket and a square corner on the tool collide at the end of every pass - one reason the corner radius family appears in so many pocket programs. None of this changes per fashion; it is the load-bearing logic of the whole trade.
Materials bend the strategy, not break it
The strategies above are the frame; the material family adjusts the numbers inside it. Titanium and its heat-hungry ways push toward light radial engagement and rigid everything - the titanium selection keeps the whole conversation conservative. Aluminum invites deep, fast slots where chip room allows, with the geometry to match. Stainless argues for steady engagement that never lets the edge rub - the work-hardening trap described in the stainless note. The frame survives every family; the starting values, as always on this blog, are adjusted to the machine, the setup and the workpiece, and nothing here is a performance promise. If you are specifying tools around particular engagement profiles - deep slots, tall walls, light-radial programs - the quote form takes those details, because length of cut, flute count and corner geometry all follow from them.