Formula for Calculating Milling Parameters
Cutting speed, feed per tooth, spindle speed, table feed, and metal removal rate: the five numbers a CAM programmer sets at the start of every milling job. This page explains where each formula comes from, which units it expects, and when the textbook answer stops matching what the machine actually does.

In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
Key takeaways
What the formula for calculating milling parameters actually controls
Five numbers decide how a milling cut behaves: cutting speed, spindle speed, feed per tooth, table feed, and depth of cut. The formula for calculating milling parameters links them in a fixed order. You pick a cutting speed from the workpiece material and tool grade, convert it to RPM using the cutter diameter, then turn feed per tooth into a table feed using the tooth count.
The order matters. Engineers who start with RPM and work backwards end up guessing at surface speed, which is the number the carbide actually feels. Surface speed controls heat at the cutting edge. Feed controls the thickness of the chip and how that heat leaves the zone. Get the pair wrong and the tool wears on the flank, chips the edge, or starts rubbing instead of cutting.
On the shop floor we treat the formulas as a starting grid, not a final answer. A 6061 aluminum job and a 17-4PH stainless job can share the same 12 mm end mill and still run at very different speeds and feeds. The formulas tell you the ratio between the variables. The machine, the fixture, and the tool holder tell you how far you can push them.
- 1Cutting speed (Vc)Relative linear speed between the cutting edge and the workpiece surface, in m/min or SFM.
- 2Spindle speed (n)Revolutions per minute, derived from Vc and the effective cutter diameter.
- 3Feed per tooth (fz)Chip load per cutting edge, the main lever on tool life and surface finish.
- 4Table feed (Vf)Programmed feed rate in mm/min, equal to fz × tooth count × RPM.
Cutting speed and spindle speed: the first formula
Cutting speed is the relative linear speed between the cutting edge tip and the workpiece. The classical definition is the product of spindle speed and the cutter circumference. Written out, Vc = π × D × n / 1000, where Vc is in m/min, D is the effective cutter diameter in millimeters, and n is spindle speed in RPM.
Rearranged for the number you actually type into the control, n = 1000 × Vc / (π × D). A 12 mm end mill running at 120 m/min in aluminum gives roughly 3,180 RPM. The same cutter in 4140 steel at 90 m/min gives about 2,390 RPM. Diameter sits in the denominator, so small tools spin fast and large face mills spin slow. That single relationship explains most of the RPM spread across a shop.
Use the effective diameter, not the shank. For a face mill the effective diameter is the cutting circle. For a corner-radius or ball tool, the effective diameter at the point of contact shrinks as the axial depth drops. If you use nominal diameter on a shallow ball-nose pass, the programmed RPM will be far below what the edge needs, and the tool will rub.
Keep unit systems separate. Imperial work uses SFM: n = 3.82 × SFM / D. Metric work uses m/min: n = 318 × Vc / D. Mixing the two constants is the single most common arithmetic error we see in customer programs.
- 1Metricn = 1000 × Vc / (π × D), with Vc in m/min and D in mm.
- 2Imperialn = 3.82 × SFM / D, with D in inches.
- 3Watch the diameterUse the effective cutting diameter at the contact zone, not the nominal size.
Feed per tooth and table feed: from chip load to programmed feed
Feed per tooth is the thickness of material each cutting edge removes in one pass through the arc of cut. It is the variable most directly tied to surface finish and edge wear. Too low and the edge rubs, work-hardens stainless, and burns through the coating. Too high and the edge chips or the spindle stalls.
The conversion to programmed feed is simple: Vf = fz × z × n, where Vf is table feed in mm/min, fz is feed per tooth, z is the number of flutes, and n is RPM. A 4-flute cutter at 3,180 RPM and 0.08 mm per tooth gives 1,018 mm/min. The same cutter at 0.03 mm per tooth gives 382 mm/min. Both may cut, but they will not sound the same or last the same number of parts.
Flute count changes the arithmetic. More flutes raise table feed for the same chip load, which is why aluminum roughing often uses 3 flutes and steel finishing uses 5 or 6. The chip evacuation channel narrows as flute count rises, so deep pockets in gummy material still favor fewer flutes.
Radial engagement modifies the real chip thickness. When radial depth of cut is below about half the cutter diameter, the chip thins at the entry and exit. The programmed feed per tooth then understates the load on the edge. In finishing passes at 5 to 10 percent radial engagement, most CAM systems apply a chip thinning factor to bring the actual load back to the target.
- 1Table feedVf = fz × z × n, in mm/min.
- 2Chip thinningLight radial passes need a higher programmed feed to keep the real chip load.
- 3Flute countMore teeth means more feed at the same chip load, but tighter chip room.
Depth of cut, radial engagement, and metal removal rate
Depth of cut does not appear in the speed and feed formulas, but it sets whether they are usable. Axial depth and radial width together define the cross-section of the chip and the load on the spindle. Metal removal rate is the product of all three: MRR = ap × ae × Vf, where ap is axial depth in mm, ae is radial width in mm, and Vf is table feed in mm/min.
A 12 mm cutter at 1 mm axial, 6 mm radial, and 1,018 mm/min removes about 6,100 mm³/min. That number is a demand on spindle power and on the fixture. A 7.5 kW spindle in aluminum can carry it comfortably. The same cut in 4140 steel will overload the spindle and deflect the tool.
There are two working strategies. Traditional roughing uses high radial engagement and low axial depth, which spreads load across the full flute. High-efficiency roughing uses low radial engagement, often 5 to 15 percent of diameter, with axial depth up to one or two times diameter. The second approach reduces radial cutting force and lets the machine run faster, but it needs a rigid holder and a CAM toolpath that keeps engagement constant.
This is where the formula for calculating milling parameters meets the machine. The arithmetic gives a rate. The spindle meter, the sound of the cut, and the chip color tell you whether that rate is real on this setup.
- 1MRRap × ae × Vf, reported in mm³/min.
- 2Traditional roughingHigh ae, low ap, load spread across the flute length.
- 3High-efficiency roughingLow ae, high ap, lower radial force, needs a rigid setup.
Where the formulas stop working
The formulas assume a rigid setup, a sharp edge, and a cutter that stays on center. Real jobs break those assumptions. Long tool overhangs, thin walls, and weak workholding absorb cutting force as deflection instead of chip load. The programmed feed still reads correct, but the tool is not removing the chip you calculated.
Tool overhang is the usual culprit. A 12 mm cutter at 60 mm gauge length deflects far more than the same cutter at 25 mm. The fix is not a new formula. It is a shorter holder, a smaller step-over, or a slower feed until the cut sounds stable again.
Thermal limits also override the arithmetic. Titanium and Inconel conduct heat poorly, so the edge runs hotter than the cutting speed table suggests. Reducing speed and increasing feed per tooth keeps the heat in the chip instead of the tool. That combination looks wrong on paper and works on the machine.
Machine dynamics set a ceiling too. Every spindle has a stability lobe: certain RPM ranges chatter and others run quiet. When a calculated RPM lands in a bad zone, shifting speed by 10 to 15 percent often fixes the cut without touching feed. That is a practical step, not a theoretical one.
- 1DeflectionLong overhang converts cutting force into tool bend, not chip removal.
- 2HeatPoorly conducting alloys need lower speed and higher chip load.
- 3ChatterMove RPM out of the unstable lobe before changing anything else.
Step by step: from material to a programmed cut
- 1Identify the workpiece and toolNote material condition and tool substrate, coating, diameter, and flute count. These four inputs drive everything else.
- 2Pick the cutting speedStart from a supplier table or the chart above. For 6061 aluminum, 300–500 m/min; for 304 stainless, 50–90 m/min.
- 3Convert to spindle speedn = 1000 × Vc / (π × D). Use effective diameter at the contact zone, not nominal diameter for ball and corner-radius tools.
- 4Set feed per toothChoose a chip load the edge can survive. Typical range is 0.03–0.15 mm depending on material and tool size.
- 5Calculate table feedVf = fz × z × n. Check the number against the machine's rapid and cutting feed limits.
- 6Choose depth of cutSet ap and ae so the resulting MRR stays inside spindle power. Reduce ae first when in doubt.
- 7Apply chip thinningWhen ae is under half the cutter diameter, raise programmed feed so the actual chip load matches the target.
- 8Listen and adjustRun one pass. If chatter appears, shift RPM 10–15 percent. If the edge rubs, raise feed per tooth.
Starting cutting speeds and feeds by material
Carbide tooling, no coatings specified. Adjust for tool grade, coolant, and rigidity.
| Material | Cutting speed | Feed per tooth | Notes |
|---|---|---|---|
| 6061-T6 aluminum | 300–500 m/min | 0.08–0.15 mm | 3-flute roughing works well; watch chip weld |
| 7075 aluminum | 200–350 m/min | 0.08–0.12 mm | Harder than 6061; reduce speed slightly |
| 1018 / 1045 steel | 90–150 m/min | 0.05–0.10 mm | Use coolant; 4–5 flutes for finishing |
| 4140 / 4340 steel | 70–120 m/min | 0.04–0.08 mm | Lower speed at higher hardness |
| 304 / 316 stainless | 50–90 m/min | 0.03–0.07 mm | Never rub; work-hardening risk is high |
| 17-4PH stainless | 40–70 m/min | 0.03–0.06 mm | Condition H900 is harder to cut than H1150 |
| Ti-6Al-4V titanium | 30–60 m/min | 0.03–0.06 mm | Low speed, high coolant pressure, sharp edges |
| Inconel 718 | 20–40 m/min | 0.02–0.05 mm | Rigid setup only; expect short tool life |
When to trust the formula and when to trust the machine
For stable setups in common alloys, the calculated numbers are a good starting grid and you should run them. For long overhangs, thin walls, titanium, or Inconel, treat the formula as an upper bound and let the cut, the chip color, and the spindle load decide the final values.
Questions engineers ask about milling parameters
Should I use the cutter diameter or the effective diameter for RPM?
Use the effective diameter at the point of contact. For a square shoulder mill taking a full-width cut, that is the nominal diameter. For a ball-nose tool at shallow axial depth, the contact circle is much smaller than the nominal size, so using nominal diameter gives an RPM that is too low and the edge rubs.
For face mills, the effective diameter is the cutting circle, which is the nominal diameter. For corner-radius tools, the effective diameter sits between the corner radius contact and the full diameter, depending on axial depth.
Why does my calculated feed rate sound fine but leave a poor finish?
Feed per tooth sets the theoretical cusp height, but finish also depends on radial engagement, tool runout, and spindle condition. A cutter with 0.02 mm of runout loads one flute harder than the others, which shows up as a repeating mark on the surface.
Check runout at the cutting edge with an indicator before changing speeds and feeds. If runout is under 0.01 mm and the finish is still rough, the issue is usually chip thinning or a resonance in the setup.
Can I increase feed per tooth to make tools last longer?
Often yes, within limits. Rubbing is more damaging than cutting, and a chip load below the edge's minimum thickness generates heat without removing material efficiently. That is common in stainless and titanium.
The upper limit is edge strength and spindle torque. If feed per tooth rises past what the edge can support, chipping starts at the corner radius. Increase in small steps and watch the wear pattern.
How does chip thinning change the programmed feed?
When radial depth of cut drops below about half the cutter diameter, the chip is thinner than the nominal feed per tooth because the entry and exit arcs shorten. To keep the real chip load at target, the programmed feed must rise.
Most CAM systems calculate this factor automatically from the toolpath engagement. If you program by hand, apply a thinning factor when ae is under 50 percent of diameter, and increase it further as ae drops toward 10 percent.
What limits metal removal rate more, spindle power or tool life?
In aluminum, spindle power usually binds first because the material cuts easily and the machine runs out of torque. In steel, stainless, and superalloys, tool life binds first because the edge wears or chips before the spindle reaches its limit.
Check both. If spindle load is under 70 percent and the edge still fails early, the problem is speed, chip load, or coolant delivery rather than the removal rate itself.
Do coolant and coating change the calculated values?
They change the practical range, not the formula. A coated carbide grade can run 20 to 40 percent faster than an uncoated one in the same material. High-pressure coolant through the tool usually allows a higher feed per tooth in deep pockets.
Start from the supplier's recommended range for the exact grade, then adjust for the setup. The arithmetic stays the same; only the inputs move.
Send us the drawing and the material
We will review the geometry, pick the tooling and parameters, and return a quotation with a free DFM analysis within 12 hours.
12-hour quote100% inspectionNo minimum order quantity