SizingKitFree milling feed rate calculator
Table feed, and whether the machine can pull it
A milling feed rate is spindle speed times flute count times feed per tooth, and on its own it tells you nothing about whether the cut will run. This works out the feed, turns radial width, axial depth and feed into a removal rate, multiplies by the material's unit power to get horsepower at the cutter and at the motor, and compares that against the plate rating you type in. It is free, needs no account, and when the cut is over the rating it solves back for the depth that fits.
- 100% free
- No signup
- MRR and horsepower
- Torque at speed
- Depth that fits
Unit power 1 hp per cubic inch a minute — the only thing the material contributes to a power estimate.
Whatever the surface speed and the diameter settled on.
Table midpoint for this cutter and material: 0.0020–0.0040 in.
100% of the cutter diameter.
How far down the flutes the cut reaches. The cheapest thing to change when the power runs out.
Off the motor plate. A variable-speed head does not hold it below its base speed.
Default 0.8 — The spindle drive efficiency assumed in the power-constant method of Machinery's Handbook's Speeds and Feeds section, where required motor horsepower is spindle horsepower divided by the drive efficiency. Belt-driven knee mills with worn belts run nearer 0.65; a direct-drive spindle on a modern machining center is above 0.9. Continuous ratings also differ from the number on the motor plate, and a variable-speed spindle loses torque badly below its base speed, which is where a large cutter in steel actually runs.
TABLE FEED
F36.0
36.00 in/min · 914 mm/min
POWER AT THE MOTOR
5.63 hp
4.19 kW required
- Material removal rate
- 4.500 in³/min / 73.7 cm³/min
- Power at the cutter
- 4.50 hp / 3.36 kW
- Spindle torque at this speed
- 7.9 lbf·ft / 10.7 N·m
- Against the machine
- short by 88%
Power is the limit, not the geometry. This cut wants 5.63 hp at the motor and the plate says 3.00. Feed and speed are not the thing to change — the removal rate is width × depth × feed, and depth is the term with no side effects on chip thickness. At 0.133 in of axial depth the same feed and stepover land exactly on the rating, so take that and add a second pass. What the arithmetic cannot see is rigidity: a knee mill that has the horsepower on paper may still chatter, and chatter is what breaks cutters at these depths, not stall.
Torque is horsepower × 5,252 ÷ rev/min, so the same cut is a very different demand at 400 rev/min and at 4,000. A variable-frequency spindle holds constant power above its base speed and constant torque below it, which is why a large cutter in steel — the case that needs the most torque — is also the case where the head has the least to give.
How to check a milling cut against the machine it will run on
Three dimensions make a removal rate; one material constant turns it into horsepower.
Describe the cut in the three dimensions that make a removal rate
Radial width across, axial depth down, and the table feed along — those three multiplied together are cubic inches a minute, and cubic inches a minute is what the spindle has to find the power for. Width and depth come off the toolpath; the feed comes from spindle speed, flute count and feed per tooth, which the panel works out for you from the same table a catalog would.
Put the motor plate rating in, not the marketing figure
The number that matters is continuous horsepower at the motor, which is what the plate on the side of the head says. Peak and intermittent ratings are larger and mean nothing for a cut that lasts a minute. The panel divides the power at the cutter by the drive efficiency to get back to the motor, and that efficiency is a field: 0.8 is the assumed figure, a worn belt drive is nearer 0.65 and a direct-drive spindle is above 0.9.
If the cut is over the rating, change depth before anything else
Removal rate is a product of three terms and only one of them can be reduced without side effects. Cutting the feed thins the chip and moves the tool toward rubbing; cutting the width changes the chip thinning correction and the radial force; cutting the depth simply removes less metal per pass and costs a second pass. The panel solves for the exact axial depth that lands on your rating, so the change is a number rather than a guess.
Technical specifications
| Unit power range | 0.3 hp per in³/min for aluminum to 1.4 for annealed alloy steel, with mild steel at 1.0 — a spread of more than four to one on the same removal rate |
|---|---|
| Drive efficiency | 0.8 assumed and editable. A worn belt-driven knee mill is nearer 0.65 and a direct-drive spindle above 0.9, so the field is where the machine you have goes in |
| Removal rate | Radial width × axial depth × table feed, reported in in³/min and cm³/min |
| Torque | Spindle torque = hp × 5,252 ÷ rev/min, in lbf·ft and N·m, because a stall is a torque event and not a power one |
| Machine check | Required motor power against the plate rating, as headroom or shortfall in percent, plus the exact axial depth that would land on the rating |
| Chip thinning | Deliberately not applied here. The panel flags a stepover under half the diameter and sends you to the page that owns the correction, rather than giving one cut two different feeds |
| Where the method breaks | The constants describe a sharp tool taking a real chip: specific energy rises at light chip loads, so a finishing pass draws more power than this predicts and a roughing pass draws about what it says |
| Runs in the tab | Seven unit-power figures and the efficiency assumption are compiled into the page; the cut you describe is not transmitted and not stored |
Frequently asked questions
What is a good material removal rate?
There is no such figure, which is why the panel reports it as an output rather than asking for it as a target. Removal rate is worth exactly as much as the power available to sustain it: 3 in³/min in aluminum needs about 0.9 hp at the cutter and 5 in³/min in annealed 4140 needs 7 hp, so the same number means an easy cut in one material and an impossible one in another. What makes a rate good is that it is the largest one your machine, your workholding and your tool life will support — and on most manual and benchtop machines the binding constraint is rigidity long before it is horsepower.
Why is the power at the motor higher than the power at the cutter?
Because belts, gears, bearings and the drive itself take a share before anything reaches the metal. The unit-power method computes horsepower at the cutting edge, and the motor has to supply that plus the losses, so the required motor power is the cutter figure divided by the drive efficiency rather than multiplied by it. At the assumed 0.8, a cut needing 2.4 hp at the edge needs 3.0 hp at the motor — which is precisely the difference between fitting a 3 hp head and not.
Should I climb mill or conventional mill?
Climb, on anything with ball screws or a tight gib, and conventional on an old machine with backlash in the table screw. Climb milling starts the chip at its thickest and ends at zero, so the tooth enters cleanly instead of rubbing its way in, the heat leaves in the chip, and the surface and tool life are both better. The catch is that the cutter pulls the work in the direction of travel, and on a worn acme screw it will snatch the table through the backlash and break the cutter. That is a machine question, not a cutting one, and it does not change the removal rate or the power at all — climb and conventional cuts of the same three dimensions demand the same horsepower.
Which do I reduce when the machine cannot take the cut?
Axial depth, almost always. The three terms in a removal rate are not interchangeable: reducing feed makes the chip thinner, which is the direction that causes rubbing, work hardening and short tool life; reducing radial width changes the chip thinning factor and means the feed should go up rather than down, so it is a two-step change; reducing depth just takes less metal per pass and costs you another pass. Reducing spindle speed reduces the power too, because power is torque times speed — but it also drops the surface speed below the band, which is the worst trade of the four.
Why does a cut that runs fine at 3,000 rev/min stall at 600?
Because torque, not horsepower, is what stalls a spindle, and torque is horsepower times 5,252 divided by rev/min. A 3 hp head delivering its full rating makes 5.3 lbf·ft at 3,000 rev/min and 26 lbf·ft at 600 — but only if it can hold the rating that low, and a variable-frequency spindle cannot. Below its base speed a VFD-driven motor is a constant-torque device, so the horsepower falls in proportion to speed and the head that was 3 hp at 1,800 rev/min is 1 hp at 600. That is why a large cutter in steel, which needs low rev/min for the surface speed and high torque for the chip, is the hardest thing to ask of a modern high-speed head.
Does the unit power figure allow for a dull cutter?
No, and the gap is large enough to matter. The unit horsepower constants describe a sharp tool taking a proper chip; a worn edge with a wide flank land can take half again as much power for the same removal rate, because more of the energy goes into rubbing and plastic deformation and less into shearing a chip. The same effect appears at very light chip loads even with a sharp tool — specific energy rises as chip thickness falls — so a finishing pass consumes more power per cubic inch than this method predicts, and a roughing pass consumes about what it says.
Is a higher removal rate the same as finishing the job sooner?
Only for the part of the cycle that is actually cutting. On a job with many small features the spindle spends most of its time rapiding, ramping in, changing tools and waiting for the control to look ahead, and a 20% better removal rate buys 20% of a small fraction. Removal rate is the right thing to maximize on a roughing operation with long continuous passes, and close to irrelevant on a plate with forty holes and a light profile — where the drilling cycle and the tool changes are the whole clock.
About table feed, removal rate and the power a cut demands
Everything a mill does is a volume per unit time. Radial width across the cut, axial depth down the flutes and table feed along the path multiply out to cubic inches a minute, and that figure times a material constant is horsepower at the cutting edge. The constant is called unit power or specific cutting energy, and it varies about four to one across ordinary shop materials — 0.3 hp per cubic inch a minute in aluminum, 1.0 in mild steel, 1.4 in annealed alloy steel. It is remarkably insensitive to what tool you use, and fairly sensitive to how thin the chip is, which is the one caveat worth carrying: the method is honest for roughing and optimistic for finishing, because a thin chip costs more energy per cubic inch than a thick one. It never errs the other way, so a comfortable margin here is genuinely comfortable.
The reason to compute power at all is that it is the constraint people discover last. A feed and speed pair looks perfectly reasonable, the toolpath posts, and then a 3 hp head bogs down in the third pass of a 1/2 in slot in 4140. What the panel above does that a plain feed calculator does not is close that loop: it converts the removal rate into horsepower at the motor rather than at the cutter — those differ by the drive efficiency, which is a real 20% on the assumed figure — and then reports the answer as headroom against your plate rating. When there is not enough, it solves the removal rate equation backwards for the axial depth that lands on the rating, because depth is the one term you can reduce without changing the chip. Above the power question sits a torque question that catches people out separately: torque is horsepower times 5,252 over rev/min, so the large cutter in steel that needs the lowest spindle speed is also the one asking for the most torque, exactly where a variable-frequency head has the least to give.
Two things are outside this page on purpose. The feed per tooth itself is not corrected for radial chip thinning here, because doing it in two places is how one cut ends up with two feeds — the panel notices a light stepover and points at the chip load calculator, which owns that correction. And the spindle speed is taken as given rather than derived, because it is a surface speed question and belongs on the surface footage calculator. Nothing here is a process sign-off either: it is a power estimate from published constants, and the machine, the workholding and the person at the handle decide what actually gets cut. If what you are about to hog out of a billet matters to the material order rather than to the spindle, the weight of the stock before you start is on the stainless steel weight calculator.
The cut you described stays here
Every number on this page is worked out by JavaScript running in the tab you are reading it in. Nothing you type — loads, lengths, nameplate ratings, the rates your utility charges you — is uploaded, logged or kept, which is also why the calculators carry on working in a mechanical room with no signal.
The machine rating you enter is the closest thing on this page to information about your shop, and it is used to compute one percentage and then discarded when the tab closes. No part of the cut is sent anywhere.