Ultimate Guide to CNC Feeds, Speeds & Material Removal Rate
Every CNC milling operation in every machine shop on earth balances three numbers: spindle speed, feed rate, and the depth of material being removed. Get the combination right and the cutter sings through the workpiece, chips curl uniformly, and tool life stretches into thousands of parts. Get it wrong and you are looking at chatter marks, broken carbide, or a part pulled out of the vise. This guide walks through the theory behind the CNC Feeds, Speeds & MRR Matrix and shows exactly how the numbers on a setup sheet are derived from a handful of physical inputs.
The chain of calculations begins with surface speed, measured in surface feet per minute (SFM). Surface speed is the linear velocity of the cutting edge as it sweeps through the material, and it is a property of the work-tool material pair, not the cutter itself. Softer, thermally conductive alloys like Aluminium 6061 tolerate 400 to 800 SFM with carbide tooling, while Titanium Ti-6Al-4V, which conducts heat poorly and work-hardens aggressively, stays in a narrow 100 to 250 SFM window. Because edge speed climbs as the tool gets bigger, a larger cutter must turn fewer revolutions to maintain the same surface speed. That relationship is the classic rule: RPM = SFM × 3.82 ÷ cutter diameter in inches, where 3.82 = 12 ÷ π converts linear feet per minute of edge travel into revolutions per minute of the spindle.
Once the spindle speed is known, the next number is feed rate, expressed in inches per minute (IPM). Each flute must cut a thin chip on every revolution, and the thickness of that chip is the chip load, usually quoted in inches per tooth. Multiply spindle RPM by the flute count and the chip load and you get the linear table feed: IPM = RPM × flutes × chip load. The chip load is where material data becomes critical. A 4-flute carbide end mill slotting Aluminium 6061 runs comfortably around 0.010 to 0.012 inches per tooth, but the same geometry in Stainless Steel 304 drops to roughly 0.0035 inches per tooth. Slots engage the full cutter diameter for 180 degrees of contact, so they demand a lower chip load than light side-milling passes that engage only a fraction of the cutter.
Multiplying the width of cut (radial engagement) by the depth of cut (axial engagement) by the feed rate yields the material removal rate in cubic inches per minute: MRR = WOC × DOC × feed. MRR is the shop-floor productivity number. Doubling the depth of cut doubles the metal removed per minute, but it also doubles the load on every cutting edge and multiplies the heat sunk into the tool. This is why the practical limit on MRR is almost never the arithmetic — it is the torque the spindle can deliver, the rigidity of the machine and toolholder, and the ability of the coolant to carry heat away.
A useful secondary output is estimated spindle horsepower. Multiplying MRR by a material-specific power factor, roughly 0.3 for aluminum, 0.9 for 4140 steel, 1.1 for stainless, and up to 1.6 for titanium, estimates the horsepower demand at the spindle. A 27.5 cubic inch per minute pass in aluminum needs about 8 horsepower at the spindle, which is a reasonable ceiling for a 10 HP machine once transmission losses are counted. This single check prevents operators from programming an MRR that the machine physically cannot feed, which is exactly how spindles stall and tools snap.
The embedded chip-load table in the tool provides realistic starting values for eight common materials: Aluminium 6061 and 7075, Titanium Ti-6Al-4V, Stainless Steel 304, Alloy Steel 4140, Carbon Steel 1018, Grey Cast Iron, and Delrin. Every row stores a recommended surface speed and separate chip loads for slotting, side milling, and plunging. When you select a material and operation, the tool loads those values into the inputs automatically, and you can still override them for a particular machine or coating. The tool also warns when the radial engagement exceeds the cutter diameter, a configuration that is geometrically impossible to cut in a single pass.
Understanding chip thinning makes the numbers practical instead of theoretical. When the radial engagement is less than half the cutter diameter, the average chip becomes thinner than the programmed chip load, because each flute enters and exits the cut at a shallow angle. In that regime, the recommended practice is to raise the feed rate to restore a real chip thickness, rather than let the edge rub. Rubbing generates heat without removing metal and rapidly blunts an edge. For full slotting passes, by contrast, chip thinning does not apply and the theoretical chip thickness matches the programmed chip load directly.
Machining data is always a starting point, never a hard rule. Machine rigidity, tool overhang, coolant type and pressure, workholding quality, and the wear state of the cutter all shift the safe operating window. A common sequence is to start near the middle of the SFM range, run the first part, and listen: a clear ring indicates a stable cut, while a dull thud or harmonic squeal is chatter. Inspecting chip color and curl provides more information than any chart — blue-tinted chips in steel mean the edge is getting hot, while tightly coiled silver chips show the feed and speed are balanced.
Use the CNC Feeds, Speeds & MRR Matrix to generate a complete cutting-data block — RPM, IPM, MRR, chip thickness, feed per revolution, and estimated spindle power — then copy it straight into your setup sheet or CAM tool. The tool runs entirely in your browser, needs no account, and keeps your shop data private. Start near the middle of the recommended range, verify with a test pass, and let the machine tell you where the true ceiling sits.