How to Calculate Feeds and Speeds: Formulas and Examples

Calculating feeds and speeds comes down to two core formulas: one that sets how fast your spindle turns (RPM) and one that sets how fast the tool moves through the material (feed rate in inches per minute). You need three pieces of information to start: the recommended surface feet per minute (SFM) for the material you’re cutting, the diameter of your tool, and the chip load per tooth for that tool and material combination. Once you have those, the math takes about 30 seconds.

The Two Formulas You Need

Every feeds and speeds calculation starts with spindle speed, then uses that result to find the feed rate.

Spindle Speed: RPM = (3.82 × SFM) / D

Feed Rate: IPM = RPM × IPT × T

Here’s what each variable means:

  • SFM (Surface Feet per Minute): the ideal cutting speed for a given material. This is not something you calculate. It comes from the tool manufacturer’s recommendations or published material charts.
  • D: the diameter of your cutter, in inches.
  • IPT (Inches Per Tooth): also called chip load, this is how much material each flute removes per revolution. Like SFM, this value comes from the manufacturer or a reference chart.
  • T: the number of flutes (teeth) on your end mill.
  • IPM (Inches Per Minute): the resulting feed rate, which is how fast the tool travels through the workpiece.

The constant 3.82 converts SFM into a value that works with tool diameter in inches. If you work in metric, a different constant applies, but for inch-based shops this number is standard.

A Worked Example

Say you’re milling 6061-T6 aluminum with a 1/2″ diameter, 3-flute carbide end mill. Your tool manufacturer recommends an SFM of 1,000 for aluminum with carbide tooling, and the chip load chart lists .004″ per tooth for a 1/2″ cutter in aluminum.

First, find RPM:

RPM = (3.82 × 1,000) / 0.5 = 7,640

Then find the feed rate:

IPM = 7,640 × .004 × 3 = 91.7 inches per minute

That gives you your starting point: set the spindle to 7,640 RPM and program a feed rate of about 92 IPM. If your machine can’t reach 7,640 RPM, cap the spindle at its maximum and recalculate the feed rate using that lower RPM so the chip load stays correct.

Where to Find SFM and Chip Load Values

SFM values depend on the workpiece material and the tool material (carbide, high-speed steel, diamond, etc.). Carbide tooling runs significantly faster than HSS. Your tool manufacturer’s website or the documentation that ships with the end mill is the most reliable source. General ranges for carbide end mills: aluminum can run from roughly 500 to 1,500 SFM, mild steel from 200 to 600 SFM, and stainless steel from 100 to 350 SFM. These ranges vary widely by alloy and coating, so treat them as ballpark starting points.

Chip load values vary by both the material and the cutter diameter. Larger cutters handle a bigger bite per tooth. Here are representative carbide end mill chip loads (in inches per tooth) from Harvey Tool’s published guidelines:

  • Aluminum (6061, 7075, 2024): .001 for a 1/8″ cutter, .002 for 1/4″, .004 for 1/2″, .007 for 1″
  • Cast Iron: .0005 for 1/8″, .002 for 1/4″, .004 for 1/2″, .008 for 1″
  • Stainless Steel (304, 316): .0001 for 1/8″, .0005 for 1/4″, .0015 for 1/2″, .004 for 1″
  • Titanium (6AL-4V): .0005 for 1/8″, .0005 for 1/4″, .001 for 1/2″, .004 for 1″

Notice the dramatic difference between aluminum and stainless steel. A 1/4″ end mill in aluminum takes a chip load of .002″, while the same size cutter in stainless takes just .0005″. Running stainless at aluminum chip loads will break tools. Running aluminum at stainless chip loads will rub instead of cut, generating heat and producing a poor finish.

Adjusting for Chip Thinning

The chip load values above assume your radial depth of cut (how far the tool steps over into the material) is about 50% of the cutter diameter. When your stepover drops below 50%, something counterintuitive happens: the actual chip each flute produces gets thinner than the programmed chip load. This is called chip thinning.

Thin chips mean the tool rubs more than it cuts, which generates heat, accelerates wear, and can cause premature failure. The fix is to increase your programmed feed rate so the actual chip thickness stays at the target value. The more you reduce the stepover, the more you need to bump the feed. At a 25% radial engagement, for example, you may need to nearly double the programmed feed rate to maintain proper chip thickness.

Most CAM software and online calculators have a chip thinning adjustment built in. If you’re calculating by hand, the adjustment factor depends on the ratio of your radial depth of cut to the cutter diameter. Tool manufacturers publish chip thinning charts, or you can use the formula: divide the tool diameter by twice the radial depth of cut, take the square root, and multiply that by your baseline chip load to get the adjusted feed per tooth.

Depth of Cut Matters Too

Feeds and speeds don’t exist in isolation from how deep you’re cutting. Two common parameters work together with RPM and feed rate:

  • Axial depth of cut (ADOC): how deep the tool plunges into the material, measured along the tool’s axis. A common starting point for conventional milling is 1x the tool diameter for aluminum and 0.5x to 1x for steel.
  • Radial depth of cut (RDOC): the stepover width. Conventional roughing often uses 40% to 50% of the tool diameter.

High-speed machining strategies like adaptive clearing or trochoidal milling flip this relationship. They use a shallow radial engagement (often 10% to 15% of the tool diameter) paired with a full axial depth and much higher feed rates. This approach keeps heat low, distributes wear along the full flute length, and can actually remove material faster despite the smaller stepover. If you use this strategy, chip thinning adjustments are essential.

Real-World Adjustments

The formulas give you a theoretical starting point. Your actual machine, setup, and conditions will push those numbers up or down.

Machine rigidity: A worn or lightweight machine with play in the ways or spindle bearings can’t handle the same parameters as a rigid, well-maintained machine. If your machine has noticeable backlash or vibration, reduce speeds and feeds from the calculated values, typically by 10% to 30%, until you find where it cuts cleanly.

Tool stickout: The farther an end mill extends from the collet, the more it deflects under cutting forces. Use the shortest stickout that clears the workpiece. If you must extend the tool for reach, reduce your radial depth of cut and possibly your feed rate to compensate for the lost rigidity.

Workholding: A part that flexes or chatters in the vise absorbs cutting energy in the wrong places. Make sure your workpiece is clamped securely with minimal overhang. Thin-walled parts or long unsupported sections may need reduced cutting forces, which means lighter depths of cut rather than slower speeds.

Coolant: Flood coolant helps evacuate chips and manage heat, letting you run closer to the recommended values. Dry cutting or mist coolant may require lower speeds in heat-sensitive materials like stainless or titanium. In aluminum, air blast is often sufficient because the material conducts heat well on its own.

Putting It All Together

Here’s the process from start to finish. Identify your workpiece material and your tool (diameter, number of flutes, tool material). Look up the recommended SFM and chip load for that combination. Plug SFM and tool diameter into the RPM formula. Plug RPM, chip load, and flute count into the feed rate formula. Check whether your radial depth of cut requires a chip thinning adjustment. Then assess your machine and setup, and dial back if anything is less than ideal.

Listen to the cut once you’re running. A smooth, consistent sound with well-formed chips (small curls or C-shapes, not dust or long ribbons) means your parameters are in the right range. Chatter, squealing, or excessive vibration usually means something needs adjustment, often the speed, the depth of cut, or the workholding rather than the feed rate alone. Over time, reading the chips and the sound becomes as useful as the math itself.