CNC feed and speed calculator
Feed rate = RPM × flutes × chip load. Everything difficult about feeds and speeds is in that last term. This calculator looks up a chip-load range for your material and cutter diameter from published tooling-manufacturer data, reduces it for the depth of your pass and for bit geometries that struggle to clear chips, then multiplies out. The working is shown beneath the result so you can check it.
Your chip load is 0.2795 mm per tooth, which is within the recommended 0.229–0.330 mm range for hardwood with a 6.35 mm upcut spiral.
Hardwood. Hardwood is the material where too slow a feed shows up fastest: the cutter rubs instead of cutting, heat builds, and the edge scorches. Burn marks mean feed up or RPM down, not the reverse.
Upcut spiral. The baseline the published charts assume. Chips are lifted clear of the cut, which keeps heat out of the workpiece at the cost of tear-out on the top surface.
A worked example
Take the most common desktop setup there is: a 1/4″ two-flute upcut spiral in hardwood, spindle at 18,000 RPM, cutting 3 mm deep.
| Table range, hardwood at 6.35 mm | 0.229 – 0.330 mm/tooth |
|---|---|
| Depth factor (3 mm is 0.47× diameter) | ×1.00 |
| Geometry factor (upcut is the baseline) | ×1.00 |
| Target chip load (midpoint) | 0.2795 mm/tooth |
| Feed = 18,000 × 2 × 0.2795 | 10,062 mm/min · 396.1 IPM |
Nearly 400 IPM surprises people, and it is the single most useful thing this calculator teaches. The published charts were written for rigid industrial routers, so a desktop machine will reach the limit of its own stiffness long before it reaches that feed - which is exactly why so many hobby cuts burn. The machine cannot go fast enough to hold the right chip load, so the honest fix is to drop the spindle speed rather than crawl the feed. At 12,000 RPM the same chip load needs only 6,708 mm/min, which most desktop machines can actually deliver.
Swap that upcut for a downcut and the target falls 30% to 0.1957 mm, taking the feed to 7,043 mm/min. Take the same cut 12.7 mm deep instead - twice the cutter diameter - and the depth factor drops to 0.75, giving 7,546 mm/min. Both adjustments are shown separately in the working above the fold so you can see which one moved the number.
Pre-filled by material
Each material has its own page with the chip-load table and the failure mode specific to cutting it.
How this calculator works
Chip load - the thickness of material each cutting edge removes per revolution - is the quantity that actually determines whether a cut succeeds. Feed rate and spindle speed only matter through their effect on it. That is why two people running the same feed rate can get completely different results: change the flute count or the cutter diameter and the chip load changes underneath them.
Where the reference ranges come from
The chip-load table is anchored at three cutter diameters - 1/8″, 1/4″ and 1/2″ - for each material, using ranges published by tooling manufacturers for the corresponding cutter classes. Between those anchors the calculator interpolates linearly on diameter, because a larger tool has the core strength to take a proportionally bigger bite. Below the smallest anchor and above the largest it clamps rather than extrapolating: a straight-line extrapolation down to a 1 mm engraving cutter produces a figure no published chart would endorse. The table is versioned (currently 2026.07-1) and lives in code rather than a database, so every change to a published number is reviewable.
The two adjustments
Depth of cut. Published chip loads assume a pass roughly equal to the cutter diameter. Go deeper and more of the flute is engaged at once, so the load on the tool rises faster than the numbers suggest. The calculator reduces the target chip load by about 25% at twice the cutter diameter and about 50% at three times, interpolating between and holding a floor beyond that.
Bit geometry. The reference ranges assume an upcut spiral, which lifts chips clear of the slot. A downcut spiral does the opposite - it packs chips down into the cut - so its effective chip load has to come down about 30% to stop the cutter recutting its own swarf. Compression, straight, ball-nose and V-bit geometries each carry their own factor, listed below.
| Geometry | Factor | Why |
|---|---|---|
| Upcut spiral | ×1.00 | The baseline the published charts assume. Chips are lifted clear of the cut, which keeps heat out of the workpiece at the cost of tear-out on the top surface. |
| Downcut spiral | ×0.70 | Packs chips down into the slot instead of clearing them, so the effective chip load has to come down about 30% to stop the cutter recutting its own swarf. Buys you a clean top edge. |
| Compression spiral | ×0.90 | Upcut at the tip, downcut above it, so both faces of a sheet stay clean. Needs a first pass deep enough to reach the compression zone to work at all. |
| Straight flute | ×0.85 | No helix means no chip-clearing action at all - evacuation depends entirely on the cut being shallow or open. |
| V-bit | ×0.60 | The effective cutting diameter changes with depth, so a single chip-load figure is an approximation by definition. Treat the output as an upper bound and start below it. |
| Ball-nose | ×0.80 | The effective diameter at the tip is smaller than the shank diameter, so the tip runs slower than the chart assumes. Used for 3D finishing passes where stepover, not feed, sets the finish. |
Why the result targets the middle of the range
The recommendation is the midpoint of the adjusted window, not its maximum. Published charts are written for rigid industrial routers bolted to a floor; a desktop machine on a bench reaches the limit of its own stiffness well before it reaches the top of the chart. Starting mid-range gives you somewhere to go in both directions once you have heard how the cut sounds.
What this cannot know
The calculator has no way to see your machine's rigidity, the sharpness of your cutter, how well the work is held down, or whether your spindle actually reaches its commanded RPM under load. Those are the variables that decide whether the arithmetic survives contact with the material. Treat every number here as a starting point for a test cut in scrap, not as a setting to run a finished piece at. It is an estimate derived from published data, not a test result.
Reference values and the exact weights are documented on the methodology page. For the inverse problem - checking the chip load of a feed rate you are already running — use the chip load calculator.