CNC feeds and speeds describe how fast the cutter moves through the material and how fast it rotates. They are connected through chip load: the thickness of material removed by each cutting edge on each revolution. A useful setting is therefore not a number copied from another machine. It is a controlled combination that produces stable motion, real chips, acceptable finish and manageable heat on your exact setup.
The safest way to improve that combination is to begin with evidence for the cutter and material, run a small controlled test, observe the cut and record enough context to reproduce it.
The Four Variables in One Equation
The basic relationship is:
Feed rate = spindle speed × number of flutes × chip load
Rearranged:
Chip load = feed rate ÷ (spindle speed × number of flutes)
This equation appears simple, but it explains why changing only one number can change cutter behavior. If feed rate stays constant while RPM increases, each flute removes a thinner chip. If RPM stays constant while feed increases, each flute takes a thicker chip. Adding flutes without changing feed or RPM divides the same forward movement among more cutting edges.
Keep the units consistent
Feed rate and chip load must use compatible units. If feed is in millimeters per minute, chip load will be in millimeters per tooth. If feed is in inches per minute, the result will be inches per tooth. RPM is revolutions per minute, and the flute count is the number of cutting edges participating in the cut.
For arithmetic illustration only, a feed of 600 mm/min at 10,000 RPM with a two-flute cutter calculates to 0.03 mm per tooth. That calculation does not make those values a recommendation. The suitable result still depends on the cutter manufacturer's data, material, tool diameter, cutting engagement, machine, spindle, stick-out and workholding.
Chip load is part of a larger load system
The equation does not include depth of cut, radial engagement, entry move, cutter projection, machine rigidity or chip evacuation. A full-width slot loads a tool differently from a light finishing pass. A cutter extended farther from the collet has more leverage. A deep pocket can trap chips that a shallow profile clears easily.
Treat calculated chip load as one diagnostic value, not a guarantee. If the machine cannot maintain the commanded feed, the effective chip load can differ from the number in CAM. Tight corners and short segments may force the axes to slow down, increasing cutter contact time and heat.
Start From Tool and Material Evidence
The best starting source is the cutter manufacturer's current data for the exact tool family and intended material. A general internet chart may use a different geometry, diameter, coating or machine class. A saved setting from your own shop is useful only when its context is complete.
Identify the cutter before choosing a value
Record the cutter manufacturer, product or part number, diameter, flute count, flute direction, cutting-edge length and actual stick-out. Note whether the operation is profiling, slotting, pocketing, V-carving, surfacing or finishing. The TwoTrees CNC router bit guide can help separate bit families by task, while the current CNC router parts collection provides a place to verify available tooling and accessory categories.
Do not substitute one flute count for another without recalculating the relationship. A two-flute cutter at the same feed and RPM does not take the same chip per tooth as a single-flute cutter. Likewise, a larger diameter may tolerate a different load but can also demand more spindle power and machine rigidity.
Describe the material as a batch, not a label
“Plywood,” “acrylic” and “aluminum” are broad categories. Record sheet or stock thickness, grade or alloy when known, supplier, surface condition and batch. Plywood glue layers and voids vary. Plastics can respond differently to heat. Wood density and grain change through a board.
If the material cannot be identified, do not compensate by guessing a more aggressive setting. Use known test stock and obtain the supplier's material information. For laser work, unknown material may create a serious fume hazard; for CNC routing, material identity still affects dust, chip behavior, cutter choice and finish.
Confirm the machine-side limits
Check the exact router's spindle range, controller behavior, maximum practical feed, workholding and dust or chip-removal setup. A programmed feed higher than the machine can sustain does not improve chip load. A spindle command is useful only if the spindle actually runs at that speed or has been measured with appropriate equipment.
Use the current manual, official wiki or TwoTrees help center for model-specific procedures. Do not alter firmware limits, motor current or safety controls merely to reach a value from a tooling chart.
Build a Controlled Test Matrix
A useful test answers one question at a time. If feed, RPM, depth and toolpath all change together, the result cannot show which change helped.
Establish a conservative baseline
Start from documented cutter and material guidance that fits the machine's supported range. Use scrap from the actual material batch. Keep the tool projection short while still clearing the work. Secure the stock, check clamp clearance, set the origin and preview the toolpath.
The first test should be small enough to stop quickly and inspect. A straight profile, shallow pocket or simple slot can reveal chip formation and finish without committing a valuable workpiece. The operation should represent the real job: a finishing pass cannot validate a deep slot, and a short engraving line cannot validate prolonged pocketing.
Change one variable in a deliberate order
Use a matrix like this rather than a universal setting chart:
| Test stage | Hold constant | Change | Observe | Decision |
|---|---|---|---|---|
| Baseline | Cutter, material, engagement, workholding | Nothing | Chips, sound, edge, heat, machine motion | Establish evidence |
| Feed test | RPM and engagement | One small feed increment | Chip thickness, finish, axis stability | Keep, reverse or refine |
| RPM test | Feed and engagement | One small RPM increment | Heat, sound, edge and chip form | Keep, reverse or refine |
| Engagement test | Feed, RPM and cutter | One depth or stepover change | Deflection, chatter, chip evacuation | Approve only for this operation |
| Confirmation | All selected variables | Repeat the candidate setting | Repeatability across two test areas | Save with conditions |
The table intentionally contains no fixed percentage or starting value. The safe size and direction of a change depend on the evidence, machine and cutter. If the cut is unstable, stop and inspect before adjusting. Do not “tune through” a loose cutter, poor workholding, excessive stick-out or mechanical play.
Read Chips, Sound, and Surface Finish
The cut leaves several kinds of evidence. None should be used alone.
Chips show whether material is leaving the cut
In many routing operations, distinct chips indicate that the cutting edges are removing material. Very fine dust can suggest a thin chip, a dull or unsuitable cutter, recutting, or a material that naturally fractures into fine particles. Melted plastic or material welded to the flute points toward excess heat and inadequate chip evacuation, but the cause may involve feed, RPM, tool geometry or a packed toolpath.
Collect observations only after the cutter stops. Never reach into a running machine. Control wood and MDF dust at the source, use appropriate eye and hearing protection, and keep hands, hair and loose clothing away from moving components.
Sound is a change detector, not a meter
A stable cut often has a consistent sound. A sudden pitch change, repeated hammering or chatter can indicate changing engagement, poor support, a loose setup or excessive load. Sound cannot provide chip load by itself, and a quiet cut is not automatically correct. Pair sound with chips, surface finish and machine motion.
Surface evidence narrows the cause
Burning on wood may point to heat, rubbing, a dirty or dull edge, slow local motion or chip recutting. Fuzzy grain may relate to cutter sharpness, flute direction, grain reversal or the need for a finishing pass. Chatter marks can come from tool stick-out, workholding, cutting load, spindle runout or machine play.
When several symptoms appear, correct basic mechanical and setup problems before parameter optimization. A feed-and-speed change cannot tighten a collet or flatten unsupported stock.
Record Settings That Are Reproducible
A number without context is not a recipe. Save enough detail for another operator—or your future self—to reproduce the setup.
Use a complete job record
Record:
- machine and spindle configuration;
- cutter manufacturer, identifier, diameter, flutes and condition;
- measured or commanded RPM, clearly labeled;
- programmed feed, plunge and entry strategy;
- material description, thickness, batch and supplier;
- axial depth, radial engagement and whether the cut is a slot;
- stick-out, workholding and spoilboard condition;
- chip evacuation or dust-collection arrangement;
- toolpath type and CAM/post-processor version;
- chip appearance, sound, finish, heat observations and any stopped test;
- date, operator and approval status.
Photos of chips and edges can make the log more useful, provided they are labeled with the test conditions. A successful production record should also note how many parts were completed before tool wear or material variation changed the result.
Separate approved recipes from experiments
Use statuses such as Draft, Tested on scrap, Approved for this material batch and Retired. This avoids treating a one-time result as a permanent standard. Store the record with the job file, not only in a notebook far from the CAM data.
For a new TwoTrees machine, build the library slowly. The current TwoTrees CNC router collection covers different machine classes; settings should not be copied from one class or spindle configuration to another without revalidation.
Scale Changes Without Losing Control
A saved recipe is a reference point, not a universal truth. When the cutter, material or engagement changes, identify which variables must be reconsidered.
This step sits within What CNC Chips Reveal About Feed, Speed, and Tool Load; continue with Measure CNC Dust Collection at the Cutter, Not the Hose and Measure CNC Spindle Runout Before Replacing the Collet when the project reaches the neighboring decision.
Changing cutter diameter or flute count
A different flute count changes the feed-RPM-chip-load relationship immediately. Diameter can change strength, internal-corner access, cutting force and recommended surface speed. Recalculate, then test. Do not preserve both feed and RPM merely because the toolpath looks similar.
Changing material or stock support
A new material batch can create different chips or finish. Thin sheet may vibrate even when the same material behaved well as thick stock. A change from perimeter clamps to tape-and-adhesive workholding changes how aggressively the part can be loaded. Reconfirm the setup on scrap.
Changing engagement or toolpath geometry
Moving from a light profile to full-width slotting raises tool engagement. Deeper pockets add chip evacuation problems. Tight corners may slow the machine. Treat each as a new operating condition, then approve it separately.
The goal is not the fastest possible cut. It is a stable, explainable process that meets the finish requirement without exceeding the machine, cutter, material or workholding limits. That method produces a settings library with real value: each entry states not only what worked, but why it was accepted and where it stops applying.
Questions About Build a CNC Feeds-and-Speeds Experiment Log You Can Reuse
What is the basic CNC chip load formula?
Chip load equals feed rate divided by RPM multiplied by the number of flutes: feed ÷ (RPM × flutes). Keep the units consistent and remember that the formula does not account for depth, radial engagement, rigidity or chip evacuation.
Should I change feed rate or spindle speed first?
Begin from verified cutter and material guidance, then change one variable at a time. The first variable depends on the observed problem and machine range; there is no universal order that replaces inspection.
Does more RPM always improve CNC finish?
No. Raising RPM while feed stays constant reduces calculated chip load and can increase rubbing or heat. Finish also depends on cutter geometry, sharpness, runout, engagement, workholding and material.
Can I copy feeds and speeds from another CNC router?
Use another record only as background evidence when the machine, spindle, cutter, material and operation match. Otherwise, return to manufacturer data and run a controlled scrap test.
Why does my CNC make dust instead of chips?
Possible causes include a thin effective chip, a dull or unsuitable cutter, recutting, poor evacuation or the material's natural behavior. Inspect the cutter and setup before changing parameters, then test one change at a time.
