CNC Terminology Explains: What the Words on a Drawing Actually Mean

• 1
  Key takeaways
• 2
  CNC terminology explains the subtractive process first
• 3
  Tolerance, allowance and the cost of a tight number
• 4
  G-code, CAM and what the machine actually reads
• 5
  Spindle speed, feed rate and surface speed
• 6
  3-axis, 4-axis and 5-axis in plain terms
• 7
  Surface finish, Ra and the difference between cut and look
• 8
  Which term actually changes your part

Before any term makes sense, fix the process in your head. A CNC machine holds a cutting tool and moves it through metal, plastic or composite. The part is what is left after the tool passes. Nothing is added, so every feature has to be reachable by a tool that has a diameter, a length and a stiffness limit.

That single idea, reachability, sits behind most of the words on a drawing. A pocket depth, a corner radius, an undercut and a thread all describe geometry that a specific tool must enter. When a design ignores tool reach, the drawing looks fine and the quotation comes back with a note.

Subtractive work also means the stock has to be held. A vise, a chuck, a fixture plate or a vacuum table clamps the blank while the tool cuts. Every clamp point is a place the tool cannot go in that setup, which is why the same part may need two or three orientations.

So when a term appears on a drawing, ask three questions. Does it change the shape? Does it change how the part is held? Does it change how the feature is measured? If the answer is no to all three, the term is descriptive rather than controlling.

• 1
  ShapeDoes the word change a dimension, a radius or a thread?
• 2
  HoldingDoes it force a new setup, a new fixture or a longer tool?
• 3
  MeasuringDoes it change how an inspector accepts or rejects the feature?

Tolerance is the total width of the acceptable zone around a nominal size. A shaft called out at Ø20.00 ±0.05 mm may measure anywhere from 19.95 mm to 20.05 mm and still pass. The number on the right of the ± sign is not a target, it is a limit.

Allowance is different. It describes the intentional gap or interference between two mating parts, such as a pin and a bore. Two parts can each sit inside their own tolerance and still produce the wrong fit if the allowance was never defined. This is where assemblies fail on the bench even though every individual part passed inspection.

Cost responds to tolerance non-linearly. Going from ±0.10 mm to ±0.05 mm usually costs nothing extra, because the machine can already hold it. Going from ±0.05 mm to ±0.005 mm changes the plan: a separate finishing pass, a temperature-stable measurement, and often a CMM report. Our shop holds ±0.005 mm (±0.0002 in) when the drawing asks for it, but we would rather see that callout only on the two or three features that actually need it.

A practical habit: tolerance the fit, not the whole part. Give the functional surfaces a tight number, leave cosmetic and clearance surfaces open, and the quote comes back lower without any loss of function.

• 1
  ±0.10 mmGeneral machining, no special plan, fastest route.
• 2
  ±0.05 mmStandard for most mating features, still routine.
• 3
  ±0.005 mmReserved for true fits, adds finishing and inspection time.

G-code is the instruction list a machine controller executes. A line such as G01 X50.0 Y20.0 F250 tells the controller to move in a straight line to that point at a feed of 250 mm/min. G00 is a rapid positioning move at full speed, G02 and G03 cut arcs, and M-codes handle non-motion actions like starting the spindle or opening coolant.

Nobody types this by hand for a complex part. CAM software reads the 3D model, lets a programmer choose tools and strategies, and posts the G-code for a specific machine and controller. The post-processor matters here. The same tool path posted for two different controllers can produce different motion, and that is a common source of first-article surprises.

The engineering meaning is simple. G-code is a consequence, not a design input. You do not need to write it, but you should know that the tool path inside it decides how the surface is generated, how much material each pass removes, and how long the cycle runs.

When a supplier says a feature is difficult, they usually mean the CAM strategy for it is slow, risky, or needs a tool that is long and thin. That is a geometry problem you can often fix in the model before the first chip is cut.

• 1
  G00Rapid move, no cutting, used between features.
• 2
  G01Linear cut at the programmed feed rate.
• 3
  G02 / G03Clockwise and counterclockwise arc cuts.
• 4
  M-codesSpindle, coolant and tool-change commands.

Spindle speed is how fast the tool rotates, in rpm. Feed rate is how fast the tool or the workpiece advances, in mm/min. They are linked by a third number: surface speed, the speed of the cutting edge against the material, in m/min. Surface speed is the number that actually governs tool life.

A Ø10 mm end mill at 8,000 rpm runs a surface speed near 250 m/min. A Ø50 mm face mill at the same 8,000 rpm would run near 1,250 m/min, far beyond what most steels tolerate. This is why large tools turn slowly and small tools turn fast, and why the same spindle speed means different things at different diameters.

Feed rate per tooth, often called chip load, decides whether the edge cuts or rubs. A four-flute cutter at 0.05 mm per tooth and 3,000 rpm feeds at 600 mm/min. Drop that chip load too far and the tool rubs, heat builds, and the finish smears. Push it too far and the tool deflects, leaving chatter marks and a dimension that drifts.

For the person reading a quote, these three numbers explain why two suppliers give different cycle times. There is a safe band for each material and tool, and the shop that runs near the top of that band removes metal faster without damaging the part.

• 1
  Surface speedCutting edge speed against material, controls tool wear.
• 2
  Chip loadFeed per tooth, controls whether the edge cuts cleanly.
• 3
  RubbingToo little chip load, generates heat and poor finish.

Axis count describes how many directions the tool and the workpiece can move relative to each other under numerical control. A 3-axis mill moves X, Y and Z. The tool always points straight down, so every face you machine has to be reachable from above.

A 4-axis machine adds rotation around one axis, usually a rotary table. A shaft with flats, holes and slots around its circumference can then be cut in one setup, because the table indexes the part to each angular position. Our rotary tables run to Ø400 mm, which covers most cylindrical work of that size.

A 5-axis machine adds a second rotary axis. The tool can tilt, so it approaches angled faces, deep cavities and contoured surfaces without the part being re-fixtured. That matters most for aerospace brackets, medical housings and impellers, where one part may have features on five different faces.

The trade-off is not that 5-axis is better. It is that 5-axis removes setups, and setups are where position error accumulates. If a part has three faces to machine, one 5-axis setup can replace three 3-axis setups and the re-fixturing between them. If a part is a flat plate with holes, 3-axis is faster and cheaper.

• 1
  3-axisPrismatic parts, flat plates, single-face features.
• 2
  4-axisCylindrical parts with features around the axis.
• 3
  5-axisAngled faces, deep cavities, contoured surfaces.

Surface roughness is written as Ra, the arithmetic mean deviation of the profile, in micrometres. It measures the average height of the tiny peaks and valleys left by the cutting edge. Lower Ra means a smoother surface, and the number is measured, not judged by eye.

The scale matters. As-machined surfaces sit around Ra 1.6–3.2 μm. A good finish pass reaches Ra 0.8–1.6 μm. Fine finishing for seals, sliding fits and optical surfaces goes to Ra 0.2–0.8 μm, and that step usually needs a smaller step-over, a sharper tool and more time.

Ra is not the same as appearance. A bead-blasted or anodized surface can look uniform while its Ra is unchanged or slightly higher. If a surface has to seal, slide or mate, control Ra. If it only has to look consistent, a finishing operation after machining is the cheaper route.

Anisotropy is the detail that catches people. Milling leaves directional marks, so roughness measured across the feed direction differs from roughness measured along it. When a drawing quotes a single Ra value, ask which direction it was measured in before you accept or reject a lot.

• 1
  Ra 1.6–3.2 μmAs-machined, general surfaces, no extra pass.
• 2
  Ra 0.8–1.6 μmFinished surfaces, mating faces, visible parts.
• 3
  Ra 0.2–0.8 μmSealing and sliding surfaces, adds cycle time.