CNC Basics: Start Here

• 1
  How a CNC machine turns a CAD file into a cut part
• 2
  What 3-axis, 4-axis, and 5-axis actually buy you
• 3
  Milling vs turning: which process suits the part
• 4
  What the material does to your tolerances and finish
• 5
  Reading a tolerance callout and a surface finish callout
• 6
  Which process and axis count fits your part

CNC machining is subtractive. You start with a solid block, plate, or bar and remove material until the remaining shape matches the drawing. There is no mold and no forming die. The cutting tool follows a path defined in software, and the machine's servos hold that path to within microns. That is the whole idea.

The chain has four links. A designer builds a 3D model in CAD. A CAM programmer picks tools, stock, and workholding, then generates a toolpath. The CAM software writes that path as G-code, a list of coordinates, feed rates, and spindle speeds. The machine controller reads the G-code and moves the axes. G-code is the language; CAM is the translator.

G-code is simpler than most people expect. A line like G01 X25.4 Y10.0 F250 tells the controller to move in a straight line to a coordinate at 250 mm/min. M-codes handle non-motion tasks: spindle on, coolant on, tool change. Everything a machine does for a full day reduces to a few thousand of these lines.

Once the controller has the program, accuracy comes from three places: the rigidity of the machine, the sharpness and geometry of the tool, and how well the part is held. Software does not create accuracy. It only stops the machine from wasting the accuracy it already has.

A 3-axis mill moves X, Y, and Z. The tool always approaches from one direction, so every feature must be reachable from that direction. Simple plates, brackets, and housings cut cleanly on 3 axes and cost the least. If your part has holes on four side faces, a 3-axis machine needs four separate setups, and each setup adds a chance for position error.

A 4-axis machine adds rotation around one axis, usually A. The part spins while the tool cuts, so you can machine a cylinder's full circumference or drill a ring of holes without re-fixturing. Shafts, couplings, and cam profiles fit here. Setup count drops and concentricity improves, because the part never leaves the chuck or fixture between features.

A 5-axis machine adds a second rotary axis, so the tool can tilt relative to the part. This is what makes undercuts, deep pockets with curved floors, and impeller blades possible in one setup. It also lets a short, stiff tool reach deep features by tilting instead of hanging out. Fewer setups, tighter true position, and shorter cycle times on complex geometry. The trade-off is programming time and hourly rate.

Pick the lowest axis count that reaches every feature. Extra axes only pay off when they remove a setup or shorten an unreachable tool.

• 1
  3-axisFlat parts, one approach direction, lowest cost per part.
• 2
  4-axisCylindrical and rotary parts; features around a single axis.
• 3
  5-axisUndercuts and contoured surfaces; one setup for five faces.

In milling, the tool spins and the workpiece stays still. In turning, the workpiece spins on a lathe and a single-point tool moves along it. The decision is geometric. If the part is mostly a body of revolution, meaning round with features around a centerline, turning is faster and holds diameter tolerance more easily. If the part is prismatic, with flat faces and pockets, milling wins.

Many real parts are both. A hydraulic manifold might start as a round blank, get turned to diameter, then get milled ports and a flat mounting face. Mill-turn centers handle this without moving the part between machines, which protects the relationship between the bore and the mounting face. A turned bore and a milled face that must be perpendicular are far easier to hold on one machine.

Cycle time separates the two more than accuracy does. Turning removes material continuously with one engaged edge. Milling removes it in passes with a tool that enters and exits. For a high-volume round part, turning can be several times faster. For a low-volume prismatic part, milling is the only practical route.

One rule holds in both: keep the tool as short and as large as the feature allows. Deflection scales with the cube of tool length. A long thin tool will chatter no matter how good the machine is.

The same drawing produces different results in different metals. Aluminum 6061 and 7075 cut freely, hold ±0.005 mm on a rigid machine, and take a fine finish. They also move under heat and can distort in thin walls, because the material is soft and springy. Climb milling and light finishing passes reduce that.

Stainless 304 work-hardens. If the tool rubs instead of cutting, the surface gets harder and the next pass cuts worse. Sharp tools, positive rake, and steady feed avoid this. Stainless 316L behaves similarly and adds galling risk, so coolant and feed matter. Titanium Ti-6Al-4V is worse: low thermal conductivity sends heat into the tool edge, so speeds drop and tool life shortens.

Plastics are a different problem. POM and ABS cut easily but melt if the tool dwells. PEEK holds dimension but is abrasive and costly. Carbon fiber reinforced plastic eats carbide edges, so diamond-coated tools are common. Each material has a window of speed and feed, and the window is narrowest for titanium and the hardest tool steels.

The practical takeaway: choose the material for the function first, then let the machinist pick the cutting parameters. Do not pick a material for its machinability unless the function allows it.

• 1
  AluminumFast to cut, good finish, watch thin-wall distortion.
• 2
  StainlessWork-hardens; needs sharp tools and steady feed.
• 3
  TitaniumHeat stays at the edge; slower speeds, shorter tool life.
• 4
  PlasticsMelt risk at low feed; abrasive grades wear tooling.

A tolerance is a band, not a target. If a drawing says ±0.05 mm on a length, the machinist aims for the middle of the band, because the middle gives the most room before either limit is breached. General tolerances on a title block often cover everything not dimensioned, and they are usually looser than the critical features. Read the title block before you assume a number.

Surface finish is measured as Ra, the arithmetic mean roughness. Ra 1.6–3.2 μm is a normal as-machined surface. Ra 0.8–1.6 μm needs a finishing pass and a sharp tool. Ra 0.2–0.8 μm usually means a smaller stepover, a slower feed, or a secondary operation such as lapping or polishing. Calling out a fine finish everywhere raises cost without adding function.

Tighter tolerance and finer finish interact. You cannot hold ±0.005 mm on a surface that is chattering, because the surface error and the dimensional error come from the same vibration. Fix the setup before chasing the dimension. A rigid fixture and a short tool solve both problems at once.

Say where the tolerance matters. Datum features, bores that receive bearings, and sealing faces deserve tight control. Clearance holes and cosmetic edges usually do not. Selective tolerancing keeps the part affordable and the critical features correct.