CNC Knowledge: How Cutting, Tolerance and Fixturing Really Work

• 1
  What CNC knowledge starts with: the cut itself
• 2
  Where tolerance is lost before the tool touches metal
• 3
  Why surface finish is a cutting outcome, not a polishing step
• 4
  How material choice changes what the machine can do
• 5
  When 3-axis is enough and when you need 5-axis
• 6
  Process fit by part characteristic

Every machining decision traces back to one event. A hard edge pushes into metal and shears it away. The chip forms in a narrow zone where pressure and heat peak in microseconds. Tool geometry, feed rate, and spindle speed decide whether that zone stays stable or turns into chatter and built-up edge.

Heat is the first constraint. Roughly 80% of cutting heat leaves with the chip in aluminum, so high feed rates can actually cool the cut. Titanium behaves the opposite way: heat lingers near the edge because the chip conducts poorly. That is why Ti-6Al-4V runs at 40–60 m/min surface speed while 6061 aluminum runs at 300–500 m/min.

Cutting force scales with chip cross-section, not with spindle speed. Doubling the feed per tooth doubles the load on the edge. Raising rpm at the same feed just spreads that heat thinner. Understanding this lets you diagnose tool breakage: look at the chip load first, then the speed.

The practical takeaway is simple. Match the cutting parameters to the material's thermal behavior, not to a generic chart. A recipe that works on 6061 will burn a 316 stainless tool in minutes.

• 1
  Chip load governs forceFeed per tooth × teeth × rpm sets the load on the cutting edge.
• 2
  Speed governs heatSurface speed controls how fast heat builds at the contact point.
• 3
  Material sets the windowAluminum tolerates high speed; titanium and stainless do not.

A drawing that calls ±0.005 mm is asking for more than a good machine. It is asking for a complete system: rigid fixturing, stable thermal conditions, and a probing routine that confirms the datum before the first pass. Most out-of-tolerance parts fail because the setup moved, not because the machine drifted.

Fixturing contributes the largest single error. A part held in a vise with 0.02 mm of jaw lift can shift during a heavy roughing pass. The finishing pass then follows that shifted datum. Soft jaws bored in place eliminate most of this. For thin walls, add support or reduce radial engagement to 5–8% of the cutter diameter.

Thermal drift matters on long cycles. A spindle running for two hours grows a few microns. On a ±0.005 mm feature, that is significant. Shops that hold tight tolerance either run finishing operations early in the shift or use in-process probing to re-zero between passes.

The machine itself is rarely the weak link. A modern 5-axis center with a Ø400 mm rotary table can position within a few microns. The question is whether the part, the fixture, and the toolpath respect that capability.

• 1
  Bore soft jaws in placeRemoves jaw lift and gives a true clamping surface.
• 2
  Probe before finishingRe-establish the datum after roughing to catch shift.
• 3
  Control the temperatureLet the spindle and part reach steady state before tight work.

Ra values come from the toolpath geometry and the edge condition, not from what happens after the part leaves the machine. A Ra 0.8–1.6 μm finish is achievable directly from a sharp carbide cutter with a corner radius and a controlled stepover. Going finer than Ra 0.2–0.8 μm usually requires a dedicated finishing pass with a small nose radius and reduced feed.

The theoretical roughness formula is Ra ≈ f² / (32 × r), where f is feed per revolution and r is the tool nose radius. Doubling the nose radius halves the roughness. Doubling the feed quadruples it. This is why a 0.8 mm corner radius at 0.1 mm/rev produces a better finish than a 0.4 mm radius at 0.15 mm/rev.

Real surfaces rarely match the formula. Tool wear, chatter, and material inclusions add roughness beyond the geometric value. A dull edge smears instead of shears, leaving a torn surface. On stainless and titanium, that smear is the main reason a finish fails inspection.

Specify finish by function. A sealing face needs a specific Ra and a flatness callout. A cosmetic panel needs an even texture that hides tool marks. A bearing bore needs roundness, not just low Ra. Chasing a low number without the right callout often adds cost and changes nothing.

• 1
  Nose radius is the leverLarger radius lowers Ra at the same feed rate.
• 2
  Feed drives roughnessCut feed per revolution to hit the target Ra.
• 3
  Wear ruins finishReplace or index the edge before the finishing pass.

Aluminum is the easy case. Alloys like 6061, 7075, and 2024 cut fast, hold tolerance well, and tolerate light fixturing. A 6061 bracket with ±0.05 mm features runs comfortably on a 3-axis mill. Switch to 7075 and the same part machines cleanly but demands sharper tools because the alloy is more abrasive.

Stainless grades split into two groups. The 303 and 304 family machines with free-cutting behavior, though 304 work-hardens if the tool rubs. The 316 and 17-4PH grades work-harden aggressively and need constant feed to stay under the hardened layer. Stopping mid-cut on 17-4PH can harden the surface enough to break the next tool.

Titanium and Inconel are the hard cases. Ti-6Al-4V has low thermal conductivity and high chemical reactivity. It welds to the cutting edge under heat. Inconel work-hardens even faster and generates extreme edge temperatures. Both require rigid setups, low surface speeds, and generous coolant. Neither is a candidate for a quick prototype on a light machine.

Plastics behave differently again. POM and PEEK machine cleanly with sharp tools and high speed. ABS and PC soften under heat and need air blast or chilled coolant to avoid melting. The cutting strategy changes with the material, not just the feed and speed numbers.

• 1
  Aluminum forgivesFast speeds, light fixturing, stable results.
• 2
  Stainless punishes rubbingKeep the edge cutting; never dwell in the cut.
• 3
  Titanium demands rigidityLow speed, high feed, flood coolant, stiff setup.

Axis count is a fixturing decision more than a technology decision. A part with features on three faces can run on a 3-axis mill with two setups. That is often the fastest and cheapest route. The same part on a 5-axis machine runs in one setup, but the cycle time may be longer because the machine moves more slowly through complex motion.

The real gain from 5-axis comes when a feature is unreachable or when setup error is the dominant tolerance risk. A hydraulic manifold with angled ports on five faces cannot be machined on a 3-axis machine without multiple fixtures. Each fixture adds error. One 5-axis setup removes that stack-up entirely.

Simultaneous 5-axis is different from 3+2 positioning. In 3+2, the rotary axes index and lock, then the cut runs like a 3-axis pass. In simultaneous mode, all five axes move together to keep the tool normal to a curved surface. That is necessary for impellers, blisks, and complex contoured surfaces. It is overkill for a flat bracket.

Use the simplest machine that reaches every feature. A 27-machine 3-axis fleet handles most brackets, plates, and housings. Save the 16 simultaneous 5-axis centers for parts where reach or setup count drives the cost.

• 1
  Count the setupsEach extra setup adds error and lead time.
• 2
  3+2 for reachIndexed rotary axes handle angled features without simultaneous motion.
• 3
  Simultaneous for contoursUse it when the tool must stay normal to a curved surface.