CNC precision SMI story: how 5-axis machining holds tolerances

• 1
  What the CNC precision SMI story is really about
• 2
  How simultaneous 5-axis motion reduces error
• 3
  When 5-axis helps and when it does not
• 4
  Datum strategy and inspection on a 5-axis part
• 5
  Design choices that keep CNC precision predictable
• 6
  What the shop floor has to support this
• 7
  Step by step: from model to first article
• 8
  Setup strategy by part geometry

The CNC precision SMI story is not a marketing anecdote. It is the pattern behind most tight-tolerance work: a part arrives with features on several faces, the tolerance stack leaves almost no room, and the geometry cannot be reached from three directions. The fix is almost never a better cutter. It is a different setup strategy.

Three-axis machining moves the tool in X, Y, and Z. That covers a large share of prismatic parts. The limit shows up when a feature sits on an angled face, under a lip, or inside a pocket with a curved wall. The part then needs a second or third fixturing, and every refixture adds position error.

Five-axis machining adds two rotary axes, usually A and C. The tool can tilt and the table can index, so the same feature can be reached without loosening the part. Fewer setups means fewer datum transfers, and datum transfers are where tolerance disappears.

This page is an explainer. It covers the mechanism, the boundary conditions, and the trade-offs. If your part is flat and open, 3-axis is cheaper and just as accurate. If it is not, the rest of this page applies.

• 1
  Setup count drives accuracyEach refixture adds position error to the stack.
• 2
  Rotary axes change reachTilting the tool reaches faces 3-axis cannot.
• 3
  Not every part needs itOpen prismatic parts run better on 3-axis.

In simultaneous 5-axis work, the linear and rotary axes move at the same time. The controller keeps the tool tip on the programmed path while the tool axis tilts. The cutting edge stays in contact with the surface at a consistent angle instead of drifting as the tool follows a curve.

That consistent angle matters most on sculpted surfaces. With a ball nose cutter on a 3-axis machine, the effective cutting speed drops to nearly zero at the tool tip. The center of the tool rubs. Surface finish suffers, and tool life shortens. Tilting the tool by 10–20° moves the contact point off center and keeps the cutting speed usable.

Shorter tools are another gain. A tilted tool holder can reach a deep feature with a stubby cutter instead of a long slender one. Deflection scales roughly with the cube of the length-to-diameter ratio, so a tool half as long is far stiffer. That is often the difference between holding ±0.005 mm and not.

The cost is programming and machine time. Simultaneous motion is slower than indexed motion, and the post-processor must handle rotary kinematics correctly. On a simple part, this extra effort buys nothing.

• 1
  Tool tilt 10–20°Keeps contact off the ball nose center.
• 2
  Shorter tool, less deflectionDeflection rises with the cube of L/D.
• 3
  Indexed is fasterUse 3+2 when the geometry allows.

Use simultaneous 5-axis when features on three or more faces must hold a tight relationship to each other, when undercuts or angled holes cannot be reached from a single direction, or when a sculpted surface needs a controlled tool angle. Impellers, housings with angled ports, and medical instruments with curved channels are typical.

Stay on 3-axis when the part is prismatic and open. A bracket, a plate, or a manifold with all features on two faces will run faster and cheaper on a 3-axis machine, often with the same tolerance. Adding rotary motion adds setup risk without adding capability.

3+2 positioning sits in between. The rotary axes index to a fixed angle, lock, and then cut like a 3-axis job. It reaches five faces in one fixturing and is usually faster than simultaneous motion. Choose it when the geometry is reachable but the part needs several angles.

Material matters too. Titanium and Inconel push cutting temperatures up, so tool tilt and coolant delivery need attention. Aluminium is forgiving. Plastics can be cut on 5-axis, but thermal growth and fixturing force usually dominate the tolerance budget before kinematics do.

• 1
  Choose simultaneous 5-axisMulti-face datums, undercuts, sculpted surfaces.
• 2
  Choose 3+2 indexingSeveral reachable angles, speed matters.
• 3
  Stay 3-axisOpen prismatic parts, two-face features.

Pick the datum before the toolpath. On a 5-axis part, the natural choice is usually a face that stays untouched through the whole cycle, plus two features that set rotation. If the datum face is machined in the first operation and then used to locate the second, the error compounds.

Probing helps. On-machine probing locates the raw stock and rotates the work coordinate system to match the actual part, not the nominal one. That absorbs casting or forging variation without adding a manual setup. It adds cycle time, so use it where stock variation is real.

Inspection should mirror the machining setup. If a position callout is measured from datum A in the drawing, it must be measured from datum A after machining. Re-datuming during inspection hides real error and creates arguments later.

We inspect 100% of parts before shipment, with raw material checks, in-process monitoring, and a final report on request. For a 5-axis part, the final report should show the datums used, not just the numbers.

• 1
  Freeze the datumChoose a face that is never remachined.
• 2
  Probe the stockAbsorb casting variation in the work offset.
• 3
  Inspect from the same datumOtherwise the report hides real deviation.

Keep wall thickness above 0.8 mm on aluminium and 1.0 mm on stainless unless there is a reason. Thin walls vibrate, and vibration shows up as chatter marks and dimensional drift. A slightly thicker wall often costs less than the extra finishing passes needed to rescue a thin one.

Radius the internal corners. A cutter has a diameter, so a sharp internal corner cannot be machined without a separate EDM step. If the corner radius matches a standard cutter, the part can be finished in one setup. This is one of the cheapest changes a designer can make.

Avoid deep, narrow slots when a wider one will do. A slot five times deeper than it is wide forces a slender tool, which deflects and chatters. Widening the slot or splitting it into two passes keeps the tool stiff and the tolerance stable.

Tolerances should be assigned where they function. Marking every dimension ±0.005 mm does not improve the part. It raises cost and slows inspection, and it hides which features actually matter. Mark the functional ones and let the rest sit at general tolerance.

• 1
  Wall thickness0.8 mm aluminium, 1.0 mm stainless minimum.
• 2
  Corner radiiMatch a standard cutter diameter.
• 3
  Slot depthKeep under 5× width where possible.

Machine capability sets the ceiling. We run 127 high-precision CNC machines across three wholly-owned plants covering 7,600 m², with 150 technicians. That includes 16 simultaneous 5-axis machining centers, 12 four-axis mills, 27 three-axis machines, and 16 mill-turn centers.

Size range runs from compact work at 500 × 310 × 200 mm up to 4,000 × 400 × 150 mm travel, with a Ø400 mm rotary table for indexed work. The largest processing envelope is 4,000 mm. Parts outside that envelope go to a partner or get split into assemblies.

Materials are routine rather than exotic: 6061, 7075, 304, 316L, 17-4PH, 4140, Ti-6Al-4V, Inconel, and engineering plastics up to PEEK. Finishing is in-house, from anodizing and electroless nickel to bead blasting and laser marking.

Quality management is certified to ISO 9001:2015, IATF 16949:2016, ISO 13485:2016, and ISO 27001:2022. Historical late-delivery probability is below 2%. No minimum order quantity, from one prototype to 10,000+ part runs.

• 1
  5-axis capacity16 simultaneous 5-axis centers.
• 2
  Size envelopeUp to 4,000 mm maximum processing size.
• 3
  Finishing in-houseAnodizing, plating, blasting, laser marking.