Is It Hard to CNC Small Parts Like This?

• 1
  Why Is It Hard to CNC Small Parts? The Real Limits
• 2
  Tool Runout and Toolholding Decide the Outcome
• 3
  Fixturing Small Parts Without Crushing Them
• 4
  When Small Parts Are Easy, and When They Are Not
• 5
  Metrology: How You Know the Small Part Is Right
• 6
  Small Part Machinability by Feature and Material

Size is not the problem. Scale is. When you shrink a part, every force and every error stays roughly the same while the material resisting it shrinks fast. A 2 mm end mill cutting 6061 aluminium sees the same cutting pressure as a 12 mm cutter, but the tool body is 36 times weaker in bending. That is the whole story in one sentence.

Deflection follows the cube of length and the fourth power of diameter. A tool hanging 30 mm out of a 3 mm collet bends 256 times more than the same tool at 15 mm out of a 6 mm holder. So the first question we ask on a small part is not about the part. It is about how far the tool must reach to get there.

Heat matters too. A small part has little mass to absorb cutting heat and a large surface area to lose it. Wall sections under 0.5 mm move after machining as the part equalises with the room. If you measure straight off the machine, you are reading a temperature, not a dimension.

• 1
  Stiffness scales down faster than forceTool deflection rises as reach cubed and as diameter to the fourth power.
• 2
  Thermal mass is smallThin walls move after cooling, so let the part settle before final inspection.
• 3
  Burrs are proportionally largeA 0.05 mm burr is nothing on a 200 mm block and fatal on a 1 mm feature.
• 4
  Handling and metrology become the bottleneckMachine time is often less than the time to fixture and measure.

On a small part, runout is not a detail. A two-flute 1 mm end mill running at 0.02 mm runout is effectively cutting with one flute. The chip load per tooth doubles, tool life drops by more than half, and the finished wall carries a chatter pattern you cannot polish out. Keep runout under 0.005 mm for tools under 3 mm.

The holder matters more than the spindle on these jobs. A shrink-fit holder or a high-precision collet chuck with a low-runout nut will beat a general-purpose holder every time. We check runout with a dial indicator on the tool shank, not on the holder face, because the shank is what actually cuts.

Cutting parameters should be conservative on depth, aggressive on speed. For a 1 mm carbide end mill in aluminium, we typically run 0.05–0.1 mm radial engagement and 0.2–0.5 mm axial depth at 12,000–20,000 rpm. Many shops slow the spindle down instead. That is backwards. Slow speeds increase the chip load and snap the tool.

• 1
  Measure runout at the shankTarget under 0.005 mm for tools below 3 mm diameter.
• 2
  Prefer shrink-fit or precision colletsA good holder removes more error than a spindle upgrade.
• 3
  Keep chip load constantRaising rpm and lowering feed per tooth protects small tools.
• 4
  Use air blast over flood coolantCoolant pressure can bend a 0.5 mm tool; air clears chips without force.

Workholding is where most small-part jobs fail. A vise strong enough to hold a 5 mm part will also deform it. Clamping force on a thin section can exceed the cutting force by an order of magnitude, so the part springs back when you release it and the dimension is wrong before you measure.

The usual fix is to support more of the part and clamp less of it. Soft jaws machined to the part profile, vacuum chucks for flat plates, and sacrificial tabs that hold the part in the stock are all standard. For a run of 500 small brackets, we often machine them in a plate and cut the tabs at the end. One setup, no re-clamping error.

For parts under 10 mm, a five-axis machine with a Ø400 mm rotary table earns its cost. You can reach five faces in one setup, so the part is clamped once. Each re-clamp adds a locating error of 0.01–0.03 mm on a small part. Removing three setups removes three chances to lose tolerance.

• 1
  Support the part, do not squeeze itSoft jaws or vacuum beat a hard vise on thin sections.
• 2
  Machine in the plate, cut tabs lastKeeps every part at the same zero point for the whole run.
• 3
  One setup beats four accurate setupsEach re-clamp adds 0.01–0.03 mm of locating error.
• 4
  Avoid clamping on a finished faceClamp on stock and machine the datum last.

A small part is easy when it is a simple prismatic shape in a free-machining material. A 6 mm aluminium spacer with two holes and a flat face is a 3-axis job. It runs in minutes, holds ±0.005 mm without drama, and needs no special fixture. Most small parts fall in this group, and most of them are not hard at all.

It gets hard when four things stack up: a feature smaller than 1 mm, a wall thinner than 0.5 mm, a tolerance tighter than ±0.005 mm, and a material that work-hardens or galls. Any one of these is manageable. Three together mean the job needs a written process plan, a first article inspection, and a real conversation about cost.

Materials change the answer. 6061-T6 aluminium and brass C36000 cut cleanly at small sizes. 316L stainless work-hardens if the tool rubs, so you must keep the cutter engaged and never dwell. Titanium Ti-6Al-4V and Inconel generate heat right at the cutting edge and need lower surface speed. PEEK and POM move with temperature, so measure them at 20 °C.

• 1
  EasyPrismatic shape, 6061 or brass, features over 1 mm, tolerance ±0.05 mm.
• 2
  ModerateOne tight tolerance or one thin wall, standard material.
• 3
  HardSub-millimetre features plus thin walls plus tight tolerance.
• 4
  Very hardAll of the above in titanium or Inconel, with a surface finish callout.

You cannot hold what you cannot measure. On a small part, a caliper is often the wrong tool. The jaw contact force alone can compress a 0.3 mm wall by more than the tolerance. For anything under 5 mm, use a vision system, a micrometer with a friction thimble, or a CMM with a small stylus and low probing force.

Temperature is the second trap. Steel grows about 11 μm per metre per °C. On a 10 mm feature that is 0.11 μm per °C, which sounds small until you compare it to a ±5 μm tolerance and a shop that swings 4 °C between morning and afternoon. We inspect in a controlled room and let parts settle before final measurement.

First article inspection is standard practice on small-part runs, not a premium option. We check the drawing, the datum scheme, and every callout on the first part, then lock the process. Reports are available on request. If the drawing has a datum that cannot be reached with a probe, we flag it before cutting, not after.

• 1
  Skip the caliper under 5 mmJaw force can exceed the tolerance on thin walls.
• 2
  Let parts reach 20 °CSteel moves about 11 μm per metre per °C.
• 3
  Lock the process after first articleChange nothing without re-verifying the first part.
• 4
  Check datums before quotingAn unreachable datum is a design problem, not a machining problem.