CNC Machining Bulk Small Parts: How the Process Actually Scales

• 1
  Why CNC Machining Bulk Small Parts Is a Different Problem
• 2
  Tolerance Drift: What Changes Between Part 1 and Part 5,000
• 3
  Fixture Design for High-Volume Small Parts
• 4
  When a Second Operation Costs More Than It Saves
• 5
  Choosing a Machine and Fixture Strategy by Part and Volume

A single small bracket takes 6 minutes of spindle time and 20 minutes of setup. At 5,000 pieces the same part takes 6 minutes of spindle time and 20 seconds of setup per part. Setup stops being the cost; everything that repeats 5,000 times becomes the cost.

That shift moves the engineering focus. On one-off work we optimize for the shortest path to the first good part. On CNC machining bulk small parts we optimize for whatever repeats: load time, chip evacuation, tool changes, and the time a probe needs to confirm a feature. A 2-second saving per part is roughly 2.8 hours of machine time across 5,000 pieces.

Geometry sets the ceiling. A part that fits inside a 100 × 100 × 50 mm envelope can often be nested several per cycle on a 500 × 500 × 450 mm table. A part that needs the Ø400 mm rotary table for a single face cannot be nested at all, and the cycle time stays flat no matter the quantity.

So the first question is not how many parts, it is how many parts per cycle. Once you know the nest count, the rest of the cost model falls into place: tool changes per part, deburr handling, and whether the part can run lights-out overnight.

• 1
  Nest count drives everythingParts per cycle, not batch size, sets the real unit cost.
• 2
  Setup amortizes fastAfter a few hundred pieces, fixturing is a rounding error.
• 3
  Small envelope, more freedomUnder 100 mm square, multi-cavity fixtures pay off quickly.

A first article measured at ±0.005 mm does not guarantee the last article holds ±0.005 mm. Thermal growth is the usual cause. A spindle running for 6 hours at 12,000 rpm warms the casting and the tool holder, and a 3 °C rise on an aluminium part 80 mm long moves about 0.005 mm on its own.

Tool wear is the second cause. On a Ø6 mm carbide end mill cutting 6061, flank wear of 0.02 mm is normal after a few hundred parts. That wear shows up as a size trend, not a sudden jump, which is why in-process probing every 50 to 100 parts catches it long before the parts go out of tolerance.

Material matters here. Aluminium 6061-T6 cuts clean and holds size well. Stainless 316L and 17-4PH work-harden, push cutting forces up, and warm the part faster. On those grades we shorten the probing interval and often add a roughing pass followed by a stress-relief dwell before finishing.

Parts with thin walls under 1 mm are the hardest case. Clamping force alone can distort them, so the fixture has to support the wall from the inside or the cut has to be a light finishing pass with near-zero radial engagement.

• 1
  Thermal growthA 3 °C rise on an 80 mm aluminium part is about 0.005 mm.
• 2
  Tool wear trend0.02 mm flank wear appears as a slow size drift, not a jump.
• 3
  Probe intervalEvery 50–100 parts on aluminium, tighter on stainless.

A vise holds one part. A soft-jaw pocket holds six. A dedicated plate with 20 pockets, a common clamp bar, and one air cylinder holds 20. The trade-off is setup time against cycle time: a 20-pocket plate may take 90 minutes to build and dial in, but it can cut per-part load time from 15 seconds to 3.

Pocket depth matters more than people expect. If the pocket is 0.05 mm deeper than the part, chips collect underneath and the part sits high on the second cycle. If it is 0.02 mm shallower, the part rocks. We usually target a pocket depth that leaves 0.5 mm of the part above the jaw face so the probe can still touch the top surface.

For parts under 20 mm, vacuum plates and adhesive film fixtures remove clamp marks entirely. They work well on flat plates and thin covers, less well on parts with holes through the middle or with heavy side milling forces, where the part can slide.

Loading ergonomics is not a soft topic. If the operator has to pick up a part with tweezers and index it by eye, the load time will be 20 seconds, not 5. Chamfered pocket edges and a visible orientation key cut that time without touching the cutting cycle.

• 1
  Multi-pocket plates20 pockets can cut load time from 15 s to 3 s per part.
• 2
  Pocket depthLeave 0.5 mm of part above the jaw face for probing.
• 3
  Vacuum and filmGood for flat thin parts, poor for parts with through holes.

Every time a part leaves the machine, it needs to be unloaded, cleaned, re-fixtured, and re-datumed. That is 30 to 90 seconds of labor plus the risk of a new error stack. On a 10,000-piece run, a second operation at 45 seconds per part is 125 hours of labor.

The alternative is a mill-turn center or a 5-axis machine that finishes five faces in one cycle. On a part with features on three faces, one 5-axis cycle at 4 minutes usually beats two 3-axis cycles at 2.5 minutes each once you count handling.

The exception is a feature that needs a different process entirely: a ground bore, a tapped hole that must be hand-started, or a surface finish below Ra 0.8 μm on a large face. Those are worth a second setup because no single machine does them well at volume.

A practical check: if the second operation is under 20 seconds of actual cutting and the feature tolerance is looser than ±0.05 mm, consider leaving it as a manual bench step rather than building a second fixture. On simple deburr and chamfer work this is often the cheaper route.

• 1
  Handling is the hidden cost45 s per part over 10,000 parts is 125 hours of labor.
• 2
  One-cycle wins5-axis at 4 min beats two 3-axis cycles at 2.5 min each.
• 3
  Keep it separateGround bores and Ra 0.2 μm faces need their own setup.