Pieces Per Day: How High-Volume CNC Capacity Actually Works

• 1
  Where pieces per day comes from
• 2
  Fixturing decides the daily rate before the spindle does
• 3
  Matching the part to 3-axis, 4-axis or 5-axis
• 4
  Inspection is part of the cycle, not a step after it
• 5
  Material and finishing keep the count steady
• 6
  When high daily output is the wrong goal
• 7
  Which machine class fits which part

A quote that says thousands of pieces per day is a claim about arithmetic, not about effort. Take one machining cycle of 4 minutes. One machine running 20 hours a day yields 300 pieces. Ten machines yield 3,000. The number is decided by cycle time, spindle hours per day and how many machines the job can occupy at once.

Cycle time itself is a sum of cutting time, tool change time, load and unload time, and any inspection pause. On a 3-axis mill cutting aluminium 6061, cutting may be 70% of the cycle. On a 5-axis job with deep pockets in 17-4PH stainless, cutting time is longer and the load-unload share shrinks. Both can still hit high daily counts; they just need different machine counts.

This is why we quote pieces per day only after seeing the drawing. Two parts with the same envelope can differ by a factor of five in cycle time because of one tight tolerance, one deep cavity, or one surface finish callout.

• 1
  Cycle timeCutting plus tool change plus load and unload.
• 2
  Spindle hoursHow many hours per day the job holds a machine.
• 3
  Machine countHow many spindles are assigned to the same part number.
• 4
  YieldGood parts out divided by parts started.

The fastest spindle cannot beat a slow load. If a part takes 90 seconds to clamp and locate, that time repeats on every piece all day. A dedicated soft jaw, a vacuum plate or a hydraulic fixture can cut it to 15 seconds. On a 3,000 piece order, that difference is measured in machine hours, not minutes.

We build workholding around the run length. For one prototype, a vise and a set of stops are enough. For 10,000 parts, we cut dedicated fixtures on one of the 27 three-axis machines, then move the fixture to the mill-turn or 5-axis cell that will run the job. The fixture becomes a tooling asset, and it is what makes the per-piece number repeatable across shifts.

Fixtures also set the tolerance floor. A part held in a soft jaw on a Ø400 mm rotary table behaves differently from the same part held in a three-point clamp. If the drawing asks for ±0.005 mm on a bore, the fixture has to support that before the program is written.

Not every part should run on a 5-axis center. If a part has features on three faces but generous tolerances, a 3-axis machine with two re-fixturings is cheaper per piece. If those three faces must be true to each other within ±0.005 mm, the re-fixturing error alone eats the tolerance. That is when a 5-axis center pays for itself.

Our floor carries 16 simultaneous 5-axis machining centers, 12 four-axis mills, 27 three-axis machines and 16 mill-turn centers. The large frame travels 4,000 × 400 × 150 mm, so long extrusions and rails can be cut in one setup. Medium frames cover 750 × 1,150 × 550 mm and 600 × 600 × 600 mm. Compact frames at 500 × 500 × 450 mm and 500 × 310 × 200 mm handle small, high-count parts where fast load cycles matter more than travel.

A part that fits a compact frame and needs no re-fixturing is the best candidate for high daily counts. A part that needs four setups and a CMM check between each one is not, no matter how many machines are free.

On a high-count run, inspection has to be built into the cycle. We check raw material before it reaches a machine, monitor the process while the job runs, and inspect 100% of parts before shipment. Reports are available on request. The point is not paperwork. It is catching a drift before 200 parts are cut to the wrong size.

A practical example: a bore held at ±0.005 mm will move as the tool wears. If the operator measures every 25th part and the tool drifts 0.003 mm over 100 parts, the last parts in that window are already at the edge. Measuring every 10th part on a tight feature costs a few seconds per piece and prevents a scrap batch.

The 99.99% qualification rate we report is a floor, not a target. It comes from process control, not from sorting bad parts out at the end. Sorting is the most expensive way to reach a quality number, because the scrap was already paid for in spindle time.

A machine that runs out of stock stops producing. Aluminium 6061, 7075 and 6082, stainless 303, 304 and 17-4PH, steel 1045 and 4140, brass C36000, and titanium TC4 are all stocked or sourced on short lead. The choice of material affects the daily count through cutting speed, not through supply alone.

Aluminium 6061 cuts fast and holds ±0.005 mm well, so it is the usual choice for high daily counts. 17-4PH stainless in the H900 condition machines cleanly but slower. Inconel and titanium TC4 need lower surface speed, more coolant and more frequent tool changes. The daily count drops, and that is a property of the material, not of the shop.

Finishing runs on a separate track. Anodizing, plating, powder coating, bead blasting, tumbling and laser marking are scheduled so the machining cell keeps cutting while finished parts move out. Laser marking needs a minimum character height of 1.5 mm; below that, the mark is not reliable across a large batch.

High volume is not always the right answer. A part that will be redesigned next month should not get a dedicated fixture and a tuned program. Cutting 200 pieces from a vise setup is cheaper than building tooling for a design that will change.

Very tight tolerances also fight volume. A ±0.005 mm callout on a feature that does not need it will slow every cycle and raise the scrap risk. We review tolerances during the free DFM analysis and flag callouts that add cost without adding function. Quotation and DFM come back within 12 hours.

Finally, volume does not fix a bad process. If a part has a chatter problem at 10 pieces, it will have the same problem at 10,000, only more often. We would rather spend an extra day on the setup and run the job clean.