Wholesale CNC machining: a guide

• 1
  Key takeaways
• 2
  What wholesale CNC machining actually means on the floor
• 3
  Which parts suit wholesale CNC machining and which do not
• 4
  What drives cost per part as quantity increases
• 5
  Fixturing and workholding: where volume runs succeed or fail
• 6
  How tolerance and inspection change at production quantity
• 7
  Why wholesale CNC machining holds tolerances differently than prototype work
• 8
  Matching quantity and geometry to the right process

Wholesale CNC machining is the production of the same part, in meaningful quantity, on CNC equipment that has been set up once and left running. The phrase describes an order structure more than a machine type. A 500-piece run of aluminium brackets and a 500-piece run of titanium housings use the same turning and milling centers, but they are planned completely differently.

The distinction matters because unit economics shift. On a one-off prototype, programming, workholding and first-article inspection dominate the price. At 1,000 pieces those costs are divided across the batch and the dominant terms become cycle time, material yield and tool wear. That is why a shop can quote a single bracket at a high price and the same bracket at a fraction of it once quantity is confirmed.

Wholesale work also changes who owns the risk. In prototype mode, a scrapped part is annoying. In a 2,000-piece run, a fixture that drifts 0.05 mm after 400 cycles turns into a batch-wide dimensional problem. Production planning is therefore less about hitting ±0.005 mm on one part and more about holding it on parts 1, 200 and 2,000.

• 1
  One setup, many partsThe machine is dialed in once and the program runs repeatedly.
• 2
  Process control over craftsmanshipConsistency comes from fixtures and offsets, not from operator attention.
• 3
  Inspection shifts to samplingFirst article and periodic checks replace measuring every feature on every part.

The best candidates are parts with stable geometry, a modest number of critical features, and material that machines predictably. Aluminium 6061, 6082 and 7075, brass C36000, and stainless 303 and 304 all cut consistently, so a program that works on part one generally works on part one thousand. Wall thickness above roughly 1.0 mm for aluminium and 1.5 mm for stainless gives the cutter enough rigidity to avoid chatter and deflection.

Parts get harder when the geometry fights the tool. Deep pockets with a depth-to-diameter ratio above 4:1 need long, slender end mills that deflect and wear quickly. Thin floors under 0.5 mm tend to spring during clamping and release. Sharp internal corners force small tools and slow feed rates, which stretches cycle time across the whole batch.

There is also a quantity threshold in the other direction. Below roughly 50 pieces, a volume-oriented process is often the wrong choice. Soft jaws get machined, programs get proven, and the setup cost per part stays high. For 1 to 50 pieces, rapid prototyping or short-run machining is usually faster and cheaper than forcing a production setup.

Finally, consider whether the part should be machined at all at volume. Die casting and vacuum casting beat CNC on unit cost once quantities climb, provided the tolerance and surface requirements allow it. Machining stays competitive when you need tight tolerances, dense features, or the flexibility of changing the design between runs.

• 1
  Good fitStable geometry, wall above 1.0 mm, fewer than 30 critical features.
• 2
  MarginalDeep pockets at 4:1, thin floors, high feature density on small parts.
• 3
  Wrong processVery low quantity, or parts better served by casting or molding.

Cycle time is the largest single term. It includes the actual metal cutting plus tool changes, rapid moves and any in-cycle probing. Reducing a 12-minute cycle to 8 minutes cuts unit cost by roughly a third on a machine-hour basis, which is why shops spend time on toolpath optimization before they start a long run.

Material yield is the second term, and it is often underestimated. Bar stock and plate are bought in standard sizes. If a part nests poorly, you pay for the offcut. On aluminium plate, a nesting change that improves yield from 70 percent to 85 percent can move the material cost per part by more than 15 percent. On titanium or Inconel, where the raw stock is expensive, the effect is much larger.

Tool life behaves differently at volume. A carbide end mill that survives 20 parts in 6061 may need indexing every 60 parts in 17-4PH stainless. Tool cost per part is small in aluminium and significant in hardened or nickel-based alloys. Shops account for this by choosing coated tools, adjusting feed and speed, or accepting more frequent changes.

Scrap rate ties everything together. A process that yields 99 percent good parts at 100 pieces is fine. The same process at 5,000 pieces means 50 rejected parts, plus the cost of detecting and replacing them. Processes are usually tightened, not loosened, as quantity grows.

Workholding is the part of wholesale machining that engineers under-plan most often. A three-jaw chuck or a single vise is fine for one part. At 500 parts, clamping has to be repeatable, fast to load, and resistant to thermal drift over a shift.

Soft jaws machined to the part profile are the standard answer for mill and lathe work. They distribute clamping force, reduce marking on finished surfaces, and allow a known reference position for every cycle. For flat plates, vacuum plates or magnetic chucks remove clamp-induced distortion entirely, which matters on thin parts that would otherwise bow.

For 5-axis runs, a Ø400 mm rotary table with modular tombstones lets you load several parts per cycle. The tradeoff is access: a tombstone makes the tool reach longer, which means more deflection on deep cuts. Shops typically balance this by reducing depth of cut per pass and increasing the number of passes.

The failure mode to watch is gradual position shift. A fixture that holds ±0.01 mm at the start of a shift may drift as chips pack into locating features or as the vise heats up. Operators handle this by cleaning locating surfaces between cycles and re-checking a reference feature at set intervals.

• 1
  Soft jawsRepeatable clamping, less marking, quick load and unload.
• 2
  Vacuum or magnetic platesNo clamp distortion on thin or flat parts.
• 3
  TombstonesMultiple parts per cycle, at the cost of tool reach.

A ±0.005 mm callout is achievable on a machined feature, but not on every feature at every quantity. The realistic question is which dimensions actually need that band. Applying a blanket tight tolerance across a drawing raises cost without improving function, because every tight dimension adds inspection time and increases the chance of a reject.

The engineering rule is to reserve tight tolerance for the features that mate, seal or locate. A bearing bore, a dowel pin hole or a sealing face may genuinely need ±0.005 mm. A mounting clearance hole at Ø8 mm does not need better than ±0.1 mm. Marking tolerances feature by feature is the single most effective way to reduce unit cost on a volume run.

Surface finish follows a similar logic. As-machined surfaces land around Ra 1.6–3.2 μm, which is fine for most functional faces. Sealing or sliding surfaces usually need Ra 0.8–1.6 μm, and optical or bearing surfaces may need Ra 0.2–0.8 μm. Each step down adds finishing time or a secondary operation, so specify the coarsest finish that works.

Inspection at volume is sampling-based. The first part is measured fully against the drawing. After that, operators check a defined set of features at set intervals, and 100 percent inspection is applied only where a defect would be safety-critical or unrecoverable. Reports are available on request, and raw material certificates are checked at incoming.

The mechanism behind consistent volume output is thermal and mechanical stability. A machine that has been running for four hours is at a stable temperature, and its ball screws and spindle have reached steady-state expansion. The first parts off a cold machine are often slightly different from parts produced two hours later. Volume runs benefit from that stabilized state.

The second mechanism is statistical. A single prototype either passes or fails. A production run produces a distribution. If the process mean sits at the nominal and the spread is narrow, few parts fall outside tolerance. If the mean drifts toward one limit, even a narrow spread produces rejects. Production control is about keeping the mean centered.

This is why shops use in-process probing on critical features and why they re-qualify fixtures after a tool change. It is also why a design that is marginal on a prototype, sitting at 90 percent of the tolerance band, becomes risky at volume. Design margin is a production asset.

The third mechanism is material consistency. Two bars of 6061 from different heats can machine slightly differently. Production runs are usually scheduled with material from a single heat or lot where possible, which removes one source of variation before the first chip is cut.