Bulk Manufacturing of Small Parts on CNC Machines
Small parts behave differently at volume. This page explains how workholding, cycle time, tool wear and inspection change when a 12 mm bracket goes from 10 pieces to 10,000. Written for design engineers and sourcing teams who have to decide whether a small part belongs on a mill, a lathe, or somewhere else entirely.
Why bulk manufacturing of small parts is hard to scale
A 40 mm aluminium housing and a 6 mm pin do not scale the same way. The big part is limited by machine travel and spindle power. The small part is limited by how fast you can load it, how much material you have to remove, and whether the operator can reach it safely.
At low volume you can afford a vise and a handful of soft jaws. At 5,000 pieces that same setup becomes the bottleneck. Every second of load and unload time is multiplied by the batch size. A 20 second improvement per part is 27 hours back on a 5,000 piece run.
Tool reach is the second constraint. A 2 mm end mill cutting a 1 mm deep pocket has an effective cutting length near 6 mm. Push past a 3:1 length-to-diameter ratio and chatter starts. Chatter means scrapped parts, and scrapped parts at volume mean real money.
So the design decision comes early. If the feature is smaller than the tool that can reach it, you are choosing between a slower cycle and a redesign. We usually flag this in DFM analysis before the first chip is cut.
How fixturing changes from 10 pieces to 10,000
At 10 pieces, a machinist clamps the part in a vise, touches off, and runs. Setup might take 45 minutes, but it only happens once. At 10,000 pieces a 45 minute setup is irrelevant next to a 4 second load time repeated 10,000 times.
The move is to multi-part fixturing. A plate holds 20, 40 or 80 small parts at once, so the spindle keeps cutting while the operator loads the next nest. This is where most of the cost saving in bulk manufacturing of small parts actually comes from, not from spindle speed.
Soft jaws machined in place hold ±0.02 mm repeatability. Hardened fixture plates with dowel pins do better, closer to ±0.01 mm, and they survive thousands of cycles without losing position. For anything above 2,000 pieces we normally build the second type.
The trade-off: a dedicated fixture costs money up front and only makes sense if the part will not change. If the design is still moving, a modular vise system with standard jaws is cheaper to modify even though the cycle is slower.
Where the seconds actually go
A small part often spends more time being moved than being cut. Roughing a 12 mm aluminium bracket might take 40 seconds of spindle time inside a 3 minute cycle. The rest is load, unload, blow-off, probe and tool change.
Tool change is the quiet cost. A 4 second tool change sounds harmless until a part needs 14 tools. That is 56 seconds per part, or 155 hours across a 10,000 piece order. Consolidating features so one tool does more work is often worth more than raising the feed rate.
Spindle speed matters less than people expect on small parts. A 6 mm carbide end mill in aluminium runs at 12,000–18,000 rpm and 1,500–3,000 mm/min. Doubling rpm rarely doubles output, because the machine still has to accelerate and decelerate between short moves.
The real lever is cutting air. Short moves with high acceleration beat long moves at high feed. Look at the toolpath, not the spec sheet.
Tool wear and the drift nobody watches
In a 100 piece run, a worn 3 mm end mill is a minor annoyance. In a 10,000 piece run it is a dimensional trend. Carbide wears gradually, and the pocket it cuts gets slowly smaller or larger depending on the direction of cut.
For aluminium, a coated carbide tool running at 15,000 rpm holds size for roughly 2,000–4,000 parts before it needs indexing. Stainless 316 is far less forgiving, closer to 300–800 parts, and 17-4PH is harder still. These numbers move with coolant, rigidity and depth of cut.
The practical fix is scheduled tool changes, not reactive ones. We index or replace on a count, then re-probe the first part off the new tool. The tolerance band stays inside ±0.005 mm instead of drifting until someone catches it at final inspection.
This is also why material choice is a volume decision. A part that works in 6061 at 10,000 pieces may not work in 316 at all if the tolerance is tight and the wall is thin.
Inspection that keeps up with the batch
Checking every part by hand does not scale. At 200 pieces a CMM programme per part is fine. At 10,000 pieces you need a plan that separates the first article from the running check from the final audit.
A typical flow: full dimensional report on the first article, in-process checks every 50–200 parts depending on tolerance, and final inspection before shipment. Gauges and fixtures do the running checks, so the CMM is free for the hard features.
For parts with a true position callout, a go/no-go gauge built to the same datum scheme as the fixture is faster and more repeatable than a CMM touch cycle. We build those when the order justifies it.
Everything ships with 100% inspection before shipment. Raw material certificates, in-process records and final reports are available on request, and we hold ISO 9001:2015, IATF 16949:2016, ISO 13485:2016 and ISO 27001:2022.
Which setup fits which batch size
Pick the row that matches your volume and tolerance.
| Batch size | Workholding | Typical cycle | Best for |
|---|---|---|---|
| 1–50 pieces | Standard vise, soft jaws | 5–20 min per part | Prototypes, design checks |
| 50–500 pieces | Multi-part plate, 10–20 nests | 1–4 min per part | Bridge and pilot runs |
| 500–5,000 pieces | Dedicated fixture, hardened pins | 30 s–2 min per part | Repeat production orders |
| 5,000+ pieces | Bar feeder or pallet changer | 10–60 s per part | Steady high-volume runs |
| Any volume, tight true position | One-op 5-axis, no re-fixture | 2–8 min per part | Complex geometry, ±0.005 mm |
When CNC is the right call, and when it is not
If the part needs ±0.005 mm, tight true position, or a material like 7075 or 17-4PH, run it on CNC at any volume. If the geometry is simple, the tolerance is looser than ±0.1 mm, and the annual quantity is above 50,000 pieces, look at die casting or injection moulding instead and keep CNC for the tooling and the first articles.
Bulk manufacturing of small parts: common questions
What is the smallest part you can machine at volume?
It depends on the smallest feature, not the overall size. A 3 mm pin with a 0.5 mm cross-hole is harder than a 30 mm plate with M3 threads.
As a working rule, features below 0.5 mm wide need micro tooling and much slower feeds, which changes the cost picture. Send the drawing and we will tell you in DFM analysis whether it is practical at your quantity.
Does bulk manufacturing of small parts cost less per piece?
Yes, but not in a straight line. Setup, programming and first-article inspection are fixed costs that spread across the batch. At 10 pieces those dominate. At 5,000 pieces they are a rounding error.
The variable costs, material and spindle time, do not fall much. Above roughly 2,000 pieces the curve flattens, so doubling the order no longer halves the unit price.
Can you hold ±0.005 mm across a 10,000 piece run?
Yes, with the right process controls. That means a hardened fixture, a scheduled tool-change count, temperature-stable coolant, and in-process checks at a defined interval.
It does not mean every part is measured on a CMM. It means the process is stable enough that sampling catches drift before it reaches the tolerance limit.
Do you have a minimum order quantity?
No minimum order quantity. We run from one prototype to 10,000+ part runs on the same equipment, so the process you qualify at prototype stage is the process that runs at volume.
Uploads are secure and confidential, and an NDA is available on request if the design is sensitive.
How fast can a bulk order start and ship?
Quotation and free DFM analysis come back within 12 hours. Production can start within 24 hours of approval, and parts typically ship in 3–5 days depending on quantity and finishing.
For repeat orders the fixture already exists, so the gap between purchase order and first chip is much shorter.
Should I use 5-axis for a small part?
Only if the geometry needs it. A 5-axis machine earns its cost when the part has features on four or more faces, or when a single setup removes a stack-up error that would otherwise push you past tolerance.
For a simple bracket with two faces, a 3-axis mill with a good fixture is faster and cheaper at any volume.
Send the drawing, get a real process plan
Upload your small part and we will return a quote, a DFM note and a suggested fixture and inspection plan within 12 hours.
12-hour quoteNo MOQ100% inspection before shipment
More process notes from the shop floor
We publish setup notes, tooling trials and inspection data from the factory floor.
