Most RFQs we receive skip past the part and go straight to the number: 40 pieces, 1,000 pieces, 10,000 pieces. That number is not really about price. It is a test the buyer is running on us - and each band tests something different. Here is what we actually have to prove at each step, and what you should ask for before you commit a five-axis order to any shop, in China or anywhere else.
At 1-10 pieces, you are buying engineering time
A one-piece order on a 5-axis center is mostly programming, proving, and first-article inspection. Machine time is a rounding error. What you should ask: how long from drawing to first article (ours runs 3-7 days), and whether the CMM report comes with the part or costs extra. A shop that prices this band on machine hours will disappoint you; a shop that prices it honestly on setup hours is the one you keep.
At 40-500 pieces, the interview is about fixturing
This is the band where quotes diverge most. Without zero-point pallet systems, a shop prices your 100-piece order like ten separate 1-piece orders - every piece re-indicated, every piece carrying setup cost. With pallets, parts move between mills without re-datum and the price collapses toward machine-hour rates. Ask the shop directly: do you use zero-point systems, and can you show a photo of a loaded pallet? We run this on our five-axis cells (100 centers, 30 of them three-plus-two, inside a 220-machine plant in Shenzhen), and it is the single biggest reason our small-batch pricing holds when buyers scale from 40 pieces upward.
At 1,000 pieces, process design replaces inspection-by-piece
At a thousand pieces nobody inspects every part and stays competitive. The shop should be talking about dedicated fixtures, tool-life management, in-process gauging, and SPC sampling plans. If your quote at this volume still prices full inspection, it is priced wrong - either for you or for them.
At 10,000 pieces, the honest answer is process migration
Here is what most shops will not tell you: at ten thousand pieces, some of your geometry should leave the 5-axis mill. Turned features migrate to Swiss-type lathes. Prismatic second operations go to 4-axis cells. The unit price follows the cheapest process that still holds the print. Our floor runs 70 Tsugami Swiss-type lathes alongside the milling centers, monthly capacity around 5 million pieces across the group - so when a buyer's forecast grows, their part moves to the right machine instead of their price doubling.
What to put in the RFQ so the quote comes back honest
- Batch ladder, not a single number: your real forecast at 1, 100, and 1,000+ pieces
- The tolerance that actually matters (over-specifying ±0.005 mm on a cosmetic bracket costs you at every band)
- Material and finish, stated as standards, not brand names
- Your incoterm. We ship EXW Shenzhen and are glad to recommend freight forwarders; knowing this upfront avoids re-quotes
One warning for the 10,000-piece buyer: ask who owns the dedicated fixtures after the first run. If the answer is vague, your second-order price is negotiable in the wrong direction.
We wrote a fuller breakdown of what our five-axis floor holds - tolerances, batch economics, and the four-axis question - here: 5 Axis CNC Machining China
This article is from CHX Precision Metal, a Shenzhen CNC machining shop (70 Tsugami Swiss-type lathes among 220 CNC machines). Technical library: https://chxmetal.com
Top comments (1)
The batch ladder is a useful framing—especially the shift from setup and fixturing at 40–500 pieces to process control at 1,000. To make quotes comparable, I’d call out the CTQ dimensions and datum scheme in the RFQ, then specify what the first-article report must include and agree on the sampling plan for each volume. If geometry moves off the 5-axis mill at 10,000 pieces, requiring a first-off approval against the same drawing revision helps keep that process change controlled. I’d also put fixture ownership, storage, and maintenance responsibility in writing before the first run; it makes repeat orders much less ambiguous.