Why Choose Custom Rapid Prototyping China: 5 Problems an Engineer Should Check First

• 1
  Prototype Problems: Symptom, Cause, Action
• 2
  Tolerance Drift on Custom Rapid Prototyping China Runs
• 3
  Wall Thickness and Warp After Finishing
• 4
  Surface Finish That Looks Wrong Under Light
• 5
  Quote and Communication Gaps
• 6
  When the Prototype Cannot Scale to Production
• 7
  How to Vet a Custom Rapid Prototyping China Supplier

Tolerance drift is the first thing we look for on a prototype that has to become a production part. A one-off bracket held at ±0.005 mm proves the machine can do it once. It does not prove the process holds. The difference matters when the same drawing goes to a 10,000-part run.

Most drift comes from heat, not from the control. A 5-axis center running a 4,000 mm part for six hours will move thermally by a few hundredths of a millimeter if the coolant and spindle warm-up are not managed. Shops that publish setup sheets will show a roughing pass, a cool-down, then a finishing pass. Shops that do not will hand you a part that measures well at 9 a.m. and badly at 3 p.m.

Ask a simple question: what is the thermal compensation routine on this machine? If the answer is vague, the tolerance on the drawing is a hope, not a process. In our plant, the 16 simultaneous 5-axis centers are scheduled with warm-up cycles before any tight-tolerance finishing cut.

The second source is fixturing. Soft jaws wear. A vise that held 50 parts will hold part 51 slightly differently. For prototypes this rarely shows. For bridge runs of 200 to 500 parts, it shows up as a gradual shift that nobody catches until inspection.

The practical fix is not a tighter drawing. It is a check plan. Specify where the critical dimensions are measured, how often, and what report format you want. That single line in a purchase order removes most of the risk.

• 1
  Watch forA quote that promises ±0.005 mm with no mention of setup or inspection
• 2
  Ask forA roughing/finishing split and a warm-up routine on tight-tolerance features
• 3
  Write into the POIn-process check frequency and report format

Warp shows up late, usually after anodizing or powder coating, when the part is already expensive. The cause is usually wall thickness combined with a finishing step that adds heat.

Aluminum below 1.2 mm wall is fragile in anodizing. The seal step runs hot, and thin sections move. Hardcoat is worse because the coating builds and the bath runs warmer. If your design has 0.8 mm ribs, say so on the drawing and expect the shop to run a cooler seal or skip hardcoat on those features.

Plastics behave differently. POM and PA absorb moisture and grow; PEEK holds dimension but machines slowly and costs more. A prototype in ABS printed or machined at nominal size will not match the same part in PC at 40 °C in service. Material selection is a prototype decision, not a production one you make later.

Carbon fibre is its own case. A carbon fibre bracket can be 40 percent lighter than aluminum at similar stiffness, but it will not hold the same tolerance on a machined edge. If the prototype is meant to validate stiffness, that is fine. If it validates a bearing bore, it is the wrong material.

One more thing: stress relief. A 7075 part hogged out of solid plate will move when you remove 70 percent of the material. Rough, stress-relieve, then finish. Shops that skip this step ship parts that are round on day one and oval on day four.

• 1
  Below 1.2 mm aluminumSpecify a cool seal or mask the feature from hardcoat
• 2
  7075 and 4340Ask for a stress-relief step between roughing and finishing
• 3
  Plastic prototypesMatch the production resin, not the cheapest one

Finish problems are easy to miss on a white table and obvious under a raking light. Ra numbers help, but they do not describe everything the eye sees.

Ra 1.6–3.2 μm is a normal as-machined finish. It is fine for a bracket that gets bolted into a frame. It is not fine for a sliding surface or a visible housing. For those, Ra 0.8–1.6 μm covers most cosmetic and light-contact work. Below that, Ra 0.2–0.8 μm, you are paying for polishing time and you should know why.

Tool marks come from stepover, not from the machine. A 12 mm cutter run at 0.5 mm stepover leaves visible scallops. The same cutter at 0.15 mm stepover looks nearly polished. If a supplier quotes a mirror finish at a roughing stepover, the finish will not be there.

Bead blasting hides small marks but changes dimensions slightly. It also changes the look of anodized color. If you blast after anodizing, you get a matte surface with exposed edges. If you blast before, the color is more uniform. Pick one and put it on the drawing.

Laser marking is the last step, and it is often forgotten. Minimum character height is 1.5 mm for a clean mark. Below that, letters fill in. If you need serial numbers or a date code, specify the field and the format on the drawing.

• 1
  Cosmetic partsName the finish class, not just the Ra value
• 2
  Blasting vs anodizing orderDecide before the parts run, not after
• 3
  Laser markingKeep characters at 1.5 mm or larger

A quote that takes a week and a DFM note that never comes are symptoms of the same problem: the shop is not reading your file. For a prototype, the DFM feedback is often worth more than the price.

We return a quotation and a free DFM analysis within 12 hours. That is not a marketing line; it is the point in the process where we find the 0.8 mm rib, the deep pocket, or the thread that will strip. Catching those before the first chip is cut saves a week.

Watch for quotes that list a price and nothing else. No material callout, no tolerance restatement, no finish specification. If the shop did not read the drawing, they will machine what they think they read.

Communication is also a scheduling issue. Production can start within 24 hours of a released order, and parts ship in 3–5 days. That speed only works if the questions are answered in the first hour, not the third day.

One practical test: send a drawing with one deliberate ambiguity, such as an unspecified corner radius. A good partner comes back with a question. A weak one machines a guess.

• 1
  Good signalDFM notes attached to the quote, with specific feature callouts
• 2
  Bad signalA price with no material, tolerance, or finish restatement
• 3
  TimingAnswer engineering questions within the first working hour

A prototype that works is not automatically a production plan. The gap shows up as a different process, a different shop, or a re-qualification you did not budget for.

The usual cause is process mismatch. A part printed in resin validates form but not strength. A part machined from billet validates strength but not die casting flow. If the production route is die casting, the prototype should at least be reviewed against draft angles and wall thickness before you commit to a tool.

Volume matters too. There is no minimum order quantity here, and runs go from one prototype to 10,000+ parts. But the economics change at the tooling boundary. Below a few hundred parts, machining is usually cheaper. Above that, casting or vacuum casting starts to win.

The third gap is documentation. A prototype shop that returns only parts leaves you to rebuild the inspection record later. Ask for the inspection report and the material certificate with the first shipment.

For regulated products, the certification path has to be planned early. ISO 9001:2015 covers general quality. IATF 16949:2016 applies to automotive. ISO 13485:2016 applies to medical devices. ISO 27001:2022 covers information security for files and drawings. If your product falls under one of these, say so at the quote stage.

• 1
  Process matchPrototype in the same family as the production process where possible
• 2
  Volume splitMachining below a few hundred parts, casting above
• 3
  PaperworkInspection report and material certificate with the first shipment