Bulk Rapid Prototyping Solution: Build 20 to 500 Parts That Match

One prototype is a drawing check. You learn whether the geometry fits, whether the wall is thick enough, whether the assembly closes. Fifty parts is a process check. Now the question is whether the second part matches the first, and whether the fiftieth matches both. That gap is where most prototype programs lose time, because a shop that machines one-off parts by hand rarely has the fixtures, tool lists, and inspection records to repeat itself.

Volume also changes the economics. Setup, programming, and first-article inspection are fixed costs. Spread across 100 parts instead of two, they shrink to a small fraction of unit price. Material buying shifts too: a bar of 7075 aluminium cut into 100 blanks costs less per part than 100 separate cuts, and remnants from one order feed the next run.

The third change is traceability. Once prototypes go to a test lab, an EMC chamber, or a customer trial, someone will ask which revision a given part came from, which heat lot, and what its measured dimensions were. A bulk rapid prototyping solution that keeps every operation under one quality system answers that question from a file. Three vendors with three numbering schemes do not.

No single process suits every part in a prototype batch. A metal bracket that carries load belongs on a CNC mill, where grain direction and wall thickness behave like production. A duct with internal channels that no tool can reach belongs on a printer. A transparent lens cover for a five-unit pilot build usually belongs in a vacuum casting silicone mold, because polishing ten CNC-machined acrylic domes costs more than casting them.

The useful rule is to sort parts by function rather than by looks. Structural parts, mating features, and anything that will be tested to failure go to machining. Non-structural covers, brackets that only locate, and ergonomic shells can go to printing or casting if the surface and tolerance allow. When a part must be both light and stiff, 3D printing with a topology-optimized shape followed by CNC finishing of the mating faces often beats either process alone.

Mixing processes inside one order also removes a coordination layer. Your engineer sends one set of drawings, one material callout, and one finish specification. The shop sequences the operations so that printed parts and machined parts arrive on the same pallet, with the same surface treatment where the design calls for it.

• 1
  CNC machiningMetal and engineering plastic parts that must hold ±0.005 mm and match production material.
• 2
  3D printingComplex internal geometry, lightweight lattices, and shapes a cutter cannot reach.
• 3
  Vacuum castingTwo to fifty copies of a housing or cover in a production-like resin.
• 4
  Sheet metalEnclosures, brackets and chassis panels formed from 0.5–6 mm stock.

Prototype budgets go wrong when every dimension carries the same tight tolerance. A ±0.005 mm callout on a cosmetic edge costs inspection time and often forces a second operation, while adding nothing to function. The dimensions that decide whether the assembly works are usually a small set: bearing bores, mating faces, locating pin holes, and any feature that sets a gap or a preload.

A practical approach is to mark those functional dimensions on the drawing and leave everything else at general tolerance. That tells the programmer where to slow the feed, where to use a finishing pass, and where a roughing cut is enough. It also tells the inspector where to spend time with a micrometer instead of a caliper.

Surface finish follows the same logic. A sealing face may need Ra 0.8–1.6 μm, a sliding surface may need Ra 0.2–0.8 μm, and a mounting face inside a housing is fine at Ra 1.6–3.2 μm as machined. Asking for a mirror polish on a part that gets powder coated is money spent on nothing.

Repeatability comes from three things: fixtures, tool life records, and in-process checks. A soft jaw cut for the first part and never re-cut will drift. A tool that has cut 300 aluminium parts no longer holds the same corner radius. A shop that measures only the first article learns nothing about part 87.

The countermeasures are not exotic. Dedicated soft jaws or a vacuum plate for every operation. A tool change interval written into the setup sheet, not left to operator judgment. Dimensional checks at fixed intervals through the run, with the results recorded against the part serial. GreatLight runs 100% inspection before shipment, combining raw material verification, in-process monitoring, and a final check, with reports available on request.

For runs above roughly 200 parts, statistical process control becomes useful rather than academic. Plotting a critical bore diameter across the run shows whether the process is centered and whether it is drifting. If it drifts, the shop adjusts the offset before the first out-of-tolerance part is made, not after.

Splitting a prototype batch across specialized vendors looks efficient on paper. In practice, the coordination cost lands on your engineering team. Three purchase orders, three lead times, three inspection formats, and three answers when a fit problem appears. Nobody owns the assembly tolerance, so nobody can fix it.

A single supplier makes sense when the batch contains parts that must fit together, when a common finish or material is required, or when the schedule is tight enough that a handoff would consume the buffer. Mixed technology under one roof also means one DFM review, so a machining limitation on one part can be designed out of a printed part before either is cut.

There are cases where splitting works. A very large casting or a specialized coating may genuinely sit outside a shop's equipment list. The test is whether the interface between vendors is simple. A part that bolts onto another part is a clean split. A part that shares a datum with another part is not.

A bulk prototyping order moves faster when the input package is complete. Send 3D models in STEP or Parasolid, 2D drawings for anything with tolerances or threads, a material callout with the temper or grade, a finish specification, and the quantity per part number. Note which dimensions are functional and which are reference. If a part has a cosmetic surface, mark the direction of the grain or the print orientation on the model.

What comes back should be more than a price. A useful DFM response flags features that will need a second operation, walls too thin for the chosen process, tolerances that cannot be measured with standard equipment, and any part where a different process would cut cost without hurting function. GreatLight returns quotation and free DFM analysis within 12 hours, and production can start within 24 hours of approval.

Uploads stay confidential, and an NDA is available on request. For programs where the prototype is the product, that paperwork should be in place before the first model file leaves your network.