Front page / How many prototypes do you need before manufacturing a product?
How many prototypes do you need before manufacturing a product?
Most products pass through three prototype stages before production tooling (Linton Group names them concept model, engineering prototype and pre-production sample), and several versions within each stage are normal.
So the honest answer to "how many prototypes" is not a single number. It is "as many as it takes to answer each stage's question", and the count depends on how complex the product is and how many surprises each round turns up. What you can control is making each prototype answer a clear question before you pay for a mould.
The three stages
Linton Group frames each stage around one question:
| Stage | Question it answers | Typical form |
|---|---|---|
| Concept model | Is the idea right? | Rough model, sketch model or 3D print |
| Engineering prototype | Does it work? | Functional parts, often printed or machined |
| Pre-production sample | Can the factory repeat it? | Parts made with production-intent processes |
Linton adds that several revisions inside each stage are normal, and warns that compressing or skipping prototype rounds to save time tends to push the same cost downstream into tooling changes or production rework.
The electronics-hardware version: EVT, DVT, PVT
Teams building electronic products often use different names for the later stages. Instrumental, which works with hardware manufacturers, describes four builds. Its quantities are for its own electronics context and other sources quote smaller numbers, so read them as one company's view, not a standard:
| Build | Purpose (Instrumental) | Typical quantity (Instrumental) |
|---|---|---|
| Proto | Test risky modules, often in several variants | 10 or fewer |
| EVT (engineering validation) | Choose the production-intent design and find what needs fixing | 100 to 1,000 |
| DVT (design validation) | One production-worthy design; qualify the first hard tool for every part | 300 to 2,000 |
| PVT (production validation) | Check yields at full production speed | 1,000 to 20,000 |
Two points from Instrumental matter for anyone getting moulded parts made. First, it says EVT units may come from soft tools but that 3D-printed parts do not count as EVT. Second, it says all DVT parts should come from hard tools or processes capable of mass production. In other words, in this scheme the production mould is already cut before design validation, so the design must be stable by then.
Matching the prototype to the question
Each prototype should answer one question, and the method should fit that question:
| Question | Prototype method that can answer it | Limit to remember |
|---|---|---|
| Does the shape feel right in the hand? | 3D print or rough model | Material and finish differ from moulded parts |
| Do the parts fit and assemble? | 3D print or machined parts | Printed parts do not shrink or warp like moulded ones |
| Does it survive real use? | Parts in the real resin (machined or moulded) | Printed materials may behave differently |
| Will the moulded part fill, shrink and look right? | Prototype mould in the real resin | Short tool life |
| Can the factory repeat it at rate? | Pre-production run from production-intent tooling | Changes now are costly |
For short runs in a real moulding process, Formlabs notes that simple low-volume 3D-printed moulds can cost as little as $100 and can produce 100 or more parts within a few days. Formlabs suggests machined aluminium moulds for mid volumes. MakerVerse puts aluminium prototype moulds at 2 to 4 weeks to make, and Fictiv offers moulded prototype parts in as little as 2 weeks; those are supplier service figures, so confirm them on your own quote.
Time to allow
Analogy Design says a simple product often takes 6 to 12 months from a tested prototype to first production, and that electronics, regulatory approvals and long-lead components add time. Analogy also describes a pilot run, often a few hundred units, made to expose assembly and quality problems before full production.
A step-by-step plan
- Write down the questions the product must answer: appearance, ergonomics, fit, strength, function, regulatory tests, assembly.
- Group them into the three stages above and decide which prototype method answers each one.
- Build the concept model; revise until the idea itself is settled.
- Build engineering prototypes; test function and fit, and revise. Expect several rounds.
- Get a manufacturability (DFM) review of the moulded parts before the design is frozen.
- Make parts in the real resin, from a prototype mould or by machining, to test what prints cannot show.
- Freeze the design, then order production tooling.
- Run pre-production samples or a pilot run and fix what it finds before ramping up.
Before approving tooling, write down what each prototype proved and what is still unproved. Anything left unproved is a risk you are carrying into the mould.
Stage names and build sizes vary by industry. Medical, children's and electrical products may need regulated test units; check what your test lab and regulator require.
Questions buyers ask
Can I skip straight from a 3D print to a production mould?
You can, but Linton Group warns that skipping prototype rounds tends to move cost into tooling changes and rework rather than remove it. A printed part does not show how the moulded part will shrink, warp or fill, so at least review the design for moulding before steel is cut.
How many units should I make in each build?
There is no single standard. Instrumental gives its own ranges for electronics builds, but other hardware guides quote smaller numbers, and a simple plastic part may need only a handful of units per round. Size each build by the tests it must feed, plus spares for breakage.
Is a prototype mould worth paying for?
It is worth it when you need parts in the real resin, made by the real process, for testing or early sales. Formlabs' figures show printed moulds suit very short runs, and aluminium moulds suit larger ones. Ask the toolmaker whether the prototype tool can be modified later or must be remade.
When is the design ready for production tooling?
When the engineering prototypes have passed the tests that matter, a DFM review has been closed, and you can list nothing important that is still unproved. If DVT-style validation needs hard-tooled parts, as Instrumental describes, the design must be stable before that build.
Sources
- Linton Group, How to Create a Product: Idea to Manufacturing: https://lintonusa.com/resources/how-to-create-a-product/
- Instrumental, Hardware engineers speak in code: EVT, DVT, PVT decoded: https://instrumental.com/resources/blog/hardware-engineers-speak-in-code-evt-dvt-pvt-decoded/
- Analogy Design, How to Manufacture a Product Idea: From Prototype to Scale: https://www.analogydesign.co/blog/a-guide-on-how-to-manufacture-your-product-idea-instantly
- Formlabs, How to Estimate Injection Molding Cost?: https://formlabs.com/blog/injection-molding-cost/
- MakerVerse, Injection Molding FAQ: https://www.makerverse.com/resources/injection-molding-faq/
- Fictiv, Injection Molding Prototypes: https://www.fictiv.com/injection-molding-prototypes