A prototype program is a sequence of validation questions, not a single build. The team that treats every prototype as "make the part" spends money on the wrong processes, while the team that asks "what must this build prove before we spend the next budget" finishes faster and cheaper. This guide is the decision framework for that question: which process to use at which stage, what each stage must prove, what budget and lead-time variables to plan, and which deliverables must survive to production. It deliberately does not repeat the full mechanics of each process; the 3D printing, CNC machining, and urethane casting pages cover those in depth. The value of this page is the framework that ties them together.
The Prototype Program as a Question Sequence
Run the program as a series of gates. Each stage answers one question, and the program advances only when the evidence exists.
| Stage | The question | What must be proven | Budget position |
|---|---|---|---|
| Concept | Does the design look and feel right? | Form, ergonomics, appearance direction | Low cost, many iterations |
| Functional | Does the mechanism and fit work? | Motion, clearances, assembly sequence | Medium, fewer iterations |
| Performance | Does it survive real load, heat, and tolerance? | Strength, thermal, dimensional data | High, production materials |
| Pre-production | Will the production process work? | Tooling, finish, yield, assembly | Highest, frozen design |
The stage table is the contract for the program. If the team cannot name the question a build answers, the build should not be ordered, because the answer determines the process and the budget. A concept question answered with a machined part wastes money; a performance question answered with a printed part produces data that cannot be trusted.
Match the Process to the Validation Question
Each prototype process has a narrow band where it is the right answer, and a list of things it cannot prove. Use the band, and move on.
| Process | Use it when | What it cannot prove |
|---|---|---|
| 3D printing | Geometry, appearance, and internal-fit iterations | Production material behavior and molded surfaces |
| CNC machining | Functional parts in production material with real tolerances | High-volume economics and molded geometry |
| Urethane casting | Molded-like appearance at small volume | Production resin mechanical and thermal properties |
| Sheet metal fabrication | Real formed geometry, fasteners, and shielding | Machined tolerances and molded detail |
| Injection molding | Molded part behavior and production validation | Low-volume economics |
The discipline is to choose the least representative process that still answers the question. Printing answers geometry questions cheaply, so print; machining answers strength and tolerance questions with production material, so machine when the data must transfer; and molding or casting answers appearance and molded-behavior questions, so use them only when the question is about the molded part. Every step up in representativeness costs more, so the process choice is a budget decision made with the stage table, not a preference.
Budget and Lead-Time Variables
Prototype budgets fail when they are set as a lump sum instead of a curve. Plan the cost per iteration at each stage, and let the stage table decide how many iterations each stage gets.
- Iteration count falls as the design matures. Concept stages should run many cheap iterations; pre-production stages run one or two expensive builds.
- Cost per iteration rises with process representativeness. Printed iterations cost the least, machined iterations cost more, and tooled iterations dominate the budget.
- Tooling nodes are the budget cliffs. A silicone mold, a prototype mold, and a production mold each create a step change in cost and lead time, and the design should be frozen before each one.
- Inspection and documentation grow with the stage. A concept model needs no report; a pre-production build needs first-article data, finish samples, and material certificates, and that work should be in the quote.
Lead time follows the same curve. Printed parts arrive in days, machined parts in a week or two depending on material, and tooled parts in weeks to months, so the schedule should be built around the tooling nodes, not the machining cycles.
The numbers vary by product and supplier, but the shape of the curve is consistent: concept iterations cost tens to low hundreds per build, machined functional parts run from hundreds to low thousands depending on material and complexity, and tooled pre-production builds dominate the budget. A program that plans for this curve avoids the two common failures: spending tooling money before the design is proven, or starving the performance stage after the cheap iterations are done. The budget table should therefore be written by stage, with the freeze point as the boundary between the iteration budget and the tooling budget.
The Design Freeze Point
The freeze point is the moment geometry, materials, and finish stop changing, and it belongs before any prototype tooling is committed. Machined and printed validation exists to make the freeze safe: functional builds confirm the mechanism and the tolerances, so the frozen design is the one already proven, not the one hoped to work.
Freeze means the drawing revision is locked, the material grades are named, and the finish and inspection requirements are written down. If a change after the freeze is inevitable, it costs tooling work and revalidation, so the freeze should be scheduled after the performance data, not before it. The practical signal to freeze is the first pre-production build: from that point, changes are change orders, not iterations.
The Deliverables That Transfer to Production
The prototype program's product is not just parts; it is data. Collect these deliverables at each stage so the production handoff starts with evidence instead of rediscovery.
- CAD with a revision number and a single source of truth
- Material grades confirmed against the production requirement
- Critical dimensions marked with datums and measurement methods
- First-article inspection data from the machined and pre-production builds
- DFM feedback records showing what was changed and why
- Finish and color samples signed against the production process
- Tolerance and thermal data from the performance stage
- A change log covering every revision from concept to freeze
A 6CProto DFM review produces part of this package at the quoting stage, and the machined builds produce the inspection data, so the team should collect it as the program runs rather than reconstructing it at the handoff. The production supplier inherits the file set, the datum scheme, and the acceptance criteria, and that is what makes the prototype program an investment instead of an expense.
The same package serves the next product in the family: the validation data from one program becomes the reference for the next, so the change log and the inspection records are assets the team keeps, not paperwork the program leaves behind.
Decision Rule
Name the question each build must answer, choose the least representative process that answers it, and freeze the design before the first tooling node. Plan budget and lead time around the iteration curve and the tooling cliffs, and collect the deliverables that transfer to production. The prototype program that works is the one where every build was ordered for the evidence it had to produce.
FAQs
What prototype data transfers to production?
The CAD with revisions, confirmed material grades, critical dimensions with datums and measurement methods, first-article inspection data, DFM feedback records, finish samples, and the change log. The production supplier inherits this file set, so the handoff starts with evidence instead of rediscovery.
How should prototype tolerances differ from production tolerances?
Prototypes should hold the tolerances that affect the validation question, not every production callout. Functional builds should match production tolerances on the features being tested, while cosmetic and non-critical dimensions can be looser. The drawing should mark which features are critical so the prototype answers the right question at the right cost.
When should a team freeze a design before prototype tooling?
Freeze after the performance data exists and before the first tooling node, because changes after tooling cost mold work and revalidation. Freeze means the drawing revision, material grades, and finish requirements are locked, and from the first pre-production build, changes are change orders, not iterations.
What files should be supplied for a prototype quote?
The CAD in a standard format with the revision, the drawing with critical dimensions, material grades, quantity, and the inspection or documentation requirements for the stage. The more complete the package, the more accurate the quote and the DFM feedback, and the fewer surprises appear in the first build.
Sources
- 6CProto Rapid Prototyping Services
- 6CProto 3D Printing Services
- 6CProto CNC Machining Services
- ISO 9001:2015 – Quality management systems
- ISO/ASTM 52900:2021 – Additive manufacturing general principles