Prototype to production

Three words, three different jobs

People use these words interchangeably. They shouldn't, because each exists to do something the others don't.

A prototype exists to learn something. You build it because there's a question you can't answer on a screen. Change is expected, and it's cheap.

A first article exists to prove that a process makes a conforming part, before you commit to repeating it. It's made the production way, on the production intent, then inspected against the drawing rather than eyeballed.

A production run exists to repeat. The design is frozen. What you're buying is sameness, and every change is a change order with a cost and a date attached.

What kind of prototype

"Prototype" covers three builds that optimize for different things. Say which one you want, because they don't look alike and they don't cost alike.

• Proof of concept. Answers one question: does this work at all? Optimizes for speed and for being wrong cheaply. It's allowed to be ugly, oversized and made of the wrong material.
• Appearance model. Looks and feels like the finished product and does nothing else. Optimizes for surface, color and proportion. Useful for photography, investor meetings and crowdfunding pages. Do not load-test it.
• Functional prototype. Behaves like the real thing under real conditions. Optimizes for material properties and fit. Costs the most and tells you the most, because it's made closer to the production way.

The trap is quietly promoting a proof of concept to a functional prototype in your head. A printed part and a machined or molded part can behave differently under load even in the same material, because the process lays the material down differently. If the process changes, the test result doesn't carry over.

The comparison

Why people skip the first article

Because it feels like paying twice for the same part, and because the prototype already worked.

That reasoning holds right up until the failed part is one of hundreds, all made identically, all wrong the same way. The prototype proved the design. The first article proves the process: the fixturing, the program, the material, the operator's reading of your drawing. Only the second one is what you're buying when you order a run.

The other reason is schedule. A first article takes a slot in the calendar and someone has to look at it. If your date came from a launch announcement rather than from a quote, that slot is the first thing to get cut, and it's the wrong thing to cut.

mid-run costs the run.

Are you production ready? Twenty-eight honest questions

Answer these yes or no. "Sort of" is a no.

Revision control

• Can you name the current revision out loud, and does everyone touching the project use that name?
• If the CAD model and the PDF drawing disagree, do you know which one wins?
• Is every change since the last build written down somewhere other than an email?

Requirements

• Are the few dimensions that control fit or function marked on the drawing, with everything else left to a general note?
• Is the material named by grade, or just by "aluminum" or "steel"?
• Is the finish specified, including which surfaces get it and which are masked?
• Is what makes a part acceptable written down, or does it live in your head?

Validation

• Has a part made the production way been tested in the real use case, not just on a bench?
• If the prototype was printed and the production part will be machined, molded or formed, has anything been re-tested since the process changed?
• Did the test include the assembly, or only the one part?
• Do you know which failure would be most expensive, and have you deliberately tried to cause it?

Quality

• Is there a definition of a good part that someone other than you could apply?
• Can that definition be checked with a measurement, or does it rely on judgment?
• Do you know what happens to a failed part: scrapped, reworked, or shipped with a note?
• Who makes that call?

Packaging

• Has anything shipped and arrived intact, or is that an assumption?
• Does the packaging protect the feature most likely to be damaged, or just fill the box?
• Is packaging in your per-unit cost, or a line item you haven't priced?
• If the product ships through a retail or marketplace channel, do you know that channel's packaging requirements?

Packaging performance is testable rather than guessed at: ISTA Procedure 3A and ASTM D4169 are the standard practices for simulating distribution hazards. If your buyer requires qualified packaging, say so at quote time, because it changes the scope.

Economics

• Do you know your per-unit cost at your actual quantity, including finishing, packaging and freight?
• Does the number still work if the quantity turns out to be half what you hoped?
• Have you priced the second run as well as the first?
• Are you counting the parts you'll scrap?

Schedule

• Is your date derived from a real quote, or from a launch you already announced?
• Does the schedule contain time for a first article and time to fix what it finds?
• Is someone waiting on your approval, and do they know it?
• If a revision lands mid-run, do you know what that does to your date?

Which gaps block. Revision control, requirements and validation are blocking. Everything downstream is built on them, so a gap there propagates into every part you make. Packaging, economics and schedule gaps are expensive rather than blocking. You can start while you close them, provided you know they're open and you've said so.

One more that isn't ours to answer: for a consumer product, the conformity obligations sit with you as the manufacturer or importer, not with your fabricator. Children's products carry additional third-party testing and labeling requirements. Find out which rules apply before the run, not after.

If your buyer requires documented validation

Some customers arrive with a supply-chain requirement rather than a drawing. It's worth knowing what the words mean.

• T0 samples with a CMM report. The first samples off the production intent, measured on a coordinate measuring machine and reported against the drawing. Available on request, at additional cost, produced through our partner network.
• EVT, DVT, PVT. Three stage gates. Engineering validation asks whether the design works at all. Design validation asks whether it meets every stated requirement when built with production-intent parts. Production validation asks whether the process makes it consistently. Available on request, at additional cost, through our partner network.
• PPAP. A production part approval package: the evidence a supplier hands a customer to show the process makes a conforming part repeatably. Common in automotive supply chains. Available on request, at additional cost, prepared through our partner network.

AS9102 is the aerospace standard that defines first article inspection as a documented practice. It's named here so you know what your buyer means when they cite it, not because it describes something we perform.

accurate measurement tools. When a drawing calls for something tight, the job goes to a partner who can supply a CMM report, which usually costs less than trying it twice.

Which one should you ask for

"I have a sketch and I'm not sure the mechanism works." Ask for a prototype, and say it's a proof of concept. Let it be rough. You're buying an answer, not a part.

"It works, and I'm about to commit money to tooling, inventory or a promise." Ask for a first article before the run, and name the revision you want proved. This is the moment the first article was invented for, and the moment people skip it.

"I've bought these before and I need more." Ask for a production run against a named revision. Send the revision name, not just the part name. If anything changed since last time, say so explicitly. A silent change discovered at inspection is the most expensive kind.

What to send us

Send the STEP file, a PDF drawing alongside it, and one sentence about which of the three you want. If you're not sure, say that too. That sentence changes how we quote it.

Uploading a file isn't a commitment. It starts a review. A person confirms feasibility, pricing, capacity and timing before we commit to anything.