Most medical device prototypes answer the wrong question well. A printed part tells you whether the thing fits in a hand; it tells you nothing about whether the molded part will hold a snap, seal a fluid path or survive a sterilisation cycle, because it is a different material made a different way. The useful discipline is naming the question first — form, function, or how the production part will actually behave — and building the cheapest prototype that answers it honestly.
01
Answers: Size, ergonomics, assembly order and whether the parts physically go together. Fast and cheap enough to iterate many times while the design is still moving.
Cannot tell you: Anything about the molded material. A print has no gate, no weld lines, no molded-in stress and different stiffness, so a snap fit or a thin wall that works in a print is evidence of nothing in PEEK or LCP.
02
Answers: How the actual grade behaves in the part — stiffness, wear, chemical exposure — and whether a critical feature can be made at all. Micro machining holds ±1 µm on machined features, so a machined prototype can test a geometry at its real size.
Cannot tell you: How the molded part will behave. A machined PEEK part has the crystallinity of the bar stock, not of a molded skin, and it has no weld lines and no gate. It is the right material made the wrong way, and the difference shows up in exactly the features micro parts depend on.
03
Answers: What the molded part actually does — shrink in each direction, where the weld lines fall, how thin walls fill, whether the snap survives. Only a molded part answers these, because they are properties of the process as much as of the material.
Cannot tell you: Stand in for validation. Molded prototypes are engineering parts for design verification and bench testing; they do not become validated product by being made from the production tool.
What this shop holds
ISO 9001:2015.
Not ISO 13485. No cleanroom. No FDA registration. Builds for clinical use, design validation or anything that will go into a patient run under the device owner’s quality system — that does not happen here.
What happens here is the engineering before that: the prototypes that settle the design, the material and the tool, so the validated build starts from a part that is already known to work.
A substitute resin shrinks differently, flows differently and fails differently. A dimension tuned on a stand-in grade has to be tuned again on the real one, and a feature that passed in the stand-in can fail in the grade that ships. If the device will be PEEK, Ultem or LCP, the molded prototype should be too — including the filled or unfilled version that production will run.
Hard-milling mold inserts directly in hardened steel means a prototype cavity can be made the same way as a production one. Then the prototype data describes the production part rather than an approximation of it, and if the design holds, the steel does not have to be re-cut and re-proven. A soft prototype tool is quicker to change and sometimes the right call early; it is the wrong call when the question is how the production part will behave.
Where the gate goes decides where the weld lines land, which direction a filled resin shrinks less in, and which thin section packs last. Those are the failure points of a micro part. Settling the gate in the prototype tool means the prototype is testing the part that will actually be made.
A prototype held to a loose tolerance proves the design works at a loose tolerance. If production needs ±10 µm on a molded feature, the prototype is the cheapest place to find out whether the resin’s shrink spread leaves room for it — and whether the feature can be measured repeatably at all.
Before a prototype tool is quoted, the shrinkage calculator will tell you whether the production tolerance survives the resin. When the question is ready, send the part or talk it through with Marve.