Your prototype looked perfect. The moulded part does not.
That is not a surprise if you treat a printed, machined or soft-tooled sample as proof of production. It is proof of geometry under a different process. The steel, the polymer and the press will behave differently. That is expected.
What the prototype actually proved
A good prototype proves fit, feel and assembly sequence. It shows whether the snap clears, whether the gasket lands, whether the board still seats when the boss is a millimetre proud. That work matters. It is not the same as proving fill, pack, warp and ejection in the production polymer.
SLA (stereolithography) and similar resins give crisp surfaces and poor guidance on shrink. SLS (selective laser sintering) nylon is closer to a polyamide family, but the density, the fibre orientation and the cooling path are not those of an injected cavity. CNC (computer numerical control) machining from solid ignores knit lines, gate blush and the way a living hinge wants to be moulded, not milled.
Soft tooling and aluminium bridges close the gap. They still do not equal hardened steel with production cooling, production cycle time and the grade on the purchase order. Close is useful. Close is not T1 (first tool trial).
Where the moulded part usually diverges
Shrinkage. The prototype was built to nominal. The cavity was cut with a shrink factor for a named grade. If the first shots run a different melt flow, a different filler load or a colour package that changes crystallinity, the part moves relative to the CAD (computer-aided design) the prototype was cut from.
Gate and flow. The printed part had no gate. The moulded part has a freeze-off, a weld line and a flow direction. Bosses that looked solid in the print can sink. Ribs that felt stiff can warp when packing is uneven. Glass fibre that was never in the prototype now refuses to sit flat on a show face.
Ejection and draft. A machined sample can leave the fixture with zero draft and a sharp undercut that a mill tolerated. The moulded part has to leave the steel. If draft was “we’ll fix it in the tool”, T1 is where that debt is collected: sticking, witness marks or a steel change you did not budget.
Condition. Nylon prototypes may have been measured dry, conditioned, or never measured at all. The moulded nylon that ships, sits in a UK warehouse and takes on moisture is a different size from the dry first-off on the try-out bench (nylon moisture and late inspection). If the prototype “fitted” in a dry lab in June, that is not the assembly the customer will see in October.
The shelf model is a conversation. The dim report is the decision.
The DFM conversation the prototype skipped
DFM (design for manufacture) is not a courtesy review after the prototype demo. It is the list of features that will not survive steel: wall thickness steps that invite sink, shut-offs that invite flash, clips with no steel life, textures that only work on the right polymer, and tolerances that only a CMM (coordinate measuring machine) can argue with after the fact.
If the prototype was signed off by industrial design and sales, and manufacturing was invited later to “make it work”, the moulded part is not failing the drawing. The drawing is failing the process.
Name the production polymer before the soft tool, not after. Balloon the dimensions that matter to function before anyone books a press. Agree how those dimensions will be measured, and in what moisture state, before the first tray is called a success.
What to do with the prototype anyway
Use it. Fit check. Assembly sequence. Soft-touch overlays. Customer feedback on size in the hand. Just do not let it become the acceptance standard for the moulded part.
Write a short bridge list: which features the prototype can honestly represent, and which must wait for moulded samples in the production grade. Living hinges, snap fits that depend on resin toughness, optical surfaces, glass-filled stiffness and any seal land that moves with shrink belong on the second list.
When the moulded samples arrive, compare them to the ballooned drawing and the agreed inspection method, not to the prototype on the shelf. The prototype was for discussion. The dim report (dimensional inspection report) is for the decision.
Soft tooling buys learning. It does not skip a named production grade.
T1 is where the story ends or continues
Soft tooling buys learning. It does not buy permission to skip a ballooned drawing or a named production grade. If the soft-tool samples are used to freeze cosmetics and fits, write that down, and still re-prove those claims when the hardened tool runs the real polymer.
If the moulded part “doesn’t match the prototype”, the useful question is not who to blame. It is which assumption was never written down: grade, shrink, gate, draft, moisture, or cosmetic standard.
Treat the prototype as a rehearsal. Treat T1 as the first real run with production steel and polymer. If the rehearsal was in a different language, do not be shocked when the audience hears something else.
If you are about to freeze a design on the strength of printed or machined samples, and the tool quote is already in motion, we can walk the gap between prototype and moulded part in a short Manufacturing Snapshot, polymer, critical features and what must pass before the date stays still.