There is a recurring mistake in development programs: validating each component separately and assuming the final assembly will work if every part meets its individual tolerance target. On paper, the logic feels correct. In real injection conditions, it often is not.
The differential shrinkage issue
When two parts made from the same resin have different geometries, their real shrinkage behavior diverges. Datasheet values are averages. Effective shrinkage depends on:
- Geometry and local wall thickness.
- Fiber orientation in reinforced materials.
- Mould temperature and thermal balance over the cycle.
- Filling, packing and cooling conditions.
The result is familiar: parts that pass dimensional checks individually but show gaps, interferences or unstable closure once assembled.
What CAD does not show
CAD models represent ideal geometry. Even with advanced simulation, there is still a gap between prediction and production reality. CAD does not physically assemble real parts coming out of real cycles with real process variation. It does not reproduce 100% of snap-fit behavior under real injection conditions. It does not fully capture accumulation effects across multiple interacting parts.
That is why there is one irreplaceable test: manufacture the full set in final material and assemble it physically.
When this validation should happen
The first full assembly validation should not happen at production mould kickoff. It should happen during prototype phase, while there is still room to:
- Adjust functional tolerances.
- Modify critical wall transitions and thickness strategy.
- Correct rib and gate positioning.
- Refine design decisions before freezing tooling.
Real cost impact: a prototype-stage correction usually means machining hours on insert tooling. The same correction in production tooling can mean weeks of delay and major unplanned cost.
The right question before production tooling launch
Before releasing production tooling for assembly parts, the key question is not only: “Does this part meet tolerances?”
The strategic question is: “Have we physically validated that the full assembly works with all parts injection-moulded in final material?”
If the answer is no, the risk is still present and has most likely been moved to the most expensive moment to fix it.