Imagine a pale band appearing near the keyway after a plastic impeller is pressed onto its shaft. It does not wipe away, but the part still turns. Treat the change as assembly evidence that needs an explanation before release. Some plastics whiten under local strain. The colour alone does not quantify remaining strength or prove that a crack has already formed. Find when it appeared, where it began and what load the assembly process put there.
Relate the mark to the load path
A mark present before assembly can result from ejection, handling, deposits or colour variation. A mark appearing during pressing at a bore corner or shoulder warrants a check for interference, a sharp edge or misalignment. Local changes in structure can scatter light. Some polymers show this strongly while others can crack without visible whitening. Do not grade different compounds by the same shade. Preserve an uncleaned sample and document both bore entries, the root of the keyway and the extent of the mark under consistent lighting and magnification.
Measure the actual mating parts. Check shaft diameter and entry geometry, the impeller bore, and whether the key sits correctly. Record temperature and conditioning where the drawing makes them relevant. Check how the pressing tool supports the hub, whether it aligns with the shaft and whether there is a smooth lead-in. When force and displacement are available, locate the change along the stroke. An early peak at the entry, a rise when the key engages and a final rise at seating can have different causes even if they reach the same maximum force.
Build a comparison that can rule things out
In an illustrative pair, a visually clean impeller presses on smoothly and retains its colour. A second unit of the same drawing revision shows a sharp rise just before seating and a pale patch on one side of the keyway. Start with key height, trapped debris and alignment at the shoulder. Those observations do not yet justify switching every impeller to a more heat-resistant compound. Have a qualified team inspect a sacrificial sample with suitable magnification, a validated nondestructive method or a section where appropriate. Preserve the original anomaly. Continued operation in customer equipment is not a crack test.
Release requires an assembly criterion, an explanation consistent with the evidence and representative validation. Check adjacent cavities, shaft lots and fixtures for the same pattern. After a tooling change, visual improvement is only one observation. Where appropriate, validate hub torque transfer, dimensions and behaviour under representative starts and running loads. Isolate and verify power off before disassembling powered equipment. A resin supplier's elongation figure does not establish the condition or life of a specific moulded hub.
Separate chemical surface changes from installation evidence. If whitening begins after a cleaner is applied or develops during storage, record the substance, contact time and whether the part was under load. Check compatibility rather than comparing a washed sample with an untouched control as if only assembly differed. At a stressed hub, colour remains a screening signal; subsequent validation should reflect relevant liquids and temperatures. Note any different lubricant or mould-release residue used during processing. Marks concentrated around a bore do not rule out those variables. An orderly comparison preserves an unassembled part of the same lot and a verified assembly reference, so the investigation can distinguish a pre-existing surface condition from a change produced by pressing.
Frequently asked questions
Can I warm the mark until it disappears?
A colour change does not establish recovery of strength. Heating can change dimensions and material condition; use an approved process.
Is whitening without a visible crack a reject?
Use the drawing and inspection plan. Isolate an unexplained mark at a highly loaded location rather than treating the absence of an obvious crack as release evidence.
Why is only part of the lot affected?
Bore and shaft dimensions, chamfers, alignment and conditioning can vary. Compare by cavity and mating lot instead of using one random assembly as a verdict.
