Fault Sampling, Documentation and Trending

Fault sampling asks where, when and how a defect occurs. It should include affected and unaffected units rather than only the most dramatic example.

Choose positions, lots, sizes and chamber locations that test the suspected mechanism. Convenience samples can hide gradients or local equipment effects.

Record product, batch, time, location, photograph, measurement method and sample condition. Descriptive words should be tied to visible or measured evidence.

Protect samples from additional drying, warming or contamination and identify any destructive testing. The sample should remain traceable to the original lot.

Group recurring events by product, supplier, line, season and cause. A low-frequency fault can become visible only when records are combined.

The trend supports priority and effectiveness checks. Data volume alone is not useful unless categories are consistent and action thresholds are defined.