What Factors Influence CCI Sample Size?

The correct sample size for a given batch isn't determined by lot size alone. Process type, equipment architecture, defect behavior, and process maturity all shape which plan applies — and how samples must be drawn to provide meaningful coverage.

How does manufacturing process type affect sampling?

Process type determines the defect model, which drives sampling architecture:

• Manual hand-crimping introduces operator-to-operator variability and fatigue effects. When multiple operators contribute to a batch, samples must be drawn proportionally from each operator's output.
• Multi-head automated systems introduce station-specific risk. Each crimping head operates independently — a worn or misaligned station produces a systematic pattern of failures invisible to any plan that doesn't guarantee per-station coverage.

Why is station stratification non-negotiable for multi-head systems?

In a 12-station rotary crimper, a defective Station 7 affects every container it processes. If the sampling interval happens to skip Station 7 — or if sampling isn't designed to cover all stations — the plan provides no protection against that failure mode. Station stratification is required, with a minimum sample from each station regardless of what the standard lot-size table dictates.

What other defect patterns require structural responses?

• Clustered / transient defects appear at startup, after machine jams, or at end-of-run. Time-stratified sampling plus event-triggered supplemental pulls (5–10 units at startup, post-jam, end-of-run, operator change, and component lot change) provide protection.
• Systematic defects are process- or station-linked and require stratification to expose.
• Random defects are adequately covered by standard systematic or random sampling within the AQL framework.

How does process maturity affect the plan?

A process in a routine validated state, following successful PPQ, justifies AQL-based release sampling. That demonstrated capability is what makes reduced sampling defensible.