Sample Size for Process Capability Studies: Minimum Requirements for Reliable Cpk
Sample size for process capability studies: the Cpk confidence interval formula, AIAG/ISO/Bosch standards, and worked numbers from n=30 to n=200.
Sample size for process capability studies: the Cpk confidence interval formula, AIAG/ISO/Bosch standards, and worked numbers from n=30 to n=200.
Honest taxonomy of free SPC software for quality engineers — trials vs OSS vs browser tools — and which fits which use case without 30-day clocks.
Reading control chart signals correctly: tell common cause from special cause, map each Nelson rule to a likely cause, and act without tampering.
CUSUM vs Shewhart charts for small shift detection: 1-sigma shift takes 44 subgroups on Shewhart vs 10 on CUSUM. Decision framework + EWMA middle ground.
Gage R&R workflow with AIAG %GRR thresholds (10/30%) and ndc criterion. Worked example computes %GRR = 30.5% and ndc = 4, plus what to do when MSA fails.
7 histogram patterns with spec limits and what they mean for process capability: centered, wide, off-center, bimodal, truncated, skewed, non-normal.
Step-by-step calculation of p-chart control limits with a worked PCB inspection example. Covers the variable sample size problem (exact, average-n, and standardized approaches), the np>=5 minimum, and the Laney p-prime chart for overdispersion.
When and how to use the I-MR (Individual Moving Range) control chart. Covers the five scenarios requiring individual charts, a complete worked example with control limit formulas (d2=1.128, D4=3.267), normality assumptions, and sensitivity comparison vs. X-bar R.
Decision tree for selecting the correct SPC control chart type. Covers variable vs. attribute data, subgroup size thresholds, p-chart and c-chart distribution requirements, CUSUM and EWMA alternatives, and common chart-type mismatches with real consequences.
Comparison of Cpk and Ppk process capability indices with worked example, formulas, and a decision table for which index to report in PPAP submissions, customer audits, and ongoing SPC monitoring. Includes diagnostic use of the Cpk-Ppk gap.
Decision framework for selecting between Western Electric Rules (4 tests, 1956) and Nelson Rules (8 tests, 1984) on SPC control charts. Covers false alarm rates, pattern recognition examples, and how to map each rule type to your reaction plan.
Step-by-step worked example for building an X-bar and R control chart from CNC manufacturing data. Includes A2, D3, D4 constants for n=5, control limit formulas, pattern rule application, and Cpk calculation.