One reading at a time? Paste them in production order and see whether the process is stable and capable, and why.
Counting defects or defectives instead of measuring? SPC Engine also covers p, np, c and u charts. Open SPC Engine
Control limits use ±3σ from the centerline, with σ estimated as MR̄/1.128 (the standard Individuals chart convention). Control chart rules follow the Nelson Rules (1-8). Anderson-Darling: AD is reported as in standard statistical software; its p-value uses the Stephens (1974) small-sample adjustment.
The Individuals and Moving Range (I-MR) chart is for one measurement at a time: one part per hour, one batch result, one reading per shift. The I chart shows each reading against its control limits; the MR chart shows how much the process moves from one reading to the next.
Enter the readings in production order. Twenty to 25 readings are a sensible minimum for judging stability, and 30 or more for capability. If you measure several parts per sample, use the X̄-R or X̄-S chart instead.
Control limits, tests for special causes and capability indices follow the AIAG SPC manual and Montgomery, and the results are tested against reference data sets analysed in specialised statistical software. Normality is checked before capability is reported.
The data behind “Try an example”: 15 readings, specification 23.0 to 25.0.
| Result | Value |
|---|---|
| I chart | CL 24.167 · UCL 25.098 · LCL 23.236 |
| Tests for special causes | No signals: stable |
| Cp · Cpk | 1.07 · 0.90 |
| Pp · Ppk | 1.29 · 1.08 |
The process is stable, but Cp is already below 1.33, so the spread has to come down before centring alone can make it capable.