Control Chart in Simple Terms
Every process varies a little. A control chart separates normal, common-cause variation from unusual, special-cause variation. Points inside the limits with no patterns mean the process is stable. A point outside the limits, or a pattern such as a long run on one side, means something has changed and should be investigated.
Control charts are the core tool of SPC. The limits, explained in UCL and LCL, usually sit three standard deviations from the centre line and are not the same as specification limits.
Common Chart Types
| Chart | Data type | Typical use |
|---|---|---|
| X-bar and R | Variable, subgroups of 2 to 9 | Machined dimensions sampled every hour |
| X-bar and S | Variable, larger subgroups | High-volume automated measurement |
| I-MR | Variable, single readings | Batch chemistry, slow processes |
| p and np | Attribute, defectives | Percent of parts rejected |
| c and u | Attribute, defect counts | Paint defects per panel |
Example
An X-bar and R chart on a shaft diameter shows seven subgroup averages in a row trending upward, although all parts are still in tolerance. The operator checks the tool and finds wear, changes the insert, and the process returns to centre before any parts are scrapped.
Common Mistakes
- Drawing specification limits on the chart and calling them control limits.
- Adjusting the process after every point, which adds variation.
- Recalculating limits too often, so shifts are absorbed.
- Charting data but not reacting to run rule signals.
The OperaHive SPC software supports X-bar, R, S and I-MR charts with alert rules. See the SPC and Cpk guide for more.