Overview
Control limits on a control chart show the range a stable process normally stays within. Points outside them signal that something has changed.
Enter the process mean, the process standard deviation, the sigma multiplier (usually 3) and the subgroup size.
Worked examples
| Inputs | Result |
|---|---|
| Process Mean42Process Std Dev6Sigma Multiplier3Subgroup Size5 | Center line: 42UCL: 50.05LCL: 33.95Subgroup SE: 2.68 |
Show 2 more examples
| Process Mean100Process Std Dev2Sigma Multiplier3Subgroup Size4 | Center line: 100UCL: 103LCL: 97Subgroup SE: 1 |
| Process Mean10Process Std Dev0.5Sigma Multiplier2Subgroup Size1 | Center line: 10UCL: 11LCL: 9Subgroup SE: 0.5 |
Center line by process mean
With the calculator's starting values, the center line is 42. The table shows how it changes with the process mean.
| Process Mean | Center line |
|---|---|
| 30 | 30 |
| 40 | 40 |
| 42starting value | 42 |
| 50 | 50 |
| 60 | 60 |
Other inputs: Process Std Dev: 6 · Sigma Multiplier: 3 · Subgroup Size: 5. Change them in the calculator above for your own numbers.
How it works
For an X-bar chart, the limits are the mean ± k × σ ÷ √n, where n is the subgroup size. With 3-sigma limits, a stable process rarely produces a point outside them, about 0.27% of the time. Control limits come from the process, not from specifications.
The formula
| k | Sigma multiplier, usually 3 |
| n | Subgroup size |
Practical tips
- A mean of 42, σ = 6 and subgroups of 5 give limits of 33.95 to 50.05.
- For individual readings, use n = 1.
- Control limits aren't spec limits. A process can be in control and still make out-of-spec parts.
Frequently asked questions
Why 3 sigma?
It balances false alarms against missing real changes, a convention from Walter Shewhart.
What's the difference between control and spec limits?
Control limits describe what the process does. Spec limits describe what the customer needs.
How should σ be estimated?
Usually from the average range or standard deviation within subgroups.