A process produces a dimension with a known mean of 50.0 mm and a standard deviation of 2.4 mm. Samples are taken in subgroups of 5. Using $UCL = \bar{x} + 3\dfrac{\sigma}{\sqrt{n}}$ and $LCL = \bar{x} - 3\dfrac{\sigma}{\sqrt{n}}$, what are the control limits for the X-bar chart?
- AUCL = 52.4 mm, LCL = 47.6 mm, using 1 standard deviation instead of 3
- BUCL = 57.2 mm, LCL = 42.8 mm, using the standard deviation directly without dividing by the square root of the subgroup size
- CUCL = 51.07 mm, LCL = 48.93 mm, using a subgroup size of 50 instead of 5
- DUCL = 53.22 mm, LCL = 46.78 mmcorrect
Show the solution
Control limits for an X-bar chart are set at three standard errors of the SAMPLE MEAN away from the process mean, not three standard deviations of individual measurements — this distinction is the entire point of using subgroup averages rather than plotting raw individual readings, since averaging reduces variability by a known factor tied to the subgroup size. Standard error of the mean: $\dfrac{\sigma}{\sqrt{n}} = \dfrac{2.4}{\sqrt{5}} = \dfrac{2.4}{2.236} = 1.073$ mm. Three standard errors: $3(1.073) = 3.22$ mm. Control limits: $UCL = 50.0+3.22=53.22$ mm, $LCL = 50.0-3.22=46.78$ mm. These limits define the natural, expected range of subgroup AVERAGES when the process is behaving exactly as it has historically — a point falling outside them signals a likely special-cause variation worth investigating, which is a different and much narrower band than the natural range of individual parts the process produces, since individual parts vary by the full $\sigma=2.4$ mm, not the reduced $1.073$ mm that subgroup averaging achieves.