Control Limits: Calculate Upper and Lower Limits

Upper and lower control limits are calculated from process data and a chart-specific formula; they show the range of variation expected when a process is stable. Enter the observed measurements or subgroup statistics, along with the number of observations per subgroup where required, to calculate the limits and center line.

Formula

There is no single formula for every control chart: limits depend on the chart type and the data structure. For individual measurements, this calculator uses an Individuals (I) chart with a moving range of two consecutive observations.

Center line: CL = x̄. Upper control limit: UCL = x̄ + 2.66 × MR̄. Lower control limit: LCL = x̄ − 2.66 × MR̄.

• x̄ is the mean of the individual observations, in the same unit as the measurements.
• MR̄ is the mean of the absolute differences between consecutive observations, in that same unit.
• 2.66 is the conventional Individuals-chart factor when each moving range contains two observations.

For this calculator, the entered subgroup size is used to identify whether the data represent individual observations: use 1 for an I chart. The formula applies to numeric measurements collected in time order. Other chart types require other inputs and constants.

Worked Example

Using the default observations, 10, 12, 11, 13, 9, 10, 12, 11, 14, and 8, with subgroup size 1:

1. Calculate the mean: x̄ = (10 + 12 + 11 + 13 + 9 + 10 + 12 + 11 + 14 + 8) / 10 = 110 / 10 = 11.
2. Find consecutive absolute differences: 2, 1, 2, 4, 1, 2, 1, 3, and 6. Their mean is MR̄ = 22 / 9 ≈ 2.4444.
3. Calculate the limits: UCL = 11 + 2.66 × 2.4444 ≈ 17.502; LCL = 11 − 2.66 × 2.4444 ≈ 4.498.

Thus, the center line is 11, the upper control limit is about 17.50, and the lower control limit is about 4.50, in the observations’ measurement unit. Limits are not rounded in the formulas; chart software may display fewer decimals.

What Control Limits Mean

Control limits describe the observed process’s statistical variation, not a desired target or an acceptable product specification. A control chart plots measurements in time order against a center line and its limits. Points beyond a limit, or certain non-random patterns within the limits, can signal a change worth investigating.

Control limits are often set near three standard deviations from the center line under the chart’s statistical assumptions. That does not mean every point outside them proves a specific cause, or every point inside them proves the process is acceptable. Interpret signals using the chart’s rules and knowledge of how the process was measured.

Choose the Right Control Chart

The I-chart formula above is suitable for individual, continuous measurements when observations are collected in sequence and are not naturally grouped. Many processes instead produce rational subgroups: small sets of measurements taken under similar conditions. For those, an X̄-R or X̄-S chart is often appropriate.

• I-MR chart: individual measurements, with a moving-range chart to track short-term variation.
• X̄-R chart: subgroup averages and ranges, commonly used for smaller subgroups.
• X̄-S chart: subgroup averages and standard deviations, often used for larger subgroups.
• Attribute charts: counts or proportions, such as defective items or defects; use p, np, c, or u charts as appropriate.

Do not apply the I-chart constant to subgroup averages or defect counts. Their sampling distributions and limit formulas differ. If the process data are counts, proportions, or measurements grouped in batches, identify the data type and sampling plan before calculating limits.

Using Limits Responsibly

Before relying on calculated limits, check that observations are in time order, use consistent units, and come from a process measured in a consistent way. A limit computed from a short or unrepresentative period may not describe the process going forward. Investigate unusual points and process changes; do not automatically delete data just because they look inconvenient.

Common mistakes include confusing limits with specification limits, mixing measurement units, shuffling observations before calculating moving ranges, and rounding intermediate values too early. In an I-MR calculation, changing the order changes the consecutive differences and therefore changes the limits. The order should reflect when the measurements occurred.

These limits also rely on a reasonably stable process and suitable data. Strong trends, shifts, autocorrelation, or an unsuitable distribution can make simple limits misleading. A control chart is a monitoring tool, not a substitute for checking measurement quality or understanding the process.

Frequently Asked Questions

What is the formula for upper and lower control limits?
It depends on the chart. For an Individuals chart with moving ranges of two, UCL = x̄ + 2.66 × MR̄ and LCL = x̄ − 2.66 × MR̄.

Can the lower control limit be negative?
Yes, some formulas produce a negative limit. For measurements that cannot be negative, such as counts, a chart-specific method may set the limit to zero. Do not change an I-chart limit without considering whether its assumptions and chart type are appropriate.

Are control limits the same as specification limits?
No. Control limits are calculated from process behavior. Specification limits are requirements set for a product or service. A stable process can fail to meet specifications, and a process meeting specifications can still be unstable.

How many data points are needed?
There is no universal minimum that guarantees reliable limits. More representative, time-ordered observations generally provide a better basis for estimating process behavior. Follow the sampling plan and charting guidance appropriate to the process.

Why do my limits change when I reorder the data?
The I-chart limits use differences between consecutive observations. Reordering changes those moving ranges, even if the mean stays the same. Keep the actual time order.

What if the lower limit is below zero for a count chart?
Count-chart methods commonly constrain a negative lower limit to zero because negative counts are impossible. Use the formula for the specific count chart rather than the Individuals-chart formula.

Related

• Aluminum Plate Weight Calculator and Formula
• Aluminium Plate Weight Calculator and Formula
• Activity Coefficient Calculator: Formula and Meaning

list data | Observations | - | 10, 12, 11, 13, 9, 10, 12, 11, 14, 8
in n | Subgroup size | items | 1 | 1