Upper Control Limit Calculator

Compute the upper and lower control limits for a process control chart using UCL = mean + L(σ/√n), with the center line and control limits shown together.

Quick Facts

Formula
UCL = mean + L(σ/√n); LCL = mean − L(σ/√n)
Standard Shewhart control limits, defaulting to three sigma (L = 3).

Your Results

Calculated
Upper control limit (UCL)
-
mean + L(σ/√n)
Center line (CL)
-
Process mean
Lower control limit (LCL)
-
mean − L(σ/√n)
Standard error (σ/√n)
-
Spread of the plotted statistic

Ready

Enter your process mean and standard deviation, then calculate.

What the upper control limit is

The upper control limit (UCL) is the top boundary line on a Shewhart control chart — the tool used in statistical process control (SPC) to tell ordinary, expected variation apart from a genuine change in a process. Together with the center line (the process mean) and the lower control limit (LCL), it defines the band inside which a stable, in-control process is expected to operate. A point that lands above the UCL, or below the LCL, is a signal that something has changed and is worth investigating.

Control limits are not the same as specification limits. Specification limits come from the customer or the engineering drawing and say what the product must be. Control limits come from the process itself — they describe what the process actually does when nothing unusual is happening.

The formula

For a control chart, the limits are placed a fixed number of standard deviations away from the center line:

  • UCL = x̄ + L · (σ / √n)
  • Center line = x̄ (the process mean)
  • LCL = x̄ − L · (σ / √n)

Here x̄ is the process mean, σ is the process standard deviation, n is the subgroup size (how many measurements are averaged for each plotted point), and L is the sigma level — the number of standard deviations to the limit. The quantity σ/√n is the standard error of the plotted statistic. When you plot individual measurements rather than subgroup averages, n = 1 and σ/√n simplifies to σ, so the familiar three-sigma limit becomes UCL = x̄ + 3σ.

Why three sigma?

By far the most common choice is L = 3, the value Walter Shewhart recommended in the 1920s. It is an economic compromise, not a statistical law. If the process is stable and roughly normal, about 99.73% of points fall within ±3σ, so only about 0.27% — roughly 1 point in 370 — will fall outside the limits by pure chance. That keeps false alarms rare while still catching real shifts. Some practitioners add warning limits at ±2σ (about 95.45% coverage) as an early caution line.

Common reference points

  • ±1σ captures about 68.27% of a normal distribution.
  • ±2σ captures about 95.45% — sometimes used as warning limits.
  • ±3σ captures about 99.73% — the standard control-limit setting.
  • Averaging a subgroup of n = 4 halves the standard error (√4 = 2), so the limits sit half as far from the mean as they would for individual points with the same σ.

Frequently Asked Questions

What is the difference between the UCL and an upper specification limit?
The upper control limit is calculated from the process's own variation (mean plus L·σ/√n) and describes what the process naturally does. An upper specification limit is set by the customer or a drawing and describes what the product must be to be acceptable. A process can be in control (inside its control limits) yet still fail specification, or vice versa — they answer different questions.
Why does the subgroup size n divide the standard deviation?
When each plotted point is the average of n measurements, the standard deviation of that average is σ/√n, not σ — averaging reduces variability. So the limits for a chart of subgroup means are tighter than for individual readings. For an individuals (I) chart, n = 1, so σ/√n = σ and the limits are the full mean ± Lσ.
Can the lower control limit be negative?
For measurements that can go negative, yes. But for counts or naturally non-negative quantities (defects, time, dimensions), a computed LCL below zero simply means there is no meaningful lower limit — it is conventionally reported as 0. This calculator shows the raw arithmetic value so you can apply that judgment yourself.

Related calculators