Skip to content
Omnikalc

EWMA control chart calculator

Enter the target, the process standard deviation, the weight λ, the width L and the subgroup size to get the EWMA control limits at start-up and in steady state.

Updated on
Your inputs

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Results

Upper limit in steady state

51.907

Lower limit in steady state
48.093
Upper limit at the chosen sample
51.638
Lower limit at the chosen sample
48.362
Width at the chosen sample compared with steady state
85.9 %
Effective memory (number of samples)
9.0

The EWMA control chart calculator gives the steady-state limits and the start-up limits of an exponentially weighted moving average chart from the target, the process standard deviation, the weight λ and the width L.

How the calculation works

  • Enter the target mean, the process standard deviation and the subgroup size.
  • Enter the weight λ and the width L.
  • Enter the sample number to get the start-up limits.
  • The calculator gives the steady-state limits, the limits at the chosen sample, their relative width and the effective memory.

Example

For a target of 50, a standard deviation of 2 and single measurements, λ = 0.2 and L = 2.86 give a steady-state half-width of 2.86 × 2 × √(0.2 / 1.8) = 1.907, so the limits are 48.093 and 51.907. At sample 3 the limits are 48.362 and 51.638, which is 85.9 % of the steady-state width.

Why it matters

The calculator helps quality engineers, laboratory staff and analysts to set the limits of a chart that detects small drifts of a mean, for example in a filling process, an assay or a service level.

Good to know

Start the plotted value at the target, z(0) = target, and update it with each new sample. If a signal comes, look for a cause of a shift of the mean, then restart the chart. Estimate the standard deviation from the within-subgroup variation or from a moving range, not from the plotted means. The effective memory of about (2 − λ) / λ samples tells you how many recent samples drive the value. An EWMA chart works with individual measurements, which is an advantage when each sample is costly, and it is robust to non-normal data when λ is small. To catch large shifts as well, add a Shewhart chart with 3 sigma limits.

The control chart limits calculator sets Shewhart limits, the CUSUM chart calculator is another approach to small shifts, the exponential moving average calculator smooths a series and the p-chart calculator handles proportions.

Frequently asked questions

What is an EWMA chart?

An exponentially weighted moving average chart plots z(t) = λ x(t) + (1 − λ) z(t−1), a smoothed value that gives more weight to recent data. A small λ gives a long memory and good detection of small shifts, while λ = 1 gives back the Shewhart chart. A signal is given when z(t) leaves the control limits.

How are the limits computed?

The limits are the target plus or minus L σ √(λ / (2 − λ) [1 − (1 − λ)^(2i)]), where σ is the standard deviation of the plotted mean. For large i the bracket tends to 1 and the limits settle at the steady-state value. Using the start-up limits makes the chart more sensitive to a shift at the beginning.

Which λ and L should I use?

Common choices are λ between 0.05 and 0.25, and L between 2.6 and 3, with smaller λ paired with smaller L to keep the false alarm rate at the same level. A λ of 0.2 with L = 2.86 is a frequent compromise, and a λ near 0.05 is used for very small shifts of about half a standard deviation.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.