Control Charts (Statistical Process Control)

Control Charts monitor whether a process is in statistical control over time. Points outside control limits or non-random patterns signal special-cause variation that warrants investigation.


Step 1 — Choose chart type

Continuous (variables) data:

Chart Use when
X-bar-R Subgrouped data, small subgroups (n=2-10)
X-bar-S Subgrouped data, larger subgroups
Individuals (I-MR) Single measurements (subgroup size = 1)

Attribute (counts) data:

Chart Use when
p chart Proportion defective, variable sample size
np chart Number defective, constant sample size
c chart Defect count per unit, constant opportunity
u chart Defects per unit, variable opportunity

Step 2 — Results

  • Centre line (CL) — process average
  • UCL / LCL — +/-3 sigma boundaries
  • Estimated process standard deviation
  • Control chart plot
  • Out-of-control violations with Nelson Rule cited
  • In-control verdict

A UniversalChatBot is available for discussion.


Statistical methods used

X-bar-R formulas:

  • UCL(X-bar) = X-double-bar + A2 * R-bar
  • LCL(X-bar) = X-double-bar - A2 * R-bar
  • UCL(R) = D4 * R-bar, LCL(R) = D3 * R-bar
  • Sigma = R-bar / d2

Individuals (I-MR):

  • UCL/LCL = x-bar +/- 2.66 * MR-bar

Attribute charts: Based on binomial (p, np) or Poisson (c, u) distributions.

Nelson Rules checked: Rule 1 (beyond 3 sigma), Rule 2 (9 consecutive same side), Rule 3 (6 trending), Rule 5 (2 of 3 beyond 2 sigma).

Control limits at +/-3 sigma give ~0.27% false alarm rate per point.

In control != meeting specifications. Use Process Capability for that assessment.