Control Charts in Manufacturing: Types, Rules, and When to Use Each

Control charts are the visual backbone of Statistical Process Control — the tool that transforms quality from detection to prevention. Developed by Walter Shewhart at Bell Labs in the 1920s, control charts have been used in manufacturing for over a century because they work. They give operators and quality engineers a simple, visual way to know whether a process is running normally or something has changed that requires attention.

This guide covers control chart types, selection criteria, interpretation rules, and practical implementation for manufacturing. For the complete quality management context, see our quality control manufacturing guide.

How Control Charts Work

A control chart plots a quality characteristic (measurement, proportion defective, defect count) over time. Three reference lines provide context:

• Upper Control Limit (UCL): Center line + 3 standard deviations
• Center Line (CL): Process average
• Lower Control Limit (LCL): Center line - 3 standard deviations

The 3-sigma limits are calculated from the process data itself — not from specifications. They represent what the process actually produces under stable conditions. When a point falls outside these limits or a non-random pattern appears, the chart signals that something has changed in the process.

Why 3-Sigma?

Three standard deviations capture 99.73% of normal process variation. This means that if the process is truly stable, a point will fall outside the control limits only about 3 times in 1,000 by random chance. When a point does fall outside, there is a 99.7% probability that something real has changed — a special cause is present.

Control Chart Types

Variable (Continuous) Data Charts

These charts monitor measurable characteristics — dimensions, weights, temperatures, pressures.

X-bar and R Charts: The most common control chart pair. The X-bar chart monitors subgroup averages (process centering) while the R chart monitors subgroup ranges (process spread). Used when you collect small subgroups (2-10 samples) at regular intervals.

X-bar and S Charts: Similar to X-bar and R, but the S chart monitors subgroup standard deviation instead of range. Preferred when subgroup size exceeds 10, as range becomes less efficient for larger subgroups.

Individual and Moving Range (I-MR) Charts: Used when subgroup size is 1 — each measurement is a single data point. Common when production is slow, testing is destructive, or only one measurement per batch is practical.

Attribute Data Charts

These charts monitor count-based quality data — defective/not defective, number of defects.

p-Chart: Monitors the proportion of defective items in a sample. Sample sizes can vary. Most versatile attribute chart.

np-Chart: Monitors the number of defective items. Requires fixed sample size. Easier for operators to understand than proportions.

c-Chart: Monitors the count of defects per inspection unit. Used when counting individual defects (scratches on a panel, solder defects on a board) rather than classifying entire units as defective.

u-Chart: Monitors defects per unit when the inspection area or opportunity varies.

Selection Guide

Interpreting Control Charts: Rules for Out-of-Control Signals

A single point beyond the control limits is the most obvious signal, but it is not the only one. The Western Electric rules (also called Nelson rules) define patterns that indicate special cause variation even when all points are within limits.

The Eight Nelson Rules

1. One point beyond 3-sigma: Immediate signal. Investigate now.
2. Nine consecutive points on one side of the center line: Process has shifted.
3. Six consecutive points trending in one direction: Process is drifting.
4. Fourteen consecutive points alternating up and down: Over-adjustment or two processes mixed.
5. Two of three points beyond 2-sigma on the same side: Likely shift.
6. Four of five points beyond 1-sigma on the same side: Likely shift.
7. Fifteen consecutive points within 1-sigma: Stratification — data may be from multiple sources.
8. Eight consecutive points beyond 1-sigma on both sides: Mixture of two processes.

For practical manufacturing use, Rules 1, 2, 3, 5, and 6 are the most important. Many shops start with just Rule 1 and add others as their SPC maturity increases.

Practical Implementation

Setting Up Control Charts

1. Verify measurement system: Run a Gage R&R study to ensure your measurements are reliable. A measurement system with more than 30% of total variation will produce unreliable control charts.

2. Collect baseline data: 20-25 subgroups under stable conditions. This means the same material lot, same operator, same machine settings. Remove any known special causes from the baseline data.

3. Calculate control limits: Use the appropriate formulas (or software) to calculate UCL, CL, and LCL from baseline data.

4. Verify baseline stability: Plot the baseline data against the calculated limits. If special causes are present in the baseline, investigate and correct them, then recollect baseline data.

5. Deploy for monitoring: Post charts at the operation (physical or digital display), train operators, and define response procedures.

Common Implementation Mistakes

Mistake: Using specification limits as control limits. Control limits are calculated from process data. Specification limits come from engineering. They serve different purposes and should never be confused.

Mistake: Adjusting the process after every point. If a process is in control, adjusting it based on normal variation (tampering) actually increases variation. Only adjust when a special cause signal is present.

Mistake: Too many charts, too few actions. A control chart that nobody responds to is decoration. Better to have 3 actively monitored charts than 30 that sit in a binder.

Mistake: Never updating control limits. When process improvements are made, recalculate limits. Old limits on an improved process will be too wide to detect deterioration.

Control Charts and Scheduling

Production scheduling and control charts interact in important ways:

• Schedule adequate setup time so operators can achieve process stability before running production
• Avoid constant job switching that prevents processes from settling into statistical control
• RMDB's finite capacity scheduling builds realistic time for first-article inspection and SPC sampling
• Track control chart events alongside schedule changes to identify scheduling-induced quality issues

Digital vs Physical Control Charts

Physical charts (posted at the workstation): High visibility, immediate operator awareness, simple to maintain. Work well for manual SPC.

Digital charts (Spreadsheet QC or dedicated software): Automatic calculations, historical trend analysis, alerting, and reporting. Better for multi-process monitoring and management review.

Most manufacturers benefit from both: digital charts for analysis and reporting, physical or large-screen displays at the workstation for operator visibility.

Frequently Asked Questions

Pair Control Charts With Smart Scheduling

Process stability starts with schedule stability. RMDB creates achievable schedules that give operators time for proper setups, SPC sampling, and process control. Track your quality data with Spreadsheet QC. Contact User Solutions to connect quality and scheduling.