The X-bar R control chart tracks whether your manufacturing process remains stable over time. By observing subgroup averages and ranges, you can identify if external factors are causing production drifts or inconsistent output, allowing you to address issues before they impact your final product quality or yield.
Understanding the Dual-Chart Approach
The X-bar R control chart utilizes two distinct visual layers to monitor your process. The X-bar chart tracks the average of each subgroup to identify shifts in the process mean, while the R chart tracks the range of each subgroup to detect changes in overall process spread or consistency.
By evaluating these two metrics simultaneously, you can differentiate between a process that has moved off-target and one that has simply become more erratic. This provides a complete picture of your process health rather than relying on a single indicator.
Detecting Instability with Nelson Rules
Lattice automatically applies the Nelson 8 rules to your data, which are established patterns used to identify non-random behavior. These rules flag signals such as a single point exceeding control limits, trends showing consecutive increasing points, or patterns suggesting measurement system issues.
When any of these rules are triggered, the platform flags the data as having an 'out of control concern.' This acts as a circuit breaker, preventing you from making incorrect assumptions about your process capability based on unstable data.
The Importance of Data Normality
Standard control charts rely on the assumption that your underlying process follows a normal distribution. Lattice performs an Anderson-Darling test to verify this assumption before generating your charts.
If your data fails the normality test, the platform will flag it with a 'non-normal concern.' In such cases, the standard control limits may not accurately represent your process, and you may need to consider alternative analytical approaches.
Ensuring Reliable Estimates
For control limits to be accurate, you need a sufficient number of subgroups. We recommend a minimum of 25 subgroups to ensure that the process variation is captured correctly. Using too few subgroups can lead to unstable estimates, making it difficult to distinguish between natural noise and actual process changes.
If you provide fewer than 25 subgroups, Lattice will flag a 'low subgroups concern' to remind you that your control limits may shift significantly as you gather more operational data.
1 · Intent → method
An LLM picks spc_xbar_r_chart from a fixed catalog.
2 · Method → numbers
Deterministic Python engine runs the math. Same input → same output.
3 · Numbers → plain language
A second LLM translates the result into your domain’s vocabulary.
Why is the X-bar R control chart a prerequisite for calculating Cpk?
Cpk calculations assume your process is stable and predictable. If the X-bar R control chart shows that your process is 'out of control' due to特殊原因 (special causes), your Cpk results will be misleading and unreliable.
What do I do if my data triggers a Nelson rule violation?
A Nelson rule violation indicates that your process is no longer behaving randomly and is likely experiencing a specific disturbance. You should investigate the source of the variation—such as tool wear, operator changes, or raw material shifts—before proceeding with further process analysis.
Tool input schema
Schema for spc_xbar_r_chart not exported yet (run pnpm export:registry).