Scenario
You're staring at an X-bar R control chart, seeing the 14th point shoot past the UCL, and the system alarm blares. The line leader rushes over, asking what to do, and the site is in chaos. A voice in your head says, "This machine has been running fine lately, could it be a false alarm?" You then mark that point as an "abnormal sample," exclude it, and continue production.
This action led to three extra hours of defective products.
Plain Language Explanation
The logic of control charts is simple: In a normal process, data points will fluctuate randomly between the UCL and LCL. Once a pattern emerges, it means an "external force" has intervened.
UCL (Upper Control Limit) and LCL (Lower Control Limit) are not specification limits. These two lines mean: "If the process is not abnormal, the probability of a data point falling outside these lines is only 0.27%."
So when you see a red dot on a control chart, it's telling you: Something has changed, and it's not random.
Common external forces include: shift change, change in material batch, temperature and humidity changes, tool wear, operator change.
Practical Diagnosis
Besides exceeding limits, there are other signals you cannot ignore:
| Pattern | Possible Cause |
|---|---|
| 9 consecutive points on the same side of the centerline | Process has a systematic shift, already drifting |
| 6 consecutive points steadily increasing or decreasing | Gradual deterioration, tool wear is most common |
| 14 consecutive points alternating up and down (zigzag pattern) | Alternating sources (two machines, two operators) |
Only looking for points exceeding limits will miss half the problems.
How InsightFab Does It
In JMP or Minitab, you have to manually set rules and remember the eight Western Electric rules for judgment. In InsightFab, after uploading data, all out-of-control signals are automatically detected and directly marked, for example, "Point 14: Upward trend (Possible cause: tool wear/environmental change)."
Golden Quote
"The red dot on a control chart is not an enemy; it's the process speaking for the first time—the question is whether you are listening."