The central idea
Replacing the burned-out part resolves the stoppage, but it does not explain why the part burned out. If the condition that damaged it is still there, the failure will come back. Root cause analysis follows the chain of causes backward until it reaches something that can be acted on in a lasting way.
Symptom, direct cause and root cause
| Level | Example in a motor |
|---|---|
| Symptom | The motor stops on a thermal trip. |
| Direct cause | The winding insulation degraded from overheating. |
| Root cause | The ventilation was blocked because the cleaning program did not cover that equipment. |
The root cause usually lies in a process, a design decision or an operating condition, not in the part.
How it is carried out
- Define the failure: what happened, when and under what conditions.
- Gather evidence before altering anything: photographs, records, condition of the parts, maintenance history.
- Propose hypotheses and test them against the evidence, discarding those that do not fit.
- Follow the causal chain to a cause on which action can be taken.
- Define corrective actions and then verify that the failure does not return.
Common tools
The five whys, the Ishikawa diagram and the fault tree are ways of organizing the reasoning. None of them replaces evidence: a cause is a root cause only if the facts support it.
When there is no single answer
Many failures have several concurrent causes. A manufacturing defect can be harmless until it coincides with an overload. In those cases, all the contributions are documented and action is prioritized by its effect and ease, rather than forcing a single cause.
A good sign that you have gone far enough is that the proposed action changes a method, a design or a condition, and does not merely ask people to “be more careful”.