Every reliability number starts with a count of failures. If two technicians disagree about what a failure is, their MTBF figures will disagree too, and nobody will trust either one. So the first job is not arithmetic. It is writing down, in one or two sentences, what counts as a failure for the machine you are tracking.
A workable definition for most production equipment is this: an unplanned stop that requires a maintenance intervention before the machine can run again. Planned cleaning, scheduled changeovers, and product-driven adjustments are not failures. A jam the operator clears in under a minute without tools is usually treated as a minor stop, not a failure, and is tracked separately.
Write the definition on the log sheet or in the maintenance system so it travels with the data. When a borderline event happens, decide once, record the decision, and apply it the same way next time. Consistency matters more than the exact wording.
Standard Procedure
- 1Write a one-sentence failure definition for the machine and post it where the log is kept.
- 2List the common stop types for that machine and mark each as failure, planned, or minor stop.
- 3Review the log at the end of each week and confirm every stop has been classified.
- 4When a new type of event appears, classify it once and add it to the list.
Lesson Summary
- ✓Agree on the failure definition before collecting any data.
- ✓Planned work and short operator-cleared jams are not failures and should be tracked separately.