Two different questions
MTBF (Mean Time Between Failures) asks how long a machine runs before something stops it. MTTR (Mean Time To Repair) asks how long it takes to get running again once it does. A machine can have excellent MTBF and terrible MTTR — rarely breaks down, but a four-hour wait for the right spare part when it does — and the two numbers point maintenance in completely different directions.
A worked example
One machine, one calendar month, 680 running hours:
Total running hours
- Value
- 680 hrs
Breakdowns in the month
- Value
- 4
Total repair time across those 4 breakdowns
- Value
- 12 hrs
| Input | Value |
|---|---|
| Total running hours | 680 hrs |
| Breakdowns in the month | 4 |
| Total repair time across those 4 breakdowns | 12 hrs |
MTBF = 680 ÷ 4 = 170 hours between failures. MTTR = 12 ÷ 4 = 3 hours average repair time. Read together: this machine fails roughly once a week of running time, and each failure costs about half a shift to fix — two separate levers, and two separate conversations with two different teams.
Why both numbers, not just one
A plant chasing only MTBF ends up over-investing in prevention for a machine whose actual problem is slow repairs (missing spares, undocumented fault codes). A plant chasing only MTTR optimizes repair speed on a machine that would be better served by not failing in the first place. Reading both per machine — not as a plant-wide average, which hides which specific machine is driving the number — is what points at the right fix. See the full predictive maintenancepicture for what changes once a machine's failure pattern is actually visible.
Run your own numbers in the MTBF calculator and the MTTR calculator.