MTBF vs MTTR: Definitions, Formulas, and How to Use Them
MTBF vs MTTR explained — what each reliability metric means, the formulas with worked examples, how they relate to availability, and how to improve them.
MTBF and MTTR are the two numbers that every maintenance engineer and controls engineer gets asked about first when reliability is on the table.
This guide gives you the formulas, worked examples using real fault-log data, and practical steps to improve both metrics on a PLC-controlled line.
What Is MTBF (Mean Time Between Failures)?
MTBF — Mean Time Between Failures — is the average operating time between one failure and the next on a repairable asset.
It is a measure of reliability. A higher MTBF means the equipment runs longer before it fails. A lower MTBF means failures are frequent.
MTBF applies to repairable systems: a conveyor drive, a pump, a robot axis, a PLC I/O module. It is not the right metric for consumable components that are replaced rather than repaired (use MTTF for those — more on this below).
The MTBF Formula
MTBF = Total Operating Time / Number of Failures
- Total Operating Time is the sum of all periods the asset was running and producing, not counting downtime.
- Number of Failures is the count of unplanned stoppages in the observation window.
Units: MTBF is expressed in hours. A conveyor with an MTBF of 720 h fails, on average, once every 720 operating hours — roughly once a month on a single-shift operation.
How MTBF Is Calculated from SCADA or CMMS Data
Modern SCADA and CMMS systems log fault timestamps automatically. To calculate MTBF from those records:
- Export the fault log for the asset and time period.
- Identify start and end timestamps for each failure event.
- Calculate total operating time: subtract all downtime periods from the calendar period.
- Count the number of distinct failure events.
- Divide total operating time by failure count.
Most CMMS platforms can produce this number in a standard report. If yours does not, the raw data is usually accessible via an OPC-UA historian or a SCADA trend export — see the PLC predictive maintenance guide for how to structure that data pipeline.
What Is MTTR (Mean Time To Repair)?
MTTR — Mean Time To Repair — is the average time required to restore a failed asset to full operating condition.
It measures maintenance efficiency. A lower MTTR means your team diagnoses, parts, and repairs faster. A higher MTTR means failures last longer and production losses accumulate.
MTTR covers the complete repair cycle:
| Phase | What it includes |
|---|---|
| Detection | Time from failure occurrence to alarm acknowledgement |
| Diagnosis | Fault isolation, PLC fault-code reading, troubleshooting |
| Parts retrieval | Walking to stores, waiting for a part to be shipped |
| Physical repair | Replacing or fixing the failed component |
| Verification | Testing, reset, proving the machine runs correctly |
All five phases add to MTTR. Reducing any one of them lowers the total.
The MTTR Formula
MTTR = Total Downtime / Number of Failures
- Total Downtime is the sum of all time the asset was unavailable due to failures in the observation window.
- Number of Failures is the same count used for MTBF.
Units: MTTR is expressed in hours (or minutes for fast-repair assets).
MTBF vs MTTF vs MTTR — Clearing Up the Confusion
These three acronyms are often used interchangeably but they are not the same thing.
| Metric | Full Name | Applies To | Measures |
|---|---|---|---|
| MTBF | Mean Time Between Failures | Repairable systems | Reliability (how long before next failure) |
| MTTF | Mean Time To Failure | Non-repairable items | Reliability (how long before it fails once) |
| MTTR | Mean Time To Repair | Repairable systems | Maintainability (how long to restore) |
The key distinction between MTBF and MTTF: MTBF includes the repair time in the cycle. When a machine fails, is repaired, and runs again, the full cycle is: operating time + downtime + operating time again. MTBF measures the operating portion of that cycle. MTTF is used for components like bearings or fuses that you replace rather than repair — once it fails, it is gone.
In most industrial controls and maintenance contexts, MTBF is the correct term for repairable assets such as drives, sensors, actuators, and PLCs.
Worked Examples: Calculating MTBF and MTTR from Fault Logs
Example 1: Conveyor Drive
A production conveyor ran for 30 days on a two-shift schedule (16 operating hours per day). The SCADA fault log shows five unplanned stoppages. Total accumulated downtime for all five events was 8.5 hours.
Total operating time:
30 days × 16 h/day = 480 h
480 h − 8.5 h downtime = 471.5 h operating time
MTBF:
MTBF = 471.5 h / 5 failures = 94.3 h
MTTR:
MTTR = 8.5 h / 5 failures = 1.7 h
The drive fails roughly every 94 hours and takes 1.7 hours to repair when it does.
Example 2: Robot Weld Cell
A robot weld cell ran 24/7 for 90 days (2,160 calendar hours). The CMMS recorded 12 failures with total repair time of 31.2 hours.
Total operating time:
2,160 h − 31.2 h = 2,128.8 h
MTBF:
MTBF = 2,128.8 h / 12 = 177.4 h
MTTR:
MTTR = 31.2 h / 12 = 2.6 h
This cell fails about every 177 hours (roughly every 7.4 days on a continuous schedule) and averages 2.6 hours per repair.
How MTBF and MTTR Determine Equipment Availability
Availability is the proportion of time an asset is available to produce. It is calculated directly from MTBF and MTTR:
Availability = MTBF / (MTBF + MTTR)
Using the conveyor example above:
Availability = 94.3 / (94.3 + 1.7) = 94.3 / 96 = 98.2%
Using the robot weld cell:
Availability = 177.4 / (177.4 + 2.6) = 177.4 / 180 = 98.6%
This availability figure feeds directly into the OEE availability component. If you are tracking how to calculate OEE, the availability percentage in that calculation is driven by MTBF and MTTR. A plant with 95% availability OEE component can lift that number by either increasing MTBF (fewer failures) or decreasing MTTR (faster repairs) — or both.
What Happens When You Improve One Metric
| Scenario | MTBF | MTTR | Availability |
|---|---|---|---|
| Baseline | 94 h | 1.7 h | 98.2% |
| Cut MTTR in half | 94 h | 0.85 h | 99.1% |
| Double MTBF | 188 h | 1.7 h | 99.1% |
| Both improvements | 188 h | 0.85 h | 99.6% |
Both routes to 99.1% availability require equivalent effort but very different strategies. Halving MTTR is often faster to achieve through maintenance process changes. Doubling MTBF usually requires a longer-term engineering effort.
How to Improve MTBF
Improving MTBF means making the equipment fail less often. The most effective approaches in a PLC-controlled environment are:
1. Shift from reactive to predictive maintenance Use PLC-collected data — motor current, vibration, temperature, cycle counts — to detect degradation before failure. A bearing that is starting to fail will draw more current and run hotter. Your PLC predictive maintenance program can trigger a work order before the unplanned stoppage occurs.
2. Apply reliability-centered maintenance (RCM) analysis RCM identifies the failure modes most likely to cause a stoppage and selects the most cost-effective maintenance task for each. Rather than maintaining everything on a fixed schedule, reliability-centered maintenance focuses effort where it has the most impact on MTBF.
3. Eliminate repeat failures When a failure recurs within a short window, it is usually a symptom of an unresolved root cause. Use fault-code history from your SCADA or PLC to identify assets with chronically low MTBF and perform a root-cause analysis.
4. Standardize lubrication, tensioning, and alignment The majority of mechanical failures on automated lines trace back to improper lubrication, belt tension, or shaft misalignment. Standardizing these tasks and tracking completion in a CMMS directly raises MTBF.
5. Use condition monitoring as an early warning system Technologies like vibration analysis, thermal imaging, and oil analysis detect failure precursors weeks before breakdown. Pairing condition monitoring with predictive maintenance is the fastest way to extend MTBF on high-criticality assets.
How to Improve MTTR
Improving MTTR means restoring the asset to service faster after a failure occurs. The controls and maintenance teams both have levers here.
1. Improve PLC fault diagnostics Generic fault codes like "Drive Fault" or "E-Stop" force technicians to manually diagnose the cause. Structured PLC alarm messages that include the faulting device, the fault type, and a step number dramatically reduce diagnosis time. See the alarm management best practices guide for implementation details.
2. Stage critical spare parts Parts retrieval is one of the largest hidden contributors to MTTR. Identifying the top ten failure modes per asset and pre-positioning those parts in a lineside kit reduces MTTR without any engineering change.
3. Create equipment-specific troubleshooting procedures A technician who arrives at a faulted cell should have a decision tree, not a blank notebook. One-page troubleshooting cards that map PLC fault codes to likely causes and corrective steps cut average diagnosis time significantly.
4. Use remote access for faster diagnosis PLC remote access tools allow an engineer to connect to the controller from the office or from home within minutes of an alarm. Combined with SCADA trend data showing the conditions leading up to the fault, the cause is often identified before the technician reaches the machine.
5. Track MTTR by failure mode, not just by asset Some failure modes are fast to repair; others are chronic time sinks. Breaking MTTR down by fault code reveals which specific repair procedures need standardization or tooling improvement.
For the broader strategy of how planned maintenance intervals affect these numbers, the preventive vs predictive maintenance comparison is a useful reference.
Limitations of MTBF and MTTR
MTBF and MTTR are powerful but they can mislead if applied without context.
MTBF assumes a constant failure rate. The formula produces a single average across an observation period. In reality, most equipment follows a bathtub curve: higher failure rates early in life (infant mortality), a low and roughly constant rate during useful life, and rising failure rates as the asset ages and wears out. A single MTBF number does not distinguish between these phases.
A high MTBF does not mean zero risk. An asset with MTBF of 1,000 hours has a meaningful probability of failing at hour 100. MTBF is an average, not a minimum. Do not use it as a guarantee that equipment will run to a fixed interval.
Small sample sizes produce unreliable numbers. MTBF calculated from two or three failures has wide statistical uncertainty. An asset that has failed twice in 600 hours has an estimated MTBF of 300 hours, but the true value could plausibly be anywhere from 100 to 1,000 hours. Treat low-failure-count calculations as rough indicators, not precise figures.
MTTR masks diagnosis variation. A single MTTR average can hide a bimodal distribution: most repairs take 30 minutes, but one repair per quarter takes 12 hours because a specialist is required. That rare long repair drives the average up without representing typical performance. Review the full distribution of repair times, not just the mean.
MTBF and MTTR do not capture production impact. A failure during a peak production window costs far more than the same failure during a planned maintenance window. Pair these metrics with a criticality ranking and a production-loss calculation to prioritize improvement effort correctly.
Frequently Asked Questions
What is the difference between MTBF and MTTR?
MTBF (Mean Time Between Failures) measures how long equipment operates between failures — it is a reliability metric. MTTR (Mean Time To Repair) measures how long it takes to fix the equipment after a failure — it is a maintainability metric. MTBF tells you how often problems occur; MTTR tells you how long each problem lasts. Both numbers together determine availability using the formula: Availability = MTBF / (MTBF + MTTR).
How do you calculate MTBF?
Divide total operating time by the number of failures in your observation window. Total operating time is calendar time minus all downtime. For example: if an asset ran for 500 hours and experienced 4 failures, MTBF = 500 / 4 = 125 hours. Pull the data from your CMMS work order history or your SCADA fault log for the most accurate figures.
What is a good MTBF?
There is no universal answer — a good MTBF is relative to the asset type, the criticality of the application, and your industry benchmark. A critical production bottleneck should target the highest achievable MTBF; a low-criticality auxiliary system warrants less investment. As a starting point, compare your current MTBF against the manufacturer's published MTBF for the component, and track trends over time to confirm whether your maintenance program is improving reliability.
How do MTBF and MTTR affect availability?
Availability is calculated as MTBF / (MTBF + MTTR). If MTBF is 200 hours and MTTR is 4 hours, availability is 200 / 204 = 98.0%. Doubling MTBF to 400 hours raises availability to 400 / 404 = 99.0%. Halving MTTR to 2 hours raises availability to 200 / 202 = 99.0%. Both improvements have the same effect on availability in this example, but different programs and costs are required to achieve each one.
