A number like 2,000,000 hours on a datasheet does not mean your transmitter will last 228 years. This article shows what MTBF really measures, how an MTBF calculation works on plant records and how it feeds availability and safety numbers.
MTBF is one of the most quoted and most misread numbers in maintenance and safety work. Learn the MTBF calculation, the common traps and the link to FIT and PFD, with solved examples and a free calculator.
What Is MTBF Calculation?
MTBF calculation finds the mean time between failures of a repairable item by dividing its total operating time by the number of failures in that period. The result in hours also gives the failure rate λ, the availability when combined with MTTR, and the FIT value used in safety studies.
- MTBF = total operating hours ÷ number of failures
- Failure rate λ = 1 ÷ MTBF, in failures per hour
- MTTR = total repair hours ÷ number of repairs
- Availability = MTBF ÷ (MTBF + MTTR)
- FIT = failures per 10⁹ hours, so MTBF = 10⁹ ÷ FIT
The same arithmetic is used by maintenance planners, reliability engineers and safety engineers, only with different data. A planner uses CMMS work orders, while a SIL study uses failure rates from vendor certificates, and both start from a correct mtbf calculation.
TeepTrak quotes the MIL HDBK 338B definition of MTBF as the average operating time between successive failures of a repairable item. If the item is thrown away after one failure, the right term is MTTF, not MTBF.
MTBF Calculation Formula and Related Reliability Formulas
λ = 1 ÷ MTBF
MTTR = repair hours ÷ n
A = MTBF ÷ (MTBF + MTTR)
FIT = 10⁹ ÷ MTBF
R(t) = e raised to (minus t ÷ MTBF)
Here T is the total operating time of all units, not calendar time, and n is the number of failures in that time. TeepTrak points out that the simple reciprocal λ = 1 ÷ MTBF holds only for a constant failure rate, which is the flat bottom of the bathtub curve.
| Term | Formula | Unit | Used For |
|---|---|---|---|
| MTBF | Operating hours ÷ failures | hours | Repairable items |
| MTTF | Operating hours ÷ failures | hours | Non repairable items |
| MTTR | Active repair hours ÷ repairs | hours | Maintenance speed |
| MDT | All downtime hours ÷ failures | hours | Includes waiting and spares |
| λ | 1 ÷ MTBF | per hour | Failure rate |
| FIT | 10⁹ ÷ MTBF | failures per 10⁹ h | Electronic parts, SIL data |
| A | MTBF ÷ (MTBF + MTTR) | % | Uptime share |
Würth Elektronik states in its reliability data document that MTBF is not a measure of life expectancy. Its example of 1 million chargers running 6 hours a day with 12 failures in 5 days gives 30 million operating hours and an MTBF of about 2.5 million hours.
How to Do an MTBF Calculation Step by Step
Our solved example below uses one analyser shelter running all year. These are the same numbers loaded in the calculator, so you can check every line of the mtbf calculation yourself.
Operating time T = 8760 h
Failures n = 4, repair hours = 12 h
MTBF = 8760 ÷ 4 = 2190 h
λ = 1 ÷ 2190 = 0.000457 per h
MTTR = 12 ÷ 4 = 3 h
A = 2190 ÷ 2193 = 99.863 %
For a fleet MTBF calculation, add the operating hours of every unit before you divide. Ten identical transmitters running 8760 hours each give 87,600 unit hours, so two failures mean an MTBF of 43,800 hours even though no single unit ran that long.
Use operating hours from the historian or run hour counters for any MTBF calculation, not calendar hours. ReliaMag shows that using calendar time for a line that ran 1,040 hours in a quarter more than doubled the apparent MTBF in its example.
MTBF vs MTTF vs MTTR
Mean time between failures of an item that is repaired and returned to service.
Mean time to failure of an item that is replaced, never repaired.
Mean time to repair, covering only the active repair work.
Mean down time, including detection, permits, spares and restart.
- Simple division that anyone can check
- Works with ordinary CMMS work orders
- Shows trends when repeated every year
- Converts directly to failure rate and FIT
- Assumes a constant failure rate
- Unreliable with only one or two failures
- Hides wear out and early failures
- Says nothing about repair time on its own
ReliaMag warns that mixing MTTR with MDT is a common error, because MTTR covers only hands on repair while MDT includes waiting for permits and spares. Inherent availability uses MTTR, while operational availability uses MDT and is always lower.
Failure Rate, FIT and MTBF Conversion
Electronics and safety data usually give failure rates in FIT, which Würth Elektronik defines as the predicted number of failures per 10⁹ hours during the useful life region. The conversion is a single division, so the mtbf calculation from FIT takes seconds.
λ = 500 FIT
MTBF = 10⁹ ÷ 500
MTBF = 2,000,000 h
2,000,000 ÷ 8760 = about 228 years
That 228 years figure is the trap. It means that in a large fleet of such units, about one failure happens every 2 million unit hours, not that any single unit lasts two centuries.
The 36.8 % value comes straight from R(t) with t equal to MTBF, which gives e raised to minus 1. So with a constant failure rate, almost two out of three units fail before reaching the MTBF age.
How to Calculate Availability From MTBF and MTTR
Availability is the share of time an item is ready to work. With the calculator default numbers, 2190 ÷ (2190 + 3) gives 99.863 %, which is about 12 hours of downtime in a year.
| Availability | Downtime per Year | Typical Example |
|---|---|---|
| 99 % | 87.6 h | Single pump with slow spares |
| 99.9 % | 8.76 h | Well maintained analyser |
| 99.99 % | 52.6 min | Redundant DCS controller |
| 99.999 % | 5.3 min | Redundant network and power |
Shorter repairs raise availability as much as longer MTBF does. That is why DCS controller redundancy and redundant power supplies are so effective, since they cut the effective repair time to near zero.
How MTBF Links to PFD and SIL
In functional safety work under IEC 61508 and IEC 61511, only the dangerous undetected part of the failure rate, λDU, enters the PFD figure. A simplified single channel formula is shown below, and our PFDavg calculation article covers the full method.
λDU = 200 FIT = 0.0000002 per h
Proof test interval TI = 8760 h
PFDavg = λDU × TI ÷ 2
PFDavg = 0.000876
Falls in the SIL 3 band of 0.0001 to 0.001
This shows why a plain MTBF calculation is not enough for SIL work, because safe and detected failures must be removed first. The proof test interval and coverage then decide how much of that rate stays hidden between tests.
Architecture constraints from hardware fault tolerance and SFF also apply, so a good PFD number alone does not prove a SIL. Safe failures still matter too, as they drive the spurious trip rate and MTTFS of the function.
Power supplies often fail through wear out of electrolytic capacitors, which is not a constant rate failure. Our capacitor lifetime calculator helps you see when wear out will dominate over random failures.
Common MTBF Calculation Mistakes
| Mistake | Effect on Result | Correct Practice |
|---|---|---|
| Calendar hours used as T | MTBF looks too high | Use running hours from counters |
| PM stops counted as failures | MTBF looks too low | Count only unplanned failures |
| MTTR mixed with MDT | Availability looks too high | Report both separately |
| Failure definition changes | Trends become meaningless | Fix one definition and keep it |
| Too few failures in the window | Very wide confidence range | Use a longer window or fleet data |
| Constant rate assumed for wear parts | Wrong replacement timing | Use Weibull analysis for wear out |
When you have only one or two failures, quote a confidence range, not a single MTBF. TeepTrak recommends a chi squared interval at 90 or 95 % confidence, as described in MIL HDBK 338B.
Telcordia SR 332 and MIL HDBK 217F are two common prediction methods for electronics, and they can give quite different MTBF values for the same board. Always check which method and temperature a datasheet MTBF calculation is based on.
Good records also support predictive and preventive maintenance decisions, since a falling MTBF trend is an early warning. Tools such as vibration analysis with AI can then target the assets with the worst numbers.
Field Checklist for Reliable MTBF Data
- One written definition of failure for every MTBF calculation on site
- Operating hours taken from run counters or the historian
- Planned maintenance removed from operating time, not counted as failure
- Active repair time and total downtime logged separately
- Failure mode noted on each work order for SIL data
- Fleet hours summed for identical tags
- Confidence range quoted when failures are few
With clean records, every mtbf calculation on site becomes repeatable and easy to audit. The same data also feeds LOPA and SIL assessment work when you need site specific failure rates.
The Würth Elektronik document below explains FIT, MTBF and upper confidence levels with real component data. It is a good free reference for any mtbf calculation on electronic parts.
The video below gives a short visual recap of MTBF and MTTR with simple maintenance examples before you try the FAQ.
MTBF Calculation FAQ
Divide the total operating hours of the item or fleet by the number of failures in the same period. For example, 8760 operating hours with 4 failures gives an MTBF of 2190 hours.
Use running hours rather than calendar hours, and do not count planned maintenance as failures. A clean mtbf calculation depends far more on good records than on the formula itself.
MTBF applies to items that are repaired and put back into service, such as pumps or analysers. MTTF applies to items that are replaced after their first failure, such as fuses or sealed sensors.
Both are calculated as operating time divided by the number of failures. The difference lies only in what happens to the item after it fails in service.
Divide 10⁹ by the FIT value, because one FIT means one failure per billion device hours. So a part rated at 500 FIT has an MTBF of 2,000,000 hours.
To go the other way, divide 10⁹ by the MTBF in hours. Safety engineers often prefer FIT because rates for many parts can simply be added together.
Availability equals MTBF divided by the sum of MTBF and MTTR, shown as a percentage. With an MTBF of 2190 hours and an MTTR of 3 hours, availability is 99.863 %.
Use mean down time in place of MTTR if you want operational availability. That figure includes waiting for permits, spares and restart, so it is always lower.
No, MTBF describes the failure rate of a fleet during its useful life, not the life of one unit. With a constant failure rate, about 63.2 % of units fail before reaching the MTBF age.
Service life is limited by wear out parts such as capacitors, bearings and seals. Check the rated service life on the datasheet separately from the MTBF value.
There is no single good value, because it depends on the type of equipment and how failure is defined. A simple switch, a control valve and an online analyser will have very different typical values.
The useful approach is to repeat the same mtbf calculation for each asset type in your fleet and track the trend. A falling MTBF over a few years points to a problem worth investigating.
Related Articles
- PFDavg Calculation Formula
- Spurious Trip Rate and MTTFS in SIS
- Proof Test Interval and Coverage
- Hardware Fault Tolerance and SFF
- Predictive vs Preventive Maintenance
External References
- Würth Elektronik: FIT Rate, MTBF and Reliability Data
- ReliaMag: How to Calculate MTBF and MTTR
- Wikipedia: Mean Time Between Failures
What We Learn Today
- A correct mtbf calculation divides real operating hours by unplanned failures, and the failure rate is simply the reciprocal of that MTBF figure.
- Availability combines MTBF with MTTR, while FIT converts the same rate to failures per billion hours for electronics and SIL data.
- MTBF is a fleet rate, not product life, and only the dangerous undetected share of the rate enters the PFDavg used for SIL.
