Three different answers to "how reliable is it?" — and only one of them tells you what you actually experience as uptime.
MTBF gets quoted like a single scorecard number for reliability, and it isn't one. It answers a narrower question than most people assume, it only applies to a specific class of equipment, and by itself it says nothing at all about the number everyone actually cares about: how much of the time the system is up and working.
MTTF — Mean Time To Failure.Applies to non-repairable items: a light bulb, a disposable sensor, a one-shot pyrotechnic device. Once it fails, it's replaced, not fixed. MTTF is simply the average operating time before that permanent failure.
MTBF — Mean Time Between Failures. Applies to repairable items: a pump, a PLC, a production line. It assumes that after a failure the system gets repaired and returns to service, and the clock effectively restarts — MTBF is the average operating time between one failure and the next.
Availability. The fraction of total time the system is actually up. It depends on both how often the system fails (MTBF) and how quickly it gets repaired (MTTR — Mean Time To Repair): Availability = MTBF / (MTBF + MTTR). MTBF alone — how rarely something breaks — is only half the story.
Availability = MTBF / (MTBF + MTTR) is just the up-time fraction of a repeat cycle: the average time spent up, divided by the average total time per cycle (up plus the repair that follows). Holding MTBF fixed, availability falls as MTTR grows — a system that almost never fails can still spend a large share of the calendar year down if each failure takes days to resolve. Holding MTTR fixed, availability rises as MTBF grows. Neither number alone determines availability; only the ratio of the two does. This is why maintenance organizations track spare-parts stocking, mean time to diagnose, and mean time to repair with as much rigor as they track failure rates — cutting MTTR in half improves uptime exactly as much as doubling MTBF, and it's frequently far cheaper to achieve.
No — MTBF alone says nothing about availability without also knowing MTTR, and the two-system comparison above shows exactly why. Both systems fail with identical frequency (MTBF = 500 hrs), yet System B — which is just as reliable — delivers noticeably worse actual uptime purely because its repairs take longer. A system that fails rarely but takes days to fix each time can have worse real-world availability than a system that fails more often but is repaired almost instantly. A spec sheet that advertises MTBF without also stating (or bounding) MTTR is giving you half of an availability number, not the whole thing — always ask for both before drawing any conclusion about expected uptime.
MTTF, MTBF, and Availability answer three different reliability questions. MTTF (Mean Time To Failure) is the average life of a non-repairable item before it fails permanently. MTBF (Mean Time Between Failures) is the average operating time between failures of a repairable item that returns to service after each repair. Availability is the fraction of total time a system is actually up, and it depends on both MTBF and MTTR (Mean Time To Repair): Availability = MTBF / (MTBF + MTTR). A high MTBF does not guarantee high availability if repairs are slow.
MTTF applies to items that are discarded and replaced on failure — a light bulb, a disposable pressure sensor, a fuse. There is no "clock restart" because there is no repair; the item simply reaches end of life. MTBF applies to repairable systems — a pump, a compressor, a control system — where failure is followed by repair and a return to service, and the interval being averaged is the time from one such failure to the next. Confusing the two matters because MTBF implicitly assumes a maintenance action happens between failures; quoting an MTBF for a non-repairable part (or an MTTF for a repairable system) misstates what was actually measured.
Availability = MTBF / (MTBF + MTTR) is the fraction of a repair cycle spent operating. Because MTTR sits in the denominator alongside MTBF, the same MTBF can correspond to wildly different availability figures depending on how quickly repairs happen. This is precisely why reliability engineers distinguish reliability (governed by MTBF/failure rate) from maintainability (governed by MTTR) as two separate disciplines that both feed into the single number — availability — that operations teams actually care about.
This distinction drives maintenance strategy: investment in faster diagnostics, spare-parts stocking, modular swap-and-repair designs, and cross-trained technicians all target MTTR, while investment in preventive maintenance, better components, and redundancy targets MTBF. A facility comparing two candidate systems, or evaluating a maintenance contract, should always ask for both figures (or the resulting availability directly) — an MTBF quoted without a bounded MTTR is an incomplete basis for an uptime commitment or an SLA.
MTTF describes non-repairable items (average time to a single, permanent failure). MTBF describes repairable items (average time between failures of a system that gets repaired and returns to service). The formulas can look similar, but the underlying assumption — replace vs. repair — is different, and using the wrong term implies the wrong maintenance model.
MTBF treats each repair as returning the system to an operating state from which a new failure interval begins. This is a modeling simplification (it implicitly assumes repair restores the system close to as-good-as-new); in practice, repeated repairs or aging components can shift failure rates over time, which is why more advanced reliability models track failure rate trends rather than assuming a constant MTBF indefinitely.
No. MTBF measures only the operating (up) time between failures. MTTR measures the separate repair (down) time after a failure. Some conventions define MTBF as including the repair time, but the more common and clearer convention keeps them distinct and combines them only in the availability formula.
Yes — this is exactly the scenario the misconception addresses. If MTTR is very large relative to MTBF, availability drops even though failures are rare. A system that fails once every five years but takes three weeks to repair each time can have worse availability than a system that fails monthly but is fixed in an hour.
Sum the total up-time and total down-time over an observation period. MTBF is total up-time divided by number of failures; MTTR is total down-time divided by number of repairs. Availability can then be computed directly as up-time / (up-time + down-time), which is algebraically the same as MTBF / (MTBF + MTTR).
Try our Industrial & Systems Engineering Studio
More calculators, simulators, and guides for this discipline.