Mean Time Between Failures (MTBF): An indicator
of expected system reliability calculated on a
statistical basis from the known failure rates of
various components of the system. Note: MTBF is
usually expressed in hours. Of a system, over a
long performance measurement period, the
measurement period divided by the number of failures
that have occurred during the measurement period. For a population of items, during a measurement
period, the total functioning life of the population of
items divided by the total number of failures within
the population during the measurement period.
Note 1: The total functioning life of the population
may be calculated as the summation of the operating
life of every item in the population over the
measurement period. When computing the MTBF,
any measure of operating life may be used, such as
time, cycles, kilometers, or events.
Note 2: For
example, if a total of 1,000 events, such as data
transfers, radio transmissions, or system boots, occurs
in a population of items during a measurement period
of 100 hours and there are a total of 10 failures
among the entire population, the MTBF for each item
is (1000)(100)/10 = 104 hours.
There are two Different ways to calculate MTBF. Government programs use calculations per the latest version of MIL-HDBK-217, while commercial programs use the Bellcore method [Telcordia Technologies].
Engineering Note; An MTBF calculation may result in an anticipated failure rate of once every 10 years, but remember MTBF is an average. An MTBF of once each ten years could also mean twice in five years, or two failures in the first few weeks of operation, with correct operation for the remaining years. An MTBF figure that indicates operation with out failure for years predicts some number of failures without regard to when they occur [with in the time frame].
The best design has the highest MTBF or the lowest failure rate for the longest time. Once an MTBF is calculated use additional circuitry to account for any up-coming failure. Design in an Encoding format to account for circuit failure, as in CRC
