|
|
reliability and mtbfs
12 may 1997
don kulha wrote:
>my inverter is rated at 35 years mbtf (exeltech sine wave).
an mtbf of over 300,000 hours? unbelievable. in general, the fewer the parts,
the higher the reliability and mtbf. perhaps there is nothing inside this
exeltech box except an extension cord that comes out the back, with directions
to plug it in to the ac outlet on the wall of your neighbor's house :-)
here are some typical failure rates from mil-std 217, as quoted in martin
shooman's 1990 book, _probabilistic reliability, an engineering approach_,
in %/1000 hr:
diode 0.02 (wattage derated to 40%, at low temperature)
transistor 0.05 (")
composition resistor 0.01 (")
wire-wound resistor 0.10 (")
paper capacitor 0.01 (mil-c-1415 7 b high-reliability capacitors, with
70% operated/rated voltage ratio, at 20 c, with
relative humidity < 60%)
tantalum capacitor 0.01 (same, but with 60% derating)
transformers 0.20 (at low temperatures, with best insulation)
total 0.40 %/1000 hours.
so a box containing only 7 parts, a single one of each of these fine,
extremely expensive, high-reliability military components would have an
mtbf of about 1/(0.004/1000) = 250k hours, not quite as good as the exeltech.
the tiros weather satellite contained about 5,000 parts and had an mtbf of
about 2,500 hours. the mariner venus probe had about 7,500 critical parts, with
an mtbf of only 330 hours. the mtbf for a commercial avionics vhf radio
containing about 1,000 parts is about 1,000 hours. a simple commercial
electronic power supply might have an mtbf of 50,000 hours, about 6 years.
reliable systems cost more. shooman says "a 40,000 hour mtbf is seen to be
about the best figure one can expect [for a communications satellite.] if
the reliability were at the level of the mariner or ussr venus probe, the
operating costs would be nearly $1 billion per year."
here's a problem from the book:
try to estimate on the basis of your own experience and reading,
the reliability of
(a) your tv set (b) your automobile
(c) your washing machine (d) your home heating unit
(e) a satellite rocket booster (f) your wristwatch
(g) a telephone
clearly state your assumptions concerning the usage factor of each device.
compute the mean time between failures.
nick
|
|
