On Sat, 21 Jan 2012 12:34:13 -0000, "N_Cook" wrote:

snip

replacing that single switch was any different from replacing a single
triac, and how they felt that a board that was repaired by having that
switch put in, would be any more reliable than a board that had been
repaired by having a triac put in it. There was no answer, because they
really couldn't see the logic ...

To be fair, the guy that I deal with direct, was with me on this, but his
hands were tied from further up the food chain. What bothers me most is
what
happens to all of the faulty boards. Do they just become scrap destined
for
landfill ? Bit of a waste ...

Arfa



I've come across the concept of failure-rate variability over time etc but I
think the term of a bathtub plot is new to me.
Joins the hockeystick, J, bell, and dead-cat bounce

The bathtub curve has been around for a long time. Some about WW2 era
military reliability books were filled with it. It is still present in
Military reliability calculations.

I got see an interesting example of management think about reliability
some decades ago. There was this high energy density capacitor (at the
time) some 40 uF at 2 kV in a 1" diameter, 4" long package, not
electrolytic! In characterization testing it had about a 400 hour in use
life, it also had something like a 40 percent failure rate in the first 10
to 20 hours. The management started insisting on 168 hours 100 percent
burn in regimen and could not understand why they could not achieve 250
hour operational life. Nobody could convince them, even with test data,
that a 24 hour burn in would give them the 250 hour operational life.

Some Managers.

?-)