RussellMc <apptechnz <at> gmail.com> writes:
> What say you?

I'd say failed sample analysis where NumSamples < 100 cannot lead to conclu=
sions
about the real reason. Even then is NumSamples >100 but NumProducts >10000 =
very
little can usually be inferred about what is really broken in the product o=
r in
the manufacturing process. It's like that joke about the physicist,
mathematician and engineer seeing a black sheep on a hill.

That being said, omitting the correct (!) washers in screw assemblies can c=
ause
a lot of trouble later. There is no cure against cracked castings and casti=
ngs
which were designed wrong and will crack anyway due to stress or fatigue. A=
lso
production people will frequently "tweak" their tools (screwdriver torque
limiter) because then they go "better". Some places use seals to prevent th=
at.

I think that the extensive testing which is applied to life support and so =
on
subassemblies is designed to find just exactly this kind of problem. Not qu=
ite
like the Comet, but not necessarily far from it?

-- Peter


--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.