Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.
|
|
|
|
Re: CI for non-normal bin
Posted:
Jun 14, 1996 8:27 PM
|
|
From: StatMan <statman@cdf.com>
>
> I'm fascinated. What sort of product allows for testing 100,000 pieces?
Easy: semiconductors! We produce incredible numbers of parts a month, and
are still doing 100% screening
in final test too.
Actually the customers *expect* defective parts to only occur at the 1/100,000
level nowdays. This level of
quality is, unfortunately, beyond the reach of normal SPC programs since the
"noise" there is considered to
be 3/1000 ....
I am presenting a paper in Chicago that shows how, at least in semicondustors,
to get rid of the quality
inspection and still give estimates of the outgoing quality at this level and l
ower. And no, it doesn't
depend on SPC..... And yes, if the process drifts it's still OK.... And yes
, in semiconductors I can tell
you how to reach any ppm defective level you wish - it's a piece of cake to ca
lculate. But more on that
during my 20 minutes (or is it 15?) in August.
Regards, Ed.
---
|
|
|
|
