I used to work at a semiconductor factory, and I can tell you that most (if
not all) semiconductors, from a simple diode up to a microprocessor are
extensively tested during most phases of manufacturing. In this way, by
rejecting bad units in early stages of manufacture materials and work is
saved. What good would it be to assemble a bad chip and at the end of the
process find out it is bad?
And failure rates are usually measured in ppms (parts per million). In the
plant I worked a goal was 100ppm, and we were actually doing about
180-250ppms.


Gabriel
TGO

-----Original Message-----

>Wow! I never expected that the manufacturing process for each individual
>chip involved so much testing! Is this, then, a major or maybe THE major
>share of the price of the IC? It would seem to me to be a fairly expensive
>undertaking. Also, in a typical consumer electronic device (take a
>microwave oven, TV, or clock, for example), is the initial failure rate
>less than 99.99% ??! Again, this is news to me. As a first year EE student
>and an avid electronics hobbyist, I have never had any of my designs go
>into production, but have considered trying several times. I guess that
>this is the reason why I am unfamiliar with the expectations of
>manufactured devices. It just seems that in my experience, I have seen at
>least several of the applicances which I have purchased fail prematurely or
>never work. If the rate were really 99.99%, I must be very unlucky. Then
>again, we engineers are especially cursed by murphy, right :)
>
>Sean