🚀 go-pugleaf

On Mon, 20 Jul 2009 13:48:51 -0700, "Joel Koltner"
<zapwireDASHgroups@yahoo.com> wrote:

>"John Larkin" <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote in message
>> Sure. Life is risky. Given a product with 200 different parts, two of
>> which are overstressed, there's probably more risk that one of the 198
>> other parts will be EOL'd.
>
>Have you ever been caught in the situation of one of your over-stressed parts 
>being changed by the manufacturer in such a way so as to render it unsuitable 
>for your needs?
>

Siemens used to make the CLY5 power mesfet. It was rated for about 9
volts or something but was good to 22 or so. They discontinued it and
TriQuint started making them, with a different process. The new ones
die hard at 15 volts. Luckily, the product that we used it on never
got beyond shipping a couple of prototypes.

Optek keeps changing their laser diodes, so you can't count on their
typical performance.

But in general we haven't had much trouble. EOL'd regular parts are a
bigger hassle because there are so many more "regular"-use parts in a
product. 

We do test parts to destruction to see what sorts of margins we have.
Some parts don't even survive their datasheet abs max ratings.

John