Posted by Rivs: Wed Jan 16, 2008 1:42 pm    Post subject: Electrical %shift / drift computation

guys need again your expertise I'm doing a electrical %shift / drift on our devices after reliability test and using the formula of (pre-test data - post-test data / pre-test data * 100) are you using the same formula as I am ? please help, is there any standards used for this?

Posted by FARel Engr: Thu Jan 17, 2008 3:39 am    Post subject:

Hi Rivs, what kind of device are we talking about and which particular parameter of this device is being measured for the shift/drift?

Posted by Rivs: Thu Jan 17, 2008 8:04 am    Post subject:

FARel thanks for the immediate response, I am talking about discrete devices such as FETs, Thyristors, Rectifiers etc . . . Actually all the electrical test parameters are being measured like breakdown voltages, leakage current, forward voltage, etc . . .

Posted by Paula: Thu Jan 17, 2008 9:27 am    Post subject: Re: Electrical %shift / drift computation

I can't recall but I think, logically, it should be (post-pre)/pre * 100. An increase will give you a + drift while a decrease - drift.

Back when I worked for discrete, we do drift analysis for all reliability tests. The good thing is our tester are capable of datalogging parameters, then we have a DOS (batch) program where we run this datalog and in a minutes you can get drift data of different parameters from a given reliability sample size simultaneously.

Posted by Rivs: Thu Jan 17, 2008 9:56 am    Post subject:

Paull thanks this is what I used before. I just want to have a copy of any standards that says so. Do you have or know one?

Posted by Paula: Thu Jan 17, 2008 1:24 pm    Post subject:

Try JESD86.

Posted by Rivs: Thu Jan 17, 2008 1:41 pm    Post subject:

Paula sorry for the typo error I'm in a hurry this morning, Thanks for the info I'll try to look at this