Tom Anderson
December 8th, 2000, 05:38 PM
[Originally Posted: 10/4/00--Transferred by ReliaSoft Moderator]
How would one deal with the following:
I need to study infant mortality, but cannot take interval times on the devices due to material restrictions. What I can do is used verified functional devices, run them in burn in for a period, test for pass/fail, then run the passing devices to some longer period (beyond which I should see no more infant failures) and test for a final total quantity of failures.
Test -> BI for X -> test -> additional BI -> test.
The final time will always be the same. I wish to vary X - with different sets of material (same product/processing) and establish the failure model (presumably Weibull with beta y fails, @X2 -> y2 fails (X2 > X1 y, @X2 (y1 - y) fails, etc.
Comments or I can answer questions with email
(tom325@ti.com)
Thanks
Tom Anderson
How would one deal with the following:
I need to study infant mortality, but cannot take interval times on the devices due to material restrictions. What I can do is used verified functional devices, run them in burn in for a period, test for pass/fail, then run the passing devices to some longer period (beyond which I should see no more infant failures) and test for a final total quantity of failures.
Test -> BI for X -> test -> additional BI -> test.
The final time will always be the same. I wish to vary X - with different sets of material (same product/processing) and establish the failure model (presumably Weibull with beta y fails, @X2 -> y2 fails (X2 > X1 y, @X2 (y1 - y) fails, etc.
Comments or I can answer questions with email
(tom325@ti.com)
Thanks
Tom Anderson