Reliability enhancement testing was the topic today. As an engineer who has worked in the reliability test lab for few days…. some of the points from Jones were very useful…unexpected failure modes are sign of bad test, test should be designed like a product (which we are yet to start), do not start a test just like that. The Duane test should be very useful in my day to day practices in our company, quite often we talk only in terms of MTBF to substantiate the life improvement but now some statistical evaluation is possible. Test to improve – fix effectiveness…has its roots on expert judgement…still a doubtful tool as this might be misguided due to the views of one person (more people idea might be confusing ). Testing exercise in the afternoon PAPER CLIP brought out the same issues which I have faced in the company….test results are having too much variation due to test differences,tester differences,test equipment,material…..blah blah. The life was predicted using an extreme accelerated testing (actual usage was 5deg while test was at 105deg), our team should have considered the actual usage first and then started the trials. Moreover we did not make any correlation between a customer usage and the test conducted….may be its just a syndicate exercise and this might not be required. But in a real situation this would have be IMPORTANT. Jane's doubt on assuming the the same trend on the other side of the graph ( measured at 105deg but predicting life at 5deg) was very obvious which did not gain the attention required. In the evening disucussion with TKJ on PPMC….hope will have good dreams.