Issue link: https://iconnect007.uberflip.com/i/474043
40 The PCB Magazine • March 2015 With a small sample size it may be more prudent to choose the low- er preconditioning temperature as there will be less uncertainty in the simulated results. Success (zero failures) and success- failure (a mix of successes and fail- ures) reliability tests are based upon either nonparametric (binomial) or parametric (Weibayes) distributions. The well-known nonparametric bi- nomial success testing sample size formula is: Where: n = Sample Size C = Confidence level R = Reliability target The success-failure testing sample size equation must satisfy the follow- ing relationship: Where: Bi = Cumulative binomial distribution r = Failures R L = Lower confidence limit n = Sample size C = Confidence level To solve for n, one can use Goal Seek in Microsoft Excel. Table 5 lists the 95% lower con- fidence limits for success-failure test- ing with user defined sample size and failures in the last row and column. The general formula for paramet- ric Weibayes sample size is: RELIABILITy TESTING AND STATISTICS continues figure 7: Coffin-Manson equations in an Excel spread- sheet. Feature