Issue link: https://iconnect007.uberflip.com/i/474043
34 The PCB Magazine • March 2015 data set and then plotted. The percent (Y-axis) mimics the shape (β) parameter, with the scale (η) is at the 63.21% percentile. This is shown in Figure 3. The beta (β) value is 5.4, indicating we're in the wearout zone of the product's life. If we're well beyond the design life of the parts, then we're happy. If we're within the design life, there is reason for concern as our parts are likely to fail during their useful life period. Let's look at another example. In this case, we're in the lab and have formulated a new electroless copper and want to compare it to an industry standard. Test vehicles are processed through the new electroless copper and then subsequently finished in a printed circuit board shop. Results of the thermal cycling data are shown in Table 3. As can be seen in table 3, there are no fail- ures; all 10 coupons were censored at 1,000 cycles. Because we have good historical data RELIABILITy TESTING AND STATISTICS continues figure 2: weibull plot of the thermal cycle data from Table 2. Table 3: pTh thermal cycling data on new electroless copper. on this coupon de- sign, the shape (β) and the scale (η) are well known (3.43 & 580.5 respectively) so we can use Weibayes analysis on this data set. Because we have no failures, the Weibayes analysis will give us a probability plot with only the lower 95% con- fidence interval. This can be seen in Figure 4. Interpretation of the analysis is simple. The historical scale (η) is 580.5, and our new electroless copper scale (η) is 1,421.4 (95% lower bound), nearly Feature