Issue link: https://iconnect007.uberflip.com/i/663184
April 2016 • The PCB Magazine 71 In this instance a linear and 2 nd degree polyno- mial fits were used certain the data sets. The re- gression fit used for each sample's (X-Y, Z-Core, and Z-Fill % Regression vs. Temperature, °C are presented on each plot in that order. The 1,000 hours test data can be used in a similar fashion to validate this conclusion and to forecast the lowest temperature (5,000-hour failure) for the final long-term aging test. These plots are pro- duced for each material independently. Solving the equation for y 50% yields a temperature that becomes the T4 (highest temperature) for the six-week test. Results for 50% degradation for each material are presented in Table 8. long-term thermal reliability of pCb materials Lack of degradation in dielectric strength for Laminates B and D prohibited calculation and ex- trapolation of 50% retention temperature. A test temperature 280°C was too high for Laminate E creating severe delamination. These data points were excluded from the regression analysis. 1000-Hour Testing Pre-screen analysis data was used to help guide selection of aging temperatures. The 1,000-hour test used more aggressive tempera- tures as bounded by the pre-screen data and UL746A procedure. Data was first analyzed in a similar fashion to the pre-screen data to iden- Figure 8: Plot of aging temperature vs. % retention in 500-hour pre-screen test. a: Calculated T4=181°C for laminate A. b: minimal to no degradation occurred for laminate B at temperatures selected. c: Calculated T4=212°C for laminate C. d: little to no degradation occurred for laminate D at temperatures selected. e: Calculated T4= 226°C for laminate E. a c e b d