Issue link: https://iconnect007.uberflip.com/i/271756
March 2014 • The PCB Magazine 31 SPHERICAL BEND TESTING continues processes which must be optimized for both resin cure and finished st ack height. The res- in and glass systems themselves were worthy of inspection. There is a notable difference in the bundles of glass reinforcement fibers. In some lots they were very well defined el- liptical shapes while in others the glass was far more dispersed and difficult to identify in the filled resin system. This dispersed glass is implemented by some suppliers to improve the consistency of laser ablation drilling tech- niques. The variation in glass reinforcement is displayed in Figure 6. Results: DOE1 Previous work with more compliant systems has allowed us to record either the change in resistance that signifies failure at one of the sol- der interfaces or a localized minimum in the strain profile that indicates a laminate failure in either the board or the package substrate. These very stiff "high complexity typical" sys- tems do not store enough energy in the pack- age to create these events in the strain gauge signal at low strain levels. Strain events only oc- cur at higher levels of displacement where they are related to catastrophic failures and do not represent the first damage to the system. Alter- natively events may occur at low strain levels but the systems produce enough signal noise from vibration effects to make the events unde- tectable. In this experiment Mode 10, "Pad cra- ter" was not only the dominant failure mode under this test method, it was the only fail- ure mode detected in assemblies from standard "primary attach" processes. Destructive evalu- ation of samples subjected to increasing levels of displacement (flexure) determined that for this material and test setup there are three dis- tinct displacement zones where the material re- sponse can be defined. These zones are depicted in Figure 7 and are described as: • Zone 1: A safe zone where no damage occurs • Zone 2: A mixed zone where package corners both fail and survive • Zone 3: A zone where all package corners fail The intent of further testing was to define the boundary conditions for Zone 2 and gener- ate distributions that would allow extrapolation to safe working limits. We defined a "step stress" procedure to test groups of samples to progres- sively higher peak strains at fixed displacement rates. For each data point peak strain, rising strain rate and outcome were recorded. The dis- tributions of these estimates of survivable strain were used to generate working limits. Figure 6: Variation in glass reinforcement.