Issue link: https://iconnect007.uberflip.com/i/1137649
70 SMT007 MAGAZINE I JULY 2019 These may be compromised solder joints. They may also indicate slight variance in the sol- der composition. More information would be needed to determine the reason for the slightly early failure. From Figure 25, the continuity failures cor- responded directly with the bulk failures for all boards tested. Interestingly, although these boards mostly failed at the board land pad-to- board interface, this material did not have the strongest response. From Figure 26, Sample B boards had similar behavior to the control material but one board (blue line) had an open very slightly ahead of the bulk solder failure. From Figure 27, the differences between the materials are less than the differences within each material. This indicates that there is lit- tle or no statistical difference. The slope and range of Material D needs to be studied further to determine if this is typical or an anomaly. Additional testing is underway. From Figure 28, Material B performed the best while Material D the worst. The failure modes were analyzed after dye and pry and cross-sectioning, and there were two primary failure modes. The control Sn3Ag0.5Cu material mostly failed below the Figure 25: Strain and daisy-chain response for four Sample A (control) boards. Figure 26: Strain and daisy-chain response for four Sample B boards. Figure 27: Summary of strain and daisy-chain response ranges for the four paste sample boards. Figure 28: Results after bending four boards of each paste type to failure.