Issue link: https://iconnect007.uberflip.com/i/1137649
JULY 2019 I SMT007 MAGAZINE 69 Figure 23: Strain and daisy-chain response for four Sample D boards. Figure 24: Strain and daisy-chain response for four Sample C boards. at a constant target head speed of 4 mm/sec- ond, which resulted in a strain response of between 5k με/sec (microstrains per second) and 8k με/sec. One material (Material D) had a couple of anomalous readings where the daisy chain went open ahead of the bulk material failure. This is likely a weakened or compromised sol- der joint on the periphery of the package. This small sample set was to get an early first-level approximation. Subsequent testing is currently underway. We will be testing eight boards at three times the strain rate (15K με/sec). In addition, we will be doing drop testing on all materials as well. The differences between the pastes were less than the differences within the pastes but we were still able to rank the performance. The doped Material B slightly outperformed the Sn3Ag05Cu control, and the other materi- als were comparable or slightly lower than the control. Material D had the widest range. In the stress and continuity response curves (Figures 23–26), the solid lines are the strain measurement (inverse because the sensors were placed on the bottom) and the dotted lines (colors commensurate with each board) are the daisy chains. The dotted lines (daisy chains) rapidly go open when the circuit fails. Signal sampling is at four times the specified rate of 500 per second (measured at 2K per second). From Figure 23, two of the four boards tested failed before the bulk material. This material had the highest as well as the lowest response range (9311με and 7313με respectively). Fur- ther evaluation would be needed to under- stand the early failures. From Figure 24, one board (black line) had a slightly early failure (before the bulk failure). Figure 19: Typical QFN void- ing for Sn3Ag0.5Cu Paste A. Figure 20: Typical QFN voiding for Paste D. Figure 21: Typical QFN voiding for Paste B. Figure 22: Typical QFN voiding for Paste C.