Issue link: https://iconnect007.uberflip.com/i/1185676
74 PCB007 MAGAZINE I NOVEMBER 2019 would be approximately 1,000 g. Figure 17 shows that the means for all the results are very similar. The fluctuation for the CNF gold is greater, but the data set is also bigger. Despite the pull strength being very similar in the fracture modes, they are different (Fig- ure 20). The CN 1 plated on HP performed the best with 80–90% mode 2 fractures. Mode 2 failures represent fractures that occur above the intermetallic compound. Interestingly, it can be seen in Figure 18 that two distinct performance levels have oc- curred. There is a disparity between CN Au1 and the other two immersion gold types. The disparity may be linked to the nickel types on which they are plated, as they are also distinct. In Figure 19, the types of nickel that are used are also evaluated. The results are statistically different, identified by the probability value; however, the difference is minimal in practice. The impact of the nickel was not evident in terms of pull strength but is significant in terms of fracture mode. The result for the HP nickel is superior to that of the MP nickel. The impacts of the nickel-plating solutions were probably the key contributors to the skewed results in Figure 18. Ball Shear Testing As the name implies, ball shear testing dif- fers with regard to CBP because it employs a horizontal lateral force. Additionally, the force generated is greater due to the speed of the transducer. Figure 21 represents a diagrammat- ic version of the equipment setup. Ball shear testing is an alternative to CBP that really focuses on the IMC formation. The fracture mode is determined by the amount of the IMC that remains after the shear (Table 2). SJR performance is affected by multiple fac- tors. The shear strength for ball shear testing, similar to CBP, is affected by the SRO area. Fur- ther, the test can have an expected minimum average value for a 380-μm SRO. This value Figure 17: Physical conditions for the cold ball pull test. Figure 18: Failure classification for the cold ball pull test. Figure 20: A comparison of the CBP fracture modes for the gold types. Figure 19: ANOVA, comparison of the CBP pull strength for the gold types.