Authored By:
Jeffrey ChangBing, LeeCheng-Chih Chen, Alice Lin, Dem Lee
iST- Integrated Service Technology, Inc., TAIWAN
Gary Long, Consultant, USA
Masahiro Tsuriya, INEMI, JAPAN
Summary
Cold ball pull testing is used to validate the resistance of PCB pad cratering for the different ultra-low loss dielectrics materials (Dk=3~4.2 and Df <= 0.005 @ 1GHz) in the study. The materials were fabricated in multiple PCB shops using a common test board design utilizing a coupon to result in a 16 mil nominal pad size for the pulls. After fabrication, a 20 mil SAC305 ball is SMT attached to the 16 mil nominal pads for pulling.
Each material had 3 coupons with 50 pull locations on each to generate 150 data points for statistical analysis. The peak pull force differences of the material builds can be compared to differentiate the results. As a result, the different ultra-low loss dielectric material's performance to withstand PCB pad cratering can be compared comprehensively with the cold ball pull test.
Conclusions
Cold ball pull testing allows PCB shops to compare which dielectric materials have the best performance to withstand pad cratering based on their process control so as to reduce the field return rate due to pad cratering in PCB assembly and in service life. The comparison of data from different PCB shops is still not suggested due to process deviations at the different shops.
The test method was demonstrated to be reproducible when carefully executed and produced no anomalous results, suggesting it is a good practical test. It was not concluded that the material with certain cold ball pull strength will be capable of withstanding actual PCB pad cratering during shipping and service life of product, but was able to provide differentiation between materials at a single fabricator, thus allowing the end user to select the best candidate material for that PCB shop. The next steps will be to determine peak pull force thresholds at which pad cratering happens when materials are used in real service. Relating this work to results of other methods, such as impact testing, spherical bend testing, DMA, etc. is recommended to accomplish that task. Thresholds are expected to be different depending on product design and fabricator process control.
Initially Published in the IPC Proceedings
|
