Issue link: https://iconnect007.uberflip.com/i/911509
December 2017 • SMT Magazine 77 EVALUATION OF STENCIL TECHNOLOGY FOR MINIATURIZATION cepted benchmark for acceptable paste transfer is 80% TE, with CV of 10% or less. Both TE and CV were determined and plotted for devices 0.3 mm BGA, 0.4 mm BGA and 0201s in Figures 7a and 7b. Results and Discussion For miniaturized components, many differ- ent stencil construction technologies were test- ed. A wide array of results were observed. The data indicates that the best performing stencils for the miniature components appear to be B5 and B6, and D14 and D16. Interestingly, B5 and D16 had no nanocoating on them. They were cut from name brand stainless steel on new, state-of-the-art cutters. B6 and D14 were also cut on new, state-of-the-art cutters and nano- coated with thermally cured fluoropolymer nanocoating. The electroformed stencil demonstrated the poorest size accuracy of the 18 stencils test- ed, which is in agreement with three previous studies since 2011. 1-3 Stencils that were laser cut with nickel overplate did not appear to per- form as well as laser cut SS without overplate, with or without nanocoating. New investiga- tive technologies that were tested show much promise for delivering quality prints at better price points, thereby representing better values to SMT assemblers. Continuing Work One of the top-performing stencils is cur- rently the process of record on Vicor's high-vol- ume production line. Because a large amount of production data already exists for this stencil configuration, a new one will not be ordered to run on the line. Instead, two other top perform- ers and two stencils employing promising new technology will also be run for a full week in high-volume production. The stencil with larg- est amount of slag and roughness will also be se- lected for volume runs to determine if these pa- rameters impact end-of-line yields. The metrics for the production runs are, first and foremost, SPI yields. Secondary metrics include TE and Cpk for process control, and end-of-line yields. Prior to the stencils entering production, they will be validated using SPI to measure two prints each on back and front strokes, and Figure 7a: Stencil transfer efficiency. Figure 7b: Assessment of stencil print repeatability. greater than 90% yield in the first hour of pro- duction. If print yields drop below 90% and the stencil is suspected as the root cause, it will be removed from the production line and replaced with the stencil used in the process of record to investigate the suspicion. Results from the lon- ger-term production study will be published at a later date. SMT Acknowledgements The authors would like to sincerely acknowl- edge and thank: • Nathan Taylor and Rey Molina of Benchmark Electronics, NH for their technical support in execution of print tests and SPI programming