QFN wire bonder efficiency improvement through DMAIC methodology

Being one of highest growth packages, QFN (Quad Flat No-lead) always facing a lot of challenges in maintaining high equipment efficiency in production mode. A lot of hypothesis was made on the root causes, and mostly blaming on frequent machine change-over or device conversion. This seems to be difficult to fix, simply due to the fact that QFN needs to maintain high flexibility in production. A comprehensive study was conducted to verify any possibility of equipment efficiency improvement. The study utilized systematic Six Sigma DMAIC methodology in deriving and optimizing all the major process elements, including people factor, leadframe variations, wafer robustness, machine hardware, measurement system analysis (MSA), auto recipe implementation, and optimum machine utilization. With total of 13 factors, respective evaluations revealed a lot of new knowledge and information. In the evaluations of people factor, there were a lot of findings on knowledge and skill of key technical personnel in supporting production. The comparison study was made by collecting performance data from each technical personnel, and then comparing to ideal technical setup guidelines and time allocation for each task. Despite the comparison, findings like mistakes or improper setup could be utilized in setting up more proper training material, and roll out to all other technical personnel. The other factor was regarding to MSA, whereby all relevant machine down time codes and data were being analyzed in details. With comparison using manual method, local machine statistics and on-line efficiency data, any discrepancy could be identified easily. For QFN wire bonders with capability of implementing auto recipe, advantage of minimizing human dependency might become disaster. The auto recipe implementation should not ignore any single details including machine hardware differences, if not, and then it should not be utilized at all. With the DMAIC study, all significant factors were able to be o- - ptimized systematically, resulting in significant increment in machine efficiency performance. Together with advanced optimizations, new process window could also withstand any flexible effect of new product introduction in QFN packages, making production machines really robust and stable.