Method Of Determining Solder Paste Inspection Tolerance Settings
Authors: Johnny Chen, R Sivam V Rajoo, Marco Zhao, Wei Wen, Golden Xu, Ace Ning, Michael Xie, An Qi Zhao, Wei Bing Qian, Fuqing Li, Ken Wong, Zhen (Jane) Feng Ph. D., and Murad Kurwa Company: Flextronics International Inc. Date Published: 10/16/2011
Abstract: Solder Paste Inspection (SPI) is advantageous when it enables defects to be detected early in the manufacturing process. Because of this, more Electronic Contract Manufacturing Services (EMS) companies are implementing SPI in their SMT lines to minimize repair costs. Flextronics uses SPI, AOI, AXI, ICT, and FCT to achieve and maintain an optimal Test Strategy. Flextronics puts places special emphasis on maximizing the test coverage early in the assembly sequence and on maintaining a robust Test Feedback process to minimize defect detection down the line where repairs are typically more costly. In this paper, we will share our study on how to determine reasonable SPI tolerance limits (specifications) based on data collected from different Flextronics sites. Our main goal is to establish better control of the Solder Paste process and to improve the printing quality by performing the following: ? Define the solder paste volume tolerance limits based on the solder paste volume distribution from mass production data collection. ? Deploy the new tolerance limits settings for each package and track the performance to determine the optimal solder paste volume pass or fail criteria for various package types. ? Continuously monitor the printing process by collecting data for review to confirm that the UCL (upper control limit) and LCL (lower control limit) which we selected remains optimal. Our SPI engineers created the Solder Paste Volume (and/or height/area) Standard Library for each component type with their associated tolerance settings which were imported during program development. Each Flextronics site has its own SPI database for optimizing the tolerance limit (specification) settings to improve yield and quality. We had AXI data from two AXI machines (AXI1 and AXI2) from two different vendors for this project. We will report the correlations between SPI-to-AXI1 and SPI-to-AXI2 based on > 10,000 variable data points with three special defect boards with different types of defective pins for different component packages. This study helped us gain a better understanding of SPI capabilities and will help us to continue to improve the performance of our SPI systems.
Solder Paste, SPI, AOI, AXI, Tolerance Settings, and Correlations.