Using CPK and CPK confidence intervals to evaluate stencil printing

It has been widely accepted that effective stencil printing is the strongest determinant for high end-of-line yields. So, assuming that the stencil-printed solder paste deposit is printed squarely on the pad, the volume consistency of the deposit will be very important for generating good end-of-line yields. Considering this situation, we proposed that Cpk would appear to be a desirable metric as it is sensitive to both accuracy and precision. In light of this perspective, we performed stencil printing experiments using numerous solder pastes and calculated the Cpk. After calculating all of the Cpks, the question arose as to what difference between Cpks was statistically significant. After a significant amount of research, we discovered a technique to calculate the 95% confidence intervals on the Cpk. We determined that the Cpks were significantly different if the confidence intervals did not overlap. This paper will discuss this new work as well as an Excel® spreadsheet developed to calculate the Cpks and their confidence intervals. In addition, we will cover the difference between data being statistically significant and practically significant.