Multiple Fault Diagnostics in Solder Paste Printing Process Using Blind Separation Approach

摘要: Abstract In this paper, a fault diagnosis and variation pattern identification method is presented based on blind source separation (BSS) approach which used to extract important information from the measured data estimate matrix in solder paste printing process. BSS techniques provide basis for identifying precise characteristics of each individual order facilitate their root causes. Measurements mixed signals collected simulated faulty process are decomposed by independent component analysis (ICA) results case study show that ICA has ability separate original mixtures identify using frequency domain analysis. The studied paper would benefit SMT research correcting unidentified stenciling problems.