Solder bumps are widely used in surface mount components, which provide electrical and mechanical connection between the chip/package and the substrate. As the solder bump getting smaller in dimension and pitch, it be...Solder bumps are widely used in surface mount components, which provide electrical and mechanical connection between the chip/package and the substrate. As the solder bump getting smaller in dimension and pitch, it becomes more difficult to inspect the solder defects hidden in the IC package. In this paper, an intelligent inspection method using the scanning acoustic microscopy(SAM) and the fuzzy C-means(FCM) algorithm was investigated. A flip chip package of FA10 was chosen as the test sample. The SAM tests of FA10 were carried out in C-scan mode. The sub-image of every solder bump was segmented from the SAM image. The statistical features were then calculated and adopted for clustering of solder bumps using the FCM algorithm. The recognition results of FCM reached a high accuracy of 94.3%. The intelligent system is effective for defect inspection in high density packages.展开更多
基金Supported by the Youth Foundation of Shandong Academy of Agricultural Sciences(2005YQ013)International Science and Technology Co-operation Projects(2006DFA33130)by the Chinese National Programs for High Technology and Development(2006AA100108)
基金supported by the National Natural Science Foundation of China(Grant Nos.51675250&51705203)the Natural Science Foundation of Jiangsu Province(Grant No.BK20160183)the Open Foundation of State Key Lab of Digital Manufacturing Equipment&Technology(Grant No.DMETKF2016005)
文摘Solder bumps are widely used in surface mount components, which provide electrical and mechanical connection between the chip/package and the substrate. As the solder bump getting smaller in dimension and pitch, it becomes more difficult to inspect the solder defects hidden in the IC package. In this paper, an intelligent inspection method using the scanning acoustic microscopy(SAM) and the fuzzy C-means(FCM) algorithm was investigated. A flip chip package of FA10 was chosen as the test sample. The SAM tests of FA10 were carried out in C-scan mode. The sub-image of every solder bump was segmented from the SAM image. The statistical features were then calculated and adopted for clustering of solder bumps using the FCM algorithm. The recognition results of FCM reached a high accuracy of 94.3%. The intelligent system is effective for defect inspection in high density packages.
