Integrated Circuit (IC) product test in semiconductor manufacturing industry is commonly conducted through the socket pogo pins contacting with the IC products. The socket pogo pins can be degraded due to repeatedly plugging-into and pulling-out from the socket. Degradation in socket pogo pins will greatly affect the accuracy of final test in semiconductor manufacturing, which in turn results in economic and reputation losses of manufacturers. How to rapidly and accurately detect the degraded pogo pins is still unresolved. In addition, the very huge data produced by a large amount of tester machines will further bring difficulties into degradation detection of socket pogo pins. Focusing on those existing problems in semiconductor manufacturing, this paper proposed a data driven framework with adopting data mining techniques to tackle them. This framework transforms the test data generated by manufacturing machines into human readable format and then analyzes them by data mining techniques, which empowers the manufacturing engineers to automatically detect the degraded socket pogo pins from the test data. Extensive experimental studies with real data were carried out and the results show great application prospect of the proposed framework.
[1]
William W. Melek,et al.
Comparison of trend detection algorithms in the analysis of physiological time-series data
,
2005,
IEEE Transactions on Biomedical Engineering.
[2]
Arun Ross,et al.
Score normalization in multimodal biometric systems
,
2005,
Pattern Recognit..
[3]
Katharine Lynn Gray,et al.
Comparison of Trend Detection Methods
,
2007
.
[4]
D. Ruppert.
The Elements of Statistical Learning: Data Mining, Inference, and Prediction
,
2004
.