This paper presents a nondestructive diagnosis method for defect detection of flip chips.
Forty-two features were extracted from vibration signals for analysis.
PCA-RBF was introduced for feature reduction and defect assessment.
The results demonstrate the feasibility of this approach for flip chip diagnosis.