In flip-chip (FC) packages, the solder bumps are sandwiched between the silicon die and the substrate, making them no longer visible for usual visual inspection. An inspection system capable of evaluating the quality of the hidden solder bumps in FC packages has been developed. The system pulses a laser on the top surface of a chip package to generate ultrasonic waves in the package to excite structural vibrations which can then be measured using an interferometer. Since defective solder bumps cause changes in the transient vibration response of a test sample, quality of the test sample can be assessed by correlating its vibration responses to that of a known-good reference sample. The limitation to this method is the necessity to use a known-good reference chip package, which typically involves expensive testing using alternate methods. In this paper, the development of a method without the need of known-good reference samples is presented. This method, called hybrid reference method, uses a statistical approach to find the most similar packages and then average their time domain vibration signals to generate virtual time domain signals. These virtual signals are what we call hybrid reference signals and they are correlated back with the vibration signals from the sample under inspection to obtain a quality signature. Finally, defective and nondefective chip packages are separated by estimating a beta distribution that fits the quality signature histogram of the inspected packages, and then determining a critical threshold for an acceptable quality signature. This method was applied on two types of FC packages with no available pre-established known-good reference chip package. The results of this quality assessment were validated by comparison with electrical test and X-ray technique results.
No takes yet. Share an insight, caveat, or question.
Ume et al. (2011) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context:
More from IEEE Transactions on Components Packaging and Manufacturing Technology