Studies On Test Data Compression Methods Based On Test Response Filling Technique

In the past decades,along with the rapid development of the very-large-scale integration(VLSI)technology,the density of transistors in a chip is increasing exponentially and the test for digital integrated circuit(IC)has become one of the biggest challenges.The amount of test data is rapidly increasing,w hich brings great pressure to ATE memory,extends testing time,reduces test quality,and increases the cost of testing.How to effectively compress the test data has become a n important issue of IC testing in present.This thesis mainly contributes in the following works focused on the test data compression issues.In order to further improve the test quality and test compression ratio,this thesis proposes an FDR codes method based on response filling and flip-flops polarity inversion techniques.Firstly,the test response of the previous test pattern is utilized to fill the don't care bits in current test pattern,which increases the randomness of the X-filling and ensures high test quality of test patterns.Next,making the difference between test vectors and test responses of the previous test pattern can increase the number of 0s.Then,applying the flip-flop polarity inversion technique which changes the polarity of the flip-flops will further increase the number of 0s of the differential vector set,and the test compression can be increased by using FDR codes for the difference vectors.Finally,we determine the orders of the test patterns based on the minimum hamming distance,and the flip-flops' polarity to get higher compression ratio using tabu search algorithm,which can decrease the number of conflict bits between the test response of the previous test pattern and the current test pattern,allow the compression ratio with FDR to be maximum.On the hardware implementation,the inversion polarity can be realized by connecting the scan inputs to the outputs(Q?)of the previous D flip-flops.Therefore,there is no additional hardware overhead at all compared to the previous scheme.The response filling technique applies to FDR codes,which improves the test compression and test quality.This thesis attempts to using the response filling technique apply to improved FDR codes,like EFDR and AFDR codes,and proposes a high test quality test compression method based on response filling and run-length codes.Firstly,different methods of X response filling are propos ed according to the characteristics of EFDR and AFDR codes,and then m ake the difference between test vectors and test responses,fill the Xs using different methods,which increases the randomness of the X-filling and improves the test quality of test patterns.Next,the order of the test patterns will be determined based on the minimal transitions,which can increase the length of runs.Lastly,the differential vector set is compressed by different run-length codes,like EFDR and AFDR codes,which can obt ain a better test compression.The simulation experiments of the two methods proposed show that the two schemes can achieve better compression and higher test quality.