Research On The Applications Of Transform Coding In The Test Data Compression

Nowadays the design and manufacture of the integrated circuits(IC)have been into the deep submicron scale,the complexity and integration of IC are rising rapidly.So IC testing is increasingly challenged.The huge amount of test data generated by IC has increased the capacity and transmission bandwidth requirements of the automatic test equipment(ATE),and also can lengthen the test application time.Therefore,how to reduce the test data volume has been one of the challenges in the field of IC testing.Focused on the test vector compression,this paper is to reduce the amount of the test set and improve the test compression ratio in a lower hardware cost.It is contributed to reduce the storage of ATE and shorten the test application time.The main work in this paper is as follows.(1)A test vector compression method is proposed based on the transform coding.Combining the transform coding with the traditional coding compression schemes,the original test vector is divided into two parts by using K-L transform:the main component and the residuals.The main component is realized approximately by using Hadamard Matrix that has a simple hardware structure,and is integrated in the circuit under test.The residuals are the results of XOR between the main component and the original vector.In this paper it codes with the residuals,but without the original vector.The experimental results show that the average compression ratio of the proposed method has reached 77.94%,and that of the Hadamard implementation is 71.88%.(2)A multi-component selecting method is proposed based on K-L transform,which is a further research on the first method.According to the different proportion of the total energy that the spectral components hold,the proposed method selects multiple spectral components to synthetic the main component.The experimental results show that the number of spectral components selected is contributed to improve the compression ratio.When three spectral components are selected,the average compression ratio can reach 79.08%.