Test Data Compression Coding-based Scheme Storing Integers Represented for Irrational Numbers

A compression method for test data coding is presented, by storing irrational numbers in integer form against large-amount time waste of transmission of test data and high cost of chip testing due to large-amount test data transmitted to chips for test via automatic testing device in integrated circuit test in procedures as follows. Firstly, dividing test data according to run lengths and setting the first run length as the default digit before the decimal point, other lengths as decimal part respectively, before transforming test data into its decimal form. Secondly, transforming the decimal into an irrational number that can be expressed in integer form via binary search of irrational numbers. Thirdly, storing integers corresponding to their irrational form for expressing m, l and k. The method is theoretically available to transform the storage of whole test unit into that expressed in integer form corresponding to single or several irrational numbers because it transmits the rule of test data but not the test data itself. The experimental results on larger sequential circuits in ISCAS89 benchmark circuits show that, the test data can be compressed with higher efficiency than that compressed by commonly-applied coding such as Golomb, FDR, EFDR and MFVRCVB under the same experimental circumstances.