Studies On Test Compression Efficient Scan Tree Structure Based On FDR Code

As we know, electronic systems are widely being used in our lives and it’s critical component is just integrated circuit(IC). In the past decades, along with the rapid development of the very-large-scale integration, the amount of transistors in chip are increasing exponentially and the test for integrated circuit becoming more and more challenge. Full scanning design is one of the most important of design for testability (DFT) method. Its main idea is to make all trigger controllable by adding certain logic, then lower complexity in test but need longer test application time and test data volume.FDR is a single run-length coding method, it can effectively reduce the test data volume, Scan tree is a kind of broadcasting compression method, only a scanning input can drive multiple scan chains. So it can effectively reduce test application time and test data. However, their compression effect is not very ideal.To further reduce test application time and test data volume, this paper will come up with a scan tree structure based on FDR code efficient test data compression. Firstly, the paper will analyze the high efficiency of the combination between scan tree and FDR code technology. Then, we will study the filling problem (fill0or1) of don’t care bit and consider the care bit to be0or1in the final for the influence of FDR code. This paper proposes an efficient scan tree structure to reduce the number of care bit, improve the efficiency of FDR code, further reduce the amount of test data, In addition, this paper considers the wiring overhead, on the basis of improve the compression efficiency to optimize the wiring overhead. Finally, the proposed method is been verified, The experimental results show that for the large ISCA’89circuit the method we propose which combining scan tree and FDR or ALT-FDR than with the method of using only FDR or ALT-FDR average more than respectively was reduced by16.65%,14.73%in the case of only considering compression ratio, and this method is also superior to the current coding method. With consideration of both the compression ratio and routing overhead, we can see that when wire average reduce62.27%, the compression ratio only loss an average of5.07%.