No abstract

No abstract

A power efficient BIST TPG method is proposed to reduce test power dissipation during scan testing. Before the test patterns are injected into scan chain, the test set adopts a series of preprocessed strategies including don't care bit based 2-D adjusting, Hamming Distance based 2-D reordering and test cube matrix based two transpose, all steps will be orderly executed in interspersed way. The six largest ISCAS'89 benchmark circuits verify the proposed method. Experimental results show that the switching activities are effectively reduced when the test set is loaded for on-chip scan testing. ASDFR with MT-filling scheme ensures high compression ratio, the scan-in test power dissipation is further decreased by don't care bit based 2-D adjusting and Hamming Distance 2-D reordering. In addition, the BIST TPG method with less test application time and smaller algorithm complexity can be widely applied to actual chip design without adding extra decoder area overhead.

We construct and implement a compressive sensing measurement matrix based on improved size-compatible (ISC)-array low-density parity-check (LDPC) code. First, we propose an improved measurement matrix from the array LDPC code matrix. The proposed measurement matrix retains suitable quasi-cyclic structures and supports arbitrary code lengths. It also achieves a high perfect recovery percentage compared with a Gaussian random matrix of the same size. Second, we propose a hardware scheme using cycle shift registers to design the compressive sensing measurement matrix generator. This provides simple circuit architecture during the generation of the measurement matrix. According to simulation verifications, the measurement matrix construction method is effective and entails fewer shift registers and a lower area overhead by using a simplified hardware implementation scheme. The compressive sensing measurement matrix generator can generate all of the required elements in Circuits Syst Signal Process the ISC-array LDPC code matrix with an acceptable hardware overhead. Therefore, it can be widely applied to large-scale sparse signal compressive sensing.