Abstract
This paper presents a complete analysis of the ability of March tests to detect ADOFs (Address Decoder Open Faults) and resistive-ADOFs in address decoders of embedded-SRAMs. Such faults are the primary target of this study because they are notoriously hard-to-detect. With this study, we show that standard March tests without modifications are not able to detect them and we propose to translate the algorithm presented in [1, 2] into March elements. These new March elements involve a particular address sequence and data to be written. For this purpose, we have exploited some Degrees of Freedom of the March tests (DOF I and IV) in order to generate these new March elements for ADOFs detection. Compared to the previous March solutions, these new March elements ensure the fault observation.