Study on Soft Fault Diagnosis of Analog Circuits

YANG Shi-yuan,HU Mei,WANG Hong
DOI: https://doi.org/10.19304/j.cnki.issn1000-7180.2008.01.001
2008-01-01
Abstract:This paper analyzes the importance and all kinds of existing methods of soft fault diagnosis of analog circuits. The characters of four approaches of soft fault dictionary methods based on the principle of branch screening, the locus of transfer function about component parameters, the law of invariance of the node-voltage sensitivity sequence and the linear relativity among node-voltage increments are discussed respectively; based on the neural network, combined with the fuzzy theory and the wavelet transform, the research actuality of the two approaches of modern analog soft fault diagnosis are also introduced. Furthermore, we point out the research direction of the analog soft fault diagnosis according to three different aspects: unified soft fault diagnosis method, the strategy of diagnosing large scale analog circuits and the requirement of diagnosing digital and analog mixed circuits, and finally discuss the issues remain to be resolved.
What problem does this paper attempt to address?