The Research For Test Vector Optimization And Fault Diagnosis Of Mixed-Signal Circuits

The complexity is higher of circuits in the area of aerospace, power and military. The test difficulty of analog circuits, digital circuits and mixed-signal circuits is more difficult, and scale of test is very huge. But the test vector has a lot of unnecessary redundancy vector, leading to increased testing time and cost. So choose the best fit test vector for the measured circuit--optimization test stimulus has become an important direction in the research of automation testing. At the same time, in the face of huge investment of complex large system, reliability, maintainability and availability is becoming more and more important. Therefore fault detection and diagnosis technology has been rapid development in nearly thirty year-end.First, this paper introduces the significance of the circuit board test vector optimization and fault diagnosis at the present stage. Further introduces the development status and existing circuit board incentive optimization, and the existing problems and development status of fault diagnosis technology.Second, this paper establishes the fault model of stable at "0" fault, stable at "1" fault of AD converter. This paper proposes the biggest dissimilarity model method and MCD-OSAM(the maximum code distance-one step adaptive mapping) method. The biggest dissimilarity model method can generate the minimum complete test set that can complete fault identification of stable at "0" fault, stable at "1" fault of analog-to-digital converter. The MCD-OSAM method can generate the optimal test vector set of l2space that can complete fault identification of stable at "0" fault, stable at "1" fault and wire-and logic fault of analog-to-digital converter. These test vector set can fully meet the characteristic of compactness and completeness. Meanwhile, these test vector set can greatly reduce the redundancy of test vectors and improves time efficiency.Third, according to the characteristics of the analog-to-digital converter, this paper confirm the digital output vector composed of analog-to-digital converter digital output bits belong to the domain of two element field space GF(2N). Through the analog-to-digital converter digital output orthogonal decomposition in N dimension GF(2N) space, this paper uses the spatial error value, digital characteristics of noise-bit as characteristic quantity to complete non-noise-bit and noise-bit fault identification. At the same time, this paper uses FOBOD (first order base origin distance) as characteristic quantity to complete non-noise-bit and noise-bit fault identification of analog to digital converter. The fault diagnosis method can effectively improve the correct rate of fault, and solves the confounding syndrome problem of the AD converter noise-bit. Finally, according to the provided analog-digital circuit board and power circuit, using multi bus automatic test system platform resources, this paper completed the fault diagnosis of analog-digital circuit board and power circuit, including:requirement analysis, hardware design of the adapter board, TPS program design of test and diagnosis, experimental verification.