Fault Prediction And Diagnosis Research Of PCB Circuit

With the rapid development of modern PCB (Printed Circuit Board) circuit, the complexity and intensity of PCB increase constantly, which results in increasingly strict requirement with the reliability of PCB circuit. Fault prediction and diagnosis of PCB circuit research field become more and more important. PCB circuit is one kind of mixed signal circuit consists of digital circuit and analog circuit, fault prediction is mainly used for analog circuits. Fault diagnosis of PCB digital circuit, fault prediction of PCB analog circuit and fault diagnosis of PCB analog circuit will be researched in this article. Fault diagnosis of PCB analog circuit is the focus of this article, a new method of fault diagnosis of PCB analog circuit based on multi-dimensional scaling is present in this article and the feasibility of method is verified on two typical PCB circuits. The main work is as follows:First. A crosstalk fault diagnosis method of PCB digital circuit is present according to feature of attack lines and victim lines.The method can largely reduce test times and the cost of resource of PCB digital circuit. A method of reducing crosstalk faults of PCB digital circuit is present in this article and the feasibility of method is verified on simulation circuits, the results show that this method can significantly reduce crosstalk between transmission lines.Sceond. A prediction model of PCB analog circuit based on neural network is built in this article. The accuracy of this model is verified on Sallen-Key low-pass filter circuit, the result show that the deviation between prediction value and actual value is very small,this prediction model is able to predict component fault accurately.Third.A new method of Fault diagnosis of PCB analog circuit based on multi-dimensional scaling is present in this article. Particle swarm optimization support vector machine as fault classifier to verify the reliability of this method. The method compare with fault diagnosis method of PCB analogy based on principal component analysis algorithm, the result shows that accuracy rate of method proposed in this article is higher than fault diagnosis method of PCB analogy based on principal component analysis algorithm.