基于多特征模型模拟电路可测性指标分析方法

A testability index analysis method for analog circuits based on multi-feature model

结合多信号模型与多特征分析方法,提出了一种新的模拟电路可测性指标分析方法——多特征模型分析方法。该方法根据电路输出节点的阶跃响应波形,提取出多个特征信号,并将所有的特征形成一个特征向量,然后转换成故障-测试依赖矩阵进行可测性指标分析。以双二阶低通滤波器电路为例,分别在电路参数名义值、故障值与容差变化情况下,对所提方法与传统故障字典法进行可测性指标分析对比试验。仿真试验结果表明,提出的方法不仅需要测点少,故障检测率与隔离率很高,而且适合于具有容差的模拟电路可测性指标分析。

Based on the multi-signal model and multi-feature analysis methods, a new testability index analysis meth- od for analog circuits：multi-feature model analysis method is proposed in this paper. Multi-feature signals can be ex- tracted according to the step response waveforms at the output points of the circuit. All the useful feature signals form a feature vector, which then is transformed into a fault-test dependency matrix for testability index analysis. Consider- ing the nominal values, fault values and tolerance changes of the circuit parameters, a biquad low-pass filter is used as an example to carry out comparison experiment of testability index analysis between the proposed method and the tra- ditional method based on fault dictionary. Simulation experiment results show that the proposed method requires fewer test points, and has higher fault detection rate （FDR） and fault isolation rate （FIR） compared with the traditional method. The proposed method is also suitable for the testability analysis of analog circuits with tolerances.