Parametric fault detection for instrumentation amplifier circuit

The paper presents a new approach for fault detection in the linear analog circuit using the method of transfer function coefficient estimation with the help of system identification toolbox integrated with the Matlab. The proposed approach consists of two parts, namely, simulation of the analog circuit using PSpice simulation followed by estimation of transfer function coefficient using the mentioned tool box from the simulated response. The parametric faults are modeled in terms of component tolerance specified by the user. From the netlist of the circuit under test and given the tolerance of components the fault free range of transfer function coefficients is computed under no fault condition. Using the input and output signal information of the faulty circuit the transfer function coefficients are estimated in the presence of modeled single faults. If these estimated coefficients are found to be out of their fault free range the circuit under test is detected faulty. The proposed fault detection approach is validated with the help of an instrumentation amplifier simulated with a sinusoidal signal.