The simulation results of applying the theory of artificial neural networks (ANN) to the problem of detecting catastrophic defects injected into a two-stage operational amplifier are presented in this paper.

Experiments were performed on both 0.7 and 0.35 micron technology, using HSPICE simulations. Using power supply current waveforms as inputs (both in the time domain and frequency domain), coverage ranging from 15 to 99 percent was achieved, depending on the size of the training set, and the complexity of the neural network.

I found the paper to be interesting. The experiments were well organized, and the data was presented in an easily understood format. A reader can understand the results without being an expert in ANNs or analog circuits. I have two criticisms: first, the authors do not realize their stated goal, that catastrophic defects be unconditionally detected, and, second, nearly every sentence has a grammatical defect. In spite of these shortcomings, the paper will be of interest to those working in the area of ANN research, and to those interested in detecting faults in analog circuits.