Prediction of element values of OPAmp for required specifications utilizing deep learning

The development of EDA tools for analog IC design which increase designers' productivity is an urgent need. Conventional analog EDA tools require recalculating every time required specification changes. In this paper, a prediction method of element values of OP Amp for the required specifications using deep learning is proposed. The proposed method can learn the relation between the performances and element values of a target circuit. Simulation results indicate that the proposed method has succeeded in predicting 7 element values which satisfies required 13 performances with an accuracy of 93% in average.