Automated E-circuit Designing and Characterization using Prominent Neural Network
Designing, characterization and modelling of circuits comprising of several active and passive devices play a vital role in the field of electronics and various e-product development. In the present paper, neural network (NN) based model is developed. The developed model can be used for automation of any electronic circuit design. As a test case, amplifier design using common base (CB) configuration of bipolar junction transistor (BJT) is considered. The CB configuration is highly significant and predominantly used in several applications such as preamplifiers, UHF and VHF RF amplifiers, and current buffer circuits. Henceforth, designing and characterization of CB configuration is considered and performed using prospective NN technique. A set of performance parameters are considered to frame the amplifier that incorporates voltage gain, input impedance, output impedance and collector emitter voltage. The design parameters considered are base, emitter and collector resistors. In the present work, Levenberg Marquardt (LM) method is utilized as fitting and training algorithm for developing NN based model. The developed neural network based model comprises of two hidden layers with respective count of neurons as 10 and 8. It is envisaged that neural network based model is highly accurate and can be incredibly beneficial in the field of VLSI design.
bipolar junction transistor (BJT), circuit designing, feed forward neural network (FFNN), mean square error (MSE).