带隙
计算机科学
电子工程
材料科学
光电子学
工程类
作者
Yunqi Yang,Dongdong Chen,Xianglong Wang,Di Li,Yintang Yang
摘要
ABSTRACT In this paper, a high‐efficiency codesign method for bandgap circuit (BGR) circuit by submodule optimization is proposed. In the proposed method, the BGR circuit is divided into operation amplifier module (OPM) and reference voltage generation module (RVGM) to be separately optimized to improve the optimization design efficiency. Firstly, the OPM is optimized based on neural network (NN) model and particle swarm optimization (PSO) algorithm. NN models are established to describe the relationship between design parameters and performance metrics of OPM, and then the parameters of OPM are optimized by PSO algorithm according to the established optimization function with constraints. The parameters of RVGM are optimized based on the optimized OPM. The high‐accuracy approximate equations of temperature coefficient (TC) are established based on compensation functions. The TC, total area, and total power consumption are co‐optimized by interior point method. According to the optimized parameters of OPM and RVGM, the verifications are conducted by Cadence. The results show that the proposed method can greatly improve the design efficiency of BGR circuit, whose total running time is only 552.5 s. Meanwhile, the optimized TC is 4.76 ppm/°C over −40°C–125°C at 2.5‐V supply, and the optimized area and power consumption are 0.0187 mm 2 and 288.5 μW. The Monte Carlo results show that V ref has a standard deviation of 1.503 mV, which is only 0.12% across 1000 runs, so the V ref is very stable. Therefore, the proposed codesign method can be used to effectively optimize the BGR circuit.
科研通智能强力驱动
Strongly Powered by AbleSci AI