互補式金氧半導體運算放大器最佳化設計之自動合成演算法

Abstract

運算放大器是類比電路中的基本模組，並且被廣泛地運用在各式電路中，發展運算放大器自動化設計可有效地加速類比電路設計研究，並增進類比電路產品的競爭力。本論文提出一個根據設計者所訂定的最佳化目標，自動最佳化單一顆MOSFET的概念，並藉由此概念設計一種自動化合成運算放大器(OPAMP)的流程，此自動化合成演算法可使運算放大器達到演算法定義的最佳化目標，並經由合成三種常用的運算放大器：Two Stage OPAMP、Folded Cascode OPAMP、Telescopic OPAMP驗證後，由HSPICE直接連結設計，最後由C++語言程式化實現整套架構。使用者根據其應用或性能的要求，挑選適合的運算放大器架構，並且輸入其需要的規格，由演算法計算出所需要的電路尺寸，最後使用模擬軟體驗證電路規格能滿足要求。

OPAMPs are the fundamental module in the analog circuit and used in all kinds of circuits extensively. Developing the OPAMPs automatic design can speed up the research on analog circuit design and increase the competitiveness of the analog circuit products. A new concept to automatically optimize a MOSFET size is proposed in this thesis. An automatic process to synthesize operational amplifiers (OPAMPs) is designed based on the concept, and the automatic synthesis algorithm can make operational amplifiers achieve the optimal goal defined by this algorithm. It has been implemented by three common OPAMP topologies: Two Stage OPAMP, Folded Cascode OPAMP, and Telescopic OPAMP. They are designed and simulated by HSPICE. Furthermore the algorithm is realized by the program C++. Users choose the suitable OPAMP topology according to their application or demand for properties and input the specifications they require. The whole device sizes will be calculated out by algorithm. Finally, the execution results are verified to be able to fit corresponding specifications by simulator.