為了縮短類比積體電路的設計時間，類比積體電路的自動化設計，也慢慢的受到重視，典型的類比電路自動化設計方法大致可分為三大類：知識基礎(knowledge-based)、模擬基礎(simulation-based)以及方程式基礎(equation-based)方法，但是每一種方法都有各自的優缺點，本論文提出一套新的方程式基礎的運算放大器自動化設計流程，並且以兩級式運算放大器說明gm/ID的設計方法，本論文提出的設計流程可以改善傳統幾何演算法的缺點，因此能有效減少設計中來回修改的次數。整個流程已經以C++實現，而非線性規劃(nonlinear programming)的部分用LINGO來解，並整合成一個完整的程式。從實驗數據的觀察並跟模擬退火演算法及幾何演算法的設計流程做比較，本論文所提出的方法可以在最短的時間內達到使用者所設計的規格。 In order to shorten the design cycles of analog circuits, analog design automation has become a popular research topic in recent years. Typically, analog design automation approaches can be classified into three categories: knowledge-based, simulation-based and equation-based approaches. But each approach has its own disadvantages in terms of accuracy and efficiency. In this thesis, we propose a new equation-based analog synthesis process and demonstrate on a two-stage operational amplifier(op-amps). Based on gm/ID design concept, this process can improve the accuracy of traditional GP-based approach thus reducing the design iterations. The proposed automation process has been implemented with C++ and LINGO, which is a nonlinear programming solver. As demonstrated in the experiments, the proposed approach can achieve the required specifications with the shortest computation time compared to previous SA-based and GP-based approaches.
