Microelectronics Journal 101 (2020) 104814 Contents lists available at ScienceDirect Microelectronics Journal journal homepage: www.elsevier.com/locate/mejo Automated design and optimization ﬂow for fully-diﬀerential switched capacitor ampliﬁers using recycling folded cascode OTA Mostafa N. Sabry a, ∗ , Islam Nashaat b , Hesham Omran c a Si-Vision LLC, Cairo, Egypt b Goodix Egypt, Cairo, Egypt c Integrated Circuits Lab, Faculty of Engineering, Ain Shams University, Cairo, Egypt ARTICLE INFO Index Terms: Analog design automation Switched capacitor ampliﬁer Systematic design Optimization Gm/ID methodology Recycling folded cascode OTA ABSTRACT An automated optimization ﬂow for a fully diﬀerential switched capacitor ampliﬁer design using the gm/ID methodology is presented in this paper. The operational transconductance ampliﬁer (OTA) is implemented using a recycling folded cascode OTA architecture. Accurate expressions for the ampliﬁer input capacitance in both low and high frequencies are derived and used iteratively to precisely set both DC loop gain and bandwidth while performing the optimization. Binary search and interpolation are used to split the total noise speciﬁcation between sampling and ampliﬁcation phases while meeting all other speciﬁcations. Non-linear optimization is used without invoking the simulator in the loop to reach the optimal gm/ID design point in a short time. Despite conventional wisdom, it is shown that using larger gm/ID can result in both power and area improvement. Moreover, the proposed ﬂow takes into account the eﬀect of variations and mismatch on the design. Simulation results of the synthesized ampliﬁer show very good agreement with the required speciﬁcations. 1. Introduction The analog design productivity-creativity gap (PCG) has signiﬁ- cantly increased in the last years [1]. Time-to-market (TTM) and design complexity are the main reasons behind the continuous increase in PCG. The best way to manage the trade-oﬀ between design complex- ity and TTM is to use accurate and reliable computer aided design (CAD) methodologies and design automation tools [2]. The purpose of the CAD methodologies and the associated design automation tools is not to replace the designers but to help them be more creative, highly productive and to develop optimized circuits without many design iter- ations. Several CAD methodologies had been proposed in the previous decades to minimize the PCG. These methodologies are targeting either topology selection or device-sizing. However, the optimal device-sizing is the most trending, because it is considered as a constrained non-linear optimization problem that can be solved by diﬀerent optimization algo- rithms. Automatic device-sizing approaches are divided into knowledge- based and optimization-based [3]. In the knowledge-based approach, the automated device-sizing is achieved using a predeﬁned design plan ∗ Corresponding author. E-mail addresses: mostafa.nashaat@si-vision.com (M.N. Sabry), inashaat@goodix.com (I. Nashaat), hesham.omran@eng.asu.edu.eg (H. Omran). set by expert analog designers [4–6]. This approach is well-known for its short execution time, however, it doesn’t take into account any kind of solution optimization. On the other side, the optimization- based approach was introduced to overcome the shortcoming of the knowledge-based approach. Optimization-based approach can be divided into equation-based and simulation-based methods. The equation-based method is characterized by fast execution time, how- ever, the approximate equations used to map the complex circuit model leads to a low precision design [7–9]. are the most relevant tools that used this method. In comparison, the simulation-based can reach an optimum solution using a precise evaluation by having SPICE-in-the- loop as in Ref. [10–12], however, this method might take very long- time in optimizing a moderate-size circuit because the circuit needs to be simulated at each iteration and it may not converge at the end. One of the most promising and commonly used techniques is to use the advantages of knowledge-based and optimization-based approaches to enhance speed and accuracy simultaneously. In Ref. [13], the authors proposed a minimal power design methodology for a SAR ADC by combining both of the knowledge-based and optimization- based approaches, however, they used SPICE-in-the-loop that leads to increase in the execution time signiﬁcantly. In Ref. [14], the authors https://doi.org/10.1016/j.mejo.2020.104814 Received 2 December 2019; Received in revised form 24 April 2020; Accepted 4 May 2020 Available online 11 May 2020 0026-2692/© 2020 Elsevier Ltd. All rights reserved.