Design of micropower class AB transconductors: A systematic approach

Abstract

A systematic design approach to achieve micropower class AB CMOS transconductors is presented. It includes techniques to get rail-to-rail operation and continuous transconductance tuning, based on floating and Quasi-Floating Gate transistors. Application of the proposed design approach leads to a new family of high-performance power-efficient class AB CMOS transconductors. To illustrate the feasibility of this approach, 12 transconductors derived from this common framework have been designed and fabricated in a 0.5 μm CMOS technology. Measurement results show THD values for 2 V inputs of −56 dB for a static power of 300 μW and silicon area <0.07 mm².

Introduction

Linear operational transcondance amplifiers, or transconductors, are widely used in several applications such as continuous-time filters, A/D and D/A converters, variable-gain amplifiers (VGAs), and sensor interfaces, to name a few [1]. Design of high-performance CMOS transconductors is challenging nowadays due to the limited signal swing available under the reduced supply voltages imposed by technology downscaling and low power requirements. Usually the applications mentioned demand the availability of transconductors featuring high linearity in a wide input range under the aforementioned low supply voltage requirements.

Several techniques have been proposed to improve linearity of transconductors, such as adaptive biasing or nonlinearity cancellation, but they often require generating extra currents thus increasing the area and power requirements, and/or rely on the MOS square law which is unrealistic in modern processes [1]. A very interesting alternative is the use of resistive degeneration by passive resistors for voltage-to-current (V–I) conversion. Unfortunately, continuous tuning of the resistance value is not possible in this case. The use of MOS transistors in triode region instead to implement the resistance element allows continuous tuning but degrades linearity noticeably. Besides, class AB operation is required in the transconductor to allow maximum dynamic currents beyond the bias current values, overcoming the tradeoff between static power consumption and dynamic performance of class A implementations. However, class AB transconductors usually require additional circuits to achieve class AB operation which often increase power consumption, noise level and supply voltage requirements, and decrease current efficiency. Besides, quiescent currents are sometimes not well controlled in the resulting class AB circuits.

In this paper it is shown that Quasi-Floating Gate and Floating-Gate techniques can be an interesting approach to overcome these drawbacks. Floating-gate [2], [3], [4], [5], [6], [7], [8], [9], [10] and Quasi-Floating Gate [11], [12], [13] techniques have found widespread use in analog design. Among other applications, they allow achievement of a weighted addition of input voltages in a simple and compact way. They also allow the efficient implementation of floating dc level shifters that can be used for instance to design class AB output stages.

The paper is organized as follows: Section 2 provides a brief overview of Floating-Gate (FG) and Quasi-Floating Gate (QFG) MOS transistors. Section 3 deals with the application of these techniques to achieve rail-to-rail input range, transconductance tuning, and class AB operation. A systematic approach to design class AB CMOS transconductors based on these techniques is presented in Section 4. Second-order effects and noise are analyzed in Section 5 and Appendix A. Measurement results of 12 transconductors thus generated are presented in Section 6. Conclusions are drawn in Section 7.