Accurate Prediction of Random Telegraph Noise Effects in SRAMs and DRAMs

Abstract

Abstract-With aggressive technology scaling and heightened variability, circuits such as SRAMs and DRAMs have become vulnerable to random telegraph noise (RTN). The bias dependence (i.e., non-stationarity), bi-directional coupling, and high inter-device variability of RTN present significant challenges to understanding its circuit-level effects. In this paper, we present two computer-aided design (CAD) tools, SAMURAI and MUS-TARD, for accurately estimating the impact of non-stationary RTN on SRAMs and DRAMs. While traditional (stationary) analysis is often overly pessimistic (e.g., it overestimates RTNinduced SRAM failure rates), the predictions made by SAMURAI and MUSTARD are more reliable by virtue of non-stationary analysis.

Modeling and Simulation of Low-Frequency Noise in Nano Devices: Stochastically Correct and Carefully Crafted Numerical Techniques

“…and CMOS image sensors [4], [5], [6]. Moreover, due to technology scaling, low-frequency noise in MOSFETs is emerging as a source of great concern in the design of even digital circuits such as SRAMs and DRAMs [7], [8]. The goal of this work is to introduce novel models and numerical techniques that can be used to conduct analyses of both the impact of individual traps and also the collective effect of multiple traps on circuit operation, enabling the analysis of diverse low-frequency noise phenomena.…”

“…In previous work, RTS noise analyses of digital memory cells were conducted using a stochastic technique similar to the SSA algorithm [8]. In this paper, we adopt a synergistic approach in developing noise modeling and analysis techniques for RTS and low-frequency noise in transistors and electronic circuits based on techniques for modeling ion channel noise in neurons.…”

“…In this paper, we adopt a synergistic approach in developing noise modeling and analysis techniques for RTS and low-frequency noise in transistors and electronic circuits based on techniques for modeling ion channel noise in neurons. Precursors of the work in this paper were presented in [8], [39] and [40]. Here, we incorporate and expand the works in [39] and [40] by providing a detailed explication of the proposed trap models with regard to their stochastic properties and developing carefully crafted circuit simulation techniques that are stochastically correct.…”

“…We first present stochastic, nonstationary models for charge carrier traps in semiconductors, i.e., RTS noise, by benefiting from the two decades of ion channel noise modeling and simulation work in neurobiology [31] and the extensive literature on modeling and simulation of stochastic chemical kinetics [34], [35]. In particular, we describe in Section II (i) a continuous-time, discrete-space, fine-grained Markov chain model, i.e., a stochastic automaton model [8], [39], and (ii) a continuous-time, continuous-space, coarse-grained Langevin equation model for a charge carrier trap [39]. These models are both fully nonstationary, capturing the impact of arbitrary time-varying bias conditions.…”

Modeling and analysis of nonstationary low-frequency noise in circuit simulators: Enabling non Monte Carlo techniques

“…However, due to the nano-scale device sizes in the latest technology nodes, RTS noise has become a limiting factor in the design of digital circuits, such as SRAMs and DRAMs, as well [14], [15]. Since these digital circuits and analog circuits such as oscillators and mixers do not operate in the smallsignal regime, results produced by simple stationary (or modulated stationary) RTS noise models are inaccurate due to the bias voltage dependent capture/emission rates of traps [16].…”