Modeling and Simulation of Substrate Noise in Mixed-Signal Circuits Applied to a Special VCO GOLNAR KHODABANDEHLOO, SATTAR MIRZAKUCHAKI Electrical Department Iran University of Science and Technology Narmak, Tehran IRAN Abstract: - The mixed-signal circuits with both analog and digital blocks on a single chip have wide applications in communication and RF circuits. Integrating these two blocks can cause serious problems especially in applications with fast digital circuits and high performance analog blocks. Fast switching in digital blocks generates a noise which can be introduced to analog circuits by the common substrate. This noise can decrease the performance of mixed-signal circuits; therefore, studying this noise and the way it is transmitted will lead to solutions for reducing it and improving mixed-signal circuit’s performance. In this paper, an efficient model for substrate is extracted from Green’s function in MATLAB environment, and its accuracy is demonstrated. Using a VCO and a multiplier as analog and digital blocks, respectively and connecting them with the proposed model of the substrate, the effects of substrate noise coupled to analog blocks are shown. Finally, some methods for reducing this noise are applied to the circuit, and the results are compared to each other. The results indicate that using P+ Guard Rings is the best method for reducing substrate noise in the mixed-signal circuits. Key-Words: - Mixes-Signal Circuits, Substrate Noise, Green’s Function, VCO, Phase Noise 1 Introduction Nowadays, VLSI technology makes it possible to have mixed-signal circuits on a single silicon die where fast digital circuits are integrated near high performance analog blocks. Noise isolation is one of the most important problems in these circuits which can be worse as noise increases because of higher integration density and faster digital circuits [1]. Noise coupling problems are a crippling factor in many advanced circuits. Therefore, studying this noise is inevitable in order to find solutions for reducing it. The aim of this paper is to simulate and analyze the noise transmitted through the common substrate in a mixed-signal circuit. For this purpose, first a VCO is designed to act as the analog block and then a multiplier is chosen as the high frequency digital block. The most important part of this work is modeling the substrate. Substrate should be modeled in a way that it can be used in simulators like SPICE. After designing analog and digital blocks, substrate model can be used to illustrate the path between them. Thus, switching noise in digital circuits transfers to analog circuits through the substrate. By comparing analog block characteristics after and before relating it to digital block, noise effects can be specified. Therefore, substrate noise in real mixed- signal circuits is simulated. Some methods of reducing substrate noise are applied to the designed circuit and compared to each other. In section 2, substrate noise and some injection and reception mechanisms are described. Some methods for modeling substrate and the one used in this paper are outlined in section 3. Section 4 discusses analog and digital blocks designed here. Analog and digital blocks are related by the substrate model in section 5, and some simulations are performed. In section 6 some substrate noise reduction methods are applied and the results are compared. 2 Substrate Coupled Noise A silicon substrate is generally considered to provide a good isolation between devices because of its high resistance. However, in some cases, switching devices can introduce turbulence to the substrate which can be transferred to other devices because of substrate infinite resistance. Figure 1 shows two common substrates: lightly doped or P- and heavily doped or P+ [2]. Each switching node can introduce a noise to the substrate. Infinite resistance of substrate means that this noise can be transferred to adjacent devices. There are several injection and reception mechanisms, some of which are shown in Figure 2. Proceedings of the 6th WSEAS International Conference on Simulation, Modelling and Optimization, Lisbon, Portugal, September 22-24, 2006 252