Mixed-Signal Circuits Modelling and Simulations Using Matlab

Abstract

Continuous advances in IC (integrated circuits) processing technologies offer the possibility to produce integrated circuits with increased complexity and capability at a reduced cost. In addition, integrated circuits are composed more and more of heterogeneous embedded systems with various kinds of digital, analogue, and mixed-signal circuits and sensors that are integrated on the same IC. In the future, the number of different embedded systems integrated on the same IC, as well as the number of all modules and the complexity of all modules, will increase. In addition, more and more heterogeneous modules (including analogue and mixed-signal circuits, sensors and actuators, micro-electro-mechanicalsystems) will be integrated on the same IC. Currently, well-established circuit design tools exist for the circuit level design of analogue circuits; even better design tools exist for the systematic and hierarchical design of large digital circuits, including an important aspect of any reliable digital circuit, that is, testability, which is also well covered by existing digital EDA tools (EDA stands for Electronic Design Automation tools). There are no such tools available at the moment for analogue and mixed-signal circuits except maybe Saber. Saber can help to model, design, and simulate a complete mixed signal system in a short amount of time. It may also be used in several analogue and mixed signal blocks and different kinds of integrated sensors. With Saber (Synopsis, 2004) it is possible to model such circuits; however, Matlab/Simulink is much more convenient because of many different toolboxes and libraries available. Matlab, together with Simulink, offers a hierarchical environment and computation engines that make it possible to model, simulate and design complex circuits in a very efficient way on a high hierarchical level. In addition, modules may be described, modelled, simulated, and designed at different abstraction levels and at different levels of detail. For example, analogue circuits could be modelled in a very simple way, that is, taking into consideration only system level aspects like transfer functions; conversely, more details about the implementation, such as offset voltage, gain bandwidth product, slew rate, thermal and 1/f noise, and parameters such as nonlinearity, could be added, when necessary. For digital circuits, it might be necessary to produce a bit-true model early in the design to take all aspects of the design into consideration. Appropriate decisions early in the design can improve the efficiency of the design process. Since large modern IC’s usually contain many mixed-signal functions, it is necessary to model and simulate the analogue and the digital