Optimization and AMS verification of on-chip temperature sensor on SOC level

Abstract

Analog and Mixed Signal (AMS) designs are essential components of today’s modern Integrated Circuits (ICs) used in the interface between real-world and digital signals. Analog and mixed-signal circuit designs are gaining the popularity of the semiconductor industry as the compact design demand and more functions equipment increases in the market. In the era of 5G, AI, Self-Driving Cars, and the Internet of Things (IoT), the need for Mixed-Signal System-on-Chips (MSSoCs) is increasing gradually. The Simulation Program with Integrated Circuit Emphasis (SPICE) tools is still the primary technique to verify analog circuits. However, the SPICE simulation's speed is still an issue as the complexity of the degine is increasing node by node. Other verification strategies are required to verify the complete analog and digital functionality within the time before the market. Current industrial verification methodologies, each addressing specific verification challenges, are useful for detecting and eliminating design failures. Nevertheless, decreases in first-pass silicon success rates illustrate the lack of cohesive, efficient techniques to allow a predictable verification process that leads to the highest possible confidence in AMS designs. For desire digital simulation even for sophisticated analog design with more efficiency in the speed factor, then SPICE Universal Verification Methodology (UVM) is used. This thesis will present the study of AMS and AMS verification with an On-chip CMOS Temperature sensor followed by the results. This approach is generally used to verify SoCs in the automotive industry. Alongside as the demand for AMSSoCs is increasing, On-chip thermal monitoring is becoming crucial as VLSI circuits are becoming more sophisticated with time and recent technologies. On-chip thermal monitors provide information about the power and thermal management structures, which is essential to prevent chip temperatures from destroying the device or maintain the lifeline. Several sensors are being placed on the chip to monitor the exact data of the on-chip temperature. These sensors are compatible with the technology node, not even taking large area and proves accurate data with standard process variation and required power supply variation. The studied temerature sensor is based upon a self-stabilized feedback architecture.