An 8b subthreshold hybrid thermal sensor with ±1.07°C inaccuracy and single-element remote-sensing technique in 22nm FinFET

2018 IEEE International Solid-State Circuits Conference - (ISSCC)(2018)

引用 25|浏览0
暂无评分
摘要
Thermal sensors are commonly used in modern highly dense systems-on-chip (SoC) to provide information about die temperature for thermal protection or performance optimization. To enable the deployment of multiple sensors in an SoC, the power and size of such sensors has been steadily reduced. Although most solutions are PNP-based [1-4], recently a low-power NPN-based current-mode thermal sensor [5] was implemented to meet the power and form-factor requirement with a robust architecture. However, since NPN devices are only available in a triple-well process, its use in low-cost dual-well processes is limited. This paper demonstrates a current-mode hybrid thermal sensor in an advanced 22nm FinFET process [6] based on subthreshold NMOS transistors and a parasitic PNP [7]. A simple voltage-based single-point soft-trimming was implemented to mitigate the sensitivity of the sensor to the subthreshold factor variability during manufacturing. Instead of placing the whole sensor system in multiple locations in a system, the hybrid architecture also supports single-element remote sensing. Only a 0.00021mm 2 PNP device placed in the area of interest and a signal connection to the main sensor are then needed for temperature sensing. The sensor system achieves +/−1.07°C (±3σ) precision with 0.0043mm 2 silicon area and 50uA current consumption from a 1V supply.
更多
查看译文
关键词
temperature sensing,sensor system,systems-on-chip,SoC,die temperature,thermal protection,multiple sensors,form-factor requirement,robust architecture,NPN devices,current-mode hybrid thermal sensor,subthreshold NMOS transistors,parasitic PNP,subthreshold factor variability,hybrid architecture,signal connection,sensitivity mitigation,voltage-based single-point soft-trimming,low-cost dual-well processes,triple-well process,low-power NPN-based current-mode thermal sensor,performance optimization,advanced FinFET process,single-element remote-sensing technique,subthreshold hybrid thermal sensor,PNP device,size 22.0 nm,voltage 1.0 V,current 50 muA,word length 8 bit
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要