Scholarlink

In response to the demand for electronic systems operating in extreme high-temperature environments (>300 \(^\circ \) C), driven by applications such as deep space exploration, oil/gas exploration, and other harsh conditions, high-performance silicon carbide (SiC) integrated circuits represent a critical breakthrough. This study proposes an artificial neural network (ANN)-based approach for accurate modeling of SiC MOSFETs across a wide temperature range (27 \(^\circ \) C–500 \(^\circ \) C). This data-driven model demonstrates superior prediction accuracy with a remarkably low root mean square error (RMSE) of 0.0014. Leveraging this model, an all-MOSFET SiC voltage reference circuit was designed and verified. The circuit delivers a typical output reference voltage of 1.41 V over a temperature range of 27 \(^\circ \) C–500 \(^\circ \) C, achieving a temperature coefficient of 149.8 ppm/ \(^\circ \) C. These results successfully validate the high reliability of the proposed ANN model for wide-temperature SiC device characterization and circuit simulation. Furthermore, this study provides a practical design example of a high-performance SiC voltage reference circuit, laying a crucial foundation for developing next-generation all-SiC high-temperature system-on-chip (SoC) solutions.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Design of an SiC All-MOSFET Voltage Reference Using Neural Network Modeling

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Wenhao Yang,
  • Angze Li,
  • Mengnan Qi,
  • Yuyin Sun,
  • Yimeng Zhang,
  • Yuming Zhang

摘要

In response to the demand for electronic systems operating in extreme high-temperature environments (>300 \(^\circ \) C), driven by applications such as deep space exploration, oil/gas exploration, and other harsh conditions, high-performance silicon carbide (SiC) integrated circuits represent a critical breakthrough. This study proposes an artificial neural network (ANN)-based approach for accurate modeling of SiC MOSFETs across a wide temperature range (27 \(^\circ \) C–500 \(^\circ \) C). This data-driven model demonstrates superior prediction accuracy with a remarkably low root mean square error (RMSE) of 0.0014. Leveraging this model, an all-MOSFET SiC voltage reference circuit was designed and verified. The circuit delivers a typical output reference voltage of 1.41 V over a temperature range of 27 \(^\circ \) C–500 \(^\circ \) C, achieving a temperature coefficient of 149.8 ppm/ \(^\circ \) C. These results successfully validate the high reliability of the proposed ANN model for wide-temperature SiC device characterization and circuit simulation. Furthermore, this study provides a practical design example of a high-performance SiC voltage reference circuit, laying a crucial foundation for developing next-generation all-SiC high-temperature system-on-chip (SoC) solutions.