The objective of this research is to develop an integrated solution to low-cost Johnson noise thermometers (JNT) with significantly faster speed than currently available, and to demonstrate its practical application to on-chip temperature measurement. The approach is to follow the same Nyquist noise principle that enabled other primary JNTs, but to focus on the most widely used industrial temperature range over which sensitivity and bandwidth requirements are drastically relaxed, and to reduce measurement time from hours to milliseconds for practical use by integrating several innovative circuit design and digital signal processing techniques.
The primary intellectual merit of the research is in developing the first practical integrated solution to JNTs. A further intellectual merit is in providing an accurate and inherently linear temperature sensor and/or sensor reference for embedded applications. Other intellectual merits include novel circuit design techniques for eliminating speed bottlenecks in primary JNTs, and sampling and post-processing techniques for significantly reducing measurement errors. Finally, this research introduces a new on-chip multi-temperature self-calibration scheme for other embedded temperature sensors.
An immediate broader impact of the research is the wide applicability of the integrated thermometer in science, industry and everyday uses. The new technology also has a high probability of being adopted by industry and creating a direct economic impact. Targeted approaches are planned to facilitate recruiting and training of female and minority students. Several noise and temperature measurement techniques and laboratory experiments will be incorporated into undergraduate courses. Finally, nontraditional dissemination activities will further broaden the project's impact.