Thermometer equations for low-temperature glass-ceramic capacitance thermometers

✍ Scribed by W.N. Lawless; E.A. Panchyk

• Publisher: Elsevier Science
• Year: 1972
• Tongue: English
• Weight: 449 KB
• Volume: 12
• Category: Article
• ISSN: 0011-2275
• DOI: 10.1016/0011-2275(72)90085-9

✦ Synopsis

and for a 120~type thermometer from O. 1 to 9 K (56 points) were curve-fitted to various empirical equations using regression-analysis techniques (the designations 110~ and 1200type refer to processing conditions). It was found that simple fourth-order expansions, T(C), yield the best fits to the 110~type data. The index of regression and residual standard deviation for the quartic fit were: 0.999988and 19 mK for a 1.7 to 20 K fit; O. 999951 and 78 mK for a 20 to 50 K fit," and O. 999972 and 70 mK for a 1. 7 to 50 K fit. When the 1.7 to 50 K region was divided into two regions, above and below 20 K, it was found that the residuals were less than, or commensurate with, the uncertainties in the original temperature measurement (germanium resistance thermometry). The 1205Ltype data were analysed from O. 1 to 4.2 K and from O. 1 to 9 K. In both regions, cubic expansions T(C) yielded good fits: 0.999932 and 11 mK from O. 1 to 4.2 K, and 0.999972 and 10 mK from O. 1 to 9 K, for the index of regression and residual standard deviation, respectively. However, the residuals for these fits were considerably larger than the temperature uncertainties, possibly indicating systematic measurement errors.