Temperature

RBR data loggers are calibrated to an accuracy of ± 0.002°C (ITS-90) over the range -5 to +35°C.

RBR maintains primary standards for:

• - 38.8344K mercury triple point
• +  0.0100K water triple point
• +29.7646K gallium freezing point

Extended temperature calibration, down to -40°C and up to +50°C is available. Beyond +35°C the platinum resistance thermometer is used to provide the extrapolation, and the accuracy is about ± 0.005°C.

Drift of the thermistors has been measured to be better than 0.002°C/year.

Calibration constants are stored in the logger, together with a complete calibration history.

The standard model thermistor is made by Thermometrics. The thermistor is aged and hermetically sealed in the tip of a shock resistant solid glass rod. It has a time constant of 85 msec and is housed in a durable 3mm 316 stainless steel externally mounted housing. In this housing the thermistor has a time constant of less than 3 seconds. An optional fast thermistor is also available with a 0.9mm OD housing and a time constant of less than 0.095 seconds.

Temperature measurements are made to 24-bits. Temperature measurements are taken by the XR-420 at a resolution of <0.00005°C. The measured rms noise floor is ±0.000046°C.

RBR fits the temperature measurements to the Steinhart-Hart Equation:

• 1/T = a + b(LnR) + c(LnR)3

This equation is an empirically developed polynomial which best represents the resistance-temperature relationships of NTC thermistors.The reference resistor is specified to 5 ppm/year and 0.6 ppm/°C.

RBR uses a Hart Scientific 4012 or 4037 stirred bath and a Model 1575 or 1590 Platinum thermometer. The calibration of the system is verified using our own primary reference standards: a -38.8344°C (±0.0001°C) triple point of mercury, a 0.0100°C triple point of water cell (±0.00008°C) and a 29.7646°C freezing point of gallium reference (±0.0001°C). All of the instruments are supplied with NIST traceable calibration. Standard calibration of each unit is carried out at eight temperatures from -5° to +35°C.

Temperature readings are taken at ten second intervals, and an average of 100 stable samples is used to obtain each point. This produces a fit with a typical uncertainty of <0.2 mK on each point. Sample temperature calibration graph: The NIST claims an uncertainty for thermistor calibration of little better than ±2 mK, and only suggests ±0.3 mK for a TPW cell. The NRC of Canada currently will only accredit secondary standards laboratories to an accuracy of ±6 mK. We offer calibration to our own primary standards to avoid any problems of transfer of calibration.