The aim of this paper is to present the results of an interlaboratory comparison in temperature and relative humidity, among eleven laboratories, that was coordinated by the National Institute of Metrology, Quality and Technology (Inmetro, Brazil). As transfer standard, a thermo-hygrometer was used. The intercomparison was carried out in the range from 20 %rh to 90 %rh (at 20 ºC and 30 ºC) and from 20 ºC to 40 ºC. The performance of the intercomparison was judged by calculating the normalized error. Results from the majority of the labs agreed well with those of Inmetro.

According to the ISO guide to the expression of uncertainty in measurement, the calibration process and its uncertainty evaluation should be expressed in terms of mathematical function(s) of input quantities. However, in practice, expressing measurement or calibration in a way that is fully compliant to the ISO guide might be unrealistic and requires a clear definition of the calibration process itself. Depending on the point of view on the calibration process, different modelling equation with various complexities can be written. In this paper, four different approaches are given to model the calibration process of industrial platinum resistance thermometers.

Three Si-SiC eutectic fixed-point cells have been constructed using electrically heated furnace for use in thermocouple thermometry. The first two cells were made from silicon powder mixed with carbon while the third was made from pure silicon. The first cell with a conventional graphite crucible was early broken during the preliminary test, so the result could not be recorded. The second cell with a modified graphite crucible, thicker and shorter than the previous one, was able to withstand until 36 melt-freeze cycles and was held above 1400 °C for about 143 h, but unfortunately this cell was also finally broken after the test. The third cell, made from pure silicon filled in the modified crucible, was successfully tested without any mechanical failure. This cell was held above 1400 °C for approximately 157 h and was subjected to 36 melt-freeze cycles, where 7 cycles of them was realized from room temperature. The melting plateau of the cell lasted about 40 min and the extreme of the measured inflection point differed by 0.1 °C. The repeatability, represented by its standard deviation, was 0.03 °C. The temperature difference between both of the last two cells was calculated to be 0.5 °C.

Despite the importance of resistance bridges in thermometry, an accurate assessment of their uncertainty has not been investigated much. Among various uncertainty components, nonlinearity and ratio error have been of particular interest, and a Hamon-type resistance network known to be RBC (resistance bridge calibrator) is widely used to measure those uncertainties. However, due to finite temperature coefficients of base resistors in the RBC, the evaluated uncertainty of the resistance bridges under normal operating environment is in doubt. In this work, to accurately evaluate the uncertainty of the resistance thermometry bridge, an air medium thermostatted chamber was devised, and the nonlinearity and ratio error of a resistance bridge at KRISS (ASL F900) were assessed. Along with this, repeatability and AC quadrature/frequency dependence of the resistance bridge were evaluated and total combined uncertainty of the resistance bridge at KRISS was finally assessed.

We set up the temperature - running period CM (Characteristic Modelling) of the noble metal thermocouples (S/B type). Using CM, we got the both thermocouples max error reduced, type S is 2.1 °C and type B is 1.8 °C which were 5.8 °C and 4.3 °C before using CM and the coefficients of determination in CM are 82.1 % and 72.5 %. It shows 85 % difference between CM and error less than 1.0 °C, especially in type B. So it is necessary to use CM to measure temperature accurately.