MEASUREMENT AND CONTROL OF TEMPERATURE 



to 0-001°C is not over difficult. However, while such temperature differences 

 can be measured, having the two junctions in two respective places, the 

 experimental problem becomes very much different as soon as one wants to 

 calibrate the device, or if one wishes to measure one temperature absolutely 

 to any degree of accuracy; for the only reference temperatures one can 

 obtain very accurately are those of melting ice and boiling water. Now 

 thermocouples have to be calibrated against something — for accurate work 

 a Beckmann thermometer is usual. The problem is that it is difficult to 

 obtain any metering device of which the scale can be read more accurately 

 than to 1 part in 1,000; this is one degree C full-scale deflection for a 

 sensitivity of 0-001°, while if one wished to read temperatures around 20°C 

 and had to display ice temperature to 20°C on the scale, one would not be 

 able to read more accurately than to 0-02°C. The sensitivity will corre- 

 spondingly go down as the temperature range is increased. If in fact this is 

 still of sufficient sensitivity, then a calibration curve is constructed by setting 

 a Beckmann thermometer in an ice-water mixture containing both junctions, 

 while the meter or galvanometer is set to zero; the thermometer with one 

 couple wound round the bulb is then placed in successively warmer water — 

 giving the mercury plenty of time to equilibrate, and stirring the water well. 



or 



Figure 29.4 Basic null-point and bridge circuits for thermocouple 



measurements 



If, however, it is necessary to work at higher sensitivity, then use can be 

 made of the law of intermediate temperatures, which states that if a thermo- 

 couple develops a given voltage between temperatures T^ and Tg, this will 

 equal the sum of the voltages developed between temperatures T^ and T^,, 

 and T^, and T^, where T^, is any intermediate temperature. It is thus possible 

 to calibrate a couple between 0°C and a suitable temperature above it, 

 then to place the cold junction, together with a second reference thermometer, 

 into high temperature water and continue the calibration curve in the form 

 of additional voltages for known temperature differences between the 

 mercury thermometers. 



Alternatively, one can use a potentiometer system to 'back-off' the excess 

 voltage due to the temperature difference {Figure 29.4). The couple can 

 again be calibrated in steps, and for very sensitive work it is advisable to 

 use a high resistance potentiometer to keep down the battery drain, and to 

 include a standard Weston cell, in a separate circuit, with which to check 

 the voltage of the ordinary cell (a standard cell will give a voltage of very 



388 



