MEASUREMENT OF TEMPERATURE 



great accuracy provided that it is limited to an infinitesimally small current 

 drain). 



Conclusions — The great advantage of the thermocouple is clearly its 

 small size and thermal capacity, and its remote-reading characteristic; to 

 use it as a device for the very accurate measurement of temperature is also 

 possible, but a considerable technical feat in itself. Where a sensitivity of 

 0-1 °C is all that is required, a couple of copper and constantan, soft-soldered 

 or brazed at the junctions, is convenient. It is not normal to use thermo- 

 couples in combination with amplifiers, or for recording or controlling 

 purposes, because of the low voltages and low impedances of the system (this 

 is not, of course, true of the output of industrial instruments measuring 

 temperatures of the order of 1,000°C which have substantial outputs). 



77?^ thermopile— ThQ output of a thermocouple can of course be increased 

 additively, by placing a large number of them in series. For accurate work, 

 the large number of leads between individual hot and cold junctions would 

 make the device cumbersome, and of greatly increased thermal capacity. 

 However, it has applications in detecting temperature differences at small 

 distance separations, e.g. radiation receptors, where radiation falls on one 

 set of junctions as a black body and the other set are shielded, or in some 

 forms of hygrometer, where one set measures wet bulb temperatures and 

 the other dry. 



Resistance thermometers — The essential of the resistance thermometer is 

 that it makes use of the temperature coefficient of resistance of a conducting 

 element. Leaving the special case of the thermistor (see below) till later, 

 the materials used as sensitive elements are almost exclusively the metals 

 nickel, copper or platinum. Over the temperature ranges which will interest 

 biologists the resistance of a metal changes linearly with temperature on 

 the Absolute temperature scale to an accuracy of about 4 per cent (though 

 of course on the platinum-resistance scale a platinum thermometer is self- 

 linear). 



The advantages of the resistance thermometer are: (a) remote reading; 

 (b) it can be made of inert material, e.g. platinum; (c) it is very stable and 

 shows almost no hysteresis with large rapid temperature excursions; 

 (d) its minute thermal capacity is particularly suitable for measuring gas 

 temperatures; (e) once calibrated, its reading is absolute {cf. a thermo- 

 couple's cold junction). 



Its disadvantages are : (a) the resistance of the element may be changed 

 by adsorbed contaminants; (b) it may not be used in contact with any 

 conducting liquid such as water; (c) a very thin wire element is delicate; 



(d) if heated by the measuring current, it may give incorrect readings; 



(e) calibration is difficult. 



Practical consideration — The accurate measurement of resistance is a 

 straightforward physical problem, and for the resistance thermometer 

 a Wheatstone bridge is advised {Figure 29.5). All leads must be of constantan 

 or very heavy copper, to keep their own temperature-resistance changes 

 minimal, and the comparator resistance must also be temperature stable. 

 In order to neglect any electrochemical voltages or polarization in the 

 circuit an a.c. bridge is recommended, and thus the comparator resistances 

 must be non-inductively wound. It has akeady been pointed out that the 



26 389 



