MEASUREMENT OF TEMPERATURE 



resistance change and high impedance means that it can be used with long 

 leads without problems of their actual resistance or temperature coefficients. 

 Like other resistance thermometers, measurement involves passing a current 

 through them, with the result that they dissipate power and heat themselves. 

 Their power-dissipation must be hmited to something of the order 1/10 mW, 

 and because of the great change in resistance over a reasonable temperature 

 excursion it is wise to feed them from a constant current source. It must be 

 appreciated that unhke the platinum thermometer this instrument can be 

 calibrated in water, but whereas such a high specific heat liquid will keep 

 the self-generated temperature down, air will not do so; hence unless the 

 thermistor emits very little heat indeed its calibration will change with 

 its surroundings. The thermistor has also some hysteresis if subjected to 

 rapid large temperature fluctuations, but these will not seriously affect 

 normal biological work: in general the instrument should not be subjected 

 to a greater temperature gradient than 10°C per minute, nor to a greater 

 overall excursion than 50-60 C° without re-calibration. The shift in 

 calibration due to more violent experience may be 1 per cent; under the 

 best conditions measurement to 0-0 1°C may be obtained. Thermistors of 

 any one type and resistance show individual differences of such an order 

 that each must be separately calibrated against a mercury-in-glass scale. 



Practical details — A thermistor is capacitive and should not be used in 

 an a.c. bridge, for the capacity will also have a temperature coefficient; 

 while a d.c. bridge technique might give the most sensitive results, methods 

 in which the temperature is displayed on a meter will commend themselves 

 for normal work. Two systems have been found reliable in practice, 

 depending on the application of the machine: a battery-operated unit for 

 portable field work, where the thermistor is used as an ecological probe; 

 and a mains-operated unit, particularly suitable for bench work, for the 

 measurement of air temperatures, and those of small animals and plants. 



Mains unit — The device of Beament and Machin^ has given satisfactory 

 service. Here, the thermistor, fed from a constant current source to limit 

 its power dissipation, controls the grid of a cathode follower. A second 

 cathode follower (second half of a double triode) gives selected voltages, 

 and these are compared by a microammeter between the cathodes. The 

 meter is 'protected' against reverse voltages by a rectifier in parallel, though 

 this must be chosen with care for it is temperature sensitive and its shunt 

 value will vary. High cathode resistances together with very high resistances 

 in the thermistor and comparator potentiometers reduce the effect of mains 

 voltage fluctuations. The thermistor chosen for this unit is of the highest 

 resistance range — 100 kQ at 20°C. The thermistor is calibrated against a 

 good mercury thermometer in water, and the comparator resistances may 

 have to be padded slightly in order to get overlap on the temperature scales. 

 In the design shown {Figure 29.6) the machine switches 'on' into the highest 

 temperature range, since the rectifier will always protect against low- 

 temperature overload but not against high-temperature overload. A small 

 amount of ripple in the power supplies does not seem to affect the instrument 

 adversely. 



Portable unit — This employs an unbalanced bridge technique; using the 

 lowest range thermistor (2 kQ at 20°C), the current drain is so small that 



391 



