MEASUREMENT AND CONTROL OF TEMPERATURE 



Operation of that heater. An entirely new concept of thermostatic control 

 arises if the heater could be of continuously variable power output — a 

 condition which one attempts to produce in mark-to-space systems with the 

 most efficient stirring of fluids, and a high thermal capacity of the heater 

 itself acting as a smoothing agent. A second, more crude, approach is 

 obtained in the largest installations where multiple-contact thermostats 

 control a succession of heaters, and a high degree of control is impossible. 

 Continuously variable control of heater power output is however achieved 

 in the method of Beament and Machin^. The heater resistance is made the 

 anode load of a power valve, and its heat output is thus controlled by the 

 grid voltage of the valve. The information to be fed to this device can no 

 longer be of the 'on-or-ofl" variety, for the valve requires quantitative informa- 

 tion of the amount of heat required to control the body's temperature ; the 

 thermostatic element must therefore establish a reference voltage and vary 

 this in a suitable sense with temperature fluctuation in order to obtain 

 compensation control. A resistance thermometer is clearly the most suitable 

 T.S.E. for this purpose and a thermistor has proved excellent in practice. It 

 is placed in a potential divider such that a change in its resistance controls 

 the power output of the valve in the appropriate sense. The sensitivity of the 

 system is increased, and is made variable by positive feedback, this loop 

 being within the negative feedback loop of the T.S.E. to the heater itself. 

 The device gives a large swing in a continuous power output to compensate 

 for small temperature changes. This is probably the most accurate fine 

 control for air temperatures yet devised, but although it has no inherent 

 'differential' in the sense of the discontinuous heater it is obvious that it can 

 only compensate because the temperature of the body does change. For 

 best results, everything which has been said about good insulation, circula- 

 tion, heater size, etc., must be adhered to. 



Continuously variable temperature 



Theoretically speaking, there is no greater problem in making a body 

 follow a temperature programme than maintain a desired temperature. All 

 that is needed is to ensure that the maximum rate of temperature increase 

 and decrease can be obtained, and to programme the temperature setting 

 unit of the thermostat. Practically speaking, such a requirement will need 

 complex controlling equipment and the most careful design; it should be 

 avoided if at all possible. 



Control below ambient 



There is no basic difference between using a thermostat to control a cooler 

 rather than a heater — a refrigerator does this crudely as an everyday affair. 

 The removal of heat energy is a more difficult engineering problem, and from 

 a thermostatic point of view one which is slow and cannot be switched 

 rapidly on or off, or made quickly variable. Practical methods will be 

 outlined below; no new principles are involved. 



Control both above and below ambient 



This can be approached in three ways: the entire ambient can be cooled 

 below the lowest temperature needed, and then conventional control of 



404 



