抄録
The basic difficulties of precise temperature control are investigated by controlling the temperature of a stirred water bath, as precisely as possible, which is thought most easily controllable.
Theoretical considerations show that (1) the effect of such disturbances as the room temperature change or the electric power souce potential variation can easily be diminished by ordinary continous control devices composed of an electronic PI controller and a magnetic amplifier etc. and the precision of the order of 1/10000 deg can be realized comparatively easily, but (2) the drift of temperature sensing element changes the above stated circumstances violently because the greater part of it passes through the temperature control closed loop network without attenuation and causes the gradual change of the controlled state.
The accuracy limitation for a long period of time is determined by the drift of temperature sensing elements.
So the drift of platinum resistance (or thermistor) bridge thermometer, especially the drift generated by the temperature variation of all components of the bridge, is analyzed, and the conditions that can, at worst, maintain the drift under 3/10000 deg for 24 hours are obtained.
The best results of the temperature control experiment with the platinum resistance thermometer operated under the required conditions are as follows; the accurracy for short period of time is about 1/6000 deg rms and the drift of controlled temperature is about 1/5000 deg per day.
The stirring action of water can not be so rapid that no inhomogenity of water temperature exists in the bath. The short range inhomogenity of the water temperature is considered the orgin of the short period fluctuation of controlled temperature.
