4.2 Dependence of viscosity on temperature

These experiments were intended to test the approximate viscosity-temperature relation mentioned in section 2.4.2. That is,
        v = A exp (Ev / k T)
where v is the kinematic viscosity, k Boltzmann's constant, T the temperature, and A and Ev constants. The measurements were obtained using both the viscometer of this project (hereafter referred to as the project viscometer), and a Ferranti-Shirley cone-on-plate viscometer. Therefore a useful comparison between the two viscometers was also obtained.

Two lubricants were tested: Dow Corning FS-1265 10000 cst lubricant, and Dow Coming DC200 1000 cst lubricant. These kinematic viscosities are stated by the manufacturer, however the temperature at which the viscosity is specified is not mentioned. However the aim of these experiments was not to determine absolute viscosities, rather to investigate the degree of temperature variation and correlation with an existing commercial viscometer. Hence both instruments were calibrated to read the stated viscosity at room temperature as a reference point.

For the project viscometer measurements, 50 cm3 of the test fluid was contained in a small beaker of diameter 4 cm. This beaker was suspended in a temperature controlled oil bath. The temperature of the test fluid itself was additionally monitored using a digital thermometer, with a thermocouple immersed in the liquid.

During the measurements ample time was allowed for the temperature to reach a stable value, usually slightly less than that of the oil bath due to heat losses from the test fluid. Temperature control is built in to the Ferranti-Shirley viscometer, which pumps heated oil through the instrument to heat the sample under investigation.

Viscosity was measured at temperatures ranging from room temperature to about a hundred degrees centigrade. Above this temperature the oil bath becomes intolerably smelly. Numerical results are presented in Appendix 0.

Figure 12 shows in graphical form the variation with temperature of the 10000 cst fluid. Calibration was performed at 22.5C. The data from the Ferrauti-Shirley instrument approximates the exponential curve extremely well, although that from the project viscometer displays a large spread. This was probably mainly due to the inadequacies of the spiral spring. Error bars plotted on the project viscometer measurements are for a 10% random variation. Temperature is measurable to an accuracy less than half a degree, and the Ferranti-Shirley instrument specifies a precision of under 3%. Thus in the interests of clarity error bars have not been plotted on these data points.

Figure 13 is a graph of viscosity against temperature for the 1000 cst fluid. Again the Ferranti-Shirley results appear more accurate than those of the project viscometer.

These results show that the exponential viscosity-temperature relation is well approximated within the measurement range of the experiment. This is particularly apparent for the 10000 cst FS-1265 lubricant, which displays a rapid decay of viscosity with rising temperature.
 
 

4.1.4 Conclusions

Contents

4.3 Accuracy and long term stability