To obtain DC resistance values, the open-circuit potential of the above 

 system or cell was first measured. The system was then shunted by switching 

 to the proper amperage scale and thus reducing the input resistance of the 

 electrometer. The new voltage was measured, and the DC resistance was calcu- 

 lated from the following equation: 



. E„ - E, 



n. = 



where R| = internal resistance of the cell (or the resistance of the coating) 



Eq = open circuit potential 



Eg = shunted potential 



Rj = shunt resistance (or input impedance of the electrometer) 



When the open circuit potentials were quite low, the total open circuit 

 potentials were increased by the addition of voltage from a potentiometer. 

 When measurements were made on coatings of very high resistance, some time 

 was required to obtain an equilibrium value. For DC coating resistances of 

 10" ohms, approximately 20 minutes were required to reach reasonable 

 equilibrium values. The time taken for the voltage readings to come to equi- 

 librium was often reduced by imposing, effectively at the electrometer 

 connections, charging potentials very close to the expected equilibrium voltage. 

 These voltages were imposed by the circuit shown in Figure 22. When the time 

 constants were very long, or when there were fluctuations in readings due to 

 external factors, the voltage readings versus time were plotted to arrive at better 

 equilibrium values or average values. 



The DC resistance values obtained by the above methods were recorded 

 and plotted, and the results are shown in Figures 3 to 15. The average initial 

 resistance values are indicated at the left-hand ordinate scale, and the average 

 values after about 7 days and continuing to 400 days are shown by the dashed 

 curves. It should be noted that in this semilog presentation, the DC resistance 

 values are plotted on a log scale differing in interval from that used for the AC 

 resistance values. Wherever the resistance values of any of the three panels 

 deviate from the average by more than 0.5 log unit, the averages of the remaining 

 panels are plotted, and the deviations are indicated at the abscissa or right-hand 

 ordinate. For System 1 15 (Figure 7), the values for each panel are plotted to 

 illustrate the maximum deviations that would be included in the averages. 



28 



