VOLT-SECOND CHARACTERISTICS OF SOLID HOMOGENEOUS DIELECTRICS

A. V. Astafurov, A. A. Vorob'ev, G. A. Vorob'ev,
K. M. Kevroleva

Volt-second characteristics of dielectrics make it possible to judge the rate of the dielectric-breakdown process, which is important to know when calculating the insulation of electrical apparatus.

We have determined the volt-second characteristics of a large number of solid dielectrics: single crystals of NaCl, KCl, KBr, and KJ, single crystals of Rochelle salt, ice, plastic talc containing crystallization water, porcelain, and others.

Fig. 1. Dependence of the electric strength of single crystals of NaCl, KCl, KBr, and KJ on the time of voltage application.

Fig. 2. Dependence of the electric strength of organic glass on the time of voltage application.

The breakdown voltages of different specimens of one and the same dielectric, at one and the same time of voltage application, show a large scatter. Under such conditions the problem arises of correctly constructing the volt-second characteristic. It should be taken into account that the greatest breakdown voltage corresponds to the most perfect specimen. The experimental data were statistically processed according to the probability of breakdown, and the values of the breakdown voltage were taken at a breakdown probability \(\psi = 90\%\). Volt-second characteristics at a breakdown probability of less than 90% have the same character, but the regularities are expressed more weakly. The magnitudes of the breakdown voltage at a breakdown probability \(\psi = 90\%\) are determined by the four or five highest values of the breakdown voltage, with a total number of broken-down specimens equal to 20–25 in one exposure.

Figures 1, 2, 3, and 4 present the volt-second characteristics for breakdown in a homogeneous field. Figure 5 presents the volt-second characteristics of rock salt for breakdown in a nonuniform field.

For all the dielectrics investigated, with a thickness of 0.15 mm, an increase in electric strength is observed at an exposure of \(1 \cdot 10^{-7}\) sec and

less, which indicates the identical nature of breakdown in these dielectrics, different in composition and structure. It is known that the increase in breakdown voltage at short exposures, which is observed both for gases and for liquid dielectrics, is due to the delay of the discharge.

At present the view has taken shape that electrical breakdown of solid dielectrics is an electronic process. Therefore

Fig. 3. Dependence of the electric strength of talc 1, ice 2, and gypsum 3 on the duration of voltage application.

Fig. 4. Volt-time characteristic of ice during breakdown in a homogeneous field.

the discharge-delay time in solid dielectrics, just as in gaseous ones, should in the general case consist of the statistical delay time and the time of discharge formation. In the experiments of A. F. Walter and L. D. Inge, for plates of uncolored and X-rayed

Fig. 5. Volt-time characteristic of rock salt NaCl during breakdown in a nonuniform field.

upper — point; lower + point.

Fig. 6. Graph for determining the discharge-formation time \((t_{\phi})\).

rock salt in a nonuniform field, the same delay time was obtained in both cases [1].

At an exposure of \(2 \div 3 \cdot 10^{-8}\) sec, identical values of electric strength were also obtained for non-X-rayed rock salt and for X-rayed rock salt, broken down under strong illumination [2]. Taking into account that in X-rayed crystals under strong illumination a large number of free electrons is formed, it follows that in rock salt and, apparently, in the other dielectrics investigated, the statistical delay time at an exposure of \(2 \div 3 \cdot 10^{-8}\) sec is practically absent. Therefore it should be considered that the discharge-delay time in the solid dielectrics studied is the discharge-formation time under the given conditions of voltage rise. The discharge-formation time \(t_{\phi}\) was determined from the voltage oscillogram, as shown in Fig. 6, where \(U_0\) is the breakdown voltage at that duration of application for which there is no delay of dis-

series, \(U_t\) is the breakdown voltage when there is already a discharge delay. In this, the values of the breakdown voltages were taken at a breakdown probability \(\psi = 90\%\).

Assuming that, during its formation, the discharge passes through the entire thickness of the dielectric \(d\), we determined the average velocity of discharge propagation,

\[ v_{\mathrm{av}}=\frac{d}{t_{\phi}} . \]

The calculated values of \(t_{\phi}\) and \(v_{\mathrm{av}}\) are given in Table 1.

Table 1

From Table 1 the following conclusions may be drawn:

1. The average velocity of discharge propagation in solid homogeneous dielectrics of high strength (\(\sim 10^6\) V/cm) is of the order of \(10^6\) cm/sec. Such a value may be adopted in an approximate calculation of the volt-second characteristic of homogeneous solid insulation of electrical apparatus.

Figure 7 gives the volt-second characteristic of the coil of a synchronous motor of type AMSO-1584 (600 kW, 6000 V), obtained experimentally by N. A. Kozyrev and D. V. Shishman [3] and replotted in relative units along the ordinate axis. Below is given the volt-second characteristic calculated by us from the condition that the velocity of discharge propagation is equal to \(1.5\cdot10^6\) cm/sec and the thickness of the insulation layer is \(d=1.5\) mm (10 layers of micanite 0.15 mm thick [4]). Under these conditions the delay time is

\[ t_{\phi}=\frac{d}{v_{\mathrm{av}}}=\frac{0.15}{1.5\cdot10^6}=10^{-7}\ \text{sec}. \]

The method of constructing this curve is clear from the drawing.

In Fig. 7 it is seen that the two characteristics differ considerably; moreover, the gentle slope of the experimental characteristic can apparently be explained by the nonuniformity of the insulation of the tested section of the coil and by the presence of air inclusions. Breakdown of such insulation evidently had a complex, stepwise character. At first glance it may seem that nonuniformity of the insulation is advantageous from the standpoint of lightning protection, since the experimental characteristic lies above the calculated one. However, if nonuniformities in the insulation were avoided and all voids were completely filled with solid dielectric, the experimental characteristic would be more gently sloping, but the breakdown voltages would be considerably higher.

1. The average velocity of discharge propagation in the case of a uniform field in thick layers (\(d=0.5 \div 1.5\) cm) is several times greater than in thin layers (\(d=0.15 \div 0.3\) mm).

1. The average rate of propagation of the discharge in rock salt proved to be considerably greater in the case of positive polarity of the point. The same regularity was obtained in experiments on the breakdown of KCl, KBr, KI single crystals and organic glass. This regularity shows that pulses of positive polarity are more dangerous for insulation when acting on an electrode having a large curvature.

Fig. 7. Experimental 1 and calculated 2 volt-time characteristics of a section of the AMSO-1584 coil.

1. With increasing overvoltage, the rate of discharge development increases.

Literature

1. Walter A. F. and Inge L. D., DAN SSSR, vol. 2, No. 2, 65, 1934.

2. Vorob’ev A. A. and Vorob’ev G. A., Investigation of the electrical breakdown of rock salt containing color centers. Izvestiya AN SSSR, Physical Series, XXII, 4, 397—400, 1958.

3. Kozyrev N. A. and Shishman D. V., Proceedings of the Scientific-Technical Session on Overvoltages, GEI, 1950, pp. 293—310.

4. Edited by Prof. L. I. Sirotinskii, High-Voltage Engineering, part 2, GEI, 1953, p. 185.