[Trade Journal]
Publication: General Electric Review
Schenectady, NY, United States
vol. 16, p. 136, col. 1-2
The Influence of Glaze on Insulator Strength
Physical Characteristics—Mechanical Strength with and without Glaze—
Effect of Glaze
By D. H. ROWLAND
Research Engineer, Locke Insulator Corporation
When the wet process porcelain insulator was originally developed it was only natural that its manufacture should follow the methods established by experienced potters. These workers skilled as they were in the creation of useful pottery were here faced, however, with entirely new problems. Nothing was known of the inherent characteristics of electrical porcelain and even when such problems were tackled by mechanical engineers difficulties of a far-reaching nature were encountered.
 |
| Fig. 1 |
Some of the earlier work was very misleading in its character owing to unaccounted for variables. Ordinary physical characteristics of porcelain such as tensile strength, compressive strength, modulus of elasticity, and coefficient of expansion, seemed almost impossible to arrive at. For a long time the tensile strength of good porcelain was supposed to be as low as 1500 lb. per sq. in. This tensile strength has since been determined as between 5000 and 6000 lb. per sq. in. on a cross-sectional area of one sq. in. It is known now that this low figure was due to extreme brittleness of the material, which makes it almost impossible to deliver a test load to a sample, and to distribute this load in such a manner as to avoid a progressive failure. Such a tendency is well known in the case of cast iron. In porcelain it is much more severe. When pulling specimens as shaped in Fig. 1-A, pieces of non-ductile material, with sharp corners, will break along a line cc at lower values than will rods of the same cross section rounded off as in Fig. 1-B. From a realization of this characteristic it is only a step to a realization of the tremendous effect which surface conditions have upon the mechanical strength of porcelain.
It is the object of this article to present certain test data in this connection, and to point out certain features of the mechanical design of porcelain resulting from this work.
 |
| Fig. 2 |
Hundreds of porcelain rods, 1N by 6 in., were made up and. broken by transverse load application as in Fig. 2. These rods were made with the greatest care and fired in the same saggers, the different groups being intermingled so that all received identical treatment except in the case of one variable, i.e., the surface condition. The rods were divided into 8 groups of approximately 25 to 30 rods, each group being dipped in a different type of glaze with the exception of Group 1, which was left unglazed. In all cases the glaze appeared to be normal with the exception of Group 8, which was purposely treated with a "bad-fitting" glaze. In this case crazing of the glaze after firing was apparent to the naked eye. The thickness of the glaze was in every case less than 1/64 of an inch. Table I gives the average strengths of these various groups together with a mean variation.
Taking the unglazed rod as a base it will be seen that it is possible to increase, or decrease, the strength of the porcelain almost at will by glazing. Group 8 shows a reduction in strength of over 60 per cent, whereas Group 5 shows practically an identical increase in strength over the unglazed porcelain.
 |
| Fig. 3 |
The explanation of this phenomenon is quite simple. Fig. 3 shows a sketch of a magnified section taken across a hair crack in the glaze on a piece of porcelain. If the glaze were in tension due to some external load, and if the angle A were almost zero, it is evident that the stress at the point A would be practically infinite, and the crack would extend into the material. Further, it is obvious that the less ductile the material the greater this stress would be. If the specimen were of a soft material, instead of cracking the small defect shown in Fig. 3 would smooth out without causing a marked decrease in mechanical strength. Porcelain, however, is much on the order of glass and, as with glass, the first step if it is desired to break or cut it is to make a small surface scratch.
Moreover a piece of porcelain may appear to be perfect in all respects and yet the glaze may be under a condition of tensile stress such that in a few years' time crazing will take place. Even if this crazing does not occur, the tensile stress in the glaze will be added to that produced by some external load, or to the load caused by thermal changes, and thus the ultimate strength and lasting quality of the ware will be greatly decreased. The evidence of this is shown in Table I. Group 2 and Group 5 are the same except for a slight difference in the fit of the glaze. It is also reasonable to expect if this theory of the mechanics of failures is correct that the properly glazed specimens would be stronger than the unglazed specimens with their relatively rougher surfaces. This is also borne out by experiment as shown by Groups 1 and 5. Group 1 is unglazed.
A thorough understanding of the importance of the above principles has led to many interesting practical applications in porcelain insulator designs. Important increases in mechanical strength have resulted.
It is evident that the first problem is one of ceramics. The glaze and porcelain must be so balanced as to avoid tension in the surface. All glazes should be based on such molecular formulas that changes in colors will not affect the physical characteristics.
Sometimes this is very hard to accomplish, especially since in America electrical porcelain, unlike decorative pottery, goes through only one firing. Avery small quantity of coloring oxide will often influence the strength of a glaze quite markedly. As a general rule, besides increasing the strength of porcelain, a good surface will bring the high and low values of the separate pieces much closer together. This, of course, is greatly to be desired.
It is impossible at present to make separate trials on a particular glaze and porcelain, and predict absolutely how these two materials will work together. They must be combined and tested for strength as has been described above. The preparation and firing of the samples Must be done with extreme care in order to avoid the introduction of unaccounted for variables. In designs where sand fired in the glaze is used for cement grips, it is of extreme importance to get enough glaze between the sand particles otherwise the surface of the porcelain will be materially weakened. This point cannot be over-emphasized.
It is beyond the scope of this article to go into the methods of plant control, inspection, and testing which are an essential part of good porcelain, but it can be said with certainty that the best engineering in the world will be of no avail in producing good insulators unless the ceramic work is complete to the last detail.
Let us now consider the mechanical design of a porcelain insulator. The point at which the mechanical design of an insulator diverges from everyday routine principles, and the point at which trouble has so often occurred in the past, is the joint between the metal hardware and ceramic material. Principles which apply to much assembly work with other materials must here be thrown aside, and instead of an absolutely inflexible joint these joints must be decidedly flexible in character. They must be capable of slipping an infinitesimal amount under stress, and thus distribute the load over the entire surface. If such movement is not provided for, concentrated point or line stresses cause progressive failure of the porcelain and the insulator will fail at a very low figure.
This is nowhere better illustrated than in the design of bus bar supports. Where strength is not the main consideration the deciding factor in the design becomes leakage distance. Even so it has been found possible, without decreasing the leakage distance, to greatly increase the strength of such supports. The modification necessary to accomplish this, and correct glazing of the contact surfaces, allows the necessary amount of movement to distribute the load properly.
But where strength is a necessity, principles developed by the above cited tests come more fully into action. Supports of this type, from a mechanical standpoint, are designed on the basis of the side pull from the upper hardware which they will carry without breaking. Fig. 4-A shows the lower hardware of a typical bus bar support in section with the porcelain fitted into it. The bond between the metal and the porcelain is usually Portland cement.
 |
| Fig. 4 |
In this design very little flexibility occurs in the joint between the porcelain and the hardware with the result that side pull places a definite concentrated line of stress on the porcelain at the joint. Unless the glaze is exact in every detail this means failure at a low figure. With a glaze such as that used on the rods shown in Table I, Group 5, this piece will develop its highest possible mechanical value. This value can be still further increased by so modifying the design as to allow an infinitesimal movement of the porcelain within the hardware which distributes the load when a pull is exerted. This type is shown in Fig. 4-B.
In the above discussion the very slight movement between the metal and porcelain, necessary from the mechanical standpoint, was between the cement and insulator. It is perfectly possible to design so that the movement is between the metal and cement. The disadvantages of this is that it is more difficult to obtain uniformity owing to the relative roughness of the metal surface. In all cases the amount of flexibility in the hardware is of great importance.
 |
| Table 1 |
In the past there has been a great deal of discussion as to the possible aging of porcelain in service. It is only natural for those interested to wish to know what influence the surface condition of a piece of ware, which affects its strength so markedly, would have on its lasting ability in service. There is no definite answer to this question. We do know, however, that crazing of the glaze on insulators which are poorly made sometimes takes place after several years in service. There is no doubt that such a condition would go a long way towards producing an incipient failure due to a progressive crack gradually working through the porcelain. Undoubtedly, where the insulators are subject to much vibration or shock, the surface strength is of extreme importance.
From an electrical standpoint little difference has been observed between ceramic materials with a good glaze and those with a poor one. In this connection it must be remembered, however, that in nearly every case of failure of modern insulators that a mechanical crack is the initial cause of the electrical failure.
Summary
(1). It is only recently that electrical porcelain has reached a state very far removed from old-time pottery, from which it originated.
(2). Designers have been handicapped by not understanding the physical characteristics of porcelain.
(3). It has been found that, other things being equal, the surface condition of porcelain has a tremendous influence on its strength, especially when the surface is subjected to a tensile stress. This point most certainly has been overlooked in the past.
(4). Ceramic materials have certain characteristics similar to cast iron but much more accentuated. In taking hold of porcelain with hardware, provision must be made for proper stress distribution.
(5). There is not much doubt that the surface glaze has an effect on the so-called aging of porcelain, especially when subjected to vibration.
Surface conditions do not influence the dielectric strength of ceramic material when no concurrent mechanical load is applied. Nearly all electrical failures of insulators are preceded by a mechanical fault.
