[Trade Journal]
Publication: Journal of the American Institute of Electrical Engineers
New York, NY, United States
vol. 42, no. 10, p. 1097-1102, col. 1-2
The Production of Porcelain, or Electrical Insulation
BY FRANK H. RIDDLE.
Associate, A.I.E.E.
Champion Porcelain Company, Jeffrey-Dewitt Insulator Company
Review of the Subject.--Plant control check tests are of importance primarily to the manufacturers of a product. These tests are, however, of considerable importance to the consumer. The lower a manufacturer's losses are, the cheaper his production costs will be. As high losses are caused by defects in the product it is evident that the fewer pieces there are with noticeable defects in them, the fewer pieces there will be that have minor defects in them that will escape detection even with the most careful inspection. Tests that will pick out the pieces that have these minor defects are of great importance to the consumer.
Porcelain should be non-porous and tests which will make it possible to cull porous pieces are of paramount importance. If the fuchsine dye penetration method is to be used for this porosity test the selection of the test specimens must be made in a definite and dependable manner. If the firing of the porcelain is not done in a uniform manner so that the location of the pieces most likely to be porous are definitely known the test of any one piece would be of no value and it would obviously be impossible to test every piece.
Where a continuous car tunnel kiln is used and the cross sectional area is relatively small the selection of the proper test specimen is not only possible but has been used successfully for several years.
A recent development for testing disk insulators comprises a mechanism for subjecting each of the insulators under test to a pull test of 5000 pounds for two minutes, and simultaneously subjecting the pieces to the high-frequency oscillator test. This eliminates any doubt as to whether the porcelain would resist the two tests when applied at once.
A petrographic study of thin sections of various porcelains is of great interest, in fact it is essential if the manufacturer or consumer desires to know something of the structure of the product, the extent to which the pyrochemical reactions have progressed and what the variations are from time to time.
Several photomicrographs of various types of porcelain are illustrated. These are selected from a wide variety of wares in order to show the various steps in the development of different qualities of porcelain.
The unlike thermal expansion and contraction of various porcelain ingredients is undoubtedly the cause of some of the deterioration of aged porcelain. It is evident that if some of the particles in the porcelain are under stress due to their tendency to contract more than the surrounding glassy matrix during cooling, after firing, there will be a tendency for these particles to rupture and break down in order to relieve this local strain. One of the illustrations shows this very well.
Overfiring causes the development of gases in the body at a stage when the glassy matrix is in a molten condition. Continued firing causes expansion of the gases and results in the development of a vesicular structure. The degree of overfiring governs the size of the vesicles and the extent to which they have become interconnected. Too glassy a structure also develops brittleness and is to be avoided. Overfiring is well illustrated in Fig. 50.
CERAMIC TESTS ON FINISHED PRODUCT
IN the manufacture of porcelain, if the quality of the product is to be maintained, it is absolutely necessary that control tests be made at every step in the manufacturing process and that the tests on the finished product be reliable. If tests are to be reliable the selection of the specimen for test is of paramount importance.
Tests of this character, particularly the plant control tests, are of interest chiefly to the manufacturers themselves. The consumers' greatest interest is in the tests of the finished product. If tests on manufacturing processes are made in such a way that the plant manager can analyze the reasons for various losses, they should also be of interest to the consumer as he knows they tend toward lower cost of manufacture and also lessen the chances for defective pieces passing through the final inspection. The higher the losses are in the factory the more pieces there are going through production with minor defects which tend to lower the percentage of perfect pieces, and the greater are the number of pieces which appear perfect and yet have minor defects.
It is not the intention of the writer to deal with any of the tests of finished insulators as regards their electrical characteristics. There are several important tests, however, of a ceramic nature which are essential to the successful manufacture of vitreous porcelain.
Porcelain of this character must be free from mechanical strains, absolutely vitreous and non-absorbent and must have a high physical strength and as uniform a texture as possible.
Since it is possible to fire porcelain so rapidly that the outside is vitreous and non-absorbent and still have it porous in the heart it is necessary to devise tests to determine this condition in a positive manner. If the firing is erratic so that it is impossible to pick out the most porous pieces for test, it is evident the test is of no value unless each piece is tested. This is obviously impossible. The only thing to do under the circum-stances is to see to it that the firing is right. It is the writer's opinion that this is impossible with periodic kilns, excepting in very rare cases. It is possible, however, with properly constructed and fired tunnel kilns. Reasons for these statements have been previously given.(1)
Fig. 38 shows the method of unloading a Dressler kiln car of fired insulators and placing them in crates for inspection. It has been shown that because of the small cross sectional area and the method of firing a tunnel kiln the lowest temperature and consequently the greatest chance for underfiring occurs in the bottom of the center bung of ware on the car. It is obvious that if a piece of ware that has been fired in this location is non-porous, every other piece of ware on the car is also non-porous. In using the method of that is illustrated, the crate used is of such a cross sectional area that one horizontal layer of insulators taken from the car will nicely fill the cross-sectional area of the crate. A layer of insulators is removed from the car and laid flat in the crate and in the same relative position to the other insulators in the same layer as it was on the car. This layer is covered with a stiff piece of beaver board. Then the second layer is unloaded from the car and placed on top of the first layer in the crate in the same manner and this is re¬peated until the car is completely unloaded and the center insulator from the bottom row of ware on the car is placed in the center and on top of the ware loaded in the crate.
This insulator, which is obviously the lowest fired insulator of any on the car is selected by the inspector as the piece to test for porosity. There are several methods of accomplishing this test. In the writer's opinion the fuchsine dye penetration test is the most satisfactory. This consists of breaking the specimen into pieces and selecting those pieces for test that are the nearest the center or heart of the piece. Obviously this part of the insulator would show underfiring if any part would.
Fig. 39 illustrates the fuchsine dye penetration equipment as regularly used at one of the insulators plants. This consists of a cast steel pot of suitable size, so arranged that when the broken porcelain test pieces are placed in it and the cover is clamped on, a solution of fuchsine dye in alcohol can be pumped in and a pressure of 200 lb. per square inch maintained over a period of two hours. The pot is large enough to hold the test samples from one day's run. Each piece is properly marked before it is tested. Alcohol is capable of absorbing air so that the colored alcohol easily displaces any air entrapped in the porous part of the porcelain if there be any. When the specimens are removed from the pot after testing, they are placed in a dryer and all alcohol thoroughly dried out, thus preventing it from running over the surface when the pieces are broken to see if there has been penetration. The slightest sign of penetration can readily be seen as the porous parts are stained red. Fig. 40 shows a broken specimen of porcelain that failed in this test with a second sample that did not fail. As the porcelain is white and the dye red the contrast is much greater than the illustration indicates.
If there should be any penetration on any of the test blocks the car from which the piece was selected is "stopped," the remainder of the cars being sent on for visual inspection. All of this first row of insulators are rejected. A second test insulator is taken from the car on which the first failed or showed penetration. This second test piece is the one from the center bung of ware and in the row directly above the first test piece and it is the piece most likely to be underfired of all those in the second row. If the test specimen from the second row is free from penetration the balance of the car of ware is passed for final inspection and as¬sembly. If, however, it shows penetration, this row is also rejected and a third test made. If the third test is a failure, the entire car of eight rows is rejected.
The outside edges of the top saggers on each bung are exposed to the greatest heat, hence the skirts of the top insulators are the first ware to show overfire. Overfire naturally occurs in the most exposed and thinnest part of the insulator first and can be detected by small blisters on that part of the ware. These blisters can be seen much more easily when clear transparent glazes are used than they can when opaque glazes are used. The blisters are a surface indication.
When an overfired piece is broken open it will show many voids usually along places horizontal to the surface. The depth these voids go into the body of the porcelain depends entirely upon the degree of overfiring. When a piece of porcelain overfires it causes an evolution of gas from some of the particles comprising the porcelain. As these gases are trapped in the vitreous glasslike mass, they cause bubbles, the size of the bubbles depending on the viscosity of the mass and the pressure of the gas. The pressure of gas increases of course, as the temperature rises after the gases have been generated, due to the expansion of the gas. The bubbles are naturally larger near the surface, particularly where glazed, as the glaze tends to soften the surface of the body somewhat and make it less viscous. If overfiring continues, several small voids eventually form one large void and in some cases will find an outlet to the surface. Fig. 50 shows the formation.
The fuchsine dye penetration test is not dependable for detecting overfire unless the voids have become interconnected as the walls of the voids are vitreous and gas tight. Overfiring in itself is not so serious as underfiring; however, it should be avoided, particularly if bad, as it indicates that the pyrochemical reactions have been carried too far and the porcelain is likely to be too glasslike and brittle as compared to properly fired porcelain.
A COMBINED MECHANICAL AND ELECTRICAL TEST
Although it is not the purpose in this article to discuss any electrical tests, it is of interest to mention a recent development in testing apparatus. Fig. 41 illustrates a combined mechanical and electrical test of 5000 pounds and flashover voltage from the high-frequency oscillator for a duration of two minutes.
THE USE OF THE MICROSCOPE AS AN AID IN THE STUDY OF THE QUALITY AND STRUCTURE OF PORCELAIN
As grain size, degree of solution of the more nearly insoluble particles, the degree of development and character of sillimanite and freeness from foreign matter and voids all have an effect upon the finished porcelain, it is evident that an examination of these qualities should be made by all parties interested. This can be best done by use of the petrographic microscope either by examining the powdered porcelain, or preferably by examining thin sections cut from the porcelain. These sections are usually 0.001 inch in thickness.
The accompanying series of photomicrographs has been selected with a view of illustrating several different types of porcelain from a poor quality insulator porcelain to a high grade spark plug porcelain.
Figs. 42 and 43 are photomicrographs of a coarse-grained high-tension insulator porcelain. The average porcelain is better than these but they are shown to illustrate what will result if the fineness of grain of the raw materials is not checked where the body is prepared from coarse materials by blunging or disintegrating in water rather than by grinding.
Figs. 44 and 45 illustrate a porcelain rather above the average, particularly in its sillimanite development. Figs. 46, 47 and 48 illustrate what can be accomplished by the use of more active fluxes than feldspar, proper grinding and increased temperature.
The development of sillimanite and solution of the components of the porcelain takes place only very slightly at cone 9 to 10 even under the best of conditions. There is a definite increase in this development that starts at cone 12 down and continues rapidly as the temperature increases. At cone 17 to 18 the development is remarkable. There is a noticeable difference in insulators fired from cone 9 to 10 and those fired at cone 12 D.
The mechanical strength of a porcelain depends more upon the temperature to which it is burned than any other one thing. The higher temperatures produce the stronger bodies. It is evident from this that the tendency among the manufacturers is going to be to increase their burning temperatures and increase the fineness of grinding of the body ingredients. This will mean a greater cost but the quality of the ware will still warrant the increase.
Fig. 49 is of particular interest. Note the triangular-shaped quartz grain in the center of the field. This piece is cracked into two pieces. This is not an uncommon occurrence and it is to be attributed to the fact that the porcelain is under constant strain due to uneven expansion of various parts of the aggregate. As was illustrated in a previous chapter quartz has a very definite and pronounced volume change at 575 deg. cent. which is reversible on cooling. This volume change is much greater than the volume change due to the thermal expansion of the porcelain mass as a whole. When the porcelain is fired and still at its maturing temperature the mass as a whole must be fairly free from strains as the fluxes are mobile and will adapt themselves to the volume changes of the other particles. As the porcelain cools, however, the flux freezes and as the cooling continues, if there be any difference in contraction between the various parts, strains are bound to result. When 575 deg. cent. is reached the quartz grains contract materially while the surrounding mass does not. The result is obvious. In some cases the grains actually rupture. In other cases there are many fine relief cracks around the grains. These are only noticeable under high magnification. It is to be expected that this constant strain will bring about a certain deterioration as the porcelain ages and it is quite likely that deterioration due to ageing can be at least partially explained by this condition.
Fig. 49 is an excellent illustration of how overfiring produces gas voids in the porcelain. As this overfiring continues the increase in voids and their interconnection occurs more and more rapidly until finally some voids will run from deep down into the body of the porcelain to the surface. Severe overfiring is bound to result in failure.
The writer wishes to express his thanks to Dr. Joseph A. Jeffery for his cooperation and criticism in preparing these articles, also to Mr. A. V. Bleininger for his suggestion and criticism, and to Prof. Albert Peck for preparing the photomicrographs.
1. See Part VI.—"The Production of Porcelain for Electrical Insulation" under the caption "Uniformity of Heat Treatment".
