[Trade Journal]
Publication: Transactions of the American Institute Of Electrical Engineers
New York, NY, United States
p. 47-54, col. 1-2
Development of the Porcelain Insulator
BY K. A. HAWLEY (1)
Member, A. I. E. E.
Synopsis.—Porcelain insulators have been manufactured and used for the transmission of high-tension electric power for forty years. The first designs were of the single-piece and multipart cemented pin type. Necessity for higher safety factors against flash-over and increase in operating voltages demanded a rapid increase in the size of the insulators. This reached an economic limit at the operating voltage of 66 kv. The suspension unit overcame this temporary check of increased operating voltage.
Further study of the electrostatic capacitance of the various parts and consequent voltage distribution, made marked refinements in the pin type insulator possible. During this time the single-piece porcelain suspension unit took practically its present form.
Early improvements were the provision of proper expansion joints and the separation of the lip of the cap from the porcelain hood.
Gradual improvements have since been made resulting in a great increase in mechanical strength. These changes have been principally in hardware design. By experiment and analysis the shapes of the cap and pin have been determined to give a uniform distribution of load from the pin to the cap. Constant check tests by the quick pull and time loading methods have shown that the suspension insulator with properly designed hardware and a suitable coating on the cap to prevent the cement from adhering to the metal, has a high strength associated with electrical reliability.
Ceramic research and exact manufacturing control has made possible the production of non-absorbent, thoroughly vitrified porcelain of consistent strength. This -has centered largely about the proper firing of the clay.
Recent experiments upon the properties of the combination of porcelain and glaze has eliminated surface stress and consequently assured stronger, longer lived porcelain. Still greater uniformity has been gained by glazing the sanded surfaces.
The elimination of the abutting joint and the proper design of the cemented joint has stopped expansion troubles. Proper use of Portland cement has resulted in insulators able to withstand drastic temperature changes without harm. A recent improvement in the pin type insulator is the metal threaded pin hole. This has lessened manufacturing and construction difficulties and in addition, due to the exact fit of the insulator on the pin, overcomes hidden corona and the consequent radio interference.
PORCELAIN insulators for the transmission of high-tension electric power have been made and used for a period of approximately forty years. Progress in their design and manufacture can be divided into ten year periods.
The last decade of the nineteenth century marked practically the beginning of electric transmission in America and also the beginning of the continuous manufacture of porcelain insulators. Insulators of this decade were all of the pin type, both single-piece and multipart cemented types. Many of these old insulators are still in use, although rarely at the operating voltages for which they were originally designed. The safety factor against flash-over was rarely sufficient for practical requirements so where the voltage has remained the same larger insulators have in most cases been substituted.
During the first ten years of the twentieth century pin type insulators advanced rapidly in size. Even at this early date it has reached its maximum economical size. For 66,000-volt transmission, pin type insulators proved to be quite satisfactory, but since the cost of the insulator varies as a power of its rating larger insulators of this type were not practical from a cost standpoint. The suspension insulator was the logical outcome of this temporary check. Comparison of the present pin type insulators shows that the cost is proportional to the 2.6 power of its flash-over value. The suspension insulator, however, offers a resistance to flashover in direct proportion to the string cost. Plotting separate cost curves for equal arc-over values show that the curves cross at approximately the 66,000-volt operating range. This corresponds very closely to the usual practise and for the higher voltages the suspension insulator has almost invariably been used.
During the next decade from 1910 to 1920, marked refinements in the pin type insulator were made. The earlier insulators were little more than porcelain cups cemented together with but a slight understanding of voltage division.(2) The division of voltage is inversely proportional to the electrostatic capacities of the respective parts. These in turn are proportional to the adjacent conducting surfaces upon the porcelain, the cement being a conductor. As these surfaces were generally small on the central and top shells, these two parts carried most of the voltage. By careful study this condition was rectified and the correct balancing of the pin type insulator was one of the decided advances of this period.
During this decade the suspension insulator took practically its present form. The unit using two pieces of porcelain, while in many ways satisfactory, necessitated a larger head diameter and a correspondingly large size of metal cap. The single porcelain unit removed these unsatisfactory features and was accordingly almost universally adopted.
Experience promptly pointed out necessary refinements. Proper expansion joints were provided and the cap which originally had been allowed to rest on the porcelain was lifted, so that under all conditions it stayed clear of the hood. (Fig. 1.) With these improvements practically all field failures stopped. Insulators even of the largest size have established a perfect life record over a period of fifteen years.
During the ten years just completed there has been a decided improvement in the suspension insulator. This improvement is not the result of any radical changes but has been rather a matter of refinements of design and manufacture. Slight changes in hardware shape, for example, have resulted in a more uniform distribution of the load delivered to the shell with a consequent greater uniformity and higher average mechanical strength. The desirability of such improvements is constantly evident. Unexpected loads greatly in excess of maximum calculated loads have been encountered and the strength of the insulator has been exceeded.(3) Three such failures upon insulators of the old type have occurred in the past two years in widely separated localities. It was at one time thought that the strength of the insulator was limited by the strength of the standard cement used, breakage resulting in loading tests from crushing and shearing of the cement and resulting in unfavorable loads on the porcelain. This is not so. A standard pin when cemented into a metal ring held consistently loads as high as twice the breaking strength of the insulator.
Following this a cup of porcelain was prepared representing the head of an insulator. Into this a standard pin was cemented in the usual manner and the cup was then supported on its lower edge on a metal ring. When pulled in this manner the standard pin was broken before it could be pulled from the cup. There was no evidence of bursting the cup unsupported except at its lower edge.
Following the many efforts to improve the pin shape, attention was directed to the cap. An analysis based upon directions of stress from cap to pin indicated possibilities of improvement. Changes in caps were made by filling the cap with babbitt and turning to the desired shape and as high as fifty per cent increase in combined electrical and mechanical strength resulted.
These caps when copied in malleable iron, however, showed no such improvement in performance. A close scrutiny of the previous tests on the caps containing the babbitt showed that under strain a slight slipping took place between the cap and the cement. Evidently the malleable iron caps must also be made to yield upon the cement. This research led to the painting of the inside surfaces of the cap with a suitable coating to prevent the cement from adhering to the metal. When this was done the desired results were obtained.
It was soon recognized that strength in quick pull-off was not always an indication of reliability. High strength against pulling apart must be associated with electrical reliability. The comparison of various available standard rated insulators brought interesting results. The insulator that had been heralded as the strongest showed on time test the earliest electrical failure. The following table shows this in detail.
The last two columns in the table show the results of the previous research. The provision of a yielding abutment to the arch of the insulator with no changes otherwise had greatly strengthened the unit. By no known test could any hazard be shown.
Since then, constant checks and rechecks with slight changes in shapes have further improved the insulator. All such changes have been tried first by the quick pull (A. I. E. E. Standard 41-305; two thousand pounds increase per minute) and by time loadings. Such time tests are made in out-of-door frames subject to considerable vibration and all changes of the weather at both Victor, New York and Baltimore, Maryland. In addition to this, periodical check tests are regular routine. Sample check test report follows:
PERIODICAL CHECK TEST
A. Puncture (A. I. E. E. Standards 41; 153)
(1) 145,000; (2) 140,000; (3) 152,000 volts.
B. Mechanical Ultimate and Electrical Test. (A. I. E. E. Standards 41: 154). (1) 16,000; (2) 15,300; (3) 16,000 lb.
C. Combined thermal—tension test.
Water temperatures, 205 deg. fahr. and 39 deg. fahr.
Ten minutes in each alternately.
Initial load 4,000 lb., increase 1,000 lb. per cycle.
Two units tested. First failure at 12,000 lb.
D. Time loading tests. Out-of-door any weather.
Initial load 8,000 lb., increase 1,000 lb. each working day.
Arced over before each increment of load.
Two units tested, first failure 17,000 lb.
Efforts have been made to distribute the load upon the outside of the porcelain more uniformly by multiple stepped caps. No advantage in any test was observed and in many cases decidedly unsatisfactory results were forthcoming. The thin metal of the cap apparently yielded so that the upper step would carry more than its share of the load. Annular ridges within the caps are of the same character and are generally worse than useless.
Similar tests upon the pin, however, have always shown advantages from the multiple stepped pin, provided such steps are above the lower edge of the insulator cap. In this case, of course, the mass of the pin is sufficient to prevent any yielding.
Most of the early porcelains were thoroughly vitrified, but only too often with an excess of flux (feldspar). The comparatively thin sections allowed quick firing and rapid cooling. Such over-fluxing and rapid cooling can only result in ware that is lacking in strength, and there is little doubt that differences in performance of individual specimens of the same type and lot of the old insulators, were largely the result of variation in strength.
Over firing to the beginning of volatilizing of some of the ingredients (ceramically known as "bloat") has rarely been the cause of trouble. In fact, experience with such ware suggests that this condition may possibly be desirable rather than hazardous.
In some cases, under firing may have resulted in a short lived semi-vitrious ware which would be a cause of trouble. With the exact manufacturing control exercised today such conditions no longer exist and need not be considered in this discussion. The best porcelain today is recognized to be one of greater sturdiness, less flux, thorough vitrification, and careful cooling. Heavy sections have been demanded to withstand external violence, power arcs, punctures, and various forms of mischievousness. Lesser flux means greater mechanical and electrical strength, but with this reduction in flux in the thicker sections far greater care must be exercised in firing. Improper firing will invariably result in internal stress in the porcelain which seriously affects both the strength and life of the ware. Today constant thermal cycle checks are employed (A. I. E. E. Standard 41-250; 350). Such tests are largely a measure of the care used in kiln cooling. Regular porosity checks are made to insure thorough vitrification and proper coordination between all details of firing.
Recently greater strides have been made in the combination of porcelain and glaze. If the glaze covering the porcelain does not "fit" due to its coefficient of expansion being different from the porcelain it will be under stress. When this stress is above the elastic limit of the glaze crazing or crackling of the surface will be apparent. In the majority of cases, however, there are no visible cracks and yet the strains are there. Add an external load either of a mechanical nature or due to thermal changes to those already inherent in the surface of the dielectric, and small cracks will develop which will form the basis for a progressive failure of the porcelain.
Making and breaking in the testing machine comparative rods with various glazes pointed the way to the elimination of surface stress and the consequent development of stronger, longer lived porcelain. As an even more recent development, further strengths and greater uniformity have been gained by glazing the sanded surfaces.(4)
The abutting joint in the insulator assembly wherever found has been a source of trouble. As already explained, the separation of the cap of the suspension insulator from the horizontal hood has stopped losses in that type. Similar breakages occurred when the caps of switch type insulators were placed upon the porcelain so that the cement would bear upon the fillet between the head and the horizontal hood. Wherever this has been done, especially in climates where freezing occurs, there have been insulator losses. When the cap and the cement within it have been kept above this fillet these breakages have not occurred.
When the cap shrinks with cold or when free water between the cement and porcelain freezes there is a radial pressure against the fillet with an upward component. This reaction becomes a tension stress upon the porcelain—a stress which it is least able to carry.
If the cement is above the fillet the stress is entirely horizontal and becomes purely a compression force upon the porcelain. Against such stress porcelain is one of the strongest materials known.
Another form of the abutting joint is that which was used largely in pin type designs in the past. In this case the joint was between two porcelain parts. This, commonly known as the closed joint, was used chiefly for two purposes. First, such an insulator would be free from corona display at relatively high voltages, and second, the self-aligning features of the closed joint offered a decided advantage in ease of factory assembly. The cause of the breakages experienced with this type of insulator was the same as with the suspension or switch type already described. The outer piece, however, in this case was generally the weaker and was usually the first to break. The edges of the upper porcelain part at the joint were frequently thin and were consequently weakened by over-firing because of this thinness.
Portland cement has been blamed for many insulator failures. It is certain that the chief complaint is not against the cement but against its improper use. It has been found that cement which will successfully pass the A. S. T. M. Soundness Test (A. S. T. M. Standard Specifications and Test for Portland Cement C9-26) is fit for insulator work. As a result special stress is laid on soundness, and every lot of cement before it is released for production is carefully checked as to this characteristic. In connection with this research, cement specimens were alternately soaked in water and frozen to — 20 deg. fahr. There was a slight contraction in the specimen corresponding to such a temperature drop. There was no increase in length which would correspond to a volume increase by formation of ice crystals. This confirms statements made by cement manufacturers that water within the colloidal structure of the cement does not freeze. Any harm done by freezing must be from free water within open voids in the cement.
Let us assume that the cement and central porcelain parts completely fill a porcelain ring outside of it. A ring of porcelain placed about an unyielding central part must shrink according to the following formula to destroy itself.
Examination of the open type of cement joints shows that insulator failures have almost always occurred in those insulators with wide cement joints or with great cement masses, or with overfilled tapered joints.
All cement when drying shrinks. Mixtures which contain the smallest percentage of water are those which show the least shrinkage when drying. Such mixtures are also stronger. The use of 1% sand in the mix will cut down the shrinkage exactly 1 and make the material much more inert. When the cement shrinks against the parts within there is a tendency towards cracking of the cement. This cracking is reduced to a minimum, if not prevented entirely, if the cement joints are narrow, if the cement is dense and strong, and if it is firmly anchored in place by sanded surfaces upon both sides of it. Where insulators with these features have been found there have been no breakages.
On the other hand, where there has been an excessive cracking of this cement or where water can find a passageway between the cement and porcelain the freezing of that water and the consequent increase in volume is frequently accompanied by a failure of porcelain. Such breakages have been observed to occur immediately adjacent to a crack in the cement as a result of localized pressure.
An examination of insulators made over a long period of time shows that they can be separated into groups, one of which may be expected to fail and the other not. This dividing line has been sufficiently sharp and close that a reasonable insulator life may now be confidently predicted. The modern insulator is made with cement joints as narrow as manufacturing tolerances will allow, while the cement used is dense and contains a minimum amount of water in the mix. The joints are not overfilled and the cement is fully set and expanded by curing at a suitable temperature in a saturated atmosphere.
One of the recent improvements in the pin type insulators has been the metal threaded pin hole. It is relatively difficult to manufacture porcelains with the thick sections demanded by modern practise. These thick sections even under the best controlled conditions do not shrink uniformly. This is chiefly due to the condition of the plaster mold. The amount of moisture in the mold varies during different hours in the day due to week end and other interruptions.
While the piece is resting within the mold it partially dries, shrinks, and rests upon its shoulder. This shoulder which is opposite the outer end of the pin hole dries rapidly ahead of other sections. It takes a quick shell like set, and later on restrains the clay immediately under it from shrinking at the normal rate of the other parts of the piece. This results in a distortion of the pin hole which cannot be accurately foretold and compensated for. Realization of this led to the adoption of the metal thread. This thread is now quite uniformly used in the pin type insulators of the higher voltages. Not only does it simplify construction and save time due to the exactness of its fit upon the pin, but in addition it is of special value in overcoming a hidden source of corona which might cause hidden stress and radio interference.
While we believe that the developments in insulator art have been worthy ones, we are by no means drifting into a self-satisfied condition. Organized research is constant and consistent and all indications are that insulators will not be the restraining factors in any future developments no matter how rapid those developments may come.
Discussion
S. Withington: The increasing reliability of insulators for medium and high voltages is a source of a great deal of gratification to all who are interested in maintaining the integrity of power supply, and there is no doubt that there has been satisfactory advance in this direction, as those associated with transmission and distribution of power before the war will agree. It is perhaps an overstatement, however, to say as Mr. Hawley does, that "insulators even of the largest size have established a perfect life record over a period of fifteen years."
It is true that the initial cost of insulators is relatively small when compared with the total construction expenses involved in a transmission line, and it is also true that this cost is relatively unimportant where continuity of service and the cost of replacement are considered. Nevertheless, the fact remains that the initial cost of standard types of porcelain insulators has risen in the past few years more than the normal indexes would seem to justify, and it is to be hoped that when the expenses of development to which Mr. Hawley briefly refers have been amortized there will be a marked decrease in cost of the consumer.
It would be of interest if Mr. Hawley were to expand his reference to the metal threaded hole for pin insulators. This is a development which is especially important in connection with the curing of cement at suitable temperatures to which Mr. Hawley refers.
Steam railroad electrification presents an important field in the use of insulators. In addition to the problems presented by the smoke from steam locomotives operating over electrified tracks, the requirements in general are similar to those of standard distribution and transmission lines, but the support, and particularly the dead-ending of the catenary system, produce somewhat heavier mechanical stresses and it is necessary that insulators be designed accordingly. There are numerous occasions also where on account of close clearance and for other reasons special forms of insulators are required. Pedestal type insulators are often necessary at low bridges, tunnels, etc., and the closest attention of the manufacturers should be given to satisfactorily adapting insulator design to these general requirements. It is probable that with increasing electrification of railroads more attention can be given to this phase of the subject and a satisfactory solution reached for the various problems involved.
A. O. Austin: The general characteristics of insulators having different types of mechanical construction are shown in Table I. In the past losses, due to abnormally high working loads have been negligible although losses due to differential thermal expansion or to concentrated loads have been very high, necessitating reinsulation in many cases.
Where long life and reliability are desired, low stress in the dielectric, due to the combined working load and differential expansion, is far more important than a good factor of safety for the working load based on the maximum or ultimate on test. Mr. Hawley has pointed out that the applied load produces an outward thrust tending to expand the cap, which may be regarded as the abutments in an arch structure.
By applying the proper values to the above equation and setting up a similar equation for the deformation of the porcelain, the radial distortion may be computed. If we assume that this amounts to approximately 0.008 in. for both cap and porcelain, and that the slope of the bearing ring of the cap L Y 0 T, is such that the tangent of L Y 0 T = 0.5, the amount of slip between the cement and the bearing surface of the cap which will compensate for the distortion, can be readily computed, or the
Longitudinal slip = 0.008/0.5 = 0.016 in.
While the slipping readily compensates for the distortion, giving high test values, reference to Fig. 3 shows that with the reduction of load a heavy unbalanced radial component may result which will tend to cause shearing of the porcelain. The coating of the surfaces with asphalt, graphite or metal foil will tend to control the coefficient of friction. The chief difficulty is that the coefficient of friction is likely to change with time.
While a few of the very high ultimate insulators were made some years ago in which the cap slipped to give the highest ultimates, it is believed that the initial stress set up by the temperature of assembly was such that slipping did not take place under maximum working load. The last of the above equations shows that the increase in diameter of the cap due to a given load will vary directly as the diameter and inversely as the effective cross section of the area and the modulus of elasticity. This has long been recognized, as many of the very heavy insulators used for long spans and catenary work have had reinforcing ribs to reduce cap distortion. Heavy cap construction, however, may result in high shearing stresses with low temperatures and light loads unless the pressure is tapered off near the edge of the cap. Reference to Fig. 3 shows the general relation of stress components for the insulator shown in Fig. 5 of Mr. Hawley's paper.
With a heavy plastic material such as asphalt, the angle L during setting up the load may be much smaller than the angle 0 which is effective in restoring the porcelain and cap to normal condition. While this arrangement is beneficial in giving high test value, the shearing stress set up in the porcelain with light working loads may create a hazard. Since porcelain is relatively weak in tension and shear, unbalanced radial components which will set up shearing stresses are dangerous, regardless of the ultimate as they may cause "dough-nutting" which is due to the radial stress from the cap shearing the head of the insulator from the flange at low temperature or under a light load following a heavy load.
Fig. 2 of Mr. Hawley's paper is hardly applicable to insulator design as the large section of metal in the ring surrounding the pin reduces distortion to a negligible quantity. The very short cement strut between pin and ring will have but slight distortion compared to the very long strut made up of cement and porcelain in the insulator. Furthermore, a crack in the cement due to a displacement might not cause failure whereas a crack in the insulator due to distortion would result in dielectric failure with the possibility of mechanical failure due to an explosion from discharge through the fault.
In many of the earlier insulators equipped with sanded surfaces, the sanding was covered with glaze, the sand grains forming pyramids with the surface. At high mechanical loads there is a tendency for this surface to slip and compensate for distortion. It would seem that it is this slipping which increases the mechanical ultimate on test. This, however, is gained at the expense of resiliency in the joint.
The mechanical characteristics which may be developed in the porcelain depend to a large extent upon the method of making the test and the size of samples, the small samples giving abnormally high values which cannot be developed in the larger sections necessary in practical construction. A glaze which will tend to cause crazing will reduce the mechanical ultimate; whereas a glaze which will tend to cause shivering may increase the mechanical ultimate for some conditions. For best conditions, the linear coefficient of expansion of the glaze and body should be the same. It is believed that the latter condition will produce the best results as there are many insulators which have been in service for some years in which there is no evidence of failure. The first of the insulators of the type shown in Fig. 7 were installed in 1914 and have shown no evidence of cracking.
The limit of 267 deg. given by Mr. Hawley applies to porcelain only. Even in this case the figure must be materially reduced for concentrations of stress owing to the shape of the parts. Where porcelain and metal are used together the limits in temperature may be comparatively low. To offset this limitation the resilient joint is exceedingly valuable for either the pin type or suspension insulator.
If the conditions in Fig. 8 cause trouble, they can be largely remedied by coating the mold so as to prevent absorption of water at the shoulder, or by trimming the case hardened surface while the insulator part is soft.
If thin metal thimbles are used for threads there is always danger that the screwing of the insulator on the pin will produce a dangerous longitudinal or radial stress which will cause destruction of the insulator. It is, therefore, advisable to use heavy thimbles which will prevent damage to the insulator. With this arrangement, however, the insulator is subject to a thermal stress which is lacking in the case of the insulator having the thread made in the porcelain. The longest lived insulators are apparently those which have threaded porcelain pin holes mounted on lead tipped pins, of proper design. While the correction for distortion in the suspension insulator pointed out by Mr. Hawley is highly desirable for heavy loads, it is even more essential that the stress return to normal condition with the relief or the reduction of the load. Coatings which flow are not resilient and any unevenness in a coating with rigid parts adjacent may cause a concentration of stress and destruction. While a compensating or resilient structure for the cap or outer cement joint can be used to advantage, the cost of this type of construction is such that with the present high standard of insulators it has been impossible to commercialize the design embodying this feature. Reducing the coefficient of friction of the surface to zero will make the compensated type of construction safe over a much wider range of working conditions. It would seem, however, that this type of construction must be brought about by a roller bearing surface of joint or one equipped with tilting struts.
In the past, insulators having high ultimates have always shown the shortest life. It is, therefore, a serious mistake to use an insulator having a higher ultimate than is necessary for the maximum working load as the effective factor of safety for the maximum stress in the porcelain may be seriously lowered due to increased thermal stress from the stronger metal parts.
The time lag before defects are apparent, together with the wide variety of conditions encountered in service, makes it exceedingly difficult to predict results from design or accelerated tests.
K. A. Hawley: Mr. Withington's statement that the cost of insulators has risen more than the normal indexes would seem to justify is quite true, but it must be realized that the great improvement in the design and quality of high-voltage insulators which has occurred in the last decade has necessitated a great increase in overhead cost. The larger manufacturers maintain million-volt high-voltage laboratories which have a very large carrying capacity. The plant routine is much more extensive than ever before. For example, some years ago the routine electrical test applied to suspension insulators was simply a 60-cycle flash-over on the unassembled porcelains. Today the porcelains are flashed over at both 60 cycle and high frequency, assembled, pulled to 40 per cent of their guaranteed strength, and then flashed over again. The other routine processes of inspection and check throughout the plant have been extended in a manner similar to the electrical tests in order to produce units of the reliability which is now demanded.
Insulators are required to be nearly perfect, surface imperfections being no longer tolerated. This perfection has been reached only at a marked increase in cost which must necessarily be carried by the customer.
Mr. Austin's statement that the linear coefficient of expansion of the glaze and body must be the same does not check up with our test results. It is impossible to test a glaze and a body with respect to the coefficient of expansion separately and then to put them together and to expect thus to get the best results. The intimate contact between the porcelain and glaze causes some of the constituents of the glaze to run into the body and vice versa. We have definitely found that a glaze which will give the highest mechanical strength to a given porcelain will give the best results in insulators.
The use of metal thimbles in the pinholes of pin type insulators cannot possibly be more of a hazard to the insulator when screwing it on to the pin than would be the case with the ordinary plain pinhole and lead-covered pin. The thickness of cement between the thimble and the porcelain will go a long way towards distributing any stress which may be produced, and in preventing a concentration on the porcelain itself which might produce cracking.
It is impossible to agree with Mr. Austin's statement that insulators having high ultimate strength have always shown the shortest life. It is very true that an insulator may be designed which will give a very high ultimate strength and a low mechani¬cal and electrical strength, but if the design is correct these two points will be very close together and it may be expected that such insulators will give a much longer life in service than those of low mechanical strength, other things being equal. The thousands upon thousands of high-strength insulators that have been giving perfect service for years substantiate this.
Due to the heavier line construction which is becoming more and more prevalent, insulator designs that will be the most effective from a mechanical standpoint will be the most economi¬cal from a cost standpoint. It is hard to reconcile Mr. Austin's statement regarding insulators of high ultimate strengths when it is realized that the units which are being made today by a large majority of the manufacturers have mechanical strengths prac¬tically 50 per cent greater than units of the same size which were made a few years ago. This increase in strength has been brought about by better attention to small details and a better under¬standing of ceramics. It is believed that the modern units will give service in the field of a character which has never been. surpassed. It is only fair to give to the operating companies the economic advantages brought about by constant research work in the mechanical and electrical design of porcelain insulators.
(1) Chief Engineer, Locke Insulator Corporation, Baltimore, Md. Presented at the Pacific Coast Convention of the A. I. E. E., Portland, Oregon, Sept. 2-5, 1930.
(2)Insulator Depreciation and Effect on Operation, A. 0. Austin, A. I. E. E. TRANS., 1914, p. 1731.
(3) Transmission Line Construction in Crossing Mountain Ranges, M. T. Crawford, A. I. E. E. TRANS., 1923, p. 970.
(4) D. H. Rowland, G. E. Review, March 1929 and June 1930.
