Telluride Power Company

[Trade Journal]

Publication: Western Electrician

Chicago, IL, United States
vol. 35, no. 18, p. 352-353, col. 1-3


Pioneer Work of the Telluride Power

Company.

BY P. N. NUNN.

During the winter of 1890, the year preceding the famous Frankfort-Lauffen experiment, apparatus was installed for the first commercial, high-pressure, alternating-current power transmission of the world. From that beginning has grown the Telluride Power Company.

The mining district surrounding Telluride, San Miguel County, Colo., is at the same time one of the most rugged and one of the richest in the Rocky Mountains; but its inaccessibility and the consequent cost of producing power caused the financial failure of many important enterprises in the early days of its history. The statement made in the Annual Report of the Treasury of the United States, in 1901, that "For the growth of its mining industry San Miguel County is indebted to the Telluride Power Transmission Company more than to any other agency," is borne out by the fact that at the present time all of the important mines and mills of the district are operated by power furnished by this company.

The Gold King mill, situated at an altitude of 12,600 feet, where the cost of fuel for steam power had become prohibitive, was the first to be operated by means of this power. This property had been attached in 1888 to satisfy a continued deficit in operations. Mr. L. L. Nunn, the attorney retained by the owners, found that this deficit was largely due to the enormous cost of power, and that there would have been a handsome margin if power could have been furnished at not more than $100 per horse-power-year. Down in a deep gorge of the valley, over 2,000 feet lower, but less than three miles away, two mountain streams formed at their confluence the South Fork of the San Miguel River, offering cheap and continuous power. A stay of proceedings was secured; and, as a means of transmitting this power, cable drive, compressed air and continuous-current electricity were all investigated. The limitations of each were apparent, while the advantages of alternating current and higher pressures became gradually recognized, and a decision was reached to attempt their use. This decision was due less to the immediate saving in copper than to a keen sense of the limitation of continuous, and faith in the final success and ultimate superiority of alternating, current.

During the investigation which followed, while selecting apparatus, little but incredulity or ridicule was encountered. Eastern investors in the enterprise were annoyed by predictions of prominent engineers, and discouraged by their insistence that the experiment would prove a miserable failure and the expenditure go for naught. It was said that there was no alternating-current motor; that oil insulators must be used, and that the line must be fenced in. However, a generator and a motor for 3,000 volts and of 100 horsepower each were ready for trial in the fall of 1890. Difficulties caused by ice at 40 degrees below zero, by speed control over unusually high water pressure, by avalanche, by blizzard, by electric storms unknown in low altitudes, and scores of others, now generally forgotten but then most serious, marked every step of progress. Notwithstanding all of these, unqualified success from the beginning caused gradual and constant growth, until at the present time the Telluride company and its allied industries have six power stations and nearly a thousand miles of line in Colorado, Utah and Montana.

Following its pioneer power transmission, it made practical experiments as early as 1895 with pressures which have never, even yet, been exceeded, and for three years it operated commercially the highest pressure transmission of the world. Thus the record of its work becomes an important chapter in the history of power transmission; but it must readily be seen that the limit of this paper precludes the possibility of describing, even in the briefest terms, all, or even a substantial part, of its pioneer work.

The initial installation, purchased through Mr. F. B. H. Paine, comprised a generator installed in a rough cabin upon the site of the present Ames station and belted to a six-foot Pelton wheel, under 320-feet head, and a motor at the mill 2.6 miles distant. The two were identical Westinghouse single-phase alternators of 100 horsepower, the largest then made. The generator was separately excited, while the motor was self-exciting. Each carried a 12-part commutator and was slightly compounded through current transformers upon opposite spokes of its armature. The latter were ironclad, or "T''-toothed. wound with 12 simple coils in cells of fullerboard and mica. Switchboards were matched and shellacked pine sheathing, and the bases of instruments were dry hardwood. Only voltmeters and ammeters were used, both of the solenoid and gravity-balance type, in black-walnut cases with window-glass fronts. Circuits were closed with jaw switches and opened by arc-light plugs. The line carried two No. 3 bare copper wires, mounted upon short Western Union cross-arms and insulators. The copper cost about $700, or about one per cent. of the estimated cost for continuous current.

The main motor was brought to synchronous speed by a single-phase induction starting motor, which received its current at full line voltage. The current taken was more than full-load current of the main motor. This starting motor even required starting by hand, its torque being zero at starting, and so feeble at low speeds that when cold it could only with the greatest difficulty be persuaded to pull up to speed its belt and loose pulley. Nor could it at speed start the main motor without help, and even then it became so hot that its short-circuited secondary frequently burned out.

Another motor of 50 horsepower was soon added. While in other respects similar to the first, this motor was intended to be self-starting, with armature and field in series through a current transformer; and, on account of its frightful flashing, it was fitted with a special eight-part commutator of non-arcing metal. This feature, however, proving a failure, was soon replaced by a separate starter.

The need of wattmeter or power-factor indicator not having been at that time recognized, motor-field charge was adjusted for least main current. This current was accepted as having unity power factor, and therefore as the measure of actual power.

Everything was extremely simple, from water-wheels to motors; and, except for lightning, the plant ran smoothly and steadily 30 days and more without a stop. The report made in the East by associates of the enterprise that at Telluride a hundred horsepower was being successfully transmitted nearly three miles over No. 3 copper, with less than five per cent. loss, was received with the utmost incredulity.

During the autumn of 1892 a 600-horsepower generator of the same characteristics was installed, and a 250-horsepower motor for the mill on Bear Creek, 10 miles from the generator. Early in 1894 a 50-horsepower, and during the fall a 75-horsepower motor were placed in Savage Basin, 14 miles from the power house. The former was soon replaced by a 100-horsepower, and in 1895 a 100-horsepower motor was set up at Pandora.

Except as to size, these motors were substantially identical. The 250-horsepower motor was badly designed, and the pole-pieces were of cast iron. Its starting motor was insufficient, and was therefore soon replaced by one having split-phase secondary with external resistances. Marble, with brass trimmings, replaced wooden-base instruments, and such elegance demanded highly polished slat switchboards of paraffined oak. Imposing marble rheostats were mounted at switchboards like keyboards upon grand organs. Fuse blocks, the only protective device, became marble slabs with duplicate aluminum strips. The first synchrophone came with the 75-horsepower equipment.

Owing to its altitude and geographic position, the Telluride district is peculiarly subject to atmospheric disturbances. Over a hundred distinct discharges have been counted within a single hour, and lightning caused more discouragement than any other obstacle. A neighboring continuous-current plant, transmitting a little more than a mile, carried several extra armatures; and even then it was so frequently compelled to close down during the daily storms of the rainy season, that the company was eventually bankrupted. The alternating plant might have suffered a similar fate, had it not been for its "T''-toothed armatures and replaceable coils, eight of which were successively burned out and replaced on one motor within a single week. To get a coil into place, and its oak keys driven home, required such bending, clamping and pounding as inevitably resulted in injury to insulation, and only by the greatest care could replaced coils be made to stand a test adequate to the 3,000 volts employed. For protection from lightning several types of manufactured arresters, then various original devices were tried, ending with a simple gap in series with a score or more of fuse blocks in parallel, arranged about a radial commutator switch, turned from point to point, as the fuses were blown by successive discharges. From the first these conditions caused the greatest apprehension as to the commercial success of electric power transmission, until Mr. Alexander J. Wurts, during a stay of several months with the company, gave the protection of the now well-known non-arcing arrester.

No transformers were used between machines and line, the largest transformers at first being two-kilowatt or 40-light. Aside from the effects of lightning, even today 3,000 volts upon the winding of small high-speed armatures requires first-class insulation. Frequent grounds were prevented by deep insulating foundations of paraffined wood. To prevent short-circuits within the coils, their cells, just before placing, were poured full of shellac, and the entire armature afterward baked for several days. By this means,the 50-horsepower motor ran a full year without trouble in a room dripping with moisture.

A lighting transformer received in 1891 was rated at five kilowatts. Theretofore transformers had been rated in lights, and generators in horsepower. This transformer was immersed in engine oil, and marks an epoch in the company's history. Lightning frequently punctured it, causing its fuses to blow, without other apparent injury. It remained in service for years. All others were soon likewise immersed. Four 500-light, dry Stanley transformers, purchased in 1892 for lighting Telluride, were broken down by the thunderstorms of the following spring. When repaired these also were immersed in engine oil, and gave no further trouble during the three years they remained in service.

Alternators were paralleled at Telluride in the spring of 1893, and thereafter they were so operated with full load upon the smaller, and regulation upon the larger machine.

Manipulation at switchboards or at brushes involved direct handling of 3,000 volts, a rather high switchboard pressure even now. It was a rule that every attendant keep one hand in his pocket while working with the other. It is pleasant to record that during these years no loss of life and but few accidents occurred.

There being no other circuit-breakers, it was necessary when a motor dropped out of step to break the circuit with the single arc-light plug. This always drew a heavy, vicious arc, which on the big motor frequently held to the full length of the six-foot cable, and then sometimes required a whiff from the attendant's hat. When not broken promptly, it frequently involved the entire switchboard and shut down the plant.

Duties of this nature required considerable skill and cool heads, and in order to operate the plant continuously, night and day, 15 or 20 competent attendants were required. To fit young men for these positions a course was arranged, during which they were taught something of machinery, of shopwork in metal and wood, and of wiring, insulating and repairing, while receiving such assistance in daily study as conditions permitted. A technical library, including the electrical papers, and a conveniently fitted testing room were always open. Each student was then given a short laboratory course in graphic treatment of alternating-current theory. This is said to have been the first systematic effort made by a corporation to train its employes for responsible positions.

Although the plant as a whole was an unqualified commercial success, no explanation need here be made as to why it was replaced by the induction system as soon as the latter had been perfected. This marks the limit of the most extensive single-phase, synchronous. plant ever operated. With but one or two motors, its operation was not difficult; but each motor added to the system brought increased demand for care and skill. The causes of difficulty were not understood then as now, nor was the effect of power factor fully appreciated. Lack of both wattmeters and power-factor indicators left the adjustment of field charges to the judgment of the operators. The power factor of each motor being dependent not only upon its own adjustment but upon that of all, the closest attention and co-operation were necessary, in marked contrast with the simplicity of operation of induction motors. Disturbances due to starting motors were especially trying, and the unqualified success attained, notwithstanding defects of apparatus and system, is attributed now, far more than then, to the skill and vigilance of the operators in this new and fascinating field.

The Tesla system substituted for the synchronous in 1896, comprised two 600-kilowatt, 60-cycle, 500-volt, two-phase generators, direct-connected to wheels under 600 and 900 feet head, respectively, and an equal capacity of raising and reducing transformers and of two-phase, 220-volt induction motors. The 12 100-kilowatt, step-up transformers were connected in pairs, two-phase three-phase, for three-phase, 10,000-volt transmission. These transformers were worthless; all broke down within a year, and one or more were always undergoing repairs. Breakdowns occasionally caused sufficient explosion to lift a cover, or splash the oil. The woodwork soon became saturated, and hot metal from the nearby main fuses frequently started fires, endangering the wooden power house. A masonry transformer house in two compartments was therefore constructed, and into it the transformers were moved this being the first known case of isolation of oil transformers on account of fire risk.

The power house at Ilium, situated six miles below Ames, on the same stream and using the same water, was built in 1900, and contains one 1,200-kilowatt, revolving-field, General Electric generator, direct connected to two impulse wheels under 500 feet head. Transmission lines extend both to the Ames station and to points of distribution, providing the insurance of duplicate transmission. Any section of line can be cut out for repair, or either power house shut down, without interrupting the service. Junctions other than generating and distributing points are equipped with open air switches, mounted upon standard line insulators and operated from platforms similarly insulated, and have proven invaluable.

Junction houses at distributing centers provide for a branch line to each customer, which is equipped with switches, fuses and a set of five record-making instruments a voltmeter, two ammeters and two wattmeters. The power company thus secures upon its own property a continuous, accurate and satisfactory record of each load.

The long spans crossing canyons and divides surrounding Savage Basin may be worthy of note. These divides are bare ridges at an altitude of 13,000 feet, inaccessible in winter, and swept by frequent snowslides. Spans up to 1,150 feet are used, in order to reach safe points for supports. A number of these supports, although simple and inexpensive, have stood for years without repair. The longest span is of No. 1, hard-drawn copper, supported by half-inch plow-steel cable, both being carried by the same insulators. The deflection is approximately 35 feet, on a slope of 31 degrees. Another is of three-eighths-inch soft-iron cable, 1,120 feet long, and has been in service five years. A third, 660 feet long, is of hard-drawn copper only, having 25 feet deflection. The strain insulators in all cases are a series of the usual line insulators and pins, upon a longitudinal arm, hinged to permit adjustment to span motion. They are simple, inexpensive and entirely successful.

A 10,000-volt underground transmission was put in operation at the Gold King mine in 1896. Power was carried through an unused tunnel, 1,300 feet long, upon bare copper conductors, 12 inches apart, on standard line insulators, to a deep mining hoist equipped for electric power. The tunnel was always dripping with water, but no trouble was experienced during the several years of operation, although slight brush discharge or halo was at times observed.

An interesting installation to which power is furnished is that of the well-known Camp Bird, Limited, near Ouray. Nineteen motors and rotaries, in sizes up to 150 kilowatts, drive crushers, Huntingtons, concentrators, compressors, pumps and hoists, aggregating in all about 1,000 kilowatts. Two underground transmissions, each a mile in extent, are in operation. Continuous current at 550 volts from two rotaries and a 650-ampere-hour storage battery operate three deep-mine hoists of 150 horsepower, and an installation designed by Mr. C. S. Runner, now engineer of the Utah department, makes use of the alternating current transmitted at 10,000 volts through paper-insulated, lead-covered cable, for the purpose of operating two 50-horsepower pumps.

The success of the original plant prompted the manager of the company, Mr. L. L. Nunn, to institute a search for other waterpowers in the West, finding as a result that such powers were very remote from available markets, requiring much longer transmissions than theretofore used. Voltages higher than from 10,000 to 15,000 were not in commercial use, and were regarded as merely problematical; but two important water rights, already acquired in Utah and Montana, would have been worthless at such pressures. Mr. Nunn therefore determined, in 1895, to undertake at Telluride an experimental transmission at higher voltages, to be installed and operated as a practical test for power purposes, and to determine if possible the problems peculiar to long distances and high pressures.

Two identical 75-kilowatt, oil-insillated transformers were installed in the autumn of 1895, one at the Ames station and the other at the Gold King mill. They were designed for pressures varying from 15,000 to 60,000 volts, by convenient steps. A separate pole line was equipped with three circuits of different characteristics, upon three types of insulators.

Measurements with many special instruments, were made, embracing the different voltages, styles of insulators, conductors and distances between them, and the conditions peculiar to the various phenomena met at every step. Observations upon a wide range of atmospheric conditions were made by means of United states Weather Bureau apparatus at either end of the line. The commercial feasibility of high pressures was demonstrated by the successful operation of the Gold King mill during a great part of the year at pressures from 30,000 to nearly 60,000 volts, as well as by continuous electrification for nearly a month during dry weather of a three-mile telephone circuit, upon telegraph insulators, at pressures rising from 10,000 to 40,000 volts.

The change of the system from single to polyphase terminated actual transmission experiments. The reducing transformer was moved to the station, and another equipment designed for polyphase tests was ordered. The remaining time was devoted to open-circuit losses, and to the verification of measurements previously made. This work continued until August, 1897, when construction was begun upon the Provo plant.

Much of the data obtained from these experiments was incomplete, requiring caution in its use, due largely to the time and study required in solving, step by step, the problems and difficulties met at every stage of the work. However, that much of value was obtained is shown by the subsequent successes at Provo. Sufficient had been learned to warrant the commercial adoption for the first time of 40,000 volts, nearly thrice the voltage of any previous plant; to lead to the manufacture of transformers which, after seven years' continuous operation, are still in daily service; to determine the design of the Provo-type insulator, the method of line construction, distance between wires, and the importance of wave-form, and to make possible this great advance in long-distance, high-voltage transmission.

This experimental work, as clearly appears from the foregoing facts, was begun, carried on and finally utilized by the Telluride company in the regular and necessary course of its growing business; yet it must be added that important services were rendered by Mr. V. G. Converse, under whose direction the transformers had been designed and constructed, and who participated throughout the greater part of the work during all the experiments with actual high-pressure transmission, and subsequently by Mr. Ralph D. Mershon, in the elaborate instrumentation and laboratory practice, including a notably ingenious method of reading high-tension losses upon low-tension circuits, devised by him and used in substantiating the accuracy of the earlier measurements; also that different types of insulators were contributed by the General Electric and the Westinghouse companies and by Mr. Fred D. [sic] M. Locke, on account of their friendly interest in the work.*

The original plant at Provo contained two 750-kilowatt, 60-cycle, 800-volt, three-phase, General Electric generators, direct-connected at 300 revolutions per minute to twin horizontal turbines under 125 feet head, a six-panel Wagner switchboard, two banks of oil transformers and two outgoing circuits. All contents thus in duplicate were assembled in two complete, independent units, designed for operation independently or paralleled at both high and low pressure. Prior to the power-factor indicator, a device which answered a somewhat similar purpose was installed, consisting of a wattmeter on the low-pressure paralleling bus, with current coil in one bus and shunt across the other two. This indicated cross current, and was used in the adjustment of field charges. Transformers were each 250 kilowatts, 800 to 40,000 volts, star-connected at both high and low pressure, with. neutrals, grounded.

Triple-pole air switches and four-foot fuses formerly connected each bank of transformers with its transmission line. One form of air switch, opening six feet, contained no metal except conductors, and was composed entirely of paraffined wood and rawhide, without porcelain, glass or other insulator. Others were sliding frames, carrying line insulators.

During the first year of operation the transmission comprised a single 32-mile line to one receiving point at Mercur, where the arrangement was similar to that at the power house, save that two reducing transformers were connected two-phase three-phase, grounded neutral, for 220-volt, two-phase induction motors. The Provo-Eureka line, 42 miles long, carries seven-strand aluminum cable equivalent to No. 4 copper. The Eureka-Mercur cross line, 28 miles long, equivalent to No. 5 copper, was added to complete the triangle thus formed and permit cutting out either of the three sides without interrupting service.

The Logan plant was completed in 1901, containing two 1,000-kilowatt, revolving-field alternators, direct connected at 400 revolutions to double-discharge twin turbines under 212 feet head. This plant is connected with the Provo system by duplicate lines over 100 miles long, passing the cities of Ogden and Salt Lake. The Provo and Logan plants are thus operated in unison through nearly 200 miles of transmission. Distributing points at Mercur, Eureka, Bingham, Salt Lake and Provo are also junction points of the duplicate lines, equipped with switches in each incoming line, as well as in circuit with the transformers, so that in case of threatened trouble the patrolman can without delay have his section cut off for immediate repair without interrupting service.

The three conductors of each transmission form an equilateral triangle 76 inches between wires, carried by a seven-foot cross-arm and the top of the pole. Extra long pins raise the insulators from six to 12 inches above cross-arms, are of selected locust, kiln-dried and immersed from six to 12 hours in hard paraffin at 150 C. Cross-arms are of Oregon fir, kiln-dried and soaked in boiling bitumen. Those upon the first line were attached in the usual manner with metal braces. The burning of cross-arms and poles on account of broken insulators, during prolonged wet weather, occurred most frequently at these braces. When the next lines were built in 1899 treated wooden braces were substituted, with results so favorable that all metal braces were soon replaced. It was still observed, however, that even light leakage seemed to concentrate around the lag bolts, carbonizing the wood and finally loosening the bolts. For the Logan lines of 1900 and all later lines, therefore, the cross-arms were mortised through the poles and wedged and pinned with hardwood thus discarding all metal except conductors. This construction was originated by Mr. A. L. Woodhouse, who, upon the close of the high-pressure experimental work in Colorado, of which he had charge, became and still is superintendent of the Utah department. It has proved amply strong, not expensive, and during the four years' operation of the 400 miles thus constructed very few poles have been burned.

Provo-type glass insulators, designed by Mr. V. G. Converse, have been used throughout. Many have broken, but these have usually shown the effects of guns or stones. In fact, there has not been a single breakage, except in one lot, improperly annealed, clearly due to either internal or dielectric stresses. It is difficult to see wherein any other insulators could have done better, unless bullet-proof. College laboratory tests to the contrary notwithstanding, leakage losses are inappreciable, except during severest storms, and then not serious where insulators are unbroken. It is a mistake to suppose that Utah climate is favorable. During the rainy season it is as wet as any, and the alkali dust of the so-called saltstorms is as trying as seacoast spray. At time dense volumes of this impalpable dust from the Great Desert are accompanied by clouds or fog. In this damp, sticky state the dust completely covers to a considerable depth the under as well as the upper surfaces of insulators, as well as poles, cross-arms and pins. Over these surfaces streamers gradually creep, until, meeting at the pole, they break into an arc, like that which was photographed by Mr. C. E. Baker, the line patrolman at Mercur, and which has several times been published. A quick turn of the generator rheostat at the critical instant breaks the arc, without interrupting service of induction motors.

The arrangement of power houses and transmissions already described is such that the opening of paralleling switches may resolve the system into a single transmission from 100 to nearly 400 miles in length, with a generator at each end, yet side by side. If one generator be reversed, synchronized as a motor with the other and loaded by its waterwheel, any length of transmission may, by manipulation of a paralleling switch, be alternately cut in and out between them. Since switchboards and instruments are connected, measurements made are immediately comparable. In this manner losses and power factor may be measured, and the corrective effect of charging current observed.

Solid aluminum wire, first used in 1898, was slightly alloyed to increase strength, but proved worthless, breaking repeatedly with square, glass-like fractures. It was at once replaced with commercially pure, seven-strand cable, still in use. Similar cables have generally been employed for subsequent lines, while spans have been successfully increased to 180 and 200 feet, with less deflection than usual with copper.

The experience with oil transformers for 10,000 volts at Telluride, and the refusal of manufacturers to give any guarantees whatever for other transformers for higher pressures, led the Telluride company, when undertaking this 40,000-volt transmission, to manufacture its own. The first equipment was made at the Wagner company's works under designs and supervision of Mr. Converse. The later ones were made by the Converse Transformer Company. When erected, the oil in the tank and the transformer in an oven were slowly raised to and then maintained during 24 hours at a temperature of 125 C. The transformer was then immersed in the oil, and both continued at the same temperature for a further 24 hours. As bearing upon the question of fire risk due to oil transformers, it may be of interest to note that of the large number of these high-pressure transformers used during the last seven years, chiefly in isolated sub-stations containing much wood and seldom visited, all but four are still in operation; that these four were destroyed by fire of doubtful origin, and that only one transformer has required repair other than change of oil.

The plant at Norris, Mont., designed and constructed in 1901 by Mr. O. B. Suhr, superintendent (now resident engineer of the Ontario Power Company), contains at present two low-speed, 1,000-kilowatt units. A duplicate transmission of 60 miles conveys power to the city of Butte. These lines, as well as both raising and reducing transformers, were designed for the use of 40,000, 60,000, or 80,000 volts. Longer pins are used than in Utah, and conductors form a triangle of 108 inches. While producing the present limited amount of power, and awaiting a suitable insulator, the lower voltage has been used.

In conclusion, it may be said that the Provo plant the first transmission at more than 16,000 volts while undertaken materially in advance of the art, and not exempt from its share of troubles, has, nevertheless, been fu