COMPRESSED AIR PRODUCTION OR THE THEORY AND PRACTICE OF AIR COMPRESSION, BY W. L. SAUNDERS. Published by COMPRESSED AIR, 26 Cortlandt Street, New York, 18436 Copyrighted 1898, By W. I,. Saunders WD COPIES RECEIVED. °\ . 2nd COPY, 1898. ko o^K^cAVAS COMPRESSED AIR PRODUCTION, By W. L. SAUNDERS. Compressed air is air under pressure. It is usual to define com- pressed air as air increased in density by pressure, but we may produce compressed air by heat alone, as illustrated by the discharge of a cork from an empty bottle when heated. Though one of the oldest of the sciences, compressed air is, in its development and use, one of the young- est. Hero, of Alexandria, a century before Christ experimented and wrote upon "Pneumatics," calling special attention to the influence of heat in expanding and contracting air. It is said that Hero put into prac- tical use an invention by which the opening and closing of temple doors was effected by the alternate rarefaction and condensation of air which was brought in contact with heated and cooled surfaces of altar tops. Yet the science of pneumatics played no important part in industrial progress until scarcely more than a century ago it came into general use for diving-bells, and was later on applied by Brunei to caisson work. In 1830 the French Academy of Sciences gave a medal to Thillorier for his method of compressing gases by stages. . In 1849 the Baron von Rathen suggested the use of compressed air at 750 pounds pressure per square inch in locomotives-; It is a singular fact that the Baron, in describing the method by which he proposed to attain this high pressure, advised compound compressors with inter-coolers. The special advan- tages of cooling the air between the different stages of compression were set forth. This stage compression and inter-cooling is one of the most important recent improvements made in air-compressing machinery. Until recent years the use of compressed air in America has been confined almost exclusively to mining, tunneling, bridge building, or to work in a confined space for which no other power was available. Elec- tricity has recently become a competitor of compressed air in that it, too, may be used in confined spaces, and may be transmitted long distances and distributed. Until such competition arose the question of producing compressed air economically was but little agitated. The attention of engineers was mainly devoted to the development of an apparatus for using compressed air, it being taken for granted that air was an expensive power at best. The manufacturer sought to perfect his compressor on lines of low first cost, light weight, economy of space and p-eneral avail- ability. Dry, pure air, delivered at a sufficient pressure by a machine which could be depended upon, has been the controlling consideration. ♦Reprinted from Compressed Air. 4 COMPRESSED AIR PRODUCTION. Compressed air and air-compressing machinery have been consid- ered, and the science developed by two classes of men — the practical men and the engineers. The practical men confined their work to the ma- chine. The confusing diagrams and figures of the engineers were not considered, because they were not understood. The engineers took occa- sional plunges into compressed air theories, producing figures controvert- ing certain well-established and so-called practical facts, and almost in- variably basing the conditions of compressed air economy upon questions of thermo-dynamics. The problems produced by the engineers were too mathematical for the practical compressed air men. These men knew too little of the theory of compressed air, hence progress in the science has been slow. During recent years an impetus nas been given to compressed air development by the strides made by electricity, and by the increased use of compressed air in the arts. Electricity, although apparently a com- petitor, has really played the part of a friend in pointing out the possi- bilities of transmission and use of air in directions before unknown; thus a market has been created. The perfection of the air compressor on lines of economy naturally followed the wide use of compressed air in competition with steam and electricity. The best steam engine practice has been applied to the com- pressor. Compound condensing Corliss engines are now used in con- nection with air cylinders of new design. The whole subject of compressed air may be divided into three heads: Production. Transmission. Use. No better evidence is needed of the obscurity of the science, even among engineers, than the fact that it is the usual thing to look upon compressed air as an expensive power, because of the great loss which is suffered during transmission. The great losses and the serious diffi- culties encountered in reality do not belong to transmission. Com- pressed air power may be transmitted and distributed with no greater difficulty than the distribution and transmission of illuminating gas. It is a question of the size of pipe, volume and the pressure. There is not a properly designed compressed air installation in operation to-day that loses over five per cent, by the transmission alone. The question is alto- gether one of the size of pipe, and if the pipe is large enough the friction loss is a small item. It is undoubtedly true that there are places where a conduit has been laid for a certain volume of air, and where the supply has been increased without increasing the size of the conduit, the result of this being that more air is forced through the pipe than its sectional diameter will admit economically; hence the velocity of flow is increased, and as the friction is in direct proportion to the velocity the loss of power is also increased. The largest compressed air long-distance power plant in America is that at the Chapin Mines in Michigan, where the power is generated at Quinnesec Falls, and transmitted three miles. This is not an economical plant, but the loss of pressure, as shown by the gauge, COMPRESSED AIR PRODUCTION. 5 is only two pounds, and this is the loss which may be laid strictly to trans- mission. During the construction of the Jeddo Tunnel, near Hazelton, Pa., compressed air at 60 pounds pressure was conveyed 10,860 feet from the central station. The writer was called upon to explain a mysterious condition which existed on both ends of the line. The pressure gauge recorded the same figures and the gauges were sent to the shops for re- pairs, because everybody was convinced that " something was wrong." The result was not changed when the gauges had been " repaired/' it being evident that this apparently perfect economy of transmission was due to the fact that a large pipe (nearly six inches in diameter) was used at that time to convey so small a volume of air that the velocity in the pipe produced so small a friction loss that it could not be recorded on the gauge. Having defined compressed air, we must next define heat, for in dealing with compressed air we are brought face to face with the com- plex laws of thermo-dynamics. When we produce compressed air we produce heat, and when we use compressed air as a power we produce cold. Based on the material theory of heat, it was said that when we take a certain volume of free air and compress it into a smaller space we gtt an increase of temperature, because we have the heat of the original vol- ume occupying less space; but no one at this date accepts the material theory of heat. The science of thermo-dynamics teaches that heat and mechanical energy are only different phases of the same thing, the one being the motion of molecules and the other that of masses. This is the accepted theory of heat. In other words, we do not believe that there is any such thing as heat, but that what we call heat is only the sensible effect of motion. In the cylinder of an air compressor the energy of the piston is converted in molecular motion in the air, and the result or the equivalent is heat. A higher temperature means an increased speed of vibration, and the lower temperature means that this speed of vibration is reduced. If we hold an open cylinder in one hand and a piston in the other, and place the piston within the cylinder, we here have a confined volume of air at normal temperature and pressure. These particles of air are in motion and produce heat and pressure in proportion to that motion. Now, if we press the piston to a point in the center of the cylin- der, that is, to one-half the stroke, we here decrease the distance between the cylinder head and the piston just one-half; hence each molecule of air strikes twice- as many blows upon the piston and head in traveling the same distance, and the pressure is doubled. We have also produced heat (about 116 degrees), because we have expended a certain amount of work upon the air; the air has done no work in return, but we have increased the energy of molecular vibration in the air, and the result is heat. But what of this heat? W r hat harm does it do? If we instantly re- lease the piston which we have forced to one-half stroke, it will return to its original position less only a fractional part, due to friction. We have, therefore, recovered all or nearly all the power spent in compressing the air. We have simply pressed and released a spring, and this illus- tration shows what a perfect spring compressed air is. We see also the COMPRESSED AIR PRODUCTION. possibility of expending one horse-power of energy upon air and getting almost one horse-power in return. Such would be the case in practical work if we could use the compressed air power immediately and at the point where the compression took place. This is scarcely possible, as the heat in the air is soon lost by radiation, and we have lost power. Thirteen cubic feet of free air at normal temperature and barometric pressure weigh about I pound. In the illustration referred to about 116 degrees of heat are liberated at half stroke of a compressor. The gauge pressure at this point reaches 24 pounds. According to Mariotte's law, '*the temperature remaining constant, the volume varies inversely as the pressure," we should have 1 $ pounds gauge pressure at half stroke. The difference, 9 pounds, represents the effect of the heat of compression in increasing the relative volume of the air. The specific heat of air under constant pressure being 0.238, we have 0.238x116=27.6 heat units produced by compressing one pound, or 13 cubic feet, of free air into one-ialf its volume; 27.6x7/ '2 (Joule's equiva- lent) 3 ^ 1,3 07 foot pounds. We know that 33,000 foot pounds is one horse-power, and we see how easily about two-thirds of a horse-power in heat units may be produced and lost in compressing one pound of air. Exactly this same loss is suffered when compressed air does work in an engine without reheating, and is expanded down to its original pressure. In other words, the heat of compression and the cold of expansion are in degree equal. Figure 1 is a sketch designed to indicate graphically the effect of heat and cold in compression and expansion of air. Rt«t '^jijUx.YJJ ft 1 ■& aYh. « t.w\ *\ ExpATV*T,oivii<$ Conltviwiouft H«eftvng. ? The sketch illustrates an open cylinder, which may serve both as an air compressor and an air engine. The piston at the point shown is sup- posed to confine a volume of free air in the cylinder and at a temperature of 60 degrees; let it be pressed down until it reaches the point indicated COMPRESSED AIR" PRODUCTION. 7 by 45 pounds, and the pressure will follow the dotted lines marked "Adia- batic." This is, of course, assuming that the heat, which is invariably produced by compression, is suffered to remain in the air and to in- fluence the pressure. We here have a confined volume of compressed air at a pressure of 45 pounds and a temperature of 320 degrees. Let there be no absorption of heat and the piston if released will return to the starting point, the pressure following exactly the line indicated during compression and the temperature returning to 60 degrees. In such a case we assume, of course, that the piston is frictionless. This points to the fact that compressed air is a perfect spring, and that the heat of com- pression when utilized can be made to return its full value of energy. An Air Compressor provided with no cooling device would show a pressure line following closely that marked "Adiabatic" on Fig. 1. The hot compressed air confined in the cylinder at the 45-pound point, if transferred through pipes to an air engine and maintained hot until used, would be available for work in the same proportion (less a little friction) as we have shown in the theoretical case where the piston, used as a compressor, is driven back to the starting point. Practically, it is impossible to convey hot compressed air any distance from the com- pressor, for air, though very slow in taking up heat, has so low a specific heat that it parts with its temperature rapidly. Steam having a higher specific heat may be conveyed as a power even through naked pipes, and this fact has led to mistakes in regard to the possibilities with compressed air. Returning to Fig. 1, let us imagine that the piston has been stopped at the 45-pound point, and that the compressed air, which, as we have seen, has a temperature of 320 degrees, is transferred into a receiver and used at a point say half a mile distant. The temperature will now be re- duced to that of the surrounding medium, or to the initial temperature, which the sketch shows to be 60 degrees; and if the system is well de- signed, that is, if the pipes are large enough and there are no leaks or other irregularities, we will have nearly 45 pounds oressure on the other end of the line, as there is a direct elastic medium between the two points. But the volume will be reduced in size, because of the reduced tempera- ture, and will correspond with the space underneath the lowest dotted line in the figure marked "Volume I." If it is now used without reheating to do work in an engine, the line of reduction in pressure will follow the lower dotted line marked "Adiabatic" until it reaches the point marked 201 degrees, which represents the theoretical temperature of the air when exhausted at atmospheric pressure. We now see that the piston, instead of returning to the starting point, has only had power enough behind it to return it to a point about half way. The illustration points to the importance in compressed air economy of reducing to the lowest point practicable the temperature of the air before and during compression and conversely increasing to the highest point the temperature before and during use or expansion. If, in the case referred to, heat had been applied during expansion, 8 COMPRESSED AIR PRODUCTION. the pressure would follow the line marked "Isothermal Expansion, the piston might be returned to the starting point. Tabi,e i. — Of the volume and weight of dry air at different temperatures under a constant atmospheric pressure of 29.92 inches of mer- cury in the barometer (one atmosphere), the volume at 32 ° fahrenheit being i. and J3 £ £ V 5 a ■1-1 D <4M 3 5 .- fl-J "I V u •2 «5 .2tJ fl) C U 4J fl 1- p,- be 4) 4; U 11 IS v 1 * 1! 3 s 3 n H £ 3 J" 1 1= > 1-s 00540 1,300 3595 32 1-000 0-0807 275 1-495 00225 42 1-020 0-0791 300 1-546 0522 1,400 3-789 0-0213 52 1-041 00776 325 1-597 0506 ,1,500 3993 0-0202 62 1-061 0761 350 1-648 0490 : 1,600 4197 0192 72 1-082 00747 375 1-689 0-0477 1,700 4-401 0-0183 83 ' 1-102 00733 400 1-750 0461 1,800 4-605 0-0175 92 1122 0-0720 450 1-852 00436 1,900 4-809 0168 102 1*143 0-0707 500 1954 0413 2,000 5 012 00161 112 1163 00694 550 2 056 00381 2,100 5-216 00155 122 1-184 0682 600 2-158 00376 2,200 5-420 00149 132 1-201 00671 650 2-260 0-0357 2,300 5624 0-0142 142 1-224 0-0660 700 2 362 0-0338 2,400 5-828 0-0138 152 1-245 0-0649 750 2-464 0-0328 2,500 6-032 0-0133 162 1-255 0638 800 2566 0315 2:600 6-236 0-0130 ]?2 1-285 0-0628 850 2-668 00303 2,700 6440 0-0125 16*5 1-306 0-0618 900 2-770 00292 2,800 6 644 00121 192 1326 00609 950 2-872 0-0281 2,900 6847 0-0118 202 1-347 0-0600 1,000 2974 0268 3 V 000 7051 00114 212 1-367 00591 1,100 3-177 00254 3,100 7255 o-oiii 230 1-404 00575 1,200 3381 0-0239 3,200 7-459 0-0108 250 1-444 00559 Another case is shown by the sketch in which the air is compressed adiabatically to 45 pounds pressure and heat enough is applied during ex- pansion to maintain the temperature at 320 degrees until the air is ex- hausted at atmospheric pressure. This case is purely theoretical and illustrates the possibility of obtaining more power out of a given volume of air after compression than was expended at the compressor. Experiments made by M. Regnault and others on the influence of heat on pressures and volumes of gases have enabled us to fix the absolute zero of temperature as about — 461 degrees Fahrenheit. This point, — 461 degrees below zero, has been taken to be the theoretical point at which a volume of air is reduced to nothing. The exact figures representing absolute zero vary with different authors, but for all practical purposes — 461 degrees F. is near enough. The volume of air at different tempera- tures is in proportion to the absolute temperature, and on this basis Box has produced table 1. The effect of the heat of compression in increasing the volume, and the heat produced at different stages of compression, are shown by table 2 (Box) : A cubic foot of free air at a pressure of one atmosphere (equal to 14.7 pounds above a vacuum) at a temperature of 60 degrees when compressed COMPRESSED AIR PRODUCTION. to twenty-five atmospheres, will register 367.5 pounds above a vacuum (352.8 pounds gauge pressure), will occupy a volume of 0.1020 cubic foot, will have a temperature of 864 degrees, and the total increase of tempera- ture is 804 degrees. These tables apply to dry air only. The effect of moisture will vary the figures of temperature and to some extent will affect the pressures, but many useful deductions may be drawn from the tables. It is seen for instance by studying table 1 that a volume of dry air will be doubled if Tabte 2. — Heat PRODUCED BY COMPRESSION OF AIR. Pressure. a at 'o ■/. ■2 5 v V rt 3 10 V Pounds per S-J 11 *J D i- Cu H Pounds per Sq. Inch "" V = £ fctf g Sq. Inch above the Vfu v ^ bCtfj CH O-H s above a Atm'sph're. § t~ll < Vacuum. (Gauge "0 v jz P ** Pressure). > ft TOO 14 70 000 1-0000 60-0 oo-o 110 16-17 T47 0-9346 74-6 14-6 125 18 37 367 0-8536 94-8 34-8 1-50 22-05 7-35 0-7501 124-9 64-9 1-75 25-81 11-11 6724 15T6 91-6 200 29-40 14-70 06U7 175-8 115*8 2-50 36-70 2200 05221 218-3 158 3 3-00 44-10 29-40 0-4588 255-1 1951 350 5P40 36-70 04113 287-8 227-8 4-00 58-80 44-10 0-3741 317-4 257-4 5'00 73 50 58-80 0-319 1 360 4 309-4 6-00 88-20 73-50 0-2806 414-5 354-5 7-00 102-90 88-20 0-2516 4545 394-5 8-00 117 60 10290 0-2288 4906 430-6 900 13230 117-60 0-2105 523-7 463-4 1000 147-00 133-80 0-1953 554-0 494-0 15-00 220*50 205-80 0-1465 6810 621-0 2000 294 00 279-80 0T195 781-0 72P0 25-00 367 50 352-80 0-1020 864-0 804-0 its temperature is increased about 500 degrees, and conversely, of course, if the volume remains constant an increase of about 500 degrees in tem- perature will double the pressure. The addition of moisture serves to in- crease these figures, because moisture increases both the specific heat and the heat conducting capacity of the air. The thermal results of air compression and expansion are shown by the accompanying diagram (Fig. 2 — Frank Richards). Both the tem- perature of the air and its volume are shown at different stages of com- pression. The simplest application of this diagram is that which gives the gauge pressure represented at different points of the stroke. This is shown in the vertical lines. But in compressing air we produce heat, and it is important to know the temperature at any given pressure, also the relative volume. All of these are shown in the diagram. The initial volume of air equal to 1 is taken and divided into ten equal parts. Each division between two horizontal lines, shown by the figures at the right, representing one-tenth of the original volume. The vertical and horizontal lines are the measures of volumes, pres- sures and temperatures. The figures at the top indicate pressures in at- IO COMPRESSED AIR PRODUCTION. mospheres above a vacuum ; the corresponding figures at the bottom de- note pressures by the gauge. At the right are volumes from one-tenth to one. At tne left are degrees of temperatures from zero to 1,000 degrees Fahrenheit. The two curves which begin at the upper left hand corner and extend to the lower right are the lines of compression. The upper one being the "Adiabatic" curve, or that which represents the pressure at any point on the stroke with the heat developed by com- pression remaining in the air; the lower is the "Isothermal/' or the pres- N. *( ^ ^ ^ >» «s -5; !46t°»b6olute Zeyo. iVre&au-re. ••*• '* »B. To loo ^^ Grange Press utc A. To \5^ )S Gravg,e ?re&au-re.. ^^?_6o°«riij xl8i^ = JUA&9 Pt.Pa>| 4.B4-°-60 o =4-14-°Xl83.s« 77,804., £. ft.fi.1. cr. froTtvW l,14, (ia»)« ?Tets,«Te. JO. ftom. loo'H<} AU rfe Pressiwe. -57*+6o°-ll7 , xifcV= 2I,46« FfcfAs, -a64 +6o°~«4.VlBSri 77,00* »lWi. Trotn 6o° «\: J\Trro»sbK« < rVc »»M»uTt. to JllifcOluTle ItTo. 4-6 1°- ?7 o »4.04-°X 1 aa* m ^.l^.' FfcWk.'F. _4-»l-S64. , -97")ll»iTH J7.799 ft Tot*\.Ci»niE- 95.60* " Fk.WU- Total D...A f. 05,60a Ft «.Pi«. Fig. 3. ures show that this energy is equal to 74,134 foot pounds. This added to the available energy gives us 95,603 foot pounds as the whole energy contained in one pound of air at fifteen pounds pressure and 60 degrees temperature. D represents one pound of air at 100 pounds pressure and 60 degrees temperature. Its available energy is 77,804 foot pounds, and its intrinsic energy is i/,799 foot pounds, or the total energy is 95,603 foot pounds, which is exactly equal to the case just cited. COMPRESSED AIR PRODUCTION. 13 Theae figures show the correctness of that thermodynamic law, which states that the power of any elastic gas is in direct proportion to its height of fall. So long as the temperature is above the absolute zero, there is as much powei in the same body of air when expanded adiabatically from a moderate temperature to an extremely low one, as when expanded from a high temperature to a moderate one, and this offers to some extent a limitation to that system of reheating which increases the volume without at the same time increasing the pressure. The development of heat when air is compressed is, perhaps, the best illustration of the acknowledged thermo-dynamic principle that work and heat are interchangeable unit for unit. When air is compressed all the work done in compression is converted into heat. This heat is capa- ble of being converted back again into power. But the question is fre- quently asked, if it be true that the power applied in the steam cylinder of an air compressor is all converted into heat in the air cylinder, how is it that power still remains in the compressed air after the heat has been lost through transmission? In order to get a clear understanding of this we must know that air is a power in itself before compression: that it contains a certain capacity for work due to its elasticity. It is not, however, in a condition available for work until compressed. This energy is made available by giving it a height of fall which is represented by a difference in pressure between the compressed air on one side and the free air on the other. If we box up free air at any given temperature and under normal atmospheric conditions, we have within the enclosure a well defined amount of energy. We cannot use it to perforin work unless the pressure outside is less than that inside the enclosure. This may be accomplished by placing the closed vessel in the rarified atmosphere, such as exists at altitudes, but in any case there is a well denned quantity of intrinsic energy within the air itself, the limit being measured by the height of fall between the free air in the vessel and absolute zero. Compression as now practiced only serves the purpose of placing the natural power which we have in the air into a condition which makes it possible for us to utilize it. This points to an undeveloped science in the use of compressed air. Inasmuch as we lose all the power expended in compression and yet have a capacity for useful work equal to from 30 to 50 per cent, of that power, it is plain that there are possibilities in the science which are now misunderstood and not realized. It is theoretically possible to realize out of compressed air more power than was expended at the compressor. This has been shown in Fig. 1, but that this statement might not be confused with perpetual motion theories the sketch shown in Fig. 4 has been prepared. Sulphuric acid in its concentrated form will, when exposed in an open dish, absorb moisture from the atmosphere to the extent of about double its weight. A hypothetical assumption is made of a pump ar- ranged to discharge sulphuric acid in the direction shown by the arrow to an open dish, elevated (say) 100 feet. 14 COMPRESSED AIR PRODUCTION. The amount of power necessary to discharge a certain quantity of sulphuric acid into this dish is exactly equal to the power which the sul- phuric acid is capable of giving out when falling back again, less the fric- tion of the pump, leakage, etc. Now, let us assume that the sulphuric acid in the open dish remained there long enough to absorb moisture from the atmosphere until its weight has been doubled; it will thus ob- viously have twice the amount of power in falling back again, and if the friction and leakage losses were not too great it will be capable of driving the pump and of returning an equivalent volume of the concentrated acid Fig. 4. to the dish. If the same acid is used over again, the moisture must be driven out, and lamps are shown in the sketch provided for this purpose. The analogy between this hypothetical case of sulphuric acid and one of compressed air is that, as with acid we may draw power from moisture which is contained in the air, so with compressed air may we draw upon the intrinsic heat energy of the atmosphere. If it were practicable to compress air isothermally — that is, without heat — we might illustrate the point referred to in the foregoing by placing an air compressor in a cold room, or what would amount to the same thing, taking in cold air to the cylinder of the compressor, this air being taken, for instance, from a cold storage room, and on being compressed the temperature being maintained at the initial point. This cold com- pressed air might then be led in pipes placed on the surface of the ground and exposed, say, to the hot sun, so that when used its temperature might be largely increased above that at the compressor. COMPRESSED AIR PRODUCTINN. 15 We would here have a case of reheating by natural conditions, and the possibility of obtaining more power at the motor than was expended at the compressor is made apparent. All this appears to be but theory, yet the value of the argument lies in the fact that it points to what may be accomplished in the future and to the importance at present of low initial temperature at the compressor and high temperature at the m-tor. It is a common thing to see air compressors at work in engine rooms drawing the air into the cvlinder at the temperature of the engine room, which, in many instances, especially in winter, is 50 degrees higher in temperature than that of the air on the outside. Even where air compressors are used with concentrated inlets made for the express purpose of being connected with the atmosphere from the outside, this question of economy by low initial temperature is frequently neglected. For every 5 degrees by which the initial tempera- ture of the intake air is lowered, there is a gain of one per cent, in volume. It is not a difficult matter to construct an air compressor plant where the intake air is made to pass over refrigeration pipes. In cities where ice-making plants are in operation, the best point to place the air com- pressor is alongside the ice-making machines, the combination of the two industries being advisable. On the other end of the line reheating is of great importance. A per- fect reheater has not yet been found, though at this time reheaters are in use which give practical results. It has already been demonstrated that compressed air may be increased in temperature and thus proportionately increased in volume and efficiency by the application of heat in a very economical manner, the quantity of coal consumed in proportion to the gain in economy being very small. Mr. Robert Hardie, in his experiments with a pneumatic motor using a hot water reheater, has figured the cost of reheating at one-eighth the coal required at the compressor. Professor Haupt, referring to this heat- er, makes the following statement: "The power required to compress 1,000 cubic feet of free air to 2,000 pounds per minute would be 400 horse-power, consuming 1,200 pounds of coal per hour at a cost of $1.80 (at $3 per ton), and the cost of reheat- ing would not exceed 22 cents to double the work performed. That these statements are not simply theoretical deductions have been proved by- actual tests. The Rome air motor when using a reheater ran fourteen miles on a consumption of 308 cubic feet of free air per mile. When the air was not reheated, the consumption of air per mile was 661 cubic feet." The first loss in air compression, that due to the fact that the heat produced cannot be maintained, is an unavoidable loss. The second loss may be called the influence of the heat of compression, and is due to the fact that this heat increases ihe relative volume of the air and resists com- l6 COMPRESSED AIJj^ PRODUCTION. pression. This heat of compression has. long been the bete noire of air compressor builders. At first it seriously affected the valves and pack- ing, and this served as an argument in favor of injecting water into the cylinder, the claim of manufacturers being that by keeping down the heat of compression repairs would be less and accidents due to the destruction of parts by heat would be avoided. The injection of water into the air cylinder is usually known as the Colladon idea. Compressors built o^i this system have shown the high- est isothermal results. It is plain that the injection of cold water in the shape of a finely divided spray directly into the air during compression will lower the temperature to a greater degree than to simply surround the cylinder and parts by water jackets. Two systems are in use by which it is attempted to absorb the heat during compression. These systems divide air compressors into two classes — (i) Wet compressors. (2) Dry compressors. A wet compressor is one which introduces water directly into the cylinder during compression. A dry compressor is one which admits no water to the air during compression. Wet compressors may be subdivided into two classes — (1) Those which inject water in the form of a spray into the cylin- der during compression. (2) Those which use a water piston for forcing the air. into confine- ment. The advantages of water injection during compression are as fol- lows: (1) Low temperature of air during compression, hence a reduced mean resistance and a saving of power. (2) Increased volume of air per stroke, due to filling of clearance spaces with water, and to a cold air cylinder. (3) Low temperature of air-immediately after compression, thus con- densing moisture at the air receiver. (4) Low temperature of cylinder and valves, thus maintaining pack- ing, etc. (5) Economical results due to compression of moist air. (See Table No. 4.) The first advantage is by far the most important one, and is really the only excuse for water injection in air compression. The percentage of work of compression (dry air) which is converted into heat and lost when no cooling system is used is as follows : Compressing to 2 atmospheres, loss 9.2 per cent. Compressing to 3 atmospheres, loss 15.0 per cent. COMPRESSED AIR PRODUCTION. 17 Compressing to 4 atmospheres, loss 19.6 per cent. Compressing to 5 atmospheres, loss 21.3 per cent. Compressing to 6 atmospheres, loss 24.0 per cent. Compressing to 7 atmospheres, loss 26.0 per cent. Compressing to 8 atmospheres, loss 27.4 per cent. We see that in compressing air to five atmospheres, which is the usual practice, the heat loss is 21.3 per cent., so that if we keep down the temperature of the air during compression to the isothermal line, we save this loss. The best practice in America has brought this heat loss down ot 3.6 per cent, (old Ingersoll Injection Air Compressors), while in Europe the heat loss has been reduced to 1.6 per cent. Steam-driven air compressors are usually run at a piston speed of about 350 feet per minute, or from 60 to 80 revolutions per minute of compressors of aver- age sizes, say 18 inches diameter of cylinder. Sixty revolutions per min- ute is equal to 120 strokes, or two strokes per second. An air cylinder 18 inches in diameter filled with free air once every half second, and at each stroke compressing the air to 60 pounds, and thereby producing 309 de- grees of heat, is thus by means of water injection cooled to an extent hardly possible with mere surface contact. The specific heat of water being about four times that of air, it readily takes up the heat of com- pression. A properly designed spray system must not be confused with the numerous devices applied to air cylinders by means of which water is introduced. In some cases the water is merely drawn in through the inlet valves. In others it passes through the centre of the piston and rod, coming in contact with the interior walls of the air cylinder between the packing rings. Introducing water into the air cylinder in any other way, except in the form of a spray, has but little effect in cooling the air during compression. On the contrary, it is a most fallacious system, be- cause it introduces all the disadvantages of water injection without its isothermal influence. Water, by mere surface contact with air, takes up but little heat, while the air having a chance to increase its temperature, absorbs water through the affinity of air for moisture, and thus carries over a volume of saturated hot air into the receiver and pipes, which on cooling (as it always does in transit to the mine), deposits its moisture and gives trouble through water and freezing. It is therefore of much importance to bear in mind that unless water can be introduced during compression to such an extent as to keep down the temperature of the air in the cylinder, it had better not be introduced at all. If too little water is introduced into an air cylinder during compres- sion, the result is warm, moist air, and if too much water is used it results in a surplus of power required to move a body of water which renders no useful service. The following table deduced from Zahner's formula gives the quan- i8 COMPRESSED AIR PRODUCTION. tity of water which should be injected per cubic foot of air compressed in order to keep the temperature, down to 104 degrees Fahrenheit. Weight of Water Weight of Water Compression by atmosphere above to be injected at 68° Fah. to keep the temperature to be injected at 68° Fah. to keep the temperature at io4°Fah.in lbs. at 104 Fah. in a vacuum. of water and per lbs of water for lbs. of free air. 1 cu. ft. of free air 2 0.734 0-056 3 1664 0-089 4 1-469 0113 5 1-701 0131 6 1-891 0145 7 3(63 0158 8 3 304 0-167 9 3 329 0*179 10 3440 0188 11 3 513 0195 13 3 634 0303 13 3719 0-309 14 3798 0315 15 3-871 0233 Experiments were made under the personal supervision of Prof. Den- ton at the Stevens Institute of Technology, to determine the relative effects in air compressors of water injection and water jackets. An ex- tended controversy had been carried on in the "Engineering- and Mining Journal" between the writer and others upon the question whether or not the injection of water reduced the temperature and increased the effi- ciency. It was claimed by some that the efficient indicator cards taken from certain injection compressors were not reliable because of leakage. It was only on such grounds that the proximity of the pressure line to the isothermal could be explained away. In the .Stevens Institute tests the greatest care was exercised to secure air-tight pistons and valves, and as these experiments were unbiased and in the hands of experts, the results may be accepted as conclusive so far as indicating isothermal economy by an efficient system of injection. Fig. 6 herewith represents graphically the results obtained at the Stevens Institute.* No clearance is shown, because the purpose of the illustration is to show the comparative effect of the various methods of cooling. The air was compressed to 150 pounds gauge pressure, the work done in each case being represented by the area between the pressure curve and the rectangular lines. The pressure curve is determined in each case by the formula — (v) B In the case of isothermal compression (assuming no heat produced), the exponent N=i. In adiabatic compressidn (the full effect of the heat of compression being available), N= 1.408. It is obvious that the value From a paper by Mr. 11. A. Parke. COMPRESSED AIR PRODUCTION. 19 of the exponent N will vary between the two points 1, and 1.408. From a large number of indicator cards taken at the Stevens Institute the fol- lowing values were shown: Water jacket N=i.35 Water jet injection N=i.33 Water spray injection N— 1.25 It has been demonstrated by experiments made in France that the power required to compress dry air has been prepared from the data of M. Fig. 6. Mallard, and shows that for five atmospheres the work expended in com- pressing one pound of dry air is 58,500 foot pounds, while that for moist air is 52,500 foot pounds. In expansion also moisture in the air adds to the economy, but in both cases the saving of power is not great enough to compensate for the many disadvantages due to the presence of water. Mr. Norman Selfe, of the Engineering Association of N. S. W., has compiled a table which shows some important theoretical conditions in- volved in producing compressed air. 20 COMPRESSED AIR PRODUCTION. P ►4 O > H P fc ° 3 5 £ 5 s a ;> en -2 .ji w w « o c fc o a3 ^S aiy o; J3;bav jo 33b;u30J3 j eo -in -* i!o 6 50 •uots -ssjdxno3 nt pasn si j3}bm. jt amiBj^dmaj, l" BU J J SfSSo riSti o'S i! 2 rt « ^ « — Si o*o «W •° ■£ k/ TO U Si 3 2 * ° S " S « 2«^£$ u V Q M .2 ll-sll a fe ^o a o V uSi. • wH.2 p ■* 5D CO U3 C^l 03 Q>*o6>'oib©?0C0*'l3S SSaoeosooicxi^tOaocs HHHN«N«5l l*-i «r- ec 1- >— t -*• t- 3j HHNNNN |»0>-*OHN900 i—( lO i- o ^ COMPRESSED AIR PRODUCTION. 2 1 There are many serious objections, however, to the use of water with- in the air cylinder. These objections are so serious that it has been found to be the best practice to suffer the heat loss during compression, and thus simplify the apparatus. Some of the objections may be stated as follows: 1. The mechanical difficulties involved in introducing the water into the cylinder so intimately mixed with the air during compression as to reduce the temperature of compression immediately when produced. 2. Impurities in the water, which, through both mechanical and chemical action, destroy exposed metallic surfaces. 3. Wear of cylinder, piston and other parts, due directly to the fact that water is a bad lubricant; and as the density of water is greater than that of oil, the latter floats on the water and has no chance to lubricate the moving parts. 4. Wet air arising from insufficient quantity of water and from in- efficient means of ejection. 5. Mechanical complications connected with the water pump, and the difficulties in the way of proportioning the volume of water and its temperature to the volume, temperature and pressure of the air. 6. Loss of power required to overcome the inertia of the water. 7. Limitations to the speed of the compressor, because of the lia- bility to break the cylinder head joint by water confined in the clearance spaces. 8. Absorption of air by water. Before the introduction of condensing air receivers, wet air resulting in freezing was considered the most serious obstacle to water injection; but this difficulty no longer exists, as experience has demonstrated that a large part of the moisture in compressed air may be abstracted in the air receiver. Even in the so-called dry compressors a great deal of moist- ure is carried over with the compressed air, because the atmosphere is never free from moisture. This subject will be referred to more fully when treating of the transmission of compressed air. A serious obstacle to water injection, and that which condemns the wet compressor, is the influence of the injected water upon the air cylin- der and parts. Even when pure water is used, the cylinders wear to such an extent as to produce leakage and to require reboring. The limitation to the speed of a compressor is also an important objection. The claim made by some that the injected water does not fill the clearance spaces, but is aerated, does not hold good, except with an inefficient injection system. Whether it be water or spray which occupies the clearance space, it is impossible for air and spray to occupy the same place at the same time. The writer has increased the speed of an air compressor (cylinders 12 in. and 12 in. by 18 in., injection ten revolutions per minute) by placing his fingers over the orifice of the suction pipe of the water pump. The boiler pressure remained the same, the cut-off was not changed and the air pressure was uniform, hence this increase of speed arose from the fact that the water was restricted and the clearance spaces were filled with compressed air, which served as a cushion or spring. While the volume 2 2 (COMPRESSED AIR PRODUCTION. of compressed air furnished by this compressor would be somewhat re- duced by the restriction of the water, yet the increase in speed which was obtained without any increase of power, fully compensated lor the clear- ance loss. Unless the water of injection can be used efficiently as a cooling agent, its value for clearance does not compensate for the disadvantages attending its use. In some of the early types of air compressors water was introduced through the inlet valves during the suction stroke; but this is an objectionable plan, because it has little effect on the heat produced until the discharge of the air, and furthermore, there is the danger of in- troducing too much water, and thus reducing the volume of air and en- dangering the cylinder heads. The presence of moisture in the air reduces the heat loss, hence, as shown by Table No. 4, less power is required to compress moist than dry air. It is not necessary to inject water during compression in order to gain this advantage, as the atmosphere is usually moist. The presence of moisture in the air has an important bearing upon the compression, trans- mission and use of air. Before compressed air became generally used, its value was thought to be prohibitive, mainly because it was said that the air would freeze. This freezing was, of course, nothing more than the formation of ice due to the presence of moisture in the air, this moisture having been first deposited in the shape of water by expansion and cool- ing, and afterwards, the temperature going down below the freezing point, it became ice. This has some time since ceased to be a serious matter, and on the whole the presence of the moisture has been found to be more beneficial than otherwise, because by increasing the specific heat of the body with which it is in contact, it reduces the temperature during com- pression and tends to increase it during expansion. In transmission it is simply necessary to keep the temperature from falling below the dew point, or to put in suitable receivers for draining the pipes. Where re- heaters are used, the presence of moisture is decidedly advantageous. The amount of moisture in the atmosphere varies with the climate. Air is never perfectly dry ; never, except in rare instances, does it contain less than 25 per cent, of the moisture necessary to saturate it. It is not an uncommon thing to read in the meterological reports in the newspapers during the summer that the moisture during an oppressively hot day reached 98 per cent., and even 99 per cent. In winter it is usually 80 or 85 per cent. At 65 per cent, we consider it moderately dry; 50 per cent, being commonly called dry air. Otto Van Guericke invented the air-pump in 1650. In 1753 Holl used an air engine for raising water. At Ramsgate Harbor, Kent, in the year 1788, Smeaton invented a "pump" for use in a diving apparatus. In 1 85 1, William Cubitt, at Rochester Bridge, and a little later an engi- neer, Brunei, at Saltash, used compressed air for bridge work. In 1852, Colladon patented the application of compressed air for driving machine drills in tunnels. The first notable use of compressed air is due to Prof. Colladon, of Geneva, whose plans were adopted at the Mont Cenis tunnel. COMPRESSED AIR PRODUCTION. 23 M. Sommeiller developed the Colladon idea and constructed the com- pressed air plant illustrated in Fig. 7. The Sommeiller compressor was operated as a ram, utilizing a natural head of water to force air at 80 pounds pressure into a receiver. The column of water contained in tne long pipe on the side of the hill was started and stopped automatically, by valves controlled by engines. The weight and momentum of the water forced a volume of air with such shock against a discharge valve that it was opened and the air was discharged into the taipk; the valve was then closed, the water checked; a portion of it was allowed to discharge and the space was filled with air, which was in turn forced into the tank. The efficiency of this compressor was about 50 per cent. ffl^^^J fig. 7.-sommeitter air compressor used at the mt. cents tunnel. At the St. Gothard tunnel, begun in 1872, Prof. Colladon first intro- duced the injection of water in the form of spray into the compressor cylinder to absorb the heat of compression. Fig. 8 illustrates the air cylinder of the Dubois-Francois type of com- pressor, which was the best in use about the year 1876. This com- pressor w r as exhibited at the Centennial Exposition and was adopted by Mr. Sutro in the construction of the Sutro tunnel. A characteristic feat- ure seems to be to get as much water into the cylinder as possible. The water which flooded the bottom of the cylinder arose from the voluminous injection; this water was pushed into the end of the cylinder and some of it escaped with the air through the discharge valve. An improved pattern of this compressor is shown in Fig 9. The first air compressor used on a large scale for practical work in America is shown in Fig. 10. H COMPRESSED AIR PRODUCTION. This machine was used at the Hoosac Tunnel, being built a little prior to 1866. The design was made under the direction of the Massa- chusetts State Commission, of which Mr. John W. Brooks was chairman FIG. 8.— DUBOIS & FRANCOIS, 1876. and Mr. Thomas Doane, chief engineer for tunnel construction. It is be- lieved that Mr. Doane deserves the largest share of credit for the inven- tion and development of this compressor, and it is to the credit of these FIG. 9.— DUBOIS & FRANCOIS, 1884. early designers to note that after the completion of the Hoosac Tunnel the compressor was transferred to the Marble Quarries, at Sutherland Falls, Vt. (now called Proctor), and that it has been used continuously up CCMPRESSED AIR PRODUCTION. 25 to the present time, compressing air to about 40 pounds to the square inch. Rock drills and channeling machines of modern construction are now using this air for quarrying the beautiful marble of Vermont. The first channeling machine was tried in this quarry perhaps with compressed air furnished by the Hoosac Tunnel compressor. The compressor is so simple that it may be readily understood by looking at the plan. It consists of 4 horizontal air cylinders, the pistons of which are propelled by a turbine wheel. The cylinders are single acting, the air being admitted through poppet valves placed in the piston. Each cylinder is 13 in. in diameter by 20 in. in stroke. It was originally intended for a speed of 120 revolutions per minute, but it has been run FIG. over 70 revolutions. The cooling arrangement applied to this com- pressor was simply an injection of water through the inlet valves into the cylinders, though since its use at Hoosac Tunnel, injection has been aban- doned, and a simple stream of water from a jet is allowed to play upon the cylinders. These illustrations are interesting from an historical point of view, as indicating the line of thought which early designers of air compressing machinery followed. As the necessity for compressed air power grew, inventors turned their attention to the construction of air-compressing en- COMPRESSED AIR PRODUCTION, ghies that would combine efficiency with light weight and economy of space and cost, the trade demanded compressors at inaccessible locali- ties, and in many cases it was preferred to sacrifice isothermal results to simplicity of construction and low cost. FIG. II.— DIRECT COMPRESSION ILLUSTRATED IN THE STRAIGHT LINE AIR COMPRESSOR IN WHICH THE MOMENTUM OF THE FLY WHEEL EQUALIZES THE PRESSURE- It is evident that an air compressor which has the steam cylinder and the air cylinder on a single straight rod will apply the power in the most direct manner, and will involve the simplest mechanics in the construc- tion of its parts. It is evident, however, that this straight line, or direct construction, results in an engine which has the greatest power at a time ^} elc\v ev^ne. n c\,l. ^ S>Te&w. * 1 il 1 — CI *7 1 ft\^-' when there is no work to perform. At the beginning of the stroke, steam at the boiler pressure is admitted behind the piston, and as the air piston at that time is also at the initial point in the stroke, it has only COMPRESSED AIR PRODUCTION. 27 free air against it. The two pistons move simultaneously as the resist- ance in the air cylinder rapidly increases as the air is compressed. To get economical results it is, of course, necessary to cut off in the steam cylinder so that at the end of the stroke, when the steam pressure is low, as indicated by the dotted line (Fig. n), the air pressure is high, as similarly indicated in the other cylinder. The early direct-acting com- pressor used steam at full pressure throughout the stroke. The Westing- house pump applied to locomotives, is built on this principle, and those who have observed it work have perhaps noticed that its speed of stroke is not uniform, but that it moves rapidly at the beginning, gradually re- ducing its speed, and then seems to labor until the direction of stroke is reversed. This construction is admitted to be wasteful, but in some cases, notably that of the Westinghouse pump, economy in steam consumption is sacrificed to lightness and economy of space. Many efforts were made to equalize the power and resistance by con- structing the air compressor on the crank shaft principle, putting the cranks at various angles, and by angular positions of steam and air cylin- ders. Several types are shown in Fig. 12. Angular positions of the cylinder involve expensive construction and unsteadiness. Experience has proved that there is nothing in the appar- ent difficulty in equalizing the strains in a direct-acting engine. It is sim- ply necessary to add enough weight to the moving parts, that is, to the piston, piston rod, fly wheel, etc., to cut off early in the stroke and secure rotative speed with the most economical results and with the cheapest construction. It is obvious that the theoretically perfect air compressor is a direct-acting one with a conical air cylinder, the base of the cone being nearest the steam cylinder. This, from a practical point of view, is impossible. Mr. E. Hill, in referring to the fallacious tendencies of pneumatic engineers to equalize power and resistance in air compressors, says: "The ingenuity of mechanics has been taxed and a great variety of devices have been employed. It is usual to build on the pattern of presses which do their work in a few inches of the end of the stroke and employ heavy fly wheels, extra strong connections, and prodigious bed plates. Counterpoise weights are also attached to such machines; the steam is allowed to follow full stroke, steam cylinders are placed at awk- ward angles to the air-compressing cylinders and the motion conveyed through yokes, toggles, levers; and many joints and other devices are used, many of which are entire failures, while some are used with ques- tionable engineering skill and very poor results." Fig. 13 illustrates the theory of Duplex Air Compressors. The hydraulic piston or plunger compressor is largely used in Germany and elsewhere on the Continent of Europe, but the duplex may be said to be the standard type of European compressor at the present time. It is also largely used in America. Fig. 13 shows the four cylinders of a duplex compressor in two positions of the stroke. It will be observed that each steam cylinder has an air cylinder connected directly to the tail rod of its 28 COMPRESSED AIR PRODUCTION. piston, so that it is a direct-acting machine, except in that the strains are transmitted through a single fly wheel, which is attached to a crank shaft connecting the engines. In other words, a duplex air compressor would be identical with a duplex steam engine, except in that the air cylinders are connected to the steam piston rods. The result is, as shown in Fig. 13, that at that point of the stroke indicated in the top section, the upper right hand steam cylinder, having steam at full pressure behind its piston, is doing work through the angle of the crank shaft upon the air in the lower left hand cylinder. At this point of the stroke the opposite steam cylinder has a reduced steam pressure and is doing little or no work, be- cause the opposite air cylinder is beginning its stroke. Referring now to the lower section, it will be seen that the conditions are reversed. One crank has turned the center, and that piston which in the upper section was doing the greatest work is now doing little or nothing, while the labor of the engine has been transferred to those cylinders which a mo- ment before had been doing no work. 1 --J -.--1 1 . ?--... I..!--- r- r ■z& FIG. 13. This is the theory of the Duplex Compressor, but it can hardly be said to be true when applied in practice, because of the heavy fly wheel which is placed on the main shaft. This wheel takes up and equalizes the power imparted by the steam pistons, and in fact it is the fly wheel which really does the work of compression at or near the end of each stroke. Heavy fly wheels are therefore essential in order to produce the best results with duplex compressors. In the duplex pattern, the crank shafts being set quartering, as is the usual construction, the engine may be run at low speed without getting on the center. Each half being complete in itself, it is possible to detach the one when only half the capacity is required. Commercially the Duplex Compressor appeals to the trade in that one side, or a half duplex, is furnished with a fly wheel and outboard bear- ings designed for a complete duplex machine, so that at first, at a little more than half the expense, one side is erected and when more air is needed the capacity of the plant is doubled by adding the other half. Where large capacities are required the duplex will admit of larger air production with economical engines, and at less expense when the cost of the plant is considered in proportion to the volume of air furnished. COMPRESSED AIR PRODUCTION. 29 Mr. John Darlington, of England, gives the following particulars of a modern air compressor of European type: "Engine, two vertical cylinders, steam jacketed, with Myer's expan- sion gear. Cylinders, 16*9 inches diameter, stroke 39-4 inches; com- pressor, two cylinders, diameter of piston 23.0 inches; stroke 39.4 inches; revolutions per minute, 30 to 40; piston speed, 39 to 52 inches per second; capacity of cylinder per revolution, 20 cubic feet; diameter of valves, viz., four inlet and four outlet, 5^ inches ; weight of each inlet valve, 8 lbs. ; outlet, 10 lbs.; pressure of air, 4 to 5 atmospheres. The diagrams taken of the engine and compressor show that the work expended in compres- sing one cubic meter of air to 4:21 effective atmospheres was 38,128 lbs. According to Boyle and Mariotte's law it would be 37,534 lbs., the differ- ence being 594 lbs., or a loss of 1 '6 per cent. Or if compressed without abstraction of heat, the work -expended would in that case have been 48,158. The volume of air compressed per revolution was 0*5654 cubic meters. For obtaining this measure of compressed air, the work ex- pended was 21,557 pounds. "The work done in the steam cylinders, from indicator diagrams, is shown to have been 25,205 pounds, the useful effect being 85^ per cent, of the power expended. The temperature of air on entering the cylinder was 50 deg. Fah.; on leaving, 62 deg. Fah., or an increase of 12 deg. Fah. Without the water jacket and water injection for cooling the temperature it would have been 302 deg. Fah. The water injected into the cylinders per revolution was 0.81 gallons." We have in the foregoing a remarkable isothermal result. The heat of compression is so thoroughly absorbed that the thermal loss is only 1 '6 per cent.; but the loss by friction of the engine is 14-5 per cent., and the net economy of the whole system is no greater than that of the best American dry compressor, w r hich loses about one-half the theoretical loss due to heat of compression, but which makes up the difference by a low friction loss. The wet compressor of the second class is the water piston compressor, Fig. 14. The illustration shows the general type of this compressor, though it has been subject to much modification in different places. In America a plunger is used instead of a piston, and as it always moves in water, the result is more satisfactory. The piston, or plunger, moves horizontally in the lower part of a U-shaped cylinder. Water at all times surrounds the piston and fills alternately the upper chambers. The free air is ad- mitted through a valve on the side of each column and is discharged through the top. The movement of the piston causes the water to rise on one side and fall on the other. As the water falls the space is occu- pied by free air, which is compressed when the motion of the piston is reversed, and the water column raised. The discharge valve is so propor- tioned that some of the water is carried out after the air has been dis- charged. Hence there are no clearance losses. The chief claim for this w r ater piston compressor is that its piston is also its cooling device, and that the heat of compression is absorbed by 3° COMPRESSED AIR PRODUCTION. the water. So much confidence seems to be placed in the isothermal feat- ures of this machine that usually no water jacket or spray pump is ap- plied. Mr. Darlington, who is one of the stanch defenders of this class of compressors, has found it necessary to introduce "spray jets of water immediately under the outlet valves," the object of which is to absorb a larger amount of heat than would otherwise be effected by the simple con- tact of the air with the water-compressing column. Without such spray connections, it is safe to say that this compressor has scarcely any cooling advantages at all, so far as air cooling is concerned. Water is not a good conductor of heat. In this case only one side of a large body of air is ex- posed to a water surface, and as water is a bad conductor, the result is that a thin film of water gets hot in the early stage of the stroke, and little or no cooling* takes place thereafter. The compressed air is doubtless cooled FIG. 14— HYDRAULIC AIR COMPRESSOR. before it gets even as far as the receiver, because so much water is tum- bled over into the pipes with it; but to produce economical results, the cooling should take place during compression. Water and cast iron have about the same relative capacity for heat at equal volumes. In this water piston compressor we have only one cooling surface, which soon gets hot, while with a dry compressor, with water jacketed cylinders and heads, there are several cold metallic sur- faces exposed on one side to the heat of compression, and on the other to a moving body of cold water. But the water piston advocate brings forward the question of speed. It is said that, admitting that the cooling surfaces are equal, we have in one case. more time to absorb the heat than in the other. This is true, and here we come to an important class division in air compressing ma- chinery high speed and short stroke as against slow speed and long stroke. Hydraulic piston compressors are subject to the laws that govern piston pumps, and are, therefore, limited to a piston speed of COMPRESSED AIR PRODUCTION. 3 1 about ioo feet per minute. It is quite out of the question to run them at much higher speed than this without shock to the engine and fluctuations of air pressure due to agitation of the water piston. The quantity of heat produced — that is, the degree of temperature reached — depends entirely upon the conditions in the air itself, as to density, temperature and moist- ure, and is entirely independent of speed. We have seen that it is possible to lose 21 "3 per cent, of work when compressing air to five atmospheres without any cooling arrangements. With the best compressors of the dry system one-half of this loss is saved by water jacket absorption, so that we are left with about n per cent., which Lie slow moving compressor seeks to erase. We are quite safe in saying that the element of time alone in the stroke of an air compressor could not possibly effect a saving of more than half of this, or 5^ per cent. Now, in order to get this 5J per cent, saving, we reduce the speed of an air-compressing engine from 350 feet per minute to 100 feet per minute. We must, therefore, in one case have a piston area three and one-half times that of the other in order to get the same capacity of air, and in doing this we build an engine of enormous proportions, with heavy moving parts. W r e load it down with a large mass of water, which it must move back and forth during its work, and thus we produce a percentage of friction loss alone equal to twice or even three times the 5-^ per cent, heat loss which is responsible for all this ex- pense in first cost and in maintenance, but which really is not saved, after all, unless water injection in the form of spray also forms a part of the system. It is obvious that cost of construction and maintenance have much to do with the commercial value of an air compressor. The hydraulic piston machine not only costs a great deal more in proportion to the power it produces, but it costs more to maintain it, and it costs more to run it. It is not an uncommon thing to hear engineers speak of the hydraulic piston compressor as the "most economical" machine for the purpose, but that it is so "expensive" and takes up so much room, and requires such ex- pensive foundations that, unless persons are "willing to spend so much money," they had better take the next best thing, a high speed machine. The hydraulic piston compressor has one solitary advantage, and that is, it has no dead spaces. It was conceived at a time when dead spaces were very serious conditions. Valves and other mechanism con- nected with the cylinder of an air compressor w r ere once of such crude construction th^t it was impossible to reduce the clearance spaces to a reasonable point, and furthermore, the valves were heavy and so com- plicated that anything like a high speed would either break them or wear them out rapidly, or derange them so that leakages would occur. But we have now reduced inlet and discharge valves and all other moving parts connected with an air cylinder to a point of extreme simplicity. Clearance space is in some cases destroyed altogether by what is, as it were, an elastic air head which is brought into direct contact w T ith the piston. All this reduces clearance to so small a point that it has no in- fluence of any consequence, The moving parts are made extremely simple. 32 COMPRESSED AIR PRODUCTION. Mr. Sturgeon, of England, has applied a most ingenious and success- ful inlet valve, which is opened and closed by the friction of the air piston rod through the gland. Mr. Sergeant in America has introduced the pis- ton inlet valve, which is opened and closed by its own inertia. We have, therefore, reached a point at which high speed is made possible. In the single or straight line compressor it is difficult to equalize power and resistance with long strokes. The speed will be jerky, and when slow the fly wheel rather retards than assists in the work of com- pression. This action tends to derange the parts and makes large bear- ings a necessity. The piston in a long stroke compressor travels through considerable space before the pressure reaches a point where the dis- charge valve opens, and after reaching that point it has to go on still fur- ther against a prolonged uniform resistance. This makes rotative speed difficult in single direct acting machines. During the early part of the stroke, the energy of the steam piston must be stored up in the moving parts, to be given out when the steam pressure has been reduced through an early cut-off. With a short stroke and a large diameter of steam cylin- der we are able to get steam economy or early cut-off and expansion without compounding. In compressors of the single or direct acting type with steam and air cylinders of equal diameter it is possible to obtain a pressure of air twice as great as the boiler pressure. This apparent enigma is made plain when it is understood that at the beginning of the stroke there is no resistance in the air cylinder. The steam end at this point has its great- est power, and the supply may be cut-off and the steam expanded in pro- portion to the pressure required in the air end, and the speed of the ma- chine. The indicator card shows a large volume and low pressure in the steam end, and a smaller volume and higher pressure in the air end, so that what is made up in the air card by high pressure is represented in the steam card by greater volume, and the area of one is nearly equal to that of the other. This can be seen by referring to Fig. n. If we omit the cut-off on the steam end the pressure, instead of fol- lowing the dotted lines, will be maintained at its maximum throughout the stroke, while the air pressure, or resistance, does not reach the steam pressure until the piston has passed the centre of the cylinder; hence if there is sufficient inertia in the moving parts, there will be no difficulty in getting an air pressure higher than that of the steam. CLEARANCE. The early designers of air compressors, as shown in the Dubois & Francois illustrations (Figs. 8 and 9), mention clearance loss in air com- pressors as a very serious matter. Even at the present time some air compressor manufacturers admit water through the inlet valves into the air cylinder, not so much for the purpose of cooling as to fill up the clearance spaces. A long stroke involving expensive construction is sometimes justified by the claim that a saving is effected by reduced clear- ance losses. COMPRESSED AIR PRODUCTION. $$ Clearance in a properly designed compressor is a loss of volume only, not a loss of power. Let us assume, for the sake of illustration, that we are compressing air with a machine which is provided with so efficient a cooling device that all of the heat of compression will be ab- sorbed as soon as produced. In other words, that we can compress air isothermally. In such a machine as this there will be a slight loss of power due to clearance space, because we would have a certain volume of air in the cylinder at each stroke, and upon which work had been done and heat produced, that heat having been absorbed and the air being re- tained in the cylinder. In other words, we would have a production and abstraction of heat, Which would represent power lost. Isothermal com- pression is practically impossible; hence we do not abstract the heat from the compressed air in the clearance space, but a large portion of this heat remains, and acts expansively upon the air, imparting its power to the piston at the moment of reversal of stroke. A reasonable clearance space behind the air piston serves a useful purpose in overcoming the inertia of the piston and moving parts acting like a spring at the end of each stroke. The clearance space in modern air compressors of the best design (including the counter-bore and discharge valve clearance) varies from .002 to .0094 of the volume of free air furnished by the cylinder. The vari- ation is somew'hat dependent upon the length of stroke of the machine. At 75 lbs. pressure, and making due allowance for increased volume of air due to heat, the clearance loss of volume varies from .01 to .047, or from one to five per cent, of the air when compressed. The actual space between the piston and the head at the end of the stroke being 1-16 in. It must not be inferred that the designers of an air compressor may neglect the question of clearance; on the contrary, it is a very important consideration. If we have a large clearance space in the end of an air compressor which is used to compress air to high pressures, we may read- ily understand a condition of things that would result in no discharge of compressed air at all, because of too large a clearance space. The entire volume of the cylinder might be compressed and retained in the clear- ance space, and the compressor will take in no free air on the return stroke, because the clearance space air when expanded is sufficient to fill the cylinder at normal or atmospheric pressure. Lass in capacity of air compressors by clearance is in direct propor- tion to the pressure. Owing to the loss of capacity by clearance in high pressure com- pressors, it is important that the cylinders be compounded. By com- pounding the air in the cylinders the clearance loss is reduced because of the reduced density in the air in the clearance space. Builders of air compressors employ three methods to reduce the clearance loss: (1) By long strokes of piston, so that the percentage of cvlinder volume to that of clearance space is reduced to a minimum. (2) 2 34 COMPRESSED AIR PRODUCTION. By filling the space with water. (3) By not allowing the full reservoir pressure to accumulate in the clearance space above the inlet valve. The long stroke plan is the best for reducing clearance, except in ma- chines of the single type, where economical compression and rotative speed cannot be accomplished with a long stroke*. The use of water to fill clearance space has been referred to previously when treating of water injection. The clearance in the initial cylinder is rilled with air at a pressure less. than the receiver pressure; and as the diameter of the high pressure cylinder is small, the loss in capacity by clearance is reduced. Mr. E. Hill states that a single type of compressor should W u u . en 3-S M . > area of the steam piston of the air-pump was no longer necessary. Experiments were conducted, both with a simple pump of suitable proportions to meet the changed conditions of service and with a compound pump, designed to attain the highest steam economy subject to the peculiar limitations of the service. The following descriptions of these tests are given by Mr. Parke in his paper before the New York Railroad Club: "The design of compound pump which seemed to best meet the re- quirements has two cylinders, respectively six inches and ten inches in diameter, and each of 10 inches stroke. Live steam is admitted to the smaller or high-pressure steam cylinder throughout the entire stroke, and, upon the return stroke, it expands into' the larger or low-pressure steam cylinder, no live steam being admitted to the latter. All the ports of both steam cylinders are controlled by a single steam valve. There are two air cylinders, the diameters of which are respectively 6\ inches and 9J inches, and the stroke of each is 10 inches. Atmospheric air is drawn into the larger or low-pressure air cylinder, and compressed therefrom into the smaller or high-pressure air cylinder. In the latter, the air is further compressed and delivered thence to the main reservoir. The piston of the high-pressure steam cylinder operates that of the low-pressure air 44 COMPRESSED AIR PRODUCTION. cylinder, and the piston of the low-pressure steam cylinder operates that of the high-pressure air cylinder. The reversing valve is actuated by the high-pressure steam piston. By this arrangement, the complete stroke of both the high-pressure steam piston and the low-pressure air piston is always assured, so that the pump cannot become dead, and a cylinder full of free air is always secured. The pump is operative at any steam pres- sure, the pressure at which the air can be delivered depending of course upon the available steam pressure. The design of simple pump resulting from these experiments was what is now known as the O/J-inch pump, which has one steam and one air cylinder, each 9 \ inches in diameter, with a 10-inch stroke. In the tests of the various pumps, as to efficiency and capacity, the steam was condensed in a surface condenser, at atmospheric pressure and weighed, and the volume of air actually delivered was carefully measured. The conditions under which the tests were made were those of about the average service to be expected on the road; this is, the steam pressure used was 140 pounds and the pumps were required to deliver against an air pressure of 90 pounds. Table VI. indicates the capacities and efficiencies of the various types of air-brake pumps under these conditions. For comparison, this table also indicates the volume of free air compressed to 90 pounds and de- Volume of Free Air Compressed to 90 lbs., and Discharged. Pump. Per Min. Per Pound of Steam at 140 lbs. Pressure. 8-inch 9^-inch Compound 5-inch Duplex 7 « Simple Compressor, oper- ated by simple engine . Two-stage compiessor operated by compound non-condensing engine 36.3 cu. ft. 44.9 43 3 39.4 38.7 1.85 cu. ft. 3.49 4.89 3.06 3.43 8.80 13.70 livered per pound of steam by a single stage commercial compressor, operated by an efficient simple engine, and by a two-stage compressor, with intercooler, operated by a compound non-condensing engine. The most striking feature of this table is the low efficiency of an air- brake pump, in comparison with a suitable compressor for a commercial supply of compressed air. Steam generation in stationary boilers for ordinary power purposes, is comparatively uniform, and the amount of fuel burned is practically proportional to the quantity of steam used. Ur>der such conditions, unless the volume of compressed air required is COMPRESSED AIR PRODUCTION. 45 very small, or is required at irregular intervals and in uncertain quan- tities, it is not economy to use an air-brake pump in the place of a suitable compressor. At the conclusion of these pump tests, it was decided to place the Qj-inch pump upon the market, and, although the design of compound pump selected proves entirely satisfactory as to capacity, and requires only one-half the steam used by any air-brake pump in the market, it was decided to abandon any thought of offering it for sale. The reasons for the latter decision w T ere chiefly the following. While the only working parts added to those of the simple pump are the additional pair of pistons and two additional air valves, a long ex- perience has led to the conviction that simplification and not complica- tion of air pump construction is what the best interests of the railroads require. The increased number of parts necessarily implies greater cost of maintenance and renewals, additional glands to be kept packed, a con- siderably increased number of sources of leakage, and the additional pair of cylinders materially increses the bulk and the weight of the pump. This was regarded as the most serious objection to the introduction of a double pump. Another serious objection is heating of the air end of the pump. It has been fully explained that compounding, or compressing in stages, is the method most to be preferred, when the air is cooled between the stages. It might also have been stated that for practical reasons, com- pounding the air, without cooling between the stages, is the worst method. By this method the air is theoretically delivered by the high-pressure air-cylinder at about the same temperature as it is from the air cylinder of a simple compressor; but the temperature of the air taken into the cylinder of a simple compressor is that of the atmosphere, while that of the air taken into the high-pressure air cylinder of the compound method (without cooling) is from 200 to 300 degrees above that of the atmos- phere. It is evident, therefore, that the mean temperature of the air in the high-pressure air cylinder of the compound method is very much higher than that of the air in a simple com- pressor. Indeed, it several times occurred, during the experiments referred to, that when a pump of the two air-cylinder type was allowed to run freely, for twenty-five or thirty minutes, under the conditions stated, the maple plank upon the brick wall of the testing room to which the air cylinders were bolted, took fire. Such high temperatures of the high- pressure air cylinder are productive of serious evils. The two air cylin- ders are necessarily cast in one piece, of an irregular shape, and expansion by heat is inevitably accompanied by distortion. This distortion at the air cylinders is aggravated by the still further increased temperature resulting from binding of the pistons. Unless the air pistons are originally fitted so loosely as to permit them to leak badly, they bind and cut, causing great wear and materially increasing the cost of maintenance. The distortion of the air cylinders always causes a large amount of leak- age past the pistons, and the actual efficiency of such pumps is far below 46 COMPRESSED AIR PRODUCTION. that calculated for them. As it is wholly impracticable to introduce a cooling chamber upon a 'locomotive, in connection with the compound air cylinders, this type of pump seems to be very undesirable. The one other important reason for abandoning- the compound pump is that steam economy in an air-brake pump is not of importance or value. Such a conclusion may at first cause some surprise, but a little study of the conditions will fully support it. The steam requirements of a locomo- tive are very difficult ones to meet, as the demand for steam to operate the engine fluctuates very greatly. The steam required by an air-brake pump is not only an exceedingly small proportion of the quantity gen- erated, but it also fluctuates between limits widely apart. The air- brake pump does its heaviest duty while stading at stations and in descending grades. At such times the engines of the locomotive use no steam, and such steam as is then used by the air-brake pump would otherwise probably escape through the pop valve. It is well known that, with the most care- ful firing, it is practically impossible to prevent a waste of steam through the safety valve at such times. FIG. 19.— DIRECT ACTING STEAM AIR COMPRESSOR. In connection with a series of tests of a Baldwin compound locomo- tive upon the Baltimore & Ohio R. R., in 1890, the consumption of steam in various incidental ways was ascertained. The weights of steam which passed through the whistle, the blower and the pop valve were computed. Instead of allowing the pop valve to blow off, a vent pipe from the steam dome was carried into the tender, and the weight of steam which would pass through it per second, when a valve in the pipe was opened, was ascertained. During the runs, when the steam reached such a pressure COMPRESSED AIR PRODUCTION. 47 that the pop valve was on the point of blowing off, the valve in the vent pipe was opened, and, by noting the length of time that it remained open, the quantity of steam which so escaped was easily computed. The num- ber and lengths of blasts of the whistle were also noted, together with the time that the blower was used. The amount of steam used by the 8-inch air-brake pump was also computed. It was found that, after the train brake apparatus had become fully charged, the pump continued to run with an average speed of about 62 strokes per minute. This would seem to indicate a considerable amount of leakage in the train pipe and con- nections. The writer has observed the operation of the 8-inch pump upon a considerable number of different trains, the air brake apparatus of which was in such condition as he happened to find it. It was found that, after the brakes had been charged throughout the train, the number of strokes per minute made by the pump were from 24, for a four-car train, to 47, for a seven-car train, the average of all the cases observed being 36. During the trials of the Baldwin compound, three separate runs were made from Baltimore to Philadelphia, with the same train, making the same time, and with the same number of scheduled stops. In the report of Mr. George H. Barrus, the steam used in different ways during these runs was computed as follows: Safety Whistle. Blower. Valve. 415 lbs 153 lbs 220 lbs. 503 lbs 198 lbs 99 lbs. 425 lbs 72 lbs 302 lbs. Average, 448 lbs 141 lbs 207 lbs. The average computed steam consumption of the air pump, under the conditions of its actual operation, was 690 lbs. Under the ordinary conditions of the brake apparatus in service, with the air pump making about 36 strokes per minute to keep up the pressure after the apparatus is once charged, the steam consumption, computed upon the same basis as that used in the report of these locomotive trials, would have been 400 lbs. It will be seen from these figures, therefore, that about the same quantity of steam w^as used by the whistle as is used by the air pump, and that, even in these locomotive trials, w T here unusual care was undoubtedly exercised in firing, the loss of steam by the safety valve was about 'half that which should be used by the air pump. Under the ordinary conditions of service, although the steam used by the air pump is largely that which would otherwise be lost at the safety valve, it is most probable that the amount of steam lost at the safety valve is greater than that used by the air pump. It will thus be understood that, with the fluctuating conditions of locomotive service, the steam consumption of the air-brake pump is an insignificant factor in the economy of steam production. It has never been demonstrated, so far as the writer is aware, that a locomotive haul- ing an air-braked train, with the brake apparatus in ordinarily good con- 48 COMPRESSED AIR PRODUCTION. dition, consumes, in the long run, a pound more coal than the same locomotive, without an air pump, would use in hauling the same train when braked by hand, and there is no good reason to believe that it would. Fig. 19 illustrates the simplest type of small air compressor built on the Straight Line or Direct Acting plan. The machine is designed for light transportation and for use in places where a compact, self-contained and very simple compressor is wanted. The steam and air cylinders are connected to a crosshead which moves in guides arranged between the cylinders. The fly wheels, which are joined by connecting rods to this crosshead, are mounted in bearings close to the steam cylinder. Slide valves are used operated by eccentrics, though in some of the larger FIG. 20. sizes the adjustable cut-off serves to aid this type of machine in fuel economy. The air cylinder is provided with poppet, inlet and discharge valves, and is water jacketed. Simplicity of design is followed in machines of this type, no claims being made to high fuel economy. Fig. 20 shows another type of simple, low cost, single or direct act- ing compressor. No crosshead is used. A single fly wheel with crank shaft and connecting rod are placed between the cylinders. Fig. 21 is shown for the purpose of illustrating the English type of single or straight line compressor. The machine shown is known as "Schram's patent." A single cast iron bed plate is used, thus making the machine self-contained, and the fly wheel's crosshead and connecting rod COMPRESSED AIR PRODUCTION. 49 FIO. 21. are placed on the bed, the steam cylinder being located between these parts and the air cylinder. This construction brings the steam and air cylinders close together, but the machine is not as compact or as simple as the direct acting compressor of the American type. Another form of European compressor on the straight line plan is shown in Fig. 22. This machine is made by Messrs. Burkhardt & Co., FIG. 22. 5<> COMPRESSED AIR PRODUCTION. Basel, Switzerland. The steam and air cylinders are widely separated and are connected by a single piston rod joined at or near its centre by a crosshead, to. whidh 'the connecting rod is attached. Fig. 23 illustrates a type of straight line compressor of the larger sizes. The following" are the specifications on which these mac'hmes are built: Cylinders. — Steam Cylinder, 14 inches in diameter, by 18 inch stroke. Air Cylinder, 14 J inches in diameter, by 18 inch stroke. Cylinders made of the best cast iron suitable for this purpose, and of proper strength and thickness for carrying 100 pounds pressure, after being re-bored once. Water Jacket. — Air Cylinder and heads provided with Water Jackets. FIG. 23. Bed Plate. — Bed Plate of the Box Girder type, made in a single cast- ing. The bed plate to extend throughout the compressor connecting both cylinders and all parts, and strong enough to withstand the severest strain of air compressing work. Bearings. — Bearings to form an integral part of the bed plate, pro- vided with removable bronze boxes, adjustable, accurately bored, and scraped to a true bearing surface. Shaft. — Main Shaft of hammered steel, 6 inches diameter in the bear- ings, turned, finished and key-seated, and free from flaws or other imper- fections. Fly Wheels. — Two square rim Fly Wheels, 6 ft. o in. in diameter, with face and edges nicely turned and hubs bored >and key-seated to fit shaft. The wheels to weigh when finished not less than 3,125 pounds. COMPRESSED AIR PRODUCTION. 5 I Crank. — Crank pins of best hammered steel and securely fastened in fly wheels, and fly wheels counterbalanced. Valve Gear. — Valve Gear of the slide valve type, with Meyer adjust- able cut-off gear. Piston Rod. — Piston Rods to extend through back head of steam cylinder and front head of air cylinder, and securely fastened in cross- head. Cross-head. — Cross-head of cast steel and amply strong, provided with bronze shoes, and with adjustment for piston rods. Material Used. — Piston Rods, Valve Rods, Connecting Rods and Crank Pins, of the best forged steel. Boxes for crank and cross-head pins of composition metal. Throttle Valve. — Steam Cylinder provided with a Globe Throttle Valve, with flanges and hand wheel turned and polished. Valve provided with drain connection. Governor. — The Compressor is provided with an Automatic Pressure Regulator and unloading device. Indicator Connection. — Provision made on both steam and air cylin- ders for Indicator connections. Oiling Devices. — Oil Cups for all wearing parts ; one sight feed Lu- bricator for steam cylinder, one Ingersoll-Sergeant Lubricator for air cylinder. All oiling devices extra large. Weight. — Total weight of Compressor ready for shipment, 10,800 pounds. Where larger volumes of air are required the direct acting type gives place to the compound and duplex. It is not considered good practice to build single compressors with air cylinders larger than 26 inches in di- ameter, except where the air is compounded. Fig. 24 illustrates a popular type of Direct Acting Compound Com- pressor. The merits of this compressor are referred to by the manufactur- ers as follows: "The large air cylinder on the left determines the capacity of the Compressor, and for the illustration we have taken its piston at 100 square inches area. The small air cylinder in the centre can have an area of 33 1-3 square inches. The small piston only encounters the heaviest pressure, and atioo lbs. pressure the resistance to its advance is 3,333 lbs. The re- sistance against tne large piston is its area multiplied by the pressure which is caused by forcing the air from the large cylinder into the smaller cylinder, which is in this case 30 lbs. per square inch. But as this 30 lbs. pressure acts on the back of the small piston and hence assists the ma- chine, the net resistance of forcing the air from the large into the small cylinder is equal to the difference of the area of the two pistons multiplied by the 30 lbs. pressure. This is 66 2-3 by 30, and equals 1,999 l° s - Hence 1,999 lbs., the resistance to forcing the air from the large into the smaller cylinder plus 3,333 lbs., the resistance in the smaller cylinder to compress- ing it to 100 lbs., is the sum of all the resistances in the compound cyl- inders at the time of greatest effort. This is 5,333 lbs. The time of great- 52 COMPRESSED AIR PRODUCTION. it Fig. 24 — Compound Air Compressor. Arrows on the water pipes show the direction of water circulation. When pistons move as indicated by the arrow on the piston rod, steam and air circulate in direction shown by arrows in the cylinders. O — Air Relief Valve, to effect A- B- -Inlet Conduit for Cold Air. ■Removable Hoods oif Wood, easy starting after stopping with all C— Inlet Valve. i D — Intake Cylinder. E — Discharge Valve. F — Intercooler. G — 'Compressing Cylinder. , H — Discharge Air Pipe. J — Steam Cylinder. K — Steam Pipe. L — Exhaust Steam Pipe. 'N — Swivel Connection for Cross- head. pressure on the pipes. 1 — Cold Water Pipe to Cooling Jacket. 2 & 3 — Water Pipe. 4 — Water Overflow or Discharge. 5 — Stone on end of Foundation. 6 — Foundation. 7 — Space to get at underside of Cylinder. . 8 — Floor line. est effort is at the end of the stroke, or when the engine is passing the centre. In the single machine this resistance is 10,000 lbs., hence we see that in the compound machine the maximum strains are less by over 46 per cent., or nearly one-half. By thus reducing the work to be done at the end of the stroke, more work is done in the first part of the stroke, and the resistance is made nearly uniform for the whole stroke. "The next step is to> render the application of power also uniform for the Whole stroke. This is accomplished in a very simple and effective Compressed air production. 53 hianner. The steam and air pistons and crosshead are mounted on the same piston rod. These parts are purposely given weight enough so so that it requires most of the power of the steam over and above the air resistance at the beginning of the stroke to start them forward at the required speed. At the end of the stroke, when the steam has become weak by expansion, the power stored up in the momentum of these reciprocating parts is given out in useful work, and the parts are brought to a state of rest by expending their force upon the air in the compressing cylinders. As the energy which can thus be stored and given out by the reciprocating parts depends upon their weight, and the square of the number of revolutions, it is evi- dent that rotative speed is the most important factor. Hence very long strokes are not desirable, because at the same piston speed the machines make fewer revolutions than machines of shorter strokes. Therefore the power is not applied to< the work so uniformly, and greater strains are brcught on shafts, connecting rods and other parts, and larger fly wheels, and frequently double engines, are necessary for successful operation, especially when steam is to be used expansively. The value of rotative speed for economical steam consumption is too well known to need re- viewing here. It is of interest, however, to note that the quick rotation that is valuable for applying the power uniformly also contributes to steam economy. "Uniform resistance and uniform power both applied, as in this com- pressor, as direct end thrust and pull upon a straight steel piston rod, do not leave much work for fly-wheels to perform. Their presence, however, is necessary to regulate the steam valve motions, to control the length of stroke, to even up and balance trifling inequalities of power and resist- ance and to secure a uniform speed to the machine." Figure 25 illustrates a common type of Duplex Compressor where the steam end is of simple construction, provided with the Meyer valve gear, the cut-off being adjusted by a hand wheel while the machine is in motion. The point of cut-off is indicated iby a pointer which moves over a graduated scale. The machine is run with a wide open throttle, being controlled entirely by the cut-off, which is proportioned in accordance w r ith the steam and air pressures. An ordinary ball governor is used to regulate the speed of the engine and to prevent it from running away in case of breakage of air pipes or sudden loss of pressure. Attached to the ball governor manufacturers usually add a pressure governor or regulator which is used to reduce the speed when the air pressure reaches the max- imum. Various types of frames are used, some of them of the Corliss pattern ; but for the purpose of insuring stability, freedom from breakage, and to resist the sudden strains which are brought to bear during com- pression, the frames, bearings and fly wheels are usually heavy. The steam and air pistons are in some patterns tied together by a heavy cast iron sole plate and tie rod. The bearings are usually of phosphor bronze, are unusually broad, and are provided with means for taking up wear. The cranks are usually made of wrought iron, and the crank pins, crosshead 54 COMPRESSED AIR PRODUCTION. phi, piston rods, shafts, valve rods, links and pins, and all wearing parts, are of steel, and when possible this is hardened and ground. The heavy fly wheel gives a smooth, uniform motion, and aids in steam economy by admitting of early cut-off. When running one side of the compressor at a time, the fly wheel prevents irregularity of motion. In this type of com- pressor the air cylinders are usually fitted with poppet valves, although FIG. 25. manufacturers have succeeded in so far perfecting the positive moved valve and the piston inlet that they are applied with economy to the Meyer Duplex machines. The air cylinder is sometimes made Of hard brass, owing to the better conductivity of this material, and is as thin as can be made with safety. The cylinders of some machines are provided with jackets for water circulation, and the piston and piston rods are hollow. COMPRESSED AIR PRODUCTION. 55 A telescopic water tube is introduced at the back end of the cylinder, and cold water is supplied to aid in keeping - down the temperature during compression. The introduction of cold water in this way undoubtedly reduces the heat of compression, but it is subject to the many disad- vantages found in water injection. The water, unless free from foreign matter, is liable to destroy the cylinder, and even when pure it is dif- ficult to lubricate the parts, the water itself being a bad lubricant. This FIG. 26. applies to that type where the water passes through the piston and in contact with the walls of the cylinder between the rings and completely around the piston. It is obvious that this leaves a thin coating of mois- ture in contact with compressed air which is at a temperature much higher than that of the water, and as the volume of air is 'so much greater in proportion than that of the water, the result is an absorption of the water, which goes off into the air receiver in the form of moisture, to be afterward deposited in the pipes when the the temperature of the compressed air is re- duced in transmission. It is claimed by the makers that one of their Du- $6 COMPRESSED AIR PRODUCT I6ti. plex Meyer Cut-off Compressors was used and tested at Shaft No. 13 on the New York Aqueduct by Prof. James E. Denton of Stevens Institute of Technology, and that it produced a horse power with a consumption of. 25 pounds of steam per horse power per hour. > - J > ! . A duplex compressor can be regulated so as to run 1 at high or low speeds. It may even stop and start automatically. Where the air is used intermittently — that is, Where the use is irregular, — a governor is fur- nished which will adjust the machine to a slow speed when little air is being used, stop when no air is required, and start again when necessary, thus using the steam only in proportion as the work is done. The following are specifications of a type of Duplex Meyer Cut-off Valve Compressor: Cylinders — 2 Steam Cylinders, 14 inches diameter, by 18 in. stroke. 2 Air cylinders, 14J inches? diameter, by 18 in. stroke. Cylinders made of the best cast iron suitable for this purpose, and of sufficient thickness to safely carry 100 pounds pressure after re-boring once. Piston Inlet. — Air cylinders of the Piston Inlet type. Water Jacket. — Air cylinders and heads provided with water jackets. Frame. — Frames of the Corliss Girder type, and strong enough to withstand the severest strain of air; compressing work. Bearings — Main pillow blocks provided with removable shell boxes of best composition metal, accurately bored and scraped to a true bearing surface. ' Shaft. — Main shaft of hammered steel, 6 inches in diameter in the bearings, and 7 inches in the centre, of proper length, turned finished and key-seated, and free from flaws or other imperfections. Cranks. — Cranks of the disc pattern, of selected charcoal iron, and of ample strength and proportion for the work required. Fly Wheel. — One square rim fly wheel, 8 feet in diameter, with hub bored and key-seated to fit shaft. The wheel when finished to weigh not less than 4,500 pounds. Valve Gear. — Valve gear of the slide valve type, with Meyer ad- justable cut-off. Piston Rods. — Piston rods extended through back head of steam cylinders, and attached by. means of couplings to piston rods of air cylinders. Tie Rods. — Air cylinders securely fastened to steam cylinders with tie rods. Governor. — A fly ball governor of approved pattern placed in main steam pipe, and driven with belt from main shaft. Lagging. — Steam cylinders covered with polished black walnut. Space between lagging and cylinder to be filled with mineral wool. Material Used. — Piston Rods, Eccentric Rods, Connecting Rods, Crank Pins, Cross-head Pins and Valve Rods of the best forged steel. Boxes for main crank pins of best /composition metal. COMPRESSED AIR PRODUCTION. $1 Throttle Valve. — Both steam cylinders provided with a throttle valve, with flanges and hand wheel turned and polished. Indicator Connection. — Provision made on both steam and air cyl- inders for indicator connections. Oiling- Devices. — Graduating sight feed oil cups for all wearing parts; one sight feed lubricator for each steam cylinder; one Ingersoll- Sergeant Lubricator for each air cylinder. Weight. — Total weight of compressor ready for shipment, 22,500 pounds. FIG. 27. Figure 26 represents a type of Vertical Compressor which consists of a steam engine connected by means of a crank shaft wrfh two single- acting air pumps, all placed in an upright position. It is compactly built, and is as economical as this type of machine will admit. The cranks are so placed in relation to each other that the greatest power of the engine is applied at the time of greatest resistance in the air cylinders. Water injection is usually applied to these machines, and it is a matter of note that it is in this compressor that water injection has had its longest service, and has given its best results. One reason for 5^ COMPRESSED AIR PRODUCTION, this may be found in the fact that the cylinders being- placed vertically, are not subject to the wear and destruction which accompanies water injection machines of the horizontal type. This compressor has been largely used in the western part of the United States for mining serv- ice, though of recent years notably at the Anaconda, Copper Mines, it has been replaced by machines of a more improved and more econom- ical type on the pattern of the Duplex Corliss Compound. Figure 27 illustrates a type of vertical air compressor used in Eu- rope, known as the "Champion." Both the inlet and the outlet valves with their seats are arranged in the cylinder covers; the form of the frame is such that the valves can be readily moved without disturbing the cylinder cover joints. The inlet and outlet valves are provided with very long guides to insure their continued working without damage, and being placed vertically, the wear is more evenly distributed. The air cylinder and outlet valve boxes are surrounded by cold water jackets. The general construction of the machine is so clearly shown in the il- lustration that further description is unnecessary. The experience of American manufacturers, which has been more extensive than that of others, has proved the value of direct compression as distinguished from indirect, as shown clearly in this type of machine. By direct compression is meant the application of power to resistance through a single straight rod. The steam and air cylinders are placed tandem. Such machines naturally show a low friction loss, because of the direct application of the power to the resistance. This friction loss has been recorded as low as 5 per cent., while the best practice is about 10 per cent, with the type which conveys the power through the angle of a crank shaft to a cylinder connected to the shaft through an addi- tional rod. 9681 1^ if)i A