REESE LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF CALIFORNIA 
 
v 
 
PRACTICE AND THEORY 
 
 OF THE 
 
 INJECTOR. 
 
 BY 
 
 STRICKLAND I,. KNEASS, C.E., 
 
 MEMBER OF THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS, FRANKLIN 
 INSTITUTE, ENGINEERS' CLUB OF PHILADELPHIA. 
 
 REVISED AND ENLARGED. 
 
 NEW YORK: 
 
 JOHN WILEY & SONS. 
 
 LONDON: CHAPMAN & HALL, LIMITED. 
 
 1898. 
 
COPYRIGHT, 898, 
 
 BY 
 STRICKLAND L. KNEASS. 
 
PREFACE. 
 
 Although much has been written concerning the 
 theory and the action of the Injector, there have been 
 few books published since the appearance of Giffard's 
 own pamphlet in 1860, which have been based directly 
 upon experimental research. 
 
 It has been the object in the following pages to 
 present solutions of some of the more interesting 
 problems, with illustrations drawn from practical tests ? 
 and to describe in detail the function of the different 
 parts. To the professional engineer and to the student, 
 theoretical discussion of the Injector is a tempting 
 field, because of the beauty of its underlying princi- 
 ple and by reason of the numerous associated problems 
 of fluid motion ; the analysis of this part of the sub- 
 ject, based upon carefully conducted laboratory tests, 
 has been fully treated, and complex formulae have 
 been avoided in the mathematical discussion. 
 
 111 
 
CONTENTS. 
 
 CHAPTER I. PAGK 
 
 EARLY HISTORY i 
 
 CHAPTER II. 
 DEVELOPMENT OF THE PRINCIPLE . . . , . . 8 
 
 CHAPTER III. 
 
 DEFINITION OF TERMS DESCRIPTION OF THE IMPORTANT 
 PARTS OF THE INJECTOR 25 
 
 CHAPTER IV. 
 THE DELIVERY TUBE 30 
 
 CHAPTER V. 
 THE COMBINING TUBE 44 
 
 CHAPTER VI. 
 THE STEAM NOZZLE 53 
 
 CHAPTER VII. 
 THE ACTION OF THE INJECTOR c . 67 
 
 CHAPTER VIII. 
 
 APPLICATION OF THE INJECTOR AMERICAN AND FOREIGN 
 PRACTICE 89 
 
 CHAPTER IX. 
 DETERMINATION OF SIZE TESTS 125 
 
 CHAPTER X. 
 
 REQUIREMENTS OF MODERN RAILROAD PRACTICE REPAIRS 
 METHODS OF FEEDING L/OCOMOTIVE BOILER 149 
 
 INDEX 159 
 
 iv 
 
THE GIFFARD INJECTOR. 
 
 CHAPTER I. 
 
 EARLY HISTORY. 
 
 To HENRI JACQUES GIFFARD, an eminent French mathe- 
 matician and engineer, belongs the honor of having invented 
 the simplest apparatus for feeding boilers that has ever been 
 devised, utilizing in a novel and ingenious way the latent 
 power of a discharging jet of steam. 
 
 From the time of his graduation from LS Ecole Centrale 
 in 1849, Giffard had directed his energies to the study of 
 aeronautics and had spent much time in developing a light 
 steam motor for propelling balloons; it is, therefore, not 
 strange that he should also have attempted to devise a com- 
 pact and convenient substitute for the steam pumps then in 
 use. Already a number of patents had been granted him for 
 the application of the steam engine to aerial navigation and 
 for other correlated inventions when, on May 8, 1858, 
 letters patent were issued for L' Injecteur Automoteur. His 
 early technical education and wonderful ingenuity well fitted 
 him for breaking away from the old beaten paths and start- 
 ing out on a new line of discovery ; and in view of the 
 originality of his work he fully deserved the unqualified 
 praise accorded him by his contemporaries. 
 
 Upon purely theoretical grounds the method by which he 
 proposed to force a continuous stream of water into the 
 boiler appeared to be entirely feasible and would, if practi- 
 cable, possess many advantages over the intermittent sys- 
 tems. The difficulty seemed to lie in fulfilling the peculiar 
 
 i 
 
2 THE GIFFARD INJECTOR. 
 
 conditions required for the condensation of the steam and 
 the subsequent reduction of the velocity of the moving mass. 
 GifTard carefully considered the various phases of the ques- 
 tion and made a working drawing embodying his ideas. A 
 model was made by M. Flaud & Cie., of Paris, who found, 
 however, considerable difficulty in forming the tubes in the 
 peculiar shapes required. But in the shape and proportions 
 of the nozzles lay the element of success, and the first in- 
 strument constructed entirely fulfilled the expectation of the 
 designer. 
 
 There have been few other inventions in which the under- 
 lying principles have been so thoroughly worked out by the 
 original inventor. Giffard seems to have made a very com- 
 plete survey of the possibilities of the Injector prior to placing 
 it before the public, and in his patent specification, describes a 
 number of improvements that have since been made. In 1860 
 he published a small brochure entitled " A Theoretical and 
 Practical Paper on the Self-acting Injector," in which he 
 says : "Of all the necessary accessories of a Steam Engine, 
 perhaps the most important is the one used for feeding water 
 to the boiler ; upon its proper working depends not only the 
 regular running of the engine, but the safety, the very exis- 
 tence of those who approach the boiler ; . . . . nevertheless, 
 by a kind of fatality, the apparatus employed up to the pre- 
 sent time for feeding is, of all others, that which leaves most 
 to be desired." After reviewing the disadvantages of the 
 various methods in use, he continues, "It is important, 
 therefore, to,create a new method, free from the imperfection 
 and inconvenience pointed out," and modestly adds, " Such 
 is, it appears to me, the result obtained by the apparatus to 
 which I have given the name of Injector, because it produces 
 a veritable continuous injection. Its mode of action, extra- 
 ordinary in appearance, contrary to that which w r e are in the 
 habit of seeing or supposing, is explained by the simplest 
 laws of mechanics and has been foreseen and calculated in 
 advance." He describes his invention in detail and ex- 
 plains very fully the best proportions for its various parts, 
 and also the mechanical theory, substantially as advanced 
 
EARL Y HISTOR V. 3 
 
 by him in 1850, eight years before the construction of his ex- 
 perimental Injector. 
 
 And yet, in common with all new inventions and radical 
 improvements, great difficulty was at first experienced in ob- 
 taining a fair trial of its merits, and in many cases the ex- 
 aggerated claims of its friends interfered as much with its 
 early adoption as the openly expressed criticism of its 
 enemies. The great advantages of the new method were 
 appreciated, however, by the Academic des Sciences of 
 France, who awarded Giffard the Grand Mechanical Prize 
 for 1859. This was all the more complimentary as it was 
 entirely unsolicited. Prominent engineers presented before 
 the principal scientific societies analytical demonstrations of 
 the theoiy of the injector and allayed to a great extent the 
 suspicion in the popular mind that the inventor was en- 
 croaching dangerously near the claim for perpetual motion. 
 Combes, Bougere, Reech, Villiers, Zuber and Pochet are 
 among the most prominent scientists who made a special 
 study of the subject, and the demonstration of Pochet is 
 still frequently used in modern text books. 
 
 It must not be supposed that Giffard was alone in his 
 efforts to utilize the power of a discharging jet. For ex- 
 hausting and pumping purposes we have record that a crude 
 ejecting apparatus had been used as early as 1570 by Vitrio 
 and Philebert de Lorme. But the first device that bears any 
 similarity to the principle of the Injector was patented 
 August 15, 1818, by Mannoury de Dectot, who describes 
 "sundry motors or means for employing the power of fire, 
 of steam, of air, etc., to start the movement of machines." 
 He applied his invention for raising water and for propelling 
 boats by utilizing the expansion and condensation of steam 
 in connection with jets of water. 
 
 Ravard followed in 1840 with improved forms, but the 
 greatest advance was made by Bourdon, the celebrated in- 
 ventor of the metallic steam gauge, who approached very 
 near the results obtained by Giffard. Two patents were 
 issued to Bourdon, one in 1848 and one in 1857, but it is to 
 the latter that special reference will be made. This con- 
 
4 THE GIFFARD INJECTOR. 
 
 tained numerous combinations of convergent and divergent 
 tubes for transforming the energy of a moving jet, or for dis- 
 charging large or small quantities of liquids or gases. The 
 similarity of the form of the apparatus to that of Giffard 
 was so marked that the question of priority at once arose 
 and was exhaustively discussed by the " Societe des Inge- 
 nieur Civils" It was shown that Giffard was wholly un- 
 aware of the last improvement of Bourdon when he applied 
 for his patent, and as he had publicly presented the theory of 
 his invention nearly seven years in advance of Bourdon, full 
 credit was given him for the conception of the Injector and 
 originality in the application of the principle. 
 
 The introduction of the injector into England by Sharp, 
 Stewart & Co., of Manchester, is thus described by one 
 thoroughly familiar with its history and to whom its early 
 success in that country was in great measure due : 
 
 "In the autumn of 1859 when our representative in Paris 
 sent over to me a No. 4 Injector as a curiosity and engineer- 
 ing anomaly, he told me simply what it did, but gave no 
 instructions for fixing or working. At about the same time 
 the Paris representative of Messrs. Robert Stephenson & Co., 
 Newcastle, sent over to them a similar Injector. I set to 
 work at once, and by good luck coupled up the correct pipes 
 to their proper flanges, but was a great deal bothered what 
 to do with the overflow flange. After a few nights' work I 
 got my Injector fixed and got up steam, and to some extent 
 began clumsily experimenting as the pressure rose to 60 
 pounds, the full working pressure of the boiler. I had the 
 Injector fixed over a tank fed by a ball tap and closed by the 
 boiler. I turned steam on and was staggered by the rush of 
 water into the tank from the overflow pipe, and thought 
 something was wrong. However, I continued to turn the 
 steam spindle, and the escape from the overflow sensibly 
 diminished until I could turn no further. In the mean 
 time the ball tap started running furiously into the tank, 
 showing me that water was going somewhere and I knew it 
 could go nowhere else but into the boiler. I then began to 
 operate with the four thread screw at the side, and found 
 
EARLY HISTORY. 5 
 
 that it adjusted the water supply, and succeeded in getting 
 the overflow "dry." I then opened the peep-holes opposite 
 the space between the combining and the receiving nozzles, 
 and saw the white stream passing from one to the other on 
 its way to the boiler. I then ceased operations, and had a 
 pipe of tobacco, and let some water out of the blow-off cock ; 
 then I tied a piece of spun yarn round the glass water gauge 
 to prepare for another start, and shortly after, the senior 
 partner came round for a stroll and found me operating. I 
 stopped it, started it, and regulated it so much to his satis- 
 faction that within one week the monopoly of its manufac- 
 ture in England was secured by the firm. Unfortunately for 
 Stephenson & Co., they coupled their sample Injector up in- 
 correctly, and it would not work." 
 
 For stationary service the Injector did not at first become 
 popular ; possibly on account of the mystery that seemed to 
 surround its working, and the general skepticism as to its 
 practical wearing powers. Some of the contributions and 
 queries published in the engineering papers of the day, are 
 very amusing, and a certain writer in one of the most 
 prominent weeklies proves most conclusively to his own 
 and probably to some of his readers' satisfaction, that the 
 new method of feeding boilers was an absolute impossibility. 
 The injector was, however, adopted in many places and 
 continued to give satisfaction. In the first trip of the 
 "Great Eastern" Injectors were used in place of pumps, 
 but for some reason not explained, they were subsequently 
 removed ; this may have been owing to the temperature of 
 the feed water being too warm for efficient service, as this 
 was the weak point of the first injectors constructed. 
 
 The first injector applied to a locomotive in England was 
 by Mr. J. Cross, Superintendent of the St. Helens Railway. 
 It was successful from the start, although not large enough 
 for the purpose and therefore a No. 8 was substituted, which 
 proved to be entirely satisfactory. 
 
 The English railroads opened a wide field for the Injector; 
 upon most of the locomotives, the earliest feeding pumps 
 were worked by hand, but afterwards coupled to a special 
 
6 THE GIFFARD INJECTOR. 
 
 eccentric or to the crosshead. Stretton, in his recent work 
 on the Locomotive, says that it was a common occurrence for 
 engines with a single pair of driving wheels, to stand on 
 well greased rails with tender brakes fast locked and drivers 
 revolving, in order to fill the boiler full of water. But even 
 though the old methods were very crude, engineers in Eng- 
 land were much prejudiced against any change, even for the 
 better. By way of illustration the following incident may 
 be given of a successful attempt to convince an obstinate 
 engineer against his will of k the advantages of the injector : 
 " Permission had been obtained from the Locomotive Super- 
 intendent of one of the principal Railway Companies in 
 Great Britain to try one on a goods engine, and for me to 
 accompany it on its trial trip, with a loaded goods train, 
 about 70 miles out and 70 miles back. On the outward 
 journey I was only permitted by the driver to make short 
 intermittent trials of the Injector, he depending for his water 
 supply upon his pumps. When we got to the end of our 
 outward journey, and while driver and firemen were having 
 their mid-day meal at a local public house, I went to the 
 running shed, filled up the boiler with the injector and took 
 out the balls from the two suction clacks and put them in 
 my pocket. We had not gone many miles on our return 
 journey when water was w r anted in the boiler, but upon the 
 pumps being tried, first on one side, then on the other, and 
 naturally refusing to work without suction check clacks, I 
 was appealed to, to put my Injector on, with the result that 
 we completed our journey without delay or hitch of any 
 kind, depending solely on the one No. 8 Injector. The 
 driver consequently reported ' Pumps out of order and could 
 not have got along without that Injector.' This was a 
 grand testimonial, but I got into a jolly row for my temeiity 
 in removing the clack balls." 
 
 The Injector was introduced in the United States by Wm. 
 Sellers & Co., who commenced its manufacture in 1860 at 
 their works in Philadelphia. Of locomotive builders, 
 Matthias Baldwin, was the first to use the new instrument, 
 applying, in September 1860, a No. 8 Injector to an engine 
 
EARL Y HISTOR Y. 7 
 
 designed for the Clarksville and Louisville R. R. The fol- 
 lowing month the Detroit and Milwaukee R. R. put the 
 Injector in use on one of their locomotives, and the Penn- 
 sylvania and the Philadelphia and Reading followed in the 
 latter part of the same year. 
 
 To Jos. R. Anderson & Co., Richmond, Va., a No. 4 
 Injector bearing progressive number i, was shipped in Octo- 
 ber 1860. As indicative of the wearing qualities of these 
 early instruments Messrs. Wm. Sellers & Co. state that 
 there was returned to them, in 1887, a No. 4 Injector, pro- 
 gressive No- 7, after a nearly continuous service of 27 years, 
 and having required but few repairs ; it further is interesting 
 to note, that, owing to improvements recently introduced, 
 American Injectors are now extensively used in France, 
 and have been adopted as a standard type by several of the 
 government railroads in the country of its inventor. 
 
 It need hardly be said that the Injector is the most popular 
 boiler feeder now in use. There have been more than 500,- 
 ooo manufactured in this country for the various kinds of 
 service, and there is scarcely a locomotive in the world that 
 is not equipped with one or two Injectors. Compact, reliable 
 and economical, it still deserves the high encomium bestowed 
 upon it in 1859, by M. Ch. Combes, Inspector General and 
 Director L^ Ecole des Mines, " It is without doubt better 
 than all devices hitherto used for feeding boilers, and the 
 best that can be employed, as it is the simplest and most in- 
 genious." 
 
CHAPTER II. 
 
 DEVELOPMENT OF THE PRINCIPLE. 
 
 GIFFARD having established beyond doubt the power, of a 
 discharging jet of steam to lift a mass of feed water many 
 times its own weight and force it against the initial pressure, 
 it became necessary to prepare the constructive details of the 
 new boiler feeder. The arrangement decided upon, could not 
 in the light of subsequent events be considered as an entire 
 success, as it contained inherent defects that caused frequent 
 failures and prevented the placing of as much confidence in 
 the new boiler feeder, as the merits of the invention deserved. 
 Many locomotives that at first were equipped with two in- 
 jectors, were afterward altered so as to have a pump upon the 
 left hand side to be used in case the injector should refuse to 
 work, and it was not until 1875 or 1876 that more recent 
 improvement in construction restored the confidence that 
 the original defects had forfeited, and the pump was driven 
 from service upon locomotives in the United States ; even 
 yet upon some of the Knglish Railways, a pump is used on 
 one side of the engine, arranged somewhat in the manner of 
 the pressure or vacuum pump for the air brakes. 
 
 The curves of the tubes and nozzles as laid down by 
 GiiFard were beyond criticism, and are still used when this 
 type of injector is manufactured ; his thorough knowledge 
 of the laws governing the action of the jet and the accelera- 
 ting velocity of the moving mass, enabled him so to con- 
 struct the curves of approach and recession, that they have 
 been used as a prototype for all subsequent forms of injec- 
 tors ; except for one change advanced by our increased in- 
 formation regarding the action of steam during expansion, 
 and a few minor modifications for economy of manufacture, 
 8 
 
DEVELOPMENT OF THE PRINCIPLE. 9 
 
 or for adapting the injector to special purposes, no change 
 in the contour of the tubes has been made. 
 
 But that there has been development, cannot be denied ; 
 it may be considered as following three lines : 
 
 First. Constructive changes. 
 
 Second. Carrying out the ideas suggested by Giffard in 
 his pamphlets or patent specifications. 
 
 Third. The discoveries of new properties of the jet, or 
 the application of new principles. 
 
 Almost all important inventions follow in this natural 
 sequence during their development, and the injector was no 
 exception to the rule. Genuine mechanical ability is seldom 
 combined with inventive genius, and it almost always follows 
 that the fullest development is obtained in other hands than 
 those of the original inventor. The first improvements were 
 therefore in the line of correcting the defects that became 
 apparent after the injector had been subjected to the test of 
 actual service ; changes required to facilitate repairs, or the 
 adjustment of the positions of the tubes. In the second di- 
 vision lies' the basis of many subsequent improvements that 
 have since proved valuable, and Giffard has never been 
 given sufficient credit for his wonderfully wide grasp of the 
 possibilities or future development of the injector. Of the 
 third there will be less to relate, as the only real advance 
 has been with the discovery of the peculiar property of the 
 movjng jet by which the instrument was made self-regu- 
 lating, and with the novel arrangement of tubes by which 
 the re-starting feature was added. 
 
 The general appearance of the injector as now constructed 
 is entirely different from the original form, and it would be 
 difficult for any one not specially familiar with the subject 
 to recognize one made in 1858 ; the arrangement of the ad- 
 justing handles, peculiarly shaped body, and queer little peep 
 holes present to the modern eye a very odd appearance, while 
 the heavy flanged pipe connections and steam cock do not 
 contrast at all favorably with the neater form now used on 
 American boilers. 
 
 Figure i shows a sectional view of the earliest form of in- 
 
THE GIFFARD INJECTOR. 
 
 <o 
 
DEVELOPMENT OF THE PRINCIPLE. 11 
 
 jector manufactured for public sale, and was intended for 
 use on either stationary or locomotive boilers ; it was made 
 entirely of brass, with the body composed of three pieces, 
 screwed and bolted together, and the steam, feed, and boiler 
 connections terminating in flanges, as is still the general 
 practice in England and on the continent. The method of 
 starting and operating the injector was as follows : upon 
 opening the cock A , steam from the boiler passed freely into 
 the steam ram a' through the small drilled holes a" ; the spin- 
 dle, which had been previously forced down to a tight bearing 
 against the taper steam nozzle a, was drawn back one turn 
 of the handle F, exposing an annular area approximately 
 one-quarter that of the end of the tube b. The rapid dis- 
 charge of the steam through the combining tube b, entrained 
 the air in the suction branch B, and formed a partial vacuum 
 in the feed pipe which raised the water to the injector; the 
 spindle was then drawn fully back, and the increased dis- 
 charge of steam imparted to the surrounding body of water, 
 during its passage through the combining tube, sufficient 
 velocity to cross the overflow space d, and enter the boiler 
 through the delivery tube c. 
 
 Assuming the correct relation to exist between the dis- 
 charge area of the steam nozzle and the annular entrance 
 area for the water at the larger end of the combining tube, 
 the steam will force into the boiler many times its weight of 
 water, without tendency to spill at the overflow d ; if change 
 take place in the pressure of the steam or of the feed water, 
 one of two conditions will be introduced, depending upon 
 the direction in which the changes of pressure occur : either 
 there will be an insufficient supply of water for perfect con- 
 densation during the passage of the jet through the combin- 
 ing tube, or too great a quantity of water will enter the in- 
 jector to pass through the delivery tube c. It was to allow 
 for this, that Giffard introduced two adjustments: one for 
 the water area by giving an axial movement to the ram and 
 steam nozzle by the lever G, and the other for the steam dis- 
 charge area by the handle F on the spindle ; if the pressure 
 of the steam or the lift of feed water were increased, it was 
 
12 THE GIFFARD INJECTOR. 
 
 necessary to draw the steam ram further back to permit the 
 admission of a larger quantity of water, while a fall of pres- 
 sure would require a reverse movement. This necessity for 
 re-adjustment was so frequent upon locomotives where the 
 pressure was subject to wide fluctuation, that the packing 
 upon the steam ram soon wore loose and gave constant 
 trouble, allowing leakage to occur between the steam and 
 feed chambers, impairing seriously the efficiency of the in- 
 jector and its power of suction. This packing was formed 
 in an ingenious manner, although in service it was found in- 
 effective, and no device has yet been introduced that has 
 proved thoroughly reliable for this purpose. The ram was 
 provided with three kinds of packing : the lower part was first 
 turned slightly larger than the bore of the body, and small 
 triangular grooves s turned, not spirally, but in planes nor- 
 mal to the axis ; the ram was then driven to place, forcing 
 the tops of the sharp edges backward and reducing it to the 
 same size as its bearing. A deeper groove p' y was filled 
 with rubber or hemp, and a second groove /, with a split 
 metallic ring; while another set of V's was placed between 
 the ring and the steam nozzle. Upon the body of the in- 
 jector and directly opposite the overflow d, four peep holes 
 were placed and covered by a sliding ring with handles d' ; 
 this could be rotated until the holes in the ring were oppo- 
 site those in the body, and permitted the inspection of the 
 condition of the passing jet. When new and in good con- 
 dition this injector worked very well and its double adjust- 
 ments gave it long wearing power ; any ordinary increase 
 in the diameter of the nozzle due to the attrition of the 
 rapidly moving jet, could be counteracted by a change in 
 the areas, although at a reduction of the efficiency. 
 
 As the constructive defects of the injector were obvious, 
 the first changes introduced were improvements in the 
 packing of the steam ram, and those manufactured in this 
 country in 1860 were provided with a conical stuffing box 
 filled with a number of split rings made of metal softer than 
 the body ; but unequal wear upon the ram soon caused leak- 
 age with this system also. It is singular that the merits of 
 
DEVELOPMENT OF THE PRINCIPLE. 13 
 
 the fixed nozzle injector were not earlier appreciated. Mill- 
 holland, of Reading, Pa., patented in 1862, an injector 
 formed of non- adjustable tubes, depending upon the use of 
 external valves for regulation ; but this met with no success, 
 and does not appear to have ever gained publicity. Internal 
 adjustments seem to have been regarded as all important, 
 for the early Giffard injectors were not provided with check 
 valves upon the overflow, and any reduction of the water 
 supply would cause considerable indraft of air to the boiler 
 unless counteracted by corresponding reduction in the steam 
 discharge ; the avoidance of the check valve may have been 
 intentional, and due to the weak form of lifting jet em- 
 ployed, as the power of suction was affected by the slightest 
 increase in resistance to the free discharge of the steam. 
 Subsequently, fixed nozzle injectors came into extensive use 
 and are now applied to many kinds of service. 
 
 As the problem of successfully packing the steam ram 
 seemed impossible of solution, Robinson & Gresham of 
 Manchester, England, attempted in 1864, the plan of regu- 
 lating the water area by moving the combining and delivery 
 tubes toward or from a fixed steam nozzle ; a pinion meshing 
 into a rack cut upon the exterior of the tube was actuated 
 by a hand wheel on the body, and a long cylindrical bearing 
 upon the outside of the delivery tube prevented any serious 
 leak from the boiler pipe. A similar device was introduced 
 in 1868, by Samuel Rue of Philadelphia, who applied stuff 
 ing boxes to the ends of each tube, and regulated their 
 position by means of a hand-lever ; the fundamental problem, 
 was, however, to make the adjustments of the water supply 
 automatic, and eliminate as far as possible the necessity for 
 watchful care on the part of the engineer. GifFard had 
 already described in his patent specification two methods by 
 which this could be obtained ; one by placing a spring be- 
 hind a movable steam nozzle in such a manner that an 
 increase of pressure would cause compression of the spring, 
 and elongate the distance between the tubes, while the re- 
 silience of the spring would regulate properly for a converse 
 condition. This was never put into practical operation, and 
 
14 THE GIFFARD INJECTOR. 
 
 it is easy to see difficulties that would discourage the 
 attempt. The other suggestion was to vary the pressure of 
 the entering water by subdividing the steam jet and using 
 the first or subsidiary apparatus, for feeding the second or 
 forcing set of tubes under a pressure that would vary in pro- 
 portion to the requirements. A still simpler method was 
 that patented by Win. Sellers of Philadelphia in August 
 1865, by which any change in the internal condition of the 
 jet would actuate a set of movable tubes so that the correct 
 proportion of the water to the steam would always be ob- 
 tained. 
 
 Ever since the introduction of the Injector, the advan- 
 tage of an automatic adjustment was realized, and a series 
 of experiments had been carried on by \Vm. Sellers & 
 Co. , for the purpose of attaining this end ; after numerous 
 experimental devices had been thrown aside the injector 
 shown in section in Fig. 2, was tested and found to work 
 satisfactorily. The basis of the invention was the discovery 
 of the strong vacuum produced by the moving jet in a 
 closed overflow chamber when the water supply was reduced 
 below the maximum, and that an excess of feed water would 
 cause the pressure in a confined overflow chamber to rise 
 above that of the atmosphere before the jet would break. Now 
 as the most efficient performance of an injector is when the 
 pressure of the jet while crossing the overflow is the same 
 as that of the atmosphere, indicating that the density is ap- 
 proximately unity and that the steam entirely condensed, 
 it only remained to compel the condition of the overflow to 
 influence the water entrance to the combining tube. Any 
 variation of the steam and water supply would then first 
 affect the absolute pressure in the overflow chamber, and 
 this would immediately respond upon the water admission 
 by a proper movement of the combining tube. 
 
 Fig. 2 gives a sectional view of one of the earliest experi- 
 mental forms of this device with constructive details inten- 
 tionally omitted. Steam, feed, and boiler connections are 
 indicated by the letters A, B, and C, and the steam nozzle, 
 combining tube and delivery tube, by a, b, and c ; D is a closed 
 
DEVELOPMENT OF THE PRINCIPLE. 
 
 15 
 
 overflow chamber, communicating wilh the interior of the 
 tubes through the overflow d, situated slightly above the 
 minimum diameter of the bore of the tube. The delivery 
 tube c passes completely through the delivery chamber C, 
 and is packed by stuffing boxes g and h of equal diameter, 
 and is therefore balanced with reference to any pressure 
 carried in the boiler, so that the tubes are free to move under 
 the influence of any pressure within the chamber D, acting 
 upon the differential areas of the piston b' and outside of the 
 delivery tube c. As discharge from the overflow chamber 
 was not permitted, the injector was started by means of a 
 valve placed in the boiler branch Cand not shown in the illus- 
 tration. Two years later it was discovered that the pressure 
 in D was strong enough to obviate the use of the balanced 
 
 delivery tube, and the simpler form shown in Fig. 3 was de- 
 vised. It could be placed in any position, horizontally or 
 vertically, as gravity had no appreciable effect upon its 
 action. - Using the same nomenclature as before, a is the 
 steam nozzle, b the combining tube and c the delivery tube ; 
 the operation was as follows : the starting valve D, placed 
 below the delivery tube was opened and free discharge 
 given for the lifting steam jet issuing from a small hole 
 drilled in the steam spindle. This gave a strong suction 
 vastly superior to the lifting power of the original Giffard 
 injector, and being in excess of the requirements of the or- 
 dinary height of lift, produced a lower pressure in the closed 
 chamber d than in the feed pipe. The combining tube 
 would therefore move down to the lower end of its stroke 
 exposing full admission area for the water, until the rising of 
 
16 
 
 GIFFARD INJECTOR. 
 FlG. 3. 
 
 D 
 
DEVELOPMENT OF THE PRINCIPLE. 17 
 
 the water and condensation of the issuing steam jet caused a 
 nearly perfect vacuum, which drew the tube to the other end 
 of its stroke. As the spindle was drawn slowly back, the 
 proportion of steam became too great for immediate conden- 
 sation, and the contraction of the jet continued while crossing 
 the overflow ; the ensuing vacuum in the overflow chamber, 
 acting upon the piston head of the combining tube drew the 
 tube away from the steam nozzle, admitting an increased 
 supply of water. It follows that regulation would occur with 
 any change in the steam supply, whether that change be 
 produced by a variation of the pressure of the boiler, or of 
 the effective area of the steam nozzle. When the spindle is 
 drawn fully back, the starting valve D is closed and the 
 feeding of the boiler established. It follows, therefore, that 
 a reduction in capacity of the injector may be effected by 
 running the spindle part way down into the steam nozzle, 
 and the water supply will be automatically reduced in pro- 
 portion ; at any fixed steam pressure the temperature of the 
 water delivered to the boiler will be nearly the same for 
 both the maximum and minimum capacity ; a change in the 
 height of lift would diminish the quantity of feed water, 
 until a downward motion of the combining tube, due to the 
 increased vacuum in the closed overflow chamber, would 
 compensate by increased area, the loss of velocity due to 
 the reduction in the difference of pressure between the in- 
 terior of the combining tube and the height of lift. This 
 method of regulation gives exceedingly good results over a 
 wide range of pressure, and with an expenditure of steam 
 that is proportional to the work performed. 
 
 From an inspection of Fig. 3, it will be seen that the per- 
 formance of the injector depends upon the tightness of the 
 joint between the delivery tube b, and its sleeve bearing, 
 just as the original form relied upon the packing of the ad- 
 justable steam nozzle. Unequal wear of this part permitted 
 leakage between the boiler and the overflow chamber, 
 although not as rapidly nor to so fatal an extent as in the 
 former case. 
 
 This led to the application of the discovery previously 
 2 
 
18 . THE GIFFARD INJECTOR. 
 
 made by Gresham, that the overflow space between the com- 
 bining and delivery tubes could be lengthened without 
 materially affecting the efficiency, and in 1873 this form of 
 injector was still further improved by separating the com- 
 bining tube from the delivery tube, screwing the latter into 
 the body, and permitting the combining tube free movement 
 toward and from the steam nozzle ; this altered the length 
 of the overflow space with every change in the position of the 
 tube; with low pressure, the combining tube was drawn 
 close to the steam nozzle, exposing the maximum length of 
 overflow space ; as the pressure rose, this distance was re- 
 duced and the length of the overflow at the higher pressure 
 was less than that used in the older form of injector. 
 
 The principle and development of this invention has been 
 explained in detail on account of its merit and originality. 
 In an improved and modified form this injector is exten- 
 sively used at the present time, and still preserves the pecu- 
 liar principle that made it such an advance over the earlier 
 instruments employed. 
 
 The other method of automatically regulating the water 
 supply, by subdividing the actuating steam jet, was intro- 
 duced in the United States in 1876-7, when two American 
 patents were issued; one to Ernest Korting of Hanover, 
 Prussia, and the other to John Hancock of Boston ; foreign 
 patents had already been granted to the former, who had 
 been a manufacturer and experimenter for many years. The 
 devices were practically alike ; the invention consisted in 
 combining two sets of tubes, and each set was proportioned 
 to the function it had to perform. A small steam nozzle 
 discharged into a large combining tube, which delivered the 
 feed water under a slight pressure to the entrance of the 
 smaller combining tube of the forcing set, where, meeting 
 the steam issuing from a larger nozzle, the water received 
 an additional impulse sufficient to force it into the boiler ; 
 the increase in feeding capacity of the first set of tubes with 
 a rise in the steam pressure, obviated the necessity for any 
 regulation of the water supply to allow for fluctuation in 
 the pressure carried in the boiler. 
 
DEVELOPMENT OF THE PRINCIPLE. 
 
 19 
 
 A section of this injector is shown in Fig. 4 where a and b 
 are the tubes of the second, or forcing set, a and b\ of the 
 lifting set ; an overflow cock is placed beyond each delivery 
 tube for the purpose of starting, as no overflow aperture 
 separates the convergent combining tube from the divergent 
 delivery tube. The proportions in which the sets differ are 
 apparent; the first set is primarily, an ejector, as the 
 diameter of the steam nozzle is less than that of its delivery 
 tube. In the other set the conditions are reversed, and an 
 injector is formed which is capable of forcing the feed water 
 against a pressure considerably above that of the initial jet. 
 A separate valve is provided for each steam nozzle, so that 
 
 by admitting steam first to a f , the feed water is raised and 
 flows from the waste cock c? ; this is now closed, compelling 
 the water to pass through the smaller combining tube b' , 
 the steam nozzle a is now opened and the jet established 
 through the forcing combining tube, aided by the velocity 
 of the entering water due to the pressure produced by the 
 delivery tube of the lifting set ; the closing of the final waste 
 cock d compels the water to enter the boiler. 
 
 The self-adjusting principle of the Double Jet Injector 
 depends upon the variation of feed pressure when the water 
 entrance to the combining tube is constant. Numerous 
 changes have since been made in the special devices by 
 which the necessary sequence of opening and closing the 
 
20 THE GIFFARD INJECTOR. 
 
 various steam and waste valves could be obtained. In its 
 modern form, this injector is very convenient to operate, and 
 the principle of its construction is incorporated in many of 
 the best known injectors of the present day. 
 
 In tracing the development of the self-adjusting principle, 
 several other improvements of the more elementary forms 
 have necessarily been omitted ; these will now be taken up 
 as nearly as possible in the order of their introduction. 
 
 In this country the self-adjusting injector was so early 
 perfected, and it so largely superseded the original form, that 
 few other improvements were made until the approach of 
 the expiration of Giffard's patents. In England and France, 
 the form of the injector as manufactured by Sharp, Stewart 
 & Co., and H. Flaud & Cie., was modified in various ways; 
 Stewart, Robinson and Gresham, brought out several meri- 
 torious inventions. Among the most important of these 
 were the adjustable combining tube, which has already been 
 described, and the separation of the lifting apparatus from 
 the injector proper; in the latter case an ejector, whose only 
 function was to raise the feed water, was connected with the 
 overflow aperture of the combining tube, and by its aid the 
 injector could be started with great facility. In 1863 the 
 same firm patented an arrangement by which the injector 
 was placed below the level of the water in the tank of a 
 locomotive ; the waste pipe from the overflow extended up 
 into the cab above the water level, and any waste from the 
 injector could be readily seen and corrected; further, the 
 drip pipe was connected with the suction, and loss of feed 
 water from any cause prevented. 
 
 In France, the few improvements that were added, were 
 chiefly in the line of English models. Turck followed 
 Gresham in the use of a stationary steam nozzle, and in ad- 
 dition applied an enveloping nozzle, enclosing an air space, 
 in order to prevent transmission of heat from the steam to 
 the surrounding feed water. Cuau brought out an injector 
 in which all the tubes were fixed, and so proportioned as to 
 permit the use of water ot very high temperature ; no lifting 
 spindle was provided, but the feed water was raised by 
 
DEVELOPMENT OF THE PRINCIPLE. 21 
 
 an auxiliary ejector. Bouvret, Polonceau, and Delpeche, 
 followed with comparatively unimportant modifications, 
 although all have lent their name to styles used upon rail- 
 roads in France. 
 
 As no patent was allowed Giffard in Austria and Ger- 
 many, his injector was manufactured by unlicensed parties 
 soon after its introduction in France, and much original in- 
 vestigation was started. Schau discovered the advantage 
 of the divergent steam nozzle in 1869, by which the expan- 
 sion of the steam is utilized to better advantage than with 
 the convergent shape. Haswell, Korting, Pradel, and 
 Krauss, introduced special changes, many of which are still 
 in service. Korting, whose invention of the double jet in- 
 jector has already had special consideration, was granted a 
 patent in 1872 for drawing exhaust steam from the cylinder 
 into the injector, by which the temperature of the water 
 delivered to the boiler was considerably raised ; an opening 
 was made in the combining tube at a point where a strong 
 suction was produced by the condensation of the motive jet : 
 the inventor states in his specifications that the injector 
 could be used with live steam alone, and that the supple- 
 mentary overflow enabled it to start very easily. Friedman 
 in 1879, patented an improved fixed nozzle injector, and in 
 the form in which he placed it upon the market, used a sup- 
 plementary opening in the combining tube for the entrance 
 of an additional supply of feed water ; this system possessed 
 a number of advantages, and after undergoing subsequent 
 modification is now extensively used both in Germany and 
 the United States. 
 
 In England in 1877, Hamer, Metcalf & Davies, obtained 
 a patent for an injector in which there was a very large 
 steam nozzle, and a combining tube with a hinged section 
 extending nearly its full length ; these modifications were 
 made for the purpose of using exhaust steam for the actua- 
 ting jet, and the injector was enabled to feed boilers carrying 
 as high as 80 Ibs. pressure ; the tubes were well designed 
 and showed much ingenuity in their construction. 
 
 Returning to American improvements we find Garner C- 
 
22 
 
 THE GIFFARD INJECTOR. 
 
 Williams of - Ellenville, N. Y., made in 1880 application for 
 a patent for the first self-starting injector ; * it was very ingen- 
 iously constructed, but seems never to have appeared upon 
 the market. John Loftus, of Albany, N. Y., followed in 
 r88i with a simpler device, shown in Fig. 5, which differed 
 from all previous forms by the insertion of a cylindrical suc- 
 tion or draft tube b' ', between the steam nozzle and the com- 
 bining tube. This obviated the necessity for a special 
 priming device, as th opening of an ordinary globe valve 
 in the steam pipe would produce a vacuum in the water 
 branch sufficient for all ordinary lifts. Check valves h ti 
 were placed on the separate overflow chambers D, D' , so 
 that the condition of the jet while crossing the starting 
 
 overflow d, was independent of, and uninfluenced by, its con- 
 dition at the upper or supplemental overflow d'. Bancroft, 
 Gresham, Penberthy, Desmond, Derby and others, altered 
 and improved the arrangement of the valves and tubes of the 
 re-starting injector and increased the efficiency of its per- 
 formance. 
 
 In 1885, J. Sellers Bancroft of the firm of Wm. Sellers & 
 Co., of Philadelphia, patented a modification of the double 
 jet principle, in which the re-starting feature was introduced ; 
 starting and supplementary overflows were so placed in 
 the combining tubes of the forcing and the lifting sets, 
 that free discharge from the steam nozzles was obtained, 
 producing a strong vacuum in the suction pipe. This form 
 
 * Subsequent research shows that a re-starting injector was patented in France by 
 Vabe, Nov. 12, 1878. 
 
DEVELOPMENT OF THE PRINCIPLE. 23 
 
 of injector was not placed upon the market, as further ex- 
 periment by the same firm developed the arrangement shown 
 in Fig. 6, invented by the author in 1887. This injector 
 consists of a double set of tubes arranged in axial line, 
 giving continuous acceleration to the jet from the moment 
 the water enters the draft tube. A, B, C, are the steam, 
 feed, and boiler connections; a', an annular lifting steam 
 nozzle discharging through the annular draft tube b ', and a 
 the forcing steam nozzle from which the jet receives the final 
 impulse sufficient to enter the boiler. The overflow open- 
 ings d, d' ', d" , are all contained in a chamber Z>, communi- 
 cating with the large waste pipe D' only by the opening of 
 the waste valve h : these overflow apertures are of such pro- 
 portion and so distributed that free discharge is obtained for 
 
 the lifting steam nozzle a'. The resultant vacuum in the 
 feed pipe always tend to raise the water to the injector after 
 a temporary interruption of the water supply and restore 
 the continuity of the jet, while the arrangement of the tubes 
 and steam nozzles permit wide variation of steam pressures 
 without waste at the overflow. 
 
 In 1890 another form of locomotive injector was invented 
 by Holden and Brooks, of Salford, England, who had 
 already secured several patents for fixed nozzle exhaust 
 steam injectors ; this device combined the use of exhaust 
 and live steam, with the tubes arranged as in the double jet, 
 and so contrived for automatic action, that opening valves 
 in the feed and exhaust pipe would start the injector even 
 against the high pressures carried in locomotive boilers. 
 
24 THE GIFFARD INJECTOR. 
 
 There have been numerous other changes made upon the 
 invention of Giffard, but those that have been described, 
 represent steps in the development of the injector; they 
 have been selected from over five hundred patents granted 
 by the United States, and as many more by foreign coun- 
 tries. Many changes may seem to be merely matters of de- 
 tail, yet it has only been by continued experiment with 
 modification of minor parts, or slight changes in the pro- 
 portions or arrangement of the tubes, that progress has been 
 made. The original invention was conceived upon theoreti- 
 cal principles, the correctness of which has been so signally 
 demonstrated, but since that time the greater part of the 
 work of development has been attained only by close atten- 
 tion combined with critical observation in experimental 
 research. 
 
CHAPTER III. 
 
 DEFINITION OF TERMS DESCRIPTION OF THE IMPORTANT 
 PARTS OF THE INJECTOR THEIR FUNCTIONS. 
 
 BRIEF descriptions have already been given of the differ- 
 ent types of injectors in connection with the history of the 
 development, but as the technical names of the various parts 
 will necessarily enter into the accounts of experiments that 
 follow, the definition will now be given as concisely as 
 possible. 
 
 It is unfortunate that there has been so little accord in the 
 direction of injector nomenclature; patent specifications are 
 sometimes obscure and misleading, and the terms used often 
 confusing ; even the names ejector and injector are often 
 appropriated to entirely wrong uses. The two apparatus 
 differ in principle as well as 1 in essential detail : the former 
 can operate with any gas or any liquid in conjunction with 
 any other gas or liquid, while the action of the injector de- 
 pends upon the expansion and condensation of a gaseous 
 fluid acting within clearly defined limits. Technically, 
 
 An Injector is an apparatus in which a gaseous jet 
 impinges and is condensed by a fluid mass whose final 
 kinetic energy exceeds that of a jet of similar form and 
 density discharging under the initial pressure of the 
 motive jet. 
 
 In addition to the use of condensible gases, the existence 
 is required of certain well established relations between the 
 areas of the discharging and the receiving nozzles, which 
 will serve for a clear distinction between the two types of 
 apparatus : 
 
 An Injector is a jet apparatus in which the cross- 
 
 25 
 
26 THE GIFFARD INJECTOR. 
 
 section of the discharge nozzle of the actuating jet is 
 greater than that of the receiving or delivery tube. 
 
 An Ejector, reverses these conditions, and the gaseous 
 or liquid discharge will freely pass through the delivery 
 tube. 
 
 Drawings or cuts of injectors are usually made with the 
 delivery tube discharging toward the bottom or right hand 
 side of the page, and the terms " upper" and "lower " will 
 be thus used. The index letters refer to the various figures, 
 where the same letter always is used on the same tube. 
 
 The Delivery Tube (c] is that tube in which the maxi- 
 mum velocity of the combined mixture of water and 
 steam is attained, and subsequently reduced, by means 
 of the expanding curves and increasing cross- section, to 
 the velocity and pressure in the boiler pipe. 
 It is usual to indicate the nominal size of the injector by 
 the mininum diameter of this tube, as the amount of water 
 delivered is chiefly dependent upon this dimension. 
 
 The Combining Tube, () extends from the upper end 
 of the delivery tube (c), to the lower end of the steam 
 nozzle (a), and receives its name from the combin- 
 ing of the steam with the water that occurs within its 
 walls. 
 
 The Draft, or Suction Tube, (') facilitates the start- 
 ing of the injector, and lies between the upper over- 
 flow (d') and the lower end of the steam nozzle (a) ; its 
 lower diameter is usually larger than the minimum dia- 
 meter of the steam nozzle. 
 
 The Steam Nozzle (a), guides and directs the motion 
 of the actuating jet ; its effective area of discharge is 
 often varied by the use of a Steam Spindle (/), in order 
 to vary the amount of steam used, or for the purpose of 
 lifting the feed water. 
 
 These tubes can be used in either single or double com- 
 bination as in the single or double jet injectors. 
 
 Overflow, (d~} Primary or Lower, a narrow annular 
 vent space or drilled aperture, placed above the mini- 
 mum diameter of the delivery tube, permitting free 
 
DEFINITION OF TERMS. 27 
 
 outlet for the water and steam during the operation of 
 starting. 
 
 Upper, or Supplemental (af), is placed nearest to the 
 steam nozzle, and often made of sufficient area to per- 
 mit free exit for the full discharge of the steam nozzle. 
 Additional overflows are frequently used, see Fig. 6, page 
 
 2 Z (d" d'"). 
 
 The following terms are used as descriptive of different 
 classes of injectors : 
 
 Single Jet Injector, one in which a single set of com- 
 bining and delivery tubes is used. 
 
 Double Jet. An injector containing two sets of steam 
 jet apparatus, of which the first or lifting set receives 
 the feed water from the source of supply and delivers it . 
 to the second, or forcing set, from which it receives suf- 
 ficient impulse to enter the boiler. 
 
 Automatic or Re-Starting. An injector that is able 
 to re-establish automatically the continuity of the jet, 
 after a temporary interruption in the steam or water 
 supply. 
 
 Self-adjusting. Applied to an injector in which the 
 supply of water is automatically adjusted to suit the 
 steam supply without waste at the overflow. 
 
 Open overflow injector, has one or more apertures in 
 the combining tube, opening into one or more overflow 
 chambers, that may be closed against the admission of 
 air by the use of light check valves opening outward. 
 
 Closed overflow injector, can only start by means of 
 an opening or vent placed beyond the delivery tube, 
 which must be closed in order to divert the jet into the 
 boiler. 
 The following terms relate to the performance : 
 
 Maximum capacity. The greatest volume or weight 
 of feed water passing through the delivery tube at any 
 given steam pressure and condition of feed. It is 
 usually measured in cubic feet or pounds per hour ; the 
 use of the gallon is not to be advised, unless its value 
 is clearly stated. 
 
28 THE GIFFARD INJECTOR. 
 
 Minimum capacity. The least volume or weight of 
 water as above, that can be continuously delivered 
 against boiler pressure without waste from the overflow. 
 It is often expressed as a percentage of the maximum 
 capacity. 
 
 Range of capacities. The difference between the 
 maximum and minimum capacities expressed in terms 
 of the maximum capacity ; for example, if the max. is 
 300 cu. ft., and the min. is 200 cu. ft., the difference, 
 100, is the range, which is 33 per cent. 
 
 Overflowing temperature. Highest admissible tem- 
 perature of feed water with which the injector can oper- 
 ate without wasting, when running against boiler 
 pressure. 
 
 Overflowing pressure. Under given conditions, the 
 highest counter pressure against which the injector can 
 run without wasting. 
 
 Breaking temperature, Breaking pressure. Highest 
 admissible feed temperature, and counter pressure, when 
 the waste valve is closed. Note. With closed overflow 
 injectors, the overflowing and breaking temperatures 
 and pressures are in most cases the same. 
 
 Efficiency. This may be based upon the ratio which 
 the total heat in the feed water and in the steam, bears 
 to the heat in the delivered water plus the heat equiva- 
 lent to the work of forcing the water into the boiler. 
 This maybe called the * ' Thermodynamic Efficiency." 
 Or, the ratio of the work performed by the steam in forc- 
 ing the water into the boiler, to the total energy given 
 out by the steam during its expansion from the initial 
 boiler pressure to the pressure in the combining tube. 
 This is the " Mechanical Efficiency." Or, it may be ex- 
 pressed in terms of the weight of water delivered per unit 
 weight of steam ; this is a very simple and convenient 
 methed of comparison, and is thoroughly practical. 
 Note . The efficiency is often measured by the volume or 
 weight of water delivered per unit area of cross-section of 
 the steam nozzle or the delivery tube. 
 
DEFINITION OF TERMS. 29 
 
 The description of the tubes of the injector will be taken 
 up in the following order, Delivery tube, Combining tube, 
 and Steam Nozzle, the theory of their action reviewed, and 
 practical questions considered that have important influence 
 upon their design. 
 
CHAPTER IV. 
 THE DEUVERY TUBE. 
 
 DESCRIPTION I EFFICIENCY OF VARIOUS TYPES EFFECT OF DIFFER- 
 ENT SHAPES AND PROPORTIONS- 
 
 THE function of this tube is to change the kinetic energy 
 of the jet to potential, with the least possible loss : or, to 
 transfer the energy due to the velocity of the water and con- 
 densed steam, into pressure in the boiler pipe. 
 
 Of all its dimensions, that of the minimum diameter is the 
 most important, for it is from this base that the dimensions 
 of all the other parts are calculated. The use of this di- 
 ameter to denote the size of the instrument was first sug- 
 gested by Giffard, and this seems to be the most rational 
 method that can be used. The quantity of water delivered 
 by an injector is directly dependent upon the minimum area 
 of this tube, and under the same conditions, varies with the 
 square of the diameter ; theoretically, the capacity should 
 be equal to that quantity of water which would be dis- 
 charged from a similar orifice under a pressure or head equal 
 to the overflowing pressure of the jet, which is always in 
 excess of the pressure carried in the boiler. If there were 
 no losses, this would give an exact method of determining 
 the weight of water delivered, and as it is, the losses referred 
 to, can be closely approximated ; by comparing the calcula- 
 ted discharge with the actual test of the injector, a per- 
 centage of efficiency on this basis can be readily determined. 
 
 This basis of efficiency has, in the cases of all single jet 
 
 injectors, a decided value; the smaller the delivery tube, 
 
 the smaller can be the other orifices of the injector, reducing 
 
 the weight of steam used, and rendering its operation more 
 
 30 
 
THE DELIVERY TUBE. 31 
 
 economical. As an example, take a case from actual prac- 
 tice, from which the efficiency will be calculated. 
 
 A No. 8 Injector, having a delivery tube 8 millimetres 
 in diameter, commences to waste against 160 pounds 
 (corrected) back pressure ; the delivery temperature is 
 148 deg. Fahr. 
 
 The height of a column of water at this temperature that 
 corresponds to the pressure of one pound per square inch, is 
 2.354ft., (see Table II., page 43) and the head of water 
 equal to 160 Ibs is 2.354 X 160 = 376.64 ft. This gives a 
 velocity, 
 
 if = \^2gh 8.024 V/j = 8.024 V/376. 64 = 155.721 ft. per sec. 
 
 The area of the delivery tube in square feet is 0.000541068,' 
 and the volume that would be discharged from this tube 
 under this head 
 
 J 55-7 2 X 0.000541068 x 3600 = 33-3 T 97 cu. ft. per hour, 
 
 which must be equal to the theorectical quantity of water 
 entering against that pressure. But an actual test showed 
 that 251.5 cubic feet had been taken from the feed tank, and 
 that the weight of dry steam used was 1304.5 pounds ; there- 
 fore the total volume passing through the delivery tube per 
 hour was 
 
 The density of the jet was 
 
 and the actual velocity of the jet 
 
 155-721 
 
 The discharge per hour per square millimetre of cross- 
 section 
 
 2728 
 
 -- =5427 cubic feet. 
 50.2656 
 
 This gives a very simple means of comparing different 
 forms of injectors, and the following table is based upon 
 
32 
 
 THE GIFFARD INJECTOR. 
 
 actual tests of single jet locomotive injectors at a steam 
 pressure of 1 20 pounds to the square inch ; no Double Jet 
 Injectors are included, as the conditions under which the 
 jet passes through the delivery tube of the two types are not 
 exactly comparable : 
 
 TABLE I. 
 
 NAME. 
 
 Size. 
 
 Garfield No. 7 
 
 Mack No. 7 
 
 Monitor No. 8 
 
 Metropolitan No. 9 
 
 Original Giffard No. 8 
 
 Sellers' 1876 No. 7 
 
 Diameter 
 Delivery 
 Tube in 
 Milli- 
 metres. 
 
 Cu. Ft. per 
 hour per 
 sq. m. m. 
 
 7.26 
 
 4.8876 
 
 8.13 
 
 3.8006 
 
 8.28 
 
 5.1258 
 
 1075 
 
 3-8555 
 
 8.00 
 
 4.4H2 
 
 7.00 
 
 5.6745 
 
 TT'-Sf&i 
 
 i ' 
 
 \ 
 
 1 
 
 \ 
 
 1 
 
 i 
 
 1 
 
 
 1 
 
 
 
 The principal function of the delivery tube has already 
 been stated, but it remains to show the means by which the 
 required effect is produced. A section of a tube is shown at 
 AB in Fig. 7, where, for the sake of clearness, imagine the 
 entering jet divided into a series of short cylinders, of the 
 
THE DELIVERY TUBE. 33 
 
 same density as water ; also, that the injector is working 
 against a pressure that compels the tube to be completely 
 filled with water ; now, if the tube were cut off at the section 
 D, the jet would impinge violently against the wall of water 
 in front, and lose its energy in forming eddies and disturb- 
 ances, (see Fig. 8), in proportion as its velocity is greater 
 than that of the water ahead. But as the jet passes into the 
 tube, the cross section becomes wider, and the cylindrical 
 form changes to conical. If the volume contained between 
 any two sections as Dd f , or d' d", etc. , be the same, the dis- 
 tances a, b, r, gradually shorten as lateral motion toward the 
 walls is induced; as these distances shorten an increasing 
 pressure is exerted upon the surrounding walls and upon the 
 particles of water directly in front. In this way the momen- 
 tum of the jet, which is at the maximum at D, as indica- 
 ted by the highest point of the velocity curve, is gradually 
 reduced to that of the feed in the boiler pipe. By refer- 
 ence to the figure the pressure and velocity of the jet at 
 any point can be determined by the height of the corres- 
 ponding curve above the base line. If the density of the 
 jet be uniform, its velocity at any point of a given tube 
 can be obtained by dividing the actual volume passing 
 through, expressed in cubic feet, by the area of the tube in 
 square feet, and the side pressure upon the walls may be 
 calculated by the following formula : 
 
 144 
 
 Where^, pressure required. 
 
 P pressure in boiler pipe at end of the delivery tube. 
 v = velocity of water in boiler pipe. 
 z/ x = velocity at point where pressure = p v 
 7 = weight of one cubic foot of water at the temperature 
 of delivery. 
 
 It is of course advantageous to shape the tube so that all 
 the energy of the jet can be utilized, but this can only be 
 done by applying the laws governing the motion of fluids ; 
 the effects produced by improperly shaped tubes may be 
 seen by the following tests, where the same conditions ob- 
 3 
 
34 
 
 THE GIFFARD INJECTOR. 
 
 tained throughout, both as to steam pressure and to the 
 volume of water passing through the injector. The first 
 form experimented with is shown in Fig. 8, which corres- 
 ponds to the mouth or entrance of a delivery tube, as if the 
 part to the right of the section D, in Fig. 7 had been taken 
 entirely away; the result of this was that the injector would 
 
 only force against a pressure of 35 pounds without overflow- 
 ing, although the pressure carried in the boiler was 65 pounds 
 to the square inch, showing the enormous proportion of 
 energy dissipated. 
 
 The next tube was made cylindrical, as is indicated in 
 Fig. 9. Under the same conditions this tube showed no im- 
 provement, as only 25 pounds was reached; it was then 
 
 ffi'gJO. 
 
 reamed in the form of a divergent curve approximating in 
 section a parabola as shown in Fig. 10, and the effect pf the 
 change was at once apparent as the permissible back pres- 
 sure rose to 62 pounds ; although the general shape was 
 improved, yet the tube was obviously too short and the 
 curvature too great for the high velocity at which the jet 
 
THE DELIVERY TUBE. 
 
 35 
 
 was moving, and the tube in Fig. n was substituted. Its 
 length was 7.6 times the diameter, and developed a pressure 
 of 88 pounds. Further change in the proportion raised the 
 pressure against which the jet would work to 93 pounds, 
 without any modification of any of the other parts of the 
 injector, or increase in the pressure of the steam. 
 
 It has long been known that the divergent tube possessed 
 the peculiar property of increasing the quantity of water 
 that would flow through an orifice in a given time, and this 
 phenomenon has led to careful experimental work by Ber- 
 nouilli, Francis, Brownlee, and others. Francis in his well- 
 known and oft-quoted Lowell Hydraulic Experiments, found 
 that with an orifice having well rounded curves of approach, 
 the weight of water discharged under a constant head of 
 1.36 feet, could be increased 2.44 times by the addition 
 
 of a divergent funnel having an angle of 5 i'. Brownlee, 
 with 6 feet head of water increased the discharge 2.42 
 times by the use of a tube whose included angle was 7 5'. 
 But the heads of water and the velocities of the jets in 
 both these cases were exceedingly small compared with those 
 in use in the injector, and the changing conditions in the 
 internal action of the jet would invalidate any positive pre- 
 diction regarding the effect of a special shape of divergent 
 tube ; yet the theory under which such tubes may be de- 
 signed is interesting, and may be applied in a limited degree 
 to the case under consideration, under the assumption that 
 the jet has the same density as that of water ; this is how- 
 ever only approached in those types of injectors which are 
 most carefully designed and constructed. 
 
36 THE GIFFARD INJECTOR. 
 
 In order to obtain the most efficient form of divergent 
 tube, it has been suggested by Nagle that it be so con- 
 structed that the retardation of the motion of the jet be 
 made uniform, and that by this means the particles of water 
 could be kept in equilibrium and internal eddies avoided. 
 This would require each succeeding section of the delivery 
 tube to be increased in such proportion that the difference 
 between the squares of the velocities of any two equidistant 
 sections would be constant; and the particles of water 
 always in contact with its walls. It should be remembered 
 that the velocity of the entering jet must be always equal to 
 that of a jet of similar density discharging from the tube 
 under a head equal to the difference between the maximum 
 pressure against which the jet is capable of working, and 
 the internal pressure of the jet at the time of entrance ; it 
 follows therefore that the absolute pressure of the jet at 
 this point would influence the amount of water passing 
 through the tube, if the construction of the other parts of 
 the apparatus were such as to permit an additional quantity 
 of water to be drawn from the supply. 
 
 The reason for the lowering of the pressure at the point 
 of minimum diameter may thus be explained : a jet of water 
 discharging freely into air, will retain its full velocity and 
 the same cross-secton for some distance beyond the mouth 
 of the tube ; but by reason of the gradually expanding 
 curves of an enveloping tube, the jet adheres to the surroun- 
 ing walls and its section is increased ; but the energy tends 
 to remain the same as before, and therefore an effort is ex- 
 erted to draw the preceding particles forward, increasing 
 their velocity, and consequently the weight of water dis- 
 charged. If the tube is open to the atmosphere, a lack of 
 equilibrium results as air is drawn in at the lower end, but 
 if immersed, the effect reacts upon the particles in the rear, 
 and, if the tube is correctly proportioned, a perfect vacuum 
 will be formed at the throat of the tube. 
 
 The formulas for uniform retardation are similar to those 
 for uniform acceleration, and may thus be expressed : 
 
 Fbe the velocity of the jet and D the diameter of the 
 
THE DELIVERY TUBE. 
 
 37 
 
 tube at the throat ; let v and d have the same values at the 
 lower end of the tube ; then 
 
 V EP 
 
 V : v : : d* : D 2 and v = .-?- 
 a 2 
 
 (2) 
 
 If S is the length to the tube, the negative acceleration, or 
 retardation is 
 
 or 
 
 If the lower end of the tube be very large the last term of 
 (3) becomes so small that it may be neglected without sensi- 
 ble error, and may be reduced to 
 
 The equation of the curve of the half section of a tube 
 that would fulfill these conditions may thus be found ; in 
 Fig. 1 2 let A B be the half section of a delivery tube whose 
 centre line is Y Y\ vertical ordinates denoted by x, and 
 abcissas by y parallel to Y Y\ substituting 2 x for d, and y 
 for S, in (4) and altering its form, 
 
 By way of example, take a delivery tube 8 millimetres in 
 diameter and the length, S 1 taken as 20. (This means 20 
 spaces of equal length and each space may be any desired 
 
38 THE GIFFARD INJECTOR. 
 
 distance although it is advantageous to make the total length 
 as great as possible.) The entrance velocity of the jet may 
 be assumed as 156 feet per second, and the final velocity, z/, 
 as o. Substituting in (4) gives the retardation, and (5) re- 
 duces to __ 
 
 * = 4y nro^y 
 
 froni which the diameter can be determined for any part of 
 the tube ; for the following values of _y, d has been calcula- 
 ted, and it will be noticed that the tube widens very slowly 
 at first, but as the lower end is approached the curve becomes 
 very steep and finally tangent at B. 
 
 Forj = o d = 8.000 
 
 = 5 = 8.590 
 
 = 10 = 9.512 
 
 = 15 = 10.912 
 
 = 20 00 
 
 Looking at these formulas from a practical point of view, it 
 will be seen that they depend upon the value of V, the entrance 
 velocity, which in the case of the injector must vary with 
 every steam pressure and change in the density of the jet, 
 although it is at the maximum capacity, where the density 
 most nearly approaches unity, that the full efficiency is 
 needed. Giffard realized the advantage of using differently 
 shaped tubes for high and low pressure, and advised the use 
 of circular arcs whose radii in the former case were 300 times 
 the diameter of the tube, and in the latter, 200 times. This 
 altered the proportion of the tube and the mean angle of 
 divergence of the tube for each size instrument, but this 
 seems for many reasons to be theoretically correct, and an 
 improvement upon the straight taper adopted by most manu- 
 facturers. 
 
 The change in the velocity of the jet and the gradually 
 increasing pressure upon the walls of the tube was clearly 
 shown for the theoretical case assumed in Fig. 7 ; as an illus- 
 tration of the actual changes of pressure occuring within 
 
THE DELIVERY TUBE. 39 
 
 the tube under different conditions, the diagram shown in 
 Fig. 13, is presented through the courtesy of Wm. Sellers 
 & Co., of Philadelphia, who have, ever since their introduc- 
 tion of the injector into the United States, made experi- 
 mental work an important part of their system. These tests 
 were made with a Self- Acting Injector of 1887, at 65, 100, 
 and 120 pounds pressures. The delivery tube was pierced 
 at four points, A, B, C, D, by small drilled holes one-six- 
 teenth irfch in diameter, upon which gauges w r ere placed to 
 indicate the internal pressure of the jet. Under a head of 
 30 feet, water was allowed to discharge through the tube 
 and the vacuum at the throat of the tube, as indicated by 
 a U mercury gauge, was found to be 27^ inches; as meas- 
 ured by actual test under a head of i8J^ feet, the amount 
 of water passing through was increased a little over 42 per 
 cent, above that discharged by a similar orifice without 
 the divergent tube. The pipes were then changed so that 
 the feed water could be lifted from a tank placed below the 
 injector, and the steam turned on; the pressures observed 
 were plotted on the vertical lines above the corresponding 
 point of the tube and connected by the curved lines and 
 marked with the initial steam pressure. The rise in pres- 
 sure w r hen the injector is running at its maximum capacity 
 is indicated by the full line, the minimum by the dotted 
 lines, and a capacity approximately half-way between, by the 
 alternate dot and dash. 
 
 It will be noticed that the curves for the maximum capa- 
 city rise easily from the low pressure at the throat of the 
 tube to the final boiler pressure at the end, distributing the 
 wear very evenly. At the mean capacity, the point where 
 the greatest abrasion will occur is evidently at the section 
 marked C, where the pressure line rises suddenly from 10" 
 vacuum to 95 pounds ; at other steam pressures the same 
 peculiarity was noticed, but the lines were not added to the 
 diagram as they would have sacrificed its clearness. At the 
 minimum capacity, the w r ear is chiefly between A and B, as 
 ma} r be seen by reference to the dotted curves. These facts 
 constitute a strong argument in favor of attaining the per- 
 
40 
 
 THE GIFFARD INJECTOR. 
 
THE DELIVERY TUBE. 41 
 
 feet condensation of the steam within the combining tube and 
 before the entrance to the delivery tube is reached, and one of 
 the reasons for basing the efficiency of an injector upon the 
 amount of water delivered per unit of area of delivery tube. 
 
 The disadvantages of the conical tube may be seen from 
 an examination of the light full line, marked 120 pounds; 
 the abrupt rise in pressure against the walls of the tube, 
 even on the maximum capacity and when all conditions are 
 the same as in the other experiments, shows the inferiority 
 of this form of tube. 
 
 The various causes of loss of energy of the jet can be 
 closely approximated in the most important cases that occur 
 in practice. That due to the final velocity in the delivery 
 tube, and impact of the water against the slowly moving 
 water in the delivery pipe, can be found from the expression, 
 
 
 
 Where p the loss in back pressure in pounds. 
 z/j = the velocity in the boiler pipe. 
 d l = the diameter of the boiler pipe. 
 d =- the diameter of the lower end of the delivery tube. 
 7 = the weight of a cubic foot of water at the temperature of 
 the delivery. 
 
 The loss in head due to the friction of the jet upon the 
 walls of the tube, can be calculated from the formula for 
 conical pipes ; or, if curved, by using the nearest angle or 
 angles to correspond. This loss amounts to very little, com- 
 pared with the pressures carried, and this equation is much 
 simpler than that for tubes formed from circular or para- 
 bolic arcs. In Fig. 14 let 
 
 p = loss in back pressure due to friction. 
 
 V = velocity of jet at the throat of tube. 
 
 D = diameter of the throat. 
 
 d = diameter of lower end of tube. 
 
 6 = included angle of tube. 
 
 ft = co-efficient of friction, depending upon condition of tube. 
 
 y = weight of one cubic foot of water, as before. 
 
 ....... < 
 
42 THE GIFFARD INJECTOR. 
 
 To apply these two formulas to an actual case, take a de- 
 livery tube 0.3" at the throat, 3.6" long, and a taper of i in 
 12 in diameter; the included angle is 4 46". Take Fas 
 156 feet per second, and /?, as 0.0147 ; d is found to be 0.6" ; 
 substituting, 
 
 _ i |- 
 
 whence, 
 
 P = 3-5 pounds 
 and from formula (7), 
 
 i 
 
 - = 9-8 Ibs. 
 2-354 V36 / 64.4 
 
 This loss could be entirely obviated by widening the end 
 of the tube in easy curves to the diameter of the boiler pipe. 
 
 The co-efficient, , will depend upon the condition of the 
 walls of the tube, whether smoothly reamed and offering 
 but little resistance to the motion of the jet, or whether 
 abraded by constant use or cut in circular grooves by poor 
 workmanship. When it is remembered that the velocity of 
 the jet at the smallest diameter of the tube corresponds to 
 that attained by a jet of the same density issuing under a 
 pressure even higher than that in the boiler, the disadvan- 
 tageous effect produced by the roughened surfaces of the 
 guiding tube can be easily realized. 
 
 This leads naturally to the subject of the wear of the tube. 
 From what has been said regarding the distribution of inter- 
 nal pressures, the points where the greatest wear will occur 
 under usual conditions, can be easily seen. The presence of 
 grit or dirt in the feed water will cut the mouth of the tube, 
 
THE DELIVERY TUBE. 
 
 43 
 
 and enlarge the minimum diameter, acting, by reason of the 
 high velocity of the jet, like a continuous grinding cylinder ; 
 but it is further down in the tube that the abrading effect is 
 first noticed, near the place marked C in Fig. 13, where an 
 annular groove is often worn around the tube, that seriously 
 impedes the motion of the water ; its location is always at 
 that point where the pressure suddenly rises ; as the cross- 
 section gradually widens, the velocity of the water decreases, 
 and the wear is less noticeable. Regarding the question of 
 repair, it may be said that this tube will generally be the 
 first that requires replacing, but it may often be saved by 
 carefully reaming or smoothing out the roughened places ; if 
 this be done, the minimum diameter may be somewhat in- 
 creased before the injector will cease to work at the higher 
 steam pressures. Its power at low steam will be affected 
 first by the wearing of the tube at this point, but the exercise 
 of a little judicious management will often prolong the life 
 of the injector, and save needless substitution of new parts. 
 The following table of the weight of a cubic foot of water 
 at different temperatures, and the head of water in feet cor- 
 responding to a pressure of one pound per square inch, will 
 materially assist the calculation of the performance of the 
 injector under different conditions. It has been compiled 
 from a "Table of Comparative Volumes," prepared by Mr. 
 A. F. Nagle, and published in Proc. Am. Soc. Mech. Eng.; 
 the heads of water are based upon his figures. 
 
 TABLE II. 
 
 
 1 
 
 
 
 Temp, ! TiTp-vht f Head in feet Temp. ! w-j-t,!. f 
 
 Head in feet 
 
 Water, deg., I.wf^kA = to i Ib. per Water, deff. _ ." JL"! != to i Ib per 
 
 Fahr. 
 
 square inch. 
 
 | Fahr ' | 
 
 square inch. 
 
 39-1 
 
 62.4250 
 
 2.3067 
 
 J 3 
 
 61.5320 
 
 2.3402 
 
 40 
 
 62.42398 
 
 2.3068 
 
 140 
 
 61.3432 
 
 2-3474 
 
 50 
 
 62.40735 
 
 2.3074 
 
 150 
 
 61.1413 
 
 2.355 2 
 
 60 
 
 62.36975 
 
 2.3088 
 
 160 
 
 60.9266 
 
 2-3635 
 
 70 
 
 62.31015 
 
 2.3IIO 
 
 170 
 
 60.6988 
 
 2.3723 
 
 80 
 
 62.2283 
 
 2.3140 
 
 180 
 
 60.4608 
 
 2.3818 
 
 90 
 
 62.1253 
 
 2.3179 
 
 190 
 
 60.2128 
 
 2-39*5 
 
 IOO 
 
 62.0033 
 
 2.3224 
 
 200 
 
 59.9569 
 
 2.4017 
 
 no 
 
 61.8626 
 
 2.3277 
 
 210 
 
 59.6935 
 
 2.4123 
 
 120 
 
 61.7053 
 
 2.3336 
 
 212 
 
 59.6400 
 
 2.4144 
 
CHAPTER V. 
 
 THE COMBINING TUBE. 
 
 IN describing the action of the jet within the delivery 
 tube, certain theories were given which seemed to corre- 
 spond closely to the most important conditions occurring in 
 actual practice ; unfortunately the action of the steam in the 
 combining tube is not so definitely understood, and the most 
 that can be done is to describe the phenomena as observed, 
 and deduce a few general conclusions. 
 
 It is probable that more experimental work has been re- 
 quired to perfect this tube than any other part of the injec- 
 tor, and each investigator has adopted a special form to suit 
 pre- determined conditions, that seem to him to be most effec- 
 tive. The best shape can be decided by experiment only, 
 and therefore the special conditions under which the tests 
 are made govern the result; it follows, therefore, that there 
 may be as many different forms as there are manufacturers, 
 and to a great extent this is true, each type of injector oper- 
 ating more or less successfully under a certain range of con- 
 ditions. 
 
 The first requisite condition that the tube must fill, is that 
 the water must be sustained during the impact of the steam, 
 and the second, that the mixture of the water and the steam 
 be made as intimate as possible in order that complete con- 
 densation may take place during the passage of the jet 
 through this tube ; this can only be done by using correct 
 proportions at the upper end, and then giving to the lower or 
 convergent part the same shape that the jet would assume 
 during the process of condensation. 
 
 It is this tube, in great measure, that governs the mechani- 
 44 
 
THE COMBINING TUBE. 45 
 
 cal efficiency of the injector. Of course each tube has its 
 own function, yet all are inter-dependent ; but the process of 
 condensation, that differentiates the injector from other simi- 
 lar apparatus, occurs within the walls of this tube. In the 
 first place, assuming a constant head of feed water, the 
 cross-section of the upper end, by regulating the quantity of 
 water that may enter, determines the proportion of water to 
 steam in the resultant mixture, and the temperature of the 
 delivery. The vacuum in the tube is dependent upon this 
 temperature, and, therefore, the expansion of the steam and 
 its velocity at the moment of impact is also fixed ; further, 
 the water in the suction pipe is drawn into the tube only by 
 the internal vacuum ; the influence of the entrance area ex- 
 tends also to the velocity of the feed water, as it approaches 
 in a thin sheet, the actuating steam. From the laws of im- 
 pact, it is found that the greater the difference between the 
 velocities, the greater will be the loss of actual energy at the 
 moment of impact ; as the w r hole transfer of the mechanical 
 energy of the steam to the water is by impact, it follows 
 that there is great advantage in giving to the entering water 
 the highest possible velocity ; this can only be obtained by 
 maintaining the pressure in the combining tube as low as 
 possible, and reducing the water entrance to a mininum. 
 The lower this pressure the higher will be the velocity 
 of the entering w r ater, and the greater the proportion of 
 water to steam in the mixture. 
 
 Secondly, the convergent taper or curve extending from 
 the end of the steam nozzle to the lower overflow, should, in 
 order to obtain the maximum efficiency from the injector, 
 conform closely to the rate of condensation of the steam, and 
 its length be modified for every variation in the pressure 
 of the steam or the temperature of the feed. The same con- 
 dition obtains to a great extent with the water entrance area, 
 whose influence upon the performance of the injector was 
 detailed in the preceding paragraph. It was appreciation 
 of these facts that led Giffard and the early experimenters to 
 lay so much stress upon the necessity for an adjustable com- 
 bining tube. The advantage of this feature can be better 
 
46 THE GIFFARD INJECTOR. 
 
 understood if we suppose an injector to be working into a 
 boiler under normal and efficient conditions, lifting the feed 
 water one foot and running at its maximum capacity ; if the 
 steam pressure now rise, more water will be required to 
 condense the increased flow of steam and preserve the 
 normal condition of the jet; this can only be effected by in- 
 creasing the head of water or widening the distance between 
 the steam nozzle and the combining tube. A reduction of 
 pressure would evidently require a reverse movement, for 
 the vacuum within the tube remaining constant, too large a 
 quantity of water would enter for the steam to force through 
 the delivery tube, and waste would occur at the overflow. 
 If the steam pressure fell much lower, the jet of steam 
 would not have sufficient power to drive the accumulating 
 mass of water through the lower end of the combining 
 tube, and the continuity of the jet would be lost. If the 
 openings in the tube, i.e., the overflows, were large enough, 
 all the water would pass out through the waste pipe instead 
 of going into the boiler, even though the sound of working 
 be the same as that with which the engineer might be famil- 
 iar; on the other hand, if the overflows are small, the 
 injector will "break," or "fly off," and steam and hot water 
 will be forced down into the suction pipe. 
 
 The following experiments with the " Little Giant Injec- 
 tor ' ' illustrate these conditions as they occur in practice ; a 
 No. 7 injector was started at 90 pounds steam on a lift of one 
 foot, and the combining tube adjusted for the full capacity ; 
 the position of the tube was measured and found to be 15 
 millimetres from the upper end of its stroke. The steam 
 pressure was then raised to 120 pounds and the temperature 
 of the delivery increased from 136 to 165 ; as this value was 
 entirely too high for efficient performance, the tube was moved 
 10 mm. further down and the temperature fell at once to 148. 
 
 Starting once more with 90 pounds steam and delivery at 
 136, the height of the lift of the feed water was increased to 
 6 feet, the result being a diminution in the capacity of the 
 injector and a delivery temperature of 150; a downward 
 movement of the combining tube of 10 mm. , brought the 
 
THE COMBINING TUBE. 47 
 
 capacity up to the standard, and reduced the delivery tem- 
 perature to 136 again; the distance between the tubes 
 under these conditions being the same as was used at 1 20 
 pounds when lifting the feed water i foot. 
 
 The following table gives the distance between the com- 
 bining tube and the steam nozzle at different pressures, both 
 for the maximum and the minimum capacities ; the height 
 of lift is i foot ; it will be noticed how much variation in the 
 position of the tube there is between high and low pressures. 
 
 Distance of Combining tube from upper 
 end of Stroke. 
 
 Steam Pressure. 
 
 Max. Capacity. 
 
 Min. Capacity. 
 
 120 
 
 24 mm. 
 
 3.5 mm. 
 
 90 
 
 15 " 
 
 1.0 " 
 
 60 
 
 7 " 
 
 0.5 " 
 
 30 J -5 " o-5 " 
 
 The wide difference between the position of the combin- 
 ing tube at 30 and at 120 pounds steam show how impossi- 
 ble it is for a single jet injector with fixed nozzles to work 
 thoroughly efficiently over a large range of pressures. The 
 only means by which this is attempted, is by assuming a 
 certain range through which the pressure may fluctuate, 
 and then adjusting the combining tube so that the injector 
 will work into the boiler, without wasting at the overflow 
 and at its maximum capacity, at the lower limit of pressure ; 
 this method sacrifices the best performance of the injector 
 at higher steam pressures, as both the overflowing tem- 
 perature and the capacity will be lower than if the tubes 
 were differently adjusted. In the previous example, if the 
 combining tube is set for a steam pressure of 40 pounds, the 
 adjustment would be correct for the minimum capacity at 
 120 pounds; therefore, as the pressure was increased, the 
 efficiency of the instrument would diminish, and a less pro- 
 portion of water be delivered per pound of steam. 
 
 This demonstrates the superiority of the adjustable or self- 
 adjusting form of injector over the fixed-nozzle type for use 
 in all places where the steam pressure is subject to much 
 fluctuation ; the correct ratio between the weight of the 
 water and the weight of the steam can always be maintained 
 
48 THE GIFFARD INJECTOR. 
 
 in the former case, and the capacity increased with the 
 steam pressure; even though a throttling valve may be 
 placed in the feed pipe, the results obtained where there is 
 considerable range of pressure, as in locomotive service, 
 cannot be as satisfactory as with the other method of regu- 
 lation. 
 
 The form of the combining tube 'as determined by various 
 experimenters differs greatly ; in the tests of the Irwin In- 
 jector by a committee of the Franklin Institute, in 1879, a 
 tube was used whose length was only four times the diame- 
 ter of the delivery tube ; this is a great contrast to that of 
 the exhaust injector, where the ratio is 18 to i ; this cor- 
 responds to the usual practice for high pressure steam, 
 where the increased quantity of heat requires provision for 
 its absorption, and this can be best accomplished by length- 
 ening the tube, and giving better opportunity for intimate 
 mixture between the water and the steam ; with the exhaust 
 injector this is necessary on account of the large volume of 
 steam used, which requires ample time for condensation, for 
 the temperature of the delivery is higher than that of a well 
 designed live steam injector working at 150 pounds. 
 
 The advantages of the adjustable combining tube for vary- 
 ing the water area are further increased if this adjustment be 
 made automatic and effected by the action of the jet itself; 
 that this has been done, was shown in the description of the 
 self-adjusting injector, accompanying Fig. 3 ; in this case the 
 automatic action was described as effected by the influence 
 of the jet passing the overflow space, and it was shown that 
 the combining tube was moved forward by the partial 
 vacuum produced by incomplete condensation of the steam, 
 or backward by the pressure due to excess of water in the 
 feed chamber. It is interesting to note the action of the 
 steam upon the upper end, which is as follows: the dis- 
 charging jet, striking against the film of water on the inside 
 of the tube, tends to impel it forward, while the rapid con- 
 densation of the steam produces a reactive effect which 
 draws the tube backward toward the steam nozzle, and these 
 two pressures almost exactly balance ; but as this action only 
 
THE COMBINING TUBE. 49 
 
 occurs in the central conical part, the rest of the piston head 
 receives the positive or negative pressure in the feed pipe, 
 just as the lower end is acted upon by the pressure in the 
 confined overflow chamber ; this has been found to be true 
 by placing a vacuum gauge upon the overflow, and noting 
 its reading as the height of the lift increases ; it is found 
 that the two readings correspond almost exactly, and that 
 the tube floats between two balancing cushions, ready to 
 respond to any change in the governing conditions. 
 
 In calculating the performance of an injector, it is often 
 desirable to know the vacuum within the combining tube. 
 This may be determined by allowing water under a constant 
 head to flow freely through the tube, special care being 
 taken to see that the feed valve is full open and that the 
 upper overflow is large enough to permit free discharge for 
 all the water that will enter ; the best proof of this is to dis- 
 connect the steam branch and observe if the water rises into 
 the steam nozzle ; if it does, the overflow space is too small 
 to permit a free discharge. If, however, this test is satisfac- 
 tory, weigh the quantity of water flowing through, and then 
 connect the steam pipe and admit steam ; without altering 
 the position of the water valve, regulate the supply of steam 
 until the injector will just run without wasting, and note the 
 new weight of water ; these two values will bear the same 
 relation to each other as the square roots of the heads. 
 
 Here is an actual test of an exhaust injector, taken under 
 rather unfavorable conditions as the feed water was 76 de- 
 grees: under a constant pressure of 4.25 pounds or 9.817 
 feet, 1987.5 pounds of water flowed through the injector; 
 with the steam valve opened and working into boiler, 
 the weight increased to 3887.5 pounds, due to the vacuum 
 in the combining tube Therefore, 
 
 5 : 3887.5 : : ^79.817 : ^37.48 = total head. 
 Subtracting, 
 
 37.48 '9.817 = 27.663 feet vacuum, = 24^" mercury, 
 
 which is the value required. 
 
50 THE GIFFARD INJECTOR. 
 
 The average vacuum within the tube cannot vary much 
 from that corresponding to the mean temperature of the feed 
 and the delivery, yet the pressure at the point where the 
 water enters is much lower ; it is there that the water is the 
 coldest, and the outside edges of the discharging jet obtains 
 its fullest expansion. This will be shown in the diagram 
 showing the discharge from the steam nozzle (see page 54), 
 and described under that heading. It is interesting to de- 
 termine as closely as possible the average pressure within 
 the tube, for it is only by this means that the final expan- 
 sion of the steam can be found, and its terminal velocity 
 calculated. 
 
 So far it has only been assumed that there was a partial 
 vacuum in the tube. As it is obvious that steam is always 
 condensed when coming in contact with a body whose tem- 
 perature is lower than that corresponding to its pressure, so 
 it is possible that steam may be condensed in an injector tube 
 even when the entering water is above 212 degrees. This is 
 most nearly approached in the double jet injector, where 
 the water entering the combining tube of the second set of 
 tubes is often as high as 180 or 190 degrees, when the tem- 
 perature of the supply is 150. It is very probable that in this 
 case only a very small percentage of the steam from the 
 second steam nozzle is condensed, but sufficient to permit 
 the combined jet of steam and water to contract sufficiently 
 to pass through the narrowest section of the delivery tube. 
 The velocity with which the water enters the tube is also 
 mucji increased, for the pressure between the two sets of 
 tubes rises at 1 20 pounds steam to 40 or 45 pounds ; so 
 that the work required of the second steam jet is not very 
 great. 
 
 There is no doubt that if an indefinite amount of time 
 were given for the condensation of the steam, feed water 
 could be taken at any temperature below that of the steam, 
 but it must be remembered that the actual time of contact 
 is only that required to traverse the tube, and that the mix- 
 ing can only be between the two conical exposed surfaces ; 
 therefore, if the difference in temperature be small, the trans- 
 
THE COMBINING TUBE. 51 
 
 mission of heat will be correspondingly slow, and the volume 
 of the steam will not be reduced enough to enter the delivery 
 tube. The effect would be similar to that of an air jet dis- 
 charging into a mass of water a dispersion and atomizing 
 of the whole mass. It is evident, therefore, that the thinner 
 the sheet of entering water and the lower the final velocity 
 of the jet, the greater will be the efficiency of the injector. 
 
 For each special purpose the tube has to be designed, and 
 the form best adapted for high temperature is not that which 
 is best suited for low pressures of steam or for high lifts, so 
 that the shape adopted for ordinary practical results is 
 usually the mean for the conditions under which the injector 
 is to operate, regarding which each designer is apt to have 
 his own opinion. 
 
 Temperatures as high as 162 degrees, and variations of 
 capacity up to 75 per cent, have been obtained in experi- 
 mental work by forms of the Sellers' Fixed Nozzle Injector 
 with varied forms of tube, neither of which the writer be- 
 lieves has ever been exceeded by any other styles of injector 
 with which he is familiar. 
 
 The advantage of maintaining the feed water pure and free 
 from lime and dirt is more apparent in the case of the com- 
 bining tube than with any other part of the injector ; as the 
 specific gravity of sand and grit is greater than that of the 
 water, all foreign particles are driven with great force 
 against the walls of the tube by the impact of the steam. 
 The tendency is not only to enlarge the diameters at differ- 
 ent points, but also to change the shape by wearing and 
 grinding off all shoulders against which the jet can strike ; 
 soft spots in the brass, caused by unequal mixing of the 
 metal before casting, soon show the effect of abrasion, and 
 this introduces a diagonal motion to the particles that wear 
 depressions upon the opposite side. Tubes that have been 
 used with impure water show this feature very strongly, and 
 the conical shape is frequently changed to that of two cylin- 
 ders, connected by an abrupt shoulder. The force of the 
 impact of the jet of steam is so severe, that a tube with walls 
 T y thick was found to have a pocket }i" deep worn in it at 
 
62 THE GIFFARD INJECTOR. 
 
 the end of the steam nozzle, and the outside was considera- 
 bly bulged by the continuous blows it had received. 
 
 But with pure water, a well designed tube will last a long 
 time, and in railroad service, where so many injectors are 
 employed, the choice of the water supply should take into 
 consideration not only the chemical analysis, but also the 
 percentage of impurities mechanically mixed with it; in 
 many cases these impurities are almost invisible to the naked 
 eye, and it is only by careful nitration that the trouble can be 
 remedied. On many of the railroads of France, this system 
 of purification is carried out very perfectly, and the effect is 
 apparent not only in a reduction in the cost of repair to the 
 locomotive boilers, but also in the longer life of the injectors. 
 
CHAPTER VI. 
 
 THE STEAM NOZZLE. 
 
 THE advantage of obtaining the highest velocity of the 
 steam at the instant it strikes the water has already been 
 shown, so that the great importance of obtaining the best 
 shape for the nozzle which guides the discharge of the actu- 
 ating jet, is at once apparent. 
 
 As it was assumed in the earliest experiments that the 
 discharge of gases followed the same laws that govern in- 
 elastic fluids, the steam nozzles of the first injectors were 
 made with a gradual convergent taper, and it was not until 
 1869 that the form was improved by the application of a 
 divergent flare, that permitted expansion of the steam within 
 the limits of the tube. The effect of this change upon 
 inelastic fluids, like water, would have been to increase the 
 volume discharged and diminish the terminal velocity ; but 
 with steam, air or other gases, the gradual lowering of the 
 pressure as the jet traverses the tube, gives an increase 
 in volume and expansion in the direction of the flow, which 
 augments the velocity of the particles of the fluid, while the 
 weight discharged remains unchanged. 
 
 In order to demonstrate more clearly the great difference 
 between the discharge of gases and liquids, some experiments 
 will be described that were made with tubes of different 
 forms, using steam under pressures of 120 and 60 pounds 
 (gauge). Sketches and instantaneous photographs were 
 taken of jets of steam discharging into the air, and the 
 form of the jet and the direction of the motion of the 
 particles were carefully noted. The results are given in 
 Fig. 15, which accurately represents the external form as- 
 
 53 
 
54 THE GIFFARD INJECTOR. 
 
 sumed by the jets, although it does not show the change 
 from transparency to whiteness that occurs shortly after the 
 steam leaves the nozzle. Four styles of tubes are given : 
 
 1. A convergent nozzle or short cylinder. 
 
 2. An aperture in a thin plate. 
 
 3. A divergent tube, straight taper. 
 
 4. A divergent tube, curved taper. 
 
 FIG. 15. 
 
 DISCHARGING TUBES 
 
 STEAM. 
 
 
 Divergent^- 
 Curve 
 
 The first tube represents the shape that would be used to 
 produce a solid water jet at a high velocity, and is similar to 
 the earliest form of the steam nozzle. The second is a thin 
 diaphragm, or any orifice where the thickness of the walls is 
 small compared with the width of the opening. The third 
 is shaped like a divergent cone, widening in the direction of 
 
THE STEAM NOZZLE. 65 
 
 the flow. The fourth has the same dimensions, but expands 
 in Curved lines from the throat to the lower end. 
 
 In all these experiments, the jet, after leaving the end of 
 the nozzle, was almost invisible for a distance of two or three 
 diameters, and of a pale bluish color, marked with light 
 lines of white, apparently produced by the entrained parti- 
 cles of water. Beyond the transparent portion, the jet ex- 
 panded to a much larger diameter, became white on the 
 surface, and was finally condensed by the cooling effect of 
 the air. 
 
 The difficulty with the first two tubes is that they permit 
 immediate diametral expansion of the steam, instead of 
 confining it and compelling expansion in the direction of 
 motion. A transparent envelope is formed, three or four 
 times the size of the orifice, through which the central jet, 
 discharging at somewhat higher velocity, can be distinctly 
 seen. This swelling of the jet is due to the fact that the 
 internal pressure at the moment of discharge is greater than 
 that of the medium into which it is flowing ; it is most ap- 
 parent in No. 2, because in this tube the internal pressure 
 of the steam at the terminal section is the highest, and in 
 all cases is more marked with high than with low steam. 
 In the divergent nozzle, the terminal pressure is the same 
 as that of the atmosphere ; if it were higher, there would 
 be the same enlargement that occurs in the previous cases, 
 and if lower, a contraction would be caused by the pressure 
 of the air. In these two nozzles, the direction cf discharge 
 of the particles is almost exactly parallel with the axis, and 
 therefore, all the energy of the steam, except the slight loss 
 due to friction against the walls, is utilized for augmenting 
 the velocity. 
 
 To find the velocity of discharge a knowledge of the in- 
 ternal condition of the jet is essential ; during the move- 
 ment of the steam toward, and through the tube, there must 
 be a reduction in its pressure, and a corresponding increase 
 in volume. If then, the area of the nozzle at different sec- 
 tions is known, it will be necessary to find the pressure at 
 those points in order to determine the volume or density 
 
56 THE GIFFARD INJECTOR. 
 
 from which the velocity may be calculated. This was done 
 when the injector was forcing water into the boiler, and also 
 when the steam was discharging freely into the air, by 
 inserting a small tube along the axis of the nozzle, and 
 observing the indications of a gauge placed on its outer end. 
 Communication was made with the interior of the jet by 
 means of a small hole drilled through the tube, so that by 
 sliding the tube backward or forward, the internal pressure 
 of the jet could be obtained within or beyond the limits of 
 the nozzle, either when the injector was working, or during 
 free discharge. By this means the pressure within the trans- 
 parent portion of the jet was ascertained, and the fall of 
 pressure during condensation. The experiment was also 
 tried of drilling minute holes through the steam nozzle 
 normal to the jet, and reading the gauge directly, but by this 
 method observations were restricted to the limits of the nozzle 
 and could not be obtained when the injector was in action. 
 
 The results of these investigations are shown in Fig. 16, 
 where A is the steam nozzle, B the combining tube, f the 
 hollow spindle with transverse hole/"', by which' the pressure 
 was communicated to the gauge. The intersections of the 
 vertical lines with the curved lines in the diagram indicate 
 the pressure at the respective points of the nozzles ; the 
 horizontal lines represent pressure above or below the at- 
 mospheric line. Observations were made at 60 and 120 
 pounds gauge pressure, the full line showing the conditions 
 when the injector is running, and the dotted line when the 
 steam nozzle is discharging into the air. 
 
 Considering first the case where the injector is working, 
 it will be noticed that as the particles of steam approach the 
 entrance to the nozzle, the., pressure falls in easy curves until 
 the smallest part of the tube is reached, when the descent is 
 more abrupt, but approaching the usual form of expansion 
 as shown by the indicator card of a steam cylinder. Just 
 bej^ond the end of the steam nozzle, at the line a' a 1 ', the prox- 
 imity of the feed water causes a quick fall of pressure that 
 is only partially recovered during its passage through the 
 combining tube. 
 
THE STEAM NOZZLE. 
 
 57 
 
 These pressures of 22" for 120 pounds, and 24" for 60 
 pounds, are found at the centre of the jet; that due to the 
 actual contact of the feed water with the steam envelope 
 would approach more nearly a perfect vacuum. 
 
 FIG. 16. 
 
 Jltmo!. 
 
 DIAGRAM 
 
 Pressure in $teawJfozzle 
 
 a,t 
 
 Jtyvctor tvorhng'. 
 
 1 1 V 
 
 i 1 i^l 
 
 Turning now to the freely discharging steam, it w r ill be 
 observed that during passage through the nozzle, the lines 
 for the two conditions almost overlie, and probably would, 
 
58 THE GIFFARD INJECTOR. 
 
 if all slight errors of observation could be excluded ; this 
 seems singular when it is considered that in one case the 
 steam is discharging into a partial vacuum, and in the other, 
 into the air ; beyond the line a' a the curves separate, yet 
 both pass below the atmospheric line, the curve of free dis- 
 charge rising and falling, owing to the unstable equilibrium 
 of the j et. Under 1 20 pounds initial pressure the j et emerging 
 from the tube at 9 pounds pressure, expands to atmospheric, 
 and then by internal condensation, n" vacuum is reached, 
 but is soon overcome and equilibrium established. This 
 peculiar phenomena of alternate rising and falling of the in- 
 ternal pressure is thoroughly- borne out by the external 
 appearance of the jet, as it presents the appearance of 
 having nodes separated by swelling curves. The 60 pound 
 pressure line shows that the steam leaves the nozzle at a 
 pressure of 4" vacuum, and, therefore, contracts instead of 
 showing the. surrounding envelope, characteristic of the 
 other case. 
 
 From the correspondence of the curves showing the fall 
 of pressure, under both conditions of discharge, it appears 
 that the outflow is unaffected by the pressure in the receiv- 
 ing chamber; when the injector was working at 120 pounds, 
 the pressure beyond the end of the steam nozzle was 22" 
 vacuum, and in the other case, 14.7 pounds, yet the con- 
 ditions within the limits of the nozzle were almost precisely 
 the same. If the experiments were carried still further, and 
 the weight of steam determined that would pass in a unit of 
 time through an orifice in a reservoir in which the initial 
 pressure was maintained constant, while the pressure in the 
 receiving tank was increased from a vacuum up to the* 
 upper limit of pressure, it would be found that the rate of 
 flow would be practically constant until the counter pressure 
 was raised to -f$ of the initial. It is true theoretically, 
 and has been also proved by careful experiment, that this 
 value of T %, corresponds closely to the relation of pressures 
 that will give maximum flow, although the actual in- 
 crease above the weight discharged against atmospheric 
 pressure is small. At 120 pounds gauge pressure, for ex- 
 
THE STEAM NOZZLE. 
 
 59 
 
 ample, a little more steam will be discharged against 66 
 pounds than there would be into the air, for, taking the 
 absolute pressures to which this ratio applies, (120 -f 15) X 
 0.6 = 8i.= the absolute, or 81 15 = 66.0 =* the gauge 
 
 FIG. 17. 
 
 pressure. As an explanation of this peculiarity of the jet, 
 Rankine has suggested that there is always a limiting sec- 
 tion at which the internal pressure is 0.58 of the initial, and 
 until that value is exceeded at a lower part of the jet, the 
 flow will not be reduced. 
 
60 THE GIFFARD INJECTOR. 
 
 With apertures in thin plates, like No. 2, Fig. 15, or short 
 cylindrical tubes, the limiting section is found directly at 
 the opening ; but with many of the steam nozzles as designed 
 for injector use, it is a little in the rear of the minimum 
 diameter ; this is true even when the tube may be designed 
 to permit immediate free expansion, as shown in Fig. 17 
 where, although the pressure at the smallest diameter is 
 below 0.58 of the initial, the velocity at that point is nearly 
 constant. The diagram shows a half section of a nozzle 
 divided by vertical dotted lines that pass through minute 
 holes that were drilled for the purpose of obtaining the 
 pressure of the jet by means of an ordinary gauge. Readings 
 were taken at the different steam pressures, and the results 
 plotted under the respective parts of the tube The pressure 
 lines are similar to those in Fig. 16, and indicate the rate of 
 expansion of the steam while traversing the nozzle. The 
 upper curves are for the velocity, and are calculated from 
 the observed pressures ; these curves almost overlie through- 
 out their whole length, and probably would, if the flare of 
 the nozzle were still further expanded. 
 
 The throat pressures, ratios and velocities are as follows : 
 
 Initial Pressure, Throat Pressure, Velocity at Throat. 
 
 Absolute. Absolute. Ratio. Feet per Sec. 
 
 135 698 0-5I7 1 612.0 
 
 45 22.8 0.508 1603.0 
 
 Although there is a difference of 95 pounds between the 
 extreme initial pressures, there is only a difference of 9 feet 
 in the two velocities. It is possible that the orifice by which 
 the pressure was recorded on the gauge in these experiments, 
 was a little below the limiting section probably not more 
 than T^J- inch as the smallest change in its position gives 
 a wide variation in the pressures. 
 
 The ordinary type of injector steam nozzle gives slightly 
 different results, for the divergent taper is straight and does 
 not allow as great an amount of transverse expansion, so 
 that the ratio of the throat pressure to the initial pressure is 
 not constant. With a tube having a divergent flare of i in 
 6 in diameter, the following results were obtained : 
 
THE STEAM NOZZLE. 61 
 
 Absolute Pressure. 
 Initial. Throat. Ratio. Velocity in Throat. 
 
 135 82.0 0.606 1407 
 
 105 61.5 .585 1448 
 
 75 42 .559 U9 1 
 
 45 24.5 .546 1504 
 
 In this case, the nozzle was formed with easy curves of 
 approach to a short cylindrical portion whose length was o. 3 
 the diameter arid into which the gauge hole was drilled; 
 from this point the tube widened with a taper of i in 6. 
 This tube corresponded to a form of steam nozzle exten- 
 sively used in injector service, and proves, from the fact that 
 neither the pressure ratios nor the velocities are constant, 
 that the actual rate of expansion is not similar to that of 
 free discharge, nor that of the widely flared nozzle shown 
 in Fig. 17. 
 
 If the tube is cylindrical the weight of steam flowing per 
 second reduces with -an increase in length ; at 75 pounds the 
 flow through a j" tube was, 
 
 }&' f long 900 Ibs. per hour. 
 
 i /x long 892 Ibs. per hour. 
 
 i^" long 864 Ibs. per hour. 
 
 To determine the weight of steam passing through an 
 orifice, it is often convenient to use a simple formula instead 
 of resorting to actual measurement. The following equation 
 was found by R. D. Napier to give results very close to 
 actual weighings, with results only about 2 per cent, low at 
 70 pounds, and i per cent, low at 120 pounds; Rankine has 
 also recommended it for approximate calculations. The 
 formula is 
 
 PX A 
 w =-^~ <S) 
 
 where 
 
 A = area of the orifice in square inches, 
 
 P = absolute pressure, = gauge pressure plus 15, 
 
 w =. weight, of steam discharged in pounds per second. 
 
 Knowing the weight of a cubic foot of steam at the pres- 
 sure in the orifice, the velocity of the jet at that point can 
 
62 THE GIFFARD INJECTOR. 
 
 easily be found. It is better, however, to apply the simple 
 approximate formula given by Rankine, which is based 
 upon the energy exerted by the steam during expansion : 
 
 V V 2g v == 8.025 v/77 ........... (10) 
 
 -- + - 2 (1105590- -5 10.2 7-0. . (ii) 
 
 Here V is the velocity in feet per second at the lower 
 pressure where the absolute temperature is T t = / -f 461.2 ; 
 and 7J the absolute initial, which at 120 pounds steam is 
 T! = 350.03 4- 461.2 = 811.23 degrees; the pressure being 
 known, the actual temperature of the steam can be obtained 
 from any steam table, and the absolute temperature, by 
 adding 461.2. The use of this formula is to calculate the 
 velocity of the jet at the instant the particles of steam strike 
 the water and the transfer of momentum is effected ; this 
 one is the simplest to apply, as it is not complicated by other 
 calculations. 
 
 The application of this formula for finding the velocity 
 requires the assumption that the work performed by steam 
 during discharge is equal to that exerted against a piston 
 moving in a cylinder, and that no heat is received from, or 
 given out to, external bodies ; in other words, the relation 
 between pressure and volume in all parts of the jet follows 
 the law of adiabatic expansion. During this process, a cer- 
 tain percentage of steam is condensed, which increases with 
 the difference between the initial and terminal pressures ; 
 for example, in expanding from 120 pounds to the atmos- 
 phere, 12^5- per cent, of the weight discharged is condensed, 
 leaving 8y T 6 ^ per cent, of gaseous steam ; when the final 
 pressure is 22" vacuum, as in the case of the injector quoted 
 on page 58, i pound of steam at the moment of impact will 
 contain 0.179 pound water and 0.821 pound steam. 
 
 These values may be calculated for any degree of expan- 
 sion by means of formula (12) : 
 
THE STEAM NOZZLE. 63 
 
 where x l and x 2 are the weights of steam at the absolute 
 initial and terminal pressures. (If initial steam dry, x l = i.) 
 TI and 7^2, = absolute temperatures = (t + 461.2), 
 TI and r 2 , = latent heat under the two conditions, 
 l and ft,, = entropy of liquid. 
 
 These terms can be obtained from Steam Tables ; the 
 values of most frequently used are, however, given in 
 Table III, page 66. 
 
 Formula (No. 12) will be found of use in calculating 
 results from experimental data, and for supplying the value 
 of x^ in (13), by which the velocity of the jet can be con- 
 veniently found : 
 
 F= 8.025 \/ 778 (*i n + ft **r t ft) ........ (13) 
 
 and the weight discharged, 
 
 _ ____ A^V^ ____ 
 
 "144 (x. t (5o.oi6) + 0.016)* ' 
 where 
 
 V velocity. 
 
 w = weight discharged, in pounds per second. 
 
 A = area of the nozzle at smallest section, in square inches. 
 
 ft and q. 2 = heat of the liquid at the terminal pressures. 
 
 5 = volume i pound steam in cubic feet, at final pressure. 
 
 By way of illustrating the use of these formulae, the velocity 
 of a jet of steam discharging from a reservoir at 120 pounds 
 pressure into the air will be calculated. The steam will be 
 supposed to be dry, and to follow the law of adiabatic ex- 
 pansion. The value of x t must be first determined by (12). 
 As the steam is supposed to be dry in its initial state, x l = 
 i ; the other terms are as follows : 
 
 6 l = 0.5027 rj = 867.3 7\ = 350.3 -f 461.2 = 8 1 1.5 q l = 321.4 
 3 =r 0.3135 r 2 = 966.07 T 2 = 2i2 +461.2 673.2 ? 2 
 
 , 867.3 
 0.5027 0.3135 + x- -- 
 
 ~ 5 = 0.876 
 
 673.2 
 
 V= 8.025 \/ 778 (867.3 + 3 2I -4 0.876 x 966.07 180.5) = 2848.2 
 Formula (10) gives 2827 feet per second, which agrees 
 fairly well with this result. 
 
 OF THB 
 
64 THE GIFFARD INJECTOR. 
 
 If dry steam at the same pressure expand to 22" 
 vacuum, or 4 pounds absolute, the value of x^ will be found 
 to be 0.8243, and the final velocity, 3446 feet, or ^3 mile per 
 second; calculated by (10), V is 3443 feet. 
 
 The presence of water in the steam is shown by the change 
 in the appearance of a jet discharging into the air ; the 
 clear bluish portion of the jet adjoining the orifice becomes 
 an opaque white. The effect upon the velocity can be found 
 by giving a lower value than unity to jqin formula (12). If 
 5 per cent, of water is entrained, x l = 0.95, and in the pre- 
 vious example x 2 becomes 0.7899, and the final velocity is 
 3223 instead of 3446 feet. 
 
 Besides retarding the velocity, entrained water has an in- 
 jurious effect upon the surfaces of the tubes, cutting and 
 abrading the metal wherever soft spots occur; dry pipes 
 should be used to prevent this trouble and all pockets or de- 
 pressions in the steam pipe avoided; for prompt starting, 
 efficient action, and general reliability of an injector are much 
 enhanced by attention to this apparently trivial condition. 
 
 The area of the steam nozzle, upon which depends the 
 weight of steam used per hour, can best be defined in terms 
 of the area of the delivery tube. This varies with the pat- 
 tern of the injector, and the purpose for which the instrument 
 is designed. From superficial considerations, it would appear 
 that if this ratio were made unity, z. e., the steam nozzle 
 area = delivery tube area, the jet would just have sufficient 
 power to sustain the initial pressure ; the nearest approach 
 to this in experimental work is with the ratio 1.007 to i-ooo, 
 but in actual practice, on account of tlje necessity of a 
 margin of -counter pressure, the ratio varies from 2 to i, to 
 3 to i ; in the exhaust injector this is increased to 1 6 to i, as 
 the ratio of the initial to the terminal pressure is i to 6, or 
 1 6 pounds to 96 (absolute). 
 
 The motive element in the injector is the steam supply, 
 and the work performed is raising the water to the level of 
 the instrument and delivering it against the pressure of the 
 boiler. From the stand point of mechanical efficiency, it is 
 obvious that the smaller the weight of steam required to 
 
THE STEAM NOZZLE. 65 
 
 perform this work, the more economical will be the opera- 
 tion. The condensation of the motive jet is the source of 
 economy in this method of feeding, as the only loss is the 
 small amount heat radiated. The employment of a larger 
 quantity of steam than is absolutely necessary, would be 
 using the instrument as a heater, or similar to the applica- 
 tion of a jet of steam to the suction pipe of a pump for the 
 purpose of warming the water. 
 
 The raising of the temperature of the feed water is bene- 
 ficial, but can best be accomplished by other means; by 
 utilizing the heat of the waste products in the smoke-box 
 or stack, or by using the heating surface of the boiler itself, 
 instead of absorbing available energy from the boiler for the 
 purpose of performing subordinate work. This is specially 
 applicable to service on locomotives, where the high rate of 
 speed employed at the present time necessitates the strictest 
 economies in the use of steam, and the application of the full 
 capacity of the boiler to its primary use, supplying steam to 
 the cylinders ; it frequently happens that the injector can only 
 be placed in operation when the excessive strain is relieved by 
 a stop at a station or during a run on a down grade. On 
 the other hand, the most economical results are obtained, 
 not by an intermittent water supply, but by maintaining a 
 constant water level and a continuous feed. 
 
 The actual amount of steam required during the feeding 
 of a locomotive boiler is not generally realized ; an evapora- 
 tion of 2,500 gallons of water per hour, requires, with some 
 patterns of injectors, 2,300 Ibs. of steam, which, at an esti- 
 mated rate of 35 Ibs. per H. P., amounts to 65 H. P. If the 
 steam were used to best advantage, this could be reduced to 
 about 45 H. P. , and the temperature of the water entering 
 the boiler would be lowered approximately 25. It is of 
 course advantageous to supply the boiler with water as hot 
 as possible, to avoid shrinkage and unequal expansion of- 
 the sheets ; but, as can be easily seen, the benefits gained in 
 that direction by raising the temperature of the feed by 
 means of a jet of high pressure steam, seldom compensate 
 for the loss of available steam supply. 
 5 
 
66 
 
 THE GIFFARD INJECTOR. 
 
 There can be no question but that the injector is superior 
 to all other devices for feeding locomotive boilers that have 
 yet been introduced, but its action should be as efficient as 
 possible, in order that the steam and fuel consumption 
 which may properly be charged to its account, can be re- 
 duced to the minimum. 
 
 A short table giving the value of 6 may be found to be 
 convenient when complete tables are not at hand : 
 
 TABLE III. 
 
 Absolute 
 pressure, 
 Ibs. persq.in. 
 
 6 
 
 Absolute 
 pressure, 
 Ibs persq. in. 
 
 e 
 
 Absolute 
 pressure, 
 Ibs. persq.in. 
 
 6 
 
 Absolute 
 pressure, 
 Ibs. per sq. in. 
 
 e 
 
 Absolute 
 pressure, 
 Ibs. persq. in. 
 
 
 
 Absolute 
 pressure, 
 Ibs. persq. in. 
 
 
 
 I. 
 
 .1329 
 
 6 
 
 .2480 
 
 14.7 
 
 3135 
 
 40 
 
 .3921 
 
 65 
 
 4337 
 
 QO 
 
 4633 
 
 2 
 
 1754 
 
 7 
 
 .2587 
 
 2O 
 
 3363 
 
 45 
 
 .4020 
 
 70 
 
 .4402 
 
 IOO 
 
 4733 
 
 3 
 
 .2013 
 
 8 
 
 .2682 
 
 25 
 
 3539 
 
 50 
 
 .4109 
 
 75 
 
 .4464 
 
 120 
 
 .49" 
 
 4 
 
 .2203 
 
 10 
 
 .2842 
 
 30 
 
 .3685 
 
 55 
 
 .419! 
 
 80 
 
 .4522 
 
 ISO 
 
 5133 
 
 5 
 
 2353 
 
 12 
 
 .2976 
 
 35 
 
 .3811 
 
 60 
 
 .4267 
 
 85 
 
 4579 
 
 2OO 
 
 .5429 
 
 An efficiency comparison of different styles of injectors 
 can be made either by rating the delivery of the injector in 
 cubic feet or pounds per hour in terms of the area of steam 
 nozzle, or by determining the ratio of the weight of steam 
 used to the weight of water forced into the boiler. In this 
 connection may be given the following tests of some of the 
 best-known patterns of locomotive injectors, all made under 
 precisely the same conditions : steam 120 pounds, feed tem- 
 perature 65 Fahr., height of lift 18". Steam dry. 
 
 Name of Injector. 
 
 Nominal Size. 
 
 Belfield No. 10 
 
 Garfield No. 7 
 
 Little Giant No. 7 
 
 Mack No. 7 
 
 Metropolitan No. 9 
 
 Monitor No. 9 
 
 Sellers' 1887 No. 8% 
 
 Weight of Water 
 
 Delivered per Ib. 
 
 of Steam. 
 
 9.69 pounds 
 
 13-53 
 12.92 
 
 13-79 
 13.16 
 11.31 
 13.80 
 
CHAPTER VII. 
 
 THE ACTION OF THE INJECTOR. 
 
 NOTWITHSTANDING the fact that much has been written 
 upon this subject, the action of the injector still appears 
 mysterious to many of those who are familiar with its opera- 
 tions. It is strange that the reason for its working is not more 
 generally understood, even by those accustomed to operate 
 it daily, especially as this method of feeding is now so uni- 
 versally employed for locomotive and stationary boilers. 
 
 The simplest method of considering the theory of the 
 injector is to eliminate the more complicated sides of the 
 question and consider it solely from a mechanical point of 
 view ; simply as an apparatus in which the momentum of a 
 jet of steam is transferred to a more slowly-moving body of 
 water, producing a resultant velocity sufficient to overcome 
 the pressure of the boiler. 
 
 The high velocity attained by a jet of steam has been cal- 
 culated, and diagrams have been given that show the fall of 
 pressure and increase in velocity as the volume is increased 
 according to the laws under which the steam expands. 
 Suppose that a nozzle connected with a reservoir containing 
 steam at 120 Ibs. pressure discharges i Ib. of steam per 
 second ; at its minimum diameter, the steam will have 
 reached a velocity of 1407 feet, but when the terminal pres- 
 sure is 22" vacuum, the velocity will be 3446 feet per second. 
 I/et us suppose that this jet flows into a combining tube, 
 which. is able, by means Of the great conductivity of its 
 walls, to abstract sufficient heat to completely condense the 
 steam at a final pressure of 22", or 4 Ibs. absolute. This 
 reduces the steam to a solid jet of water having a cross 
 
 67 
 
68 THE GIFFARD INJECTOR. 
 
 section T ^ the area of the steam while passing through the 
 steam nozzle, and yet does not in any way affect the velocity, 
 as the contraction of the jet is entirely lateral. A jet of 
 water issuing from the delivery tube, forced out by the pres- 
 sure of the boiler, would have a velocity nearly equal to 
 that due to the head, or approximately, 133 feet per second, 
 only ^ of that of the jet of condensed steam ; but an injec- 
 tor is required to perform useful work, forcing a supply of 
 feed water into the boiler ; therefore a certain weight of feed- 
 water must be added which will take the place of the cold 
 walls df the tube for the purpose of condensation. This 
 mass of water receives the energy of the moving steam, 
 condenses it, and the two fluids move along together through 
 the delivery tube with a terminal velocity greater than a jet 
 of the same density issuing from the boiler. If the weight 
 pf water supplied is too great, the steam will not have power 
 enough to give the required velocity of 133 feet; if there is 
 an insufficient supply, the volume of the steam will not be 
 reduced sufficiently to pass through the tubes, and in neither 
 case will the injector work properly. 
 
 Turning again to figures, and taking the simplest possible 
 case, we can follow the steam through its whole course 
 within the tubes, and determine the relation of the different 
 parts of the injector. As the velocity and volume of the 
 steam at the instant of passing the minimum diameter of 
 the steam nozzle are 1407 feet, and 5.05 cubic feet, respec- 
 tively, and after complete expansion in the combining tube 
 3446 feet (see page 64) and 74.2, the cross section of the 
 steam jet at that time must be, 
 
 z- X = 5-99 times the area of the steam nozzle. 
 
 3446 5-05 
 
 During complete condensation, the volume of i Ib. of steam 
 shrinks from 74.2 cubic feet to o. 16 cubic feet, so that the 
 cross section of the jet after condensation is 
 
 X - = the area of the steam nozzle, 
 
 74.2 i 774 
 
 so that the jet of condensed steam would pass through an 
 
THE ACTION OF THE INJECTOR. 69 
 
 orifice yf* the area of the steam nozzle at a velocity of 3446 
 feet per second. 
 
 The substitution of a cone of water as the condensing 
 medium, at a ratio of 13 pounds of water to the pound of 
 steam as this is a fair performance of a locomotive in- 
 jector at this pressure will increase the volume of the jet 
 (13 + i) = 14 times, and require the area of the delivery 
 tube to be enlarged to 7 VV Further, the addition of this 
 weight of water will retard the motion of the jet, and de- 
 mands a still larger orifice of entrance ; it will therefore be 
 necessary to determine the final velocity of the mixture 
 before the size of the delivery tube can be obtained. The 
 transfer of the energy of the steam to the water is similar to 
 that occurring during the impact of two inelastic bodies, 
 and owing to the fact that the particles of steam do not all 
 strike the molecules of water in the proper direction, and 
 owing to the obliquity of the entrance of the feed into the 
 combining tube, and also other causes, a loss occurs which 
 varies somewhat under different conditions, but under those 
 assumed, amounts to 40 per cent. Multiplying the weight 
 of the water by its entrance velocity of 40 feet due to the 
 22" vacuum in the combining tube, and adding the momen- 
 tum of i Ib. of steam at a velocity of 3446 feet, we have the 
 following equation of momentum, which determines the 
 velocity of the mixture : 
 
 (i.oo 0.40) x (4 X 13 -}- 3446 x i) = 14 X v 
 v = 169.97 feet per second. 
 
 The temperature of this mixture would probably be about 
 150, and i Ib. per square inch, from Table II, would be 
 
 equal to a head of 2.355 f eet > and from h = --, and h = p 
 X 2.355 we have 
 
 . 
 
 t> - = -- , - = 190.5 pounds back pressure 
 2-355 X 64.4 151.66 
 
 corresponding to this velocity ; but from this must be sub- 
 tracted the pressure in the combining tube below the atmos- 
 
70 THE GIFFARD INJECTOR. 
 
 phere, 2-2" or 1 1 Ibs. , so that the available counter pressure 
 as shown on the gauge will be 179.5 Ibs. 
 
 The velocity of the original jet is now reduced from 3446 
 feet, as in the assumed case when passing through the de- 
 livery tube, to 169.97 feet, requiring a further enlargement of 
 the area, in order that the augmented volume of the jet may 
 find entrance at this reduced velocity. Summing up the 
 two changes, one due to the decrease in velocity, and the 
 other due to increase of volume, we have, 
 
 - X = : the area of the steam nozzle ; 
 
 169.97 774 2.72 
 
 or, in other words, if the area of the delivery tube be taken 
 as unity, the area of the steam orifice will be 2.72, and the 
 ratio of the diameters will be i to 1.65, which approaches 
 closely the proportions in ordinary practice. 
 
 It is thus seen that the whole action of the injector de- 
 pends upon the fact that the velocity of a jet of steam dis- 
 charging into the combining tube, is 20 to 25 times that of a 
 jet of water issuing from a boiler under the same pressure, 
 and that the enormous reduction of the volume during con- 
 densation concentrates the momentum of the jet upon an 
 area which is but a small fractional part of the orifice from 
 which it issues, leaving a large margin of available energy 
 which may be applied to useful purposes. As condensation 
 plays such an important part in the operation, it is seen that 
 any condensible gas may be substituted for the motive steam, 
 if the inherent conditions are properly considered, but some 
 modifications of the proportions of the parts as used in the 
 steam injector might be found necessary in order to work 
 satisfactorily under the new conditions. 
 
 The principle of the action of the injector working at 120 
 pounds steam into a boiler carrying the same pressure, may 
 appear more easy of explanation than the case of an exhaust 
 injector forcing water into a boiler at 80 pounds steam ; that 
 there is no difference beyond a change in the proportions of 
 the parts, can be seen from the following example, which 
 will be worked out according to the same analysis that was 
 applied to the case of high pressure steam. 
 
THE ACTION OF THE INJECTOR. 71 
 
 Assume the steam at o (gauge), or 14.7 pounds absolute 
 pressure, containing, as it arrives direct from the cylinder, 
 about 10 per cent, of moisture. From page 49 the vacuum 
 in the combining tube will be assumed to be 24" or 3 pounds 
 ab. From equation (12), we find that i pound of steam in its 
 final condition will contain 0.832 pound of steam, and o. 168 
 pound of water ; its velocity after complete expansion will 
 be found from (13), 
 
 K=8.o25 \778 (0.90x966.07 -f- 181.60.832x1015.3109.8) = 2205 ft. 
 
 To work against a pressure of 80 pounds, the delivered 
 water must be forced out of the combining tube, in which 
 the pressure is 12.7 pounds below the atmosphere, so that 
 the total pressure against which the jet must be capable of 
 working is 80 -f 12.7 = 92.7 pounds ; this requires a terminal 
 velocity of 1 35 feet, under the assumption that the steam is 
 all condensed and that its density is unity. The feed water 
 enters under a head of 6 feet, and with a velocity corres- 
 ponding to the sum of the partial vacuum and the head, 
 approximately 48 feet per second. A somewhat greater 
 percentage of loss must be taken than in the other case, as 
 the action of the jet is not as efficient, and it does not seem 
 as if it could be made so ; taking 0.50 as the value of this 
 coefficient, we have for the equation of momentum, from 
 which the weight of feed water may be obtained, 
 
 (i X 2205 + 48 x W) xo.5o=(JF+i) x 135, 
 whence 
 
 W-= 8.74 pounds of water per pound of steam, 
 
 which corresponds to the usual practice. 
 
 In both the examples given the simplest conditions were 
 assumed and all uncertain elements avoided. The jet as it 
 passed through the delivery tube was supposed to have the 
 same density as water at the same temperature ; in fact, this 
 seldom occurs, as there is almost always a part of the steam 
 uncondensed until the mixture has passed far into the deliv- 
 ery tube, and there is, in addition, a volume of air mixed 
 with the steam that displaces a corresponding volume of 
 water and gives to the jet its white, opaque appearance. 
 
72 
 
 THE GIFFARD INJECTOR. 
 
 This quantity of air will depend upon the manner in which 
 the boiler is fed, and the condition of the suction pipes and 
 valves and stuffing-boxes of the injector or pump, as the 
 total amount of air contained in the steam is greater than 
 can be held in solution by the feed water. Experiments 
 made with an injector placed entirely- under water, so as to 
 collect all the air discharged with the delivery, showed that 
 this amount, though very variable, was by no means inap- 
 preciable ; a No. 8 injector at 120 pounds, discharged 5.51 
 cubic feet of air per hour, and at 60 pounds, 4.38 cubic feet, 
 measured at atmospheric pressure. These results do not 
 represent the maximum quantity, but the mean of several 
 tests. 
 
 The density, however, depends chiefly upon the percent- 
 age of steam condensed, and is, therefore, intimately con- 
 nected with the water ratio. The following table shows the 
 variation, and the increase in the velocity of the jet as the 
 density decreases : steam pressure, 135 pounds absolute. 
 
 TABLE IV. 
 
 
 tl 
 
 V U 
 
 c 
 
 fi' ^ 
 
 1 S S 
 
 3 gT 
 
 c 
 
 
 
 
 P. 60 
 
 "2 a. 
 
 2-S P. 
 
 en {jj 
 
 g 
 
 ^>o 
 
 
 
 
 h 
 
 &S-S 
 
 Ifcl 
 
 O ^fc . 
 
 P 
 
 .52 a> u 
 
 -i 3 3 
 
 si 
 
 rt-JJ 
 
 & 
 
 If 
 
 S 
 
 
 ** *"* " 
 
 2? ^ W-t 
 
 O .*_ cd H 
 
 ** ^5 8 w 
 
 
 
 5 * 
 
 c 
 
 <C M 
 
 
 
 
 
 
 "S C8 PU5 
 
 ^ 
 
 o oft 
 
 5 
 
 p 
 
 Q 
 
 ! 
 
 
 4 
 
 JF 
 
 ^7 
 
 p 
 
 P 
 
 p 
 
 A 
 
 A- 
 
 *) 
 
 154 
 
 I2.OO 
 
 3435- 
 
 1653 
 
 1 80. 
 
 186.1 
 
 0.788 
 
 o.55i 
 
 6>) 
 
 167 
 
 10.33 
 
 335. 
 
 1740 
 
 199. 
 
 236.4 
 
 0.542 
 
 0.524 
 
 c] 
 
 185 
 
 8.63 
 
 3228. 
 
 181.6 
 
 215. 
 
 302.8 
 
 0.360 
 
 0.493 
 
 d] 
 
 209 
 
 7.02 
 
 3058. 
 
 177-3 
 
 202. 
 
 346.3 
 
 0.261 
 
 0.432 
 
 e] 
 
 240 
 
 5 .60 
 
 2970. 
 
 147-2 
 
 H3- 
 
 294.0 
 
 0.250 
 
 0.310 
 
 f) 
 
 268 
 
 4.69 
 
 2890. 
 
 141.6 
 
 130. 
 
 305.0 
 
 0.2 1 1 
 
 0.265 
 
 The ratio of water to steam that produces the highest 
 back pressure is shown in experiment (c} as the energy of 
 the jet is then at its maximum ; with less water, the 
 density of the jet is too low, even though the velocity is 
 greater, and below that ratio the density rises, but the 
 
THE ACTION OF THE INJECTOR. 73 
 
 velocity is insufficient, on account of the large increase in 
 the weight of the mixture. 
 
 Simple formulae can be applied to find the density : if the 
 back pressure is known, find the velocity of a jet of water 
 at the temperature of delivery, which designate by v, and 
 the actual velocity by v^. Let A be the density, and H the 
 back pressure expressed in feet. 
 
 Then 
 
 If. A = the area of the delivery tube, the theoretical capa- 
 city = v x A, and we obtain the following, by substituting 
 (15) and reducing 
 
 62.4 A v^ A _ Weight of actual jet _ Actual capacity _ 
 
 A v~ Weight of jet unit density Theoret. capac- ^ 
 
 so that 
 
 A = ( __ Actualcapadty \* _ (i6) 
 
 V Theoretical capacity/ 
 
 The actual velocity in the table was found by dividing the 
 calculated velocity by the square root of the density, and 
 the changes in the velocity of the entering water and the 
 discharging steam are due to the variation of the pressure 
 within the combining tube with the temperature of the 
 delivery. 
 
 Experiments ( e ) and (/") were made with the overflow 
 closed, permitting a high pressure in the confined overflow 
 chamber, otherwise there would have been a discharge of 
 steam ; but in many cases where two or more apertures in 
 the combining tube are contained in the same chamber and 
 only closed to admission of air by a light check valve, the 
 delivery temperature may rise above the boiling point of 
 water at atmospheric pressure. In one pattern of injector, 
 where several overflows are connected, the author has seen 
 the temperature of the water going to the boiler carrying 
 140 pounds of steam, reach 250 Fahr., with the overflow 
 chamber closed only by a light check valve weighing but a 
 few ounces. 
 
74 THE GIFFARD INJECTOR. 
 
 As the proportion of water to steam is reduced, and the 
 density of the jet diminished, it will be noticed that the 
 value of the coefficient, as given in the last column, also 
 grows less ; when the water ratio was i to 13, as taken in 
 the example at the opening of this chapter, this value was 
 taken at 0.60, but when the weight of water per pound of 
 steam is 7.02, this coefficient falls to 0.432. 
 
 The actual causes of the loss covered by this term cannot 
 be definitely described ; it is very probable that the whole 
 mass of steam is not entirely condensed at the instant of 
 impact, and there remains a certain amount of elasticity 
 that produces a rebound and an interference with the motion 
 of the particles of steam following. A jet of steam dis- 
 charging into a pail of water will blow the water in all 
 directions, and only a small portion of the steam will be 
 condensed ; the warmer the water, the greater this ten- 
 dency, so that it is not at all surprising that with high tem- 
 peratures of delivery, due either to warm feed water or 
 insufficient feed supply, the coefficient of efficiency should 
 have a lower value than when the conditions of working are 
 more nearly normal. The oblique angle at which the water 
 enters the combining tube is also disadvantageous, while the 
 roughened surfaces of the tubes add their quota to the gen- 
 eral sum of losses, which will be represented by the coeffi- 
 cient K. 
 
 It is exceedingly difficult, in fact almost impossible, to 
 represent algebraically the conditions that obtain within the 
 combining tube of an injector, and to frame an equation 
 that will apply to all conditions or all types of instruments. 
 The shape of the tubes, the conditions of the surfaces, and 
 the proportions of the orifices, all introduce special consid- 
 erations that would so complicate an equation as to invali- 
 date its utility ; fundamental relations can, however, be 
 shown to exist, and simple equations given to prove the 
 mechanical theory as already outlined ; these will be fol- 
 lowed by formulae based upon the heat theory. 
 
 Taking the most elementary form of injector, one with a 
 single set of tubes and but one overflow in the combining 
 
THE ACTION OF THE INJECTOR. 75 
 
 tube, assume i pound of steam per second from the steam 
 nozzle as the actuating force. This steam will have a 
 velocity V due to the work performed in expanding from 
 the initial pressure of the boiler to the low pressure in the 
 combining tube. Its momentum will therefore be 
 
 7 .............. " 
 
 The feed water enters with a velocity v 2 due to the differ- 
 ence between its head h and the absolute pressure within the 
 combining tube ; as these pressures must be reckoned above 
 a perfect vacuum, the pressure in the tube becomes (34 ^), 
 and if the water comes to the injector under a head h, the 
 total head forcing it into the combining tube is (34 h^ -f h). 
 Calling the weight of water W, the momentum of the en- 
 tering water is 
 
 w v, = ^v/ 2 77^v+Tj 
 g g 
 
 The sum of these two equations represent the momenta of 
 the separate masses before they come in contact, but during 
 impact and the condensation of the steam jet, there is a loss 
 of momentum due to causes already outlined and indicated 
 by the coefficient K. The sum of (17) and (18) is the mo- 
 mentum of the mass as it approaches the delivery tube ; its 
 velocity must depend upon the difference between the abso- 
 lute permissible back pressure expressed in feet and the 
 pressure in the combining tube, or (H h^. Therefore 
 
 and the momentum of the combined mass is, 
 
 (y+o _____ 
 
 j- </2g(ff hd ......... (19) 
 
 The complete equation of momentum is 
 
76 THE GIFFARD INJECTOR. 
 
 Solving for W^we obtain the ratio of the weight of water 
 forced into the boiler per pound of steam, 
 
 (34- 
 
 If h v is not known and cannot be readily found, it may be 
 assumed to be about 24" vacuum or 7.3 feet absolute pres- 
 sure when the delivery temperature is 150, and 26", or 4.7 
 feet absolute, at 130, although a slight variation will not 
 materially affect the results. Upon h and h depends the 
 velocity of entrance of the water into the combining tube, 
 and these terms do not appear in their true importance in 
 the equation, as the area of entrance is not expressed, but is 
 assumed to be of the most advantageous proportion ; this 
 is only the case in the self-adjusting or adjustable combining 
 tube form of injector, shown in Figs, i and 2 or 3 ; as all the 
 water must enter through this opening, whose area we will 
 designate by B, the value of Win the ordinary fixed-nozzle 
 pattern would be 
 
 ^=62.39 B v/ 2g (34 hi -f h) ......... (22) 
 
 If the supply is lifted, h must be negative ; if the feed water 
 flows, h is positive. To show the change in the value of W 
 with different heights of lift ( h ) the following tests of a 
 fixed-nozzle injector, adjusted for its maximum capacity at 
 i foot lift, are given in Table V. Steam was constant at 
 65 pounds pressure, and the area of the water entrance to 
 the combining tube fixed ; feed and steam valves wide open. 
 
 TABLE V. 
 
 
 
 Ratio Wt. 
 
 
 
 Ratio Wt. 
 
 Lift in 
 
 Capacity Cubic 
 
 Water to 
 
 Lift in 
 
 Capacity Cubic 
 
 Water to 
 
 Feet. 
 
 Feet. 
 
 Steam. 
 
 Feet. 
 
 Feet. 
 
 Steam. 
 
 h 
 
 Q 
 
 W 
 
 h 
 
 Q 
 
 W 
 
 I 
 
 1 08 
 
 H54 
 
 8 
 
 87.8 
 
 11.83 
 
 2 
 
 107.2 
 
 H.43 
 
 10 
 
 8l.8 
 
 II. O2 
 
 3 
 
 104-9 
 
 14.13 
 
 12 
 
 76.9 
 
 9-92 
 
 4 
 
 101.9 
 
 13.73 
 
 14 
 
 66.9 
 
 8.97 
 
 5 
 
 97-9 
 
 13.18 
 
 16 
 
 60.8 
 
 8.13 
 
 6 
 
 94.8 
 
 12.76 
 
 18 
 
 583 
 
 7.72 
 
 7 
 
 90.4 
 
 12-43 
 
 20 
 
 51-6 
 
 6.89 
 
 In the self-adjusting injector, the area B changes auto- 
 matically with the steam pressure, so that an increase in the 
 
THE ACTION OF THE INJECTOR. 77 
 
 height of lift affects the capacity but slightly. A well 
 known pattern of the double-jet type at 60 pounds steam, 
 gave the following capacities at different lifts, 
 
 4 Feet Lift. 12 Feet Lift. 24. Feet Lift. 
 
 150.6 cubic feet. 134.6 cubic feet. 78.1 cubic feet. 
 
 In both (21) and (22) the head of water representing a 
 perfect vacuum was taken as 34 feet; this corresponds 
 closely to a temperature of 40, but if the temperature of the 
 feed rises, vapor of water is given off which reduces the 
 vacuum correspondingly. The following table gives values 
 that may be substituted for 34 in the above equation with 
 different temperatures of feed water : 
 
 TABLE VI. 
 
 Feed Temp. Vacuum in Feet. 
 
 t\ 
 
 4 34 
 
 70 33 
 
 90 32.2 
 
 ioo 31.4 
 
 120 29.7 
 
 130 27.3 
 
 Feed Temp. Vacuum in Feet, 
 
 'i 
 
 140 25.9 
 
 150 24.8 
 
 160 22.5 
 
 180 16.9 
 
 200 9.3 
 
 212 O.O 
 
 The chief criticism against the formula denoting the in- 
 terchange of momenta, is that there is no term in the equa- 
 tion that indicates the use of condensible gases ; as written, 
 air or any other permanent gas could be substituted for 
 steam and all the conditions satisfied; yet, as has been 
 shown, the principle upon which the injector acts, precludes 
 the use of all gases save those that may be reduced to a liquid 
 form under the conditions that obtain within the combining 
 tube of an injector. 
 
 The co-efficient K, which is such an important factor in 
 the equation of momenta (21), seems to depend somewhat 
 on /, but chiefly upon (/ 2 tj, the increase in the tempera- 
 ture of the feed water. As the heat of the water supplied 
 rises, the weight of water delivered per pound of steam de- 
 creases, and the coefficient K has also a less value as is 
 shown by Table VII ; this table also gives the capacity, the 
 value of W, and the density of the jet at 120 pounds steam, 
 in a No. 8 Self-adjusting Sellers' injector, placed i foot above 
 the feed level. 
 
78 THE GIFFARD INJECTOR. 
 
 TABLE VII. 
 
 Feed 
 Temp. 
 t\ 
 
 a) 65 
 b) 80 
 
 c] H5 
 d) 130 
 
 Delivery 
 Temp. 
 k 
 
 144-5 
 160.0 
 202.5 
 226.0 
 
 Water to 
 Steam. 
 IV 
 
 !3-53 
 13.26 
 11.63 
 10.36 
 
 Cubic Feet 
 per Hour. 
 Q 
 278.7 
 274.2 
 239.2 
 213.0 
 
 Density 
 of Jet. 
 A 
 
 0.8836 
 0.8154 
 0.7490 
 0.6430 
 
 Coeffi- 
 cient. 
 K 
 
 0-595 
 
 0.583 
 0.525 
 0-473 
 
 The falling off of the capacity with the increase in t lt is 
 caused by the reduced vacuum within the combining tube, 
 and the change in the terminal velocity of the steam jet; 
 this could be partially remedied by augmenting h in (22) as 
 occurs in the double jet pattern of injector, but yet it may be 
 stated as a general rule, that under conditions otherwise the 
 same, the lower the temperature of the feed water supplied, 
 the greater will be the capacity. In the case of the double 
 jet type, the weight of water passing through a unit section 
 of delivery tube is never as great as in a single jet injector, 
 because the warming action of the first set, reduces the rate 
 of condensation within the forcing combining tube, and 
 lowers the density of jet. 
 
 It follows, therefore, that in order to give the same capac- 
 ity as a single jet, the area of the delivery tube -of a double- 
 jet injector must be made considerably larger; the amount 
 of increase required depends upon the proportions of the 
 tubes, and may be anywhere from 13 to 80 per cent. The 
 action of the first set, although disadvantageous as far as the 
 warming of the water is concerned, yet produces a pressure 
 at the entrance to the combining tube of the forcing set that 
 increases its power of taking hot water ; this pressure rises 
 with the steam pressure and also with the temperature of the 
 supply. In a series of tests made with an injector of this 
 type, the following results were obtained, 
 
 Pressure 
 Steam. between Sets. 
 
 40 Ibs. 2 Ibs. 
 
 70 " 8.5 " 
 
 Pressure 
 Steam. between Sets. 
 
 90 Ibs. 17 Ibs. 
 
 120 " 27 " 
 
 when the feed temperature was 63 degrees, but when raised 
 to 120 degrees the intermediate pressure was 22 Ibs, at 90 
 Ibs. steam, and 40 Ibs. at 120 Ibs. steam. 
 
THE ACTION OF THE INJECTOR. 79 
 
 As the delivery of an injector depends directly upon the 
 velocity of the jet and its density at the instant of passing 
 the smallest diameter of the delivery tube, it follows that the 
 capacity will be proportional to the square root of the terminal 
 pressure plus a constant : if P is the boiler pressure as indi- 
 cated by gauge, Q the capacity in cubic feet per hour, D the 
 diameter of the delivery tube in millimetres, (i millimetre 
 = 0.0397") as many manufacturers use the French system 
 we have, approximately, 
 
 Q = 0.374 n*v P + 18 ....... . . . (23) 
 
 This applies to self-adjusting injectors up to that pressure at 
 which the maximum capacity commences to reduce, but 
 does not refer to the double jet type, although a fair ap- 
 proximation for the other class. 
 
 In the following Table is given the weights of water that 
 should be delivered per pound of steam by a well-designed 
 injector at the steam pressures ordinarily used: 
 
 TABLE VIII. 
 
 Gauge Pressure. W. 
 
 10 Ibs. 33 Ibs. 
 
 20 " 29.2 " 
 
 40 " 23 " 
 
 70 " 17.8 " 
 
 Gauge Pressure. W. 
 
 loo Ibs. 15.0 Ibs. 
 
 120 " 13.6 " 
 
 150 " 12.6 " 
 
 200 " 10.3 " 
 
 Thus far the mechanical action alone has been discussed, 
 but as the heat received by the feed water during the process 
 of injection is one of the chief advantages of the system of 
 feeding upon which the injector is based, the consideration 
 of the thermal action will now be taken up. 
 
 In order to supply the energy necessary to give the final 
 velocity to the particles of steam during expansion, an 
 amount of heat must be absorbed corresponding to the work 
 performed ; according to the theory assumed under which 
 expansion takes place, all the heat required must be fur- 
 nished by the steam itself, and as the greater part of the heat 
 contained in the steam is latent, a portion of the gas must 
 be condensed ; so that in its final state the steam is composed 
 of a mixture of steam and water, in proportions which may 
 be determined by equation (12). After expanding to the 
 
80 THE GIFFARD INJECTOR. 
 
 pressure within the combining tube, the steam strikes against 
 the cold water, its momentum is checked and its energy of 
 motion transformed to heat ; this is at once absorbed by the 
 feed, as is also the remaining sensible and latent heat in the 
 steam mixture. It will be shown that all the energy that is 
 apparently wasted, reappears in the form of heat, so that 
 the injector may be considered as a theoretically perfect de- 
 vice for boiler feeding purposes, and its thermal efficiency as 
 unity ; the actual number of heat units that are lost by 
 radiation that can be charged directly to the instrument 
 itself is inappreciable compared with those contained in the 
 delivered water. 
 
 Referring again to figures, take an injector which receives 
 steam at 120 pounds and feed water at 68, forcing 13 pounds 
 of water per pound of steam against a permissible counter- 
 pressure of 141 pounds, at a delivery temperature of 150, 
 and apply the analysis of the transfer of heat just outlined. 
 
 During the expansion of i pound of steam from 135 pounds 
 to 4 pounds (absolute) there is an amount of energy de- 
 veloped that is represented by U 'in equation (n) or, by the 
 term under the radical, in (13). Taking the latter equation, 
 and substituting the figures previously used, and retaining 
 the result in heat units, 
 
 *i r i + 4i ~ *" r >2 Q-2 2 37-07 Thermal Units (24) 
 
 The water flowing into the combining tube possesses a 
 small amount of energy which may be represented by 
 
 S X % = -W ( 34 + k ~ *'> =iVto + -9-4) 
 =0.75 Thermal Units . (25) 
 
 where 33 is the head of water in feet equal to the vacuum, h 
 the head under which the feed flows to the injector and has 
 a minus sign if the water is lifted ; h^ the absolute pressure 
 in the combining tube expressed in feet ; v. 2 is the velocity 
 of entrance; /z, i foot. 
 
 These two equations represent the motive power in the 
 injector expressed in thermal units, one of which is taken as 
 equal to 778 foot pounds of work. This energy is used for 
 
THE ACTION OF THE INJECTOR. 81 
 
 forcing the mixture into the boiler with a velocity v, which 
 
 requires 
 
 (i + W) v> __(!+ W) 
 ~T78" *^- ~7~78" '* l) 
 
 = -^ X ( 145 - 4 ) 2.355 - 5-97 Thermal Units .... (26) 
 
 H is the absolute permissible counter-pressure in feet of 
 head, and 2.355 is the head of water at a temperature of 
 150 equal to i pound per square inch from Table II. This 
 value for the useful work performed 5.97 thermal units is 
 very small compared with the energy expended ; adding (24) 
 and (25) and subtracting (26) , 
 
 ( 2 37- 7 4- -75 ) 5-97 = 231.85 Thermal Units. 
 
 which must be absorbed by the feed water. The increase in 
 temperature is therefore 231.85 -f- 13 = 17.83 per pound 
 of feed water due to the loss of actual energy of the steam 
 at the time of impact. Expressed algebraically 
 
 which represents the increase in the temperature of the de- 
 livered water due to the absorption of the remaining energy 
 of the steam jet. 
 
 When intimate contact is established between the steam 
 and the water, condensation is effected, and the latent and 
 sensible heat of the steam transferred to the feed. At this 
 time, the heat in the expanded steam is 
 
 * r* + ft + 32 t 2 ............... (28) 
 
 an equation of the same form as (24), but containing the final 
 temperature of the delivery, as the temperature of the steam 
 mixture cannot fall below that value. (The specific heat 
 of water between 32 and 212 is so nearly constant, that 
 the temperature of the feed is substituted for the ' ' heat of 
 the liquid," as the change simplifies the use of the formula, 
 and the difference in the result is inappreciable.) 
 
 From (12) and also from the calculations made on page 64, 
 we find the weight of steam remaining in i pound of initially 
 6 
 
82 THE GIFFARD INJECTOR. 
 
 dry steam after expanding from 135 to 4 pounds absolute, is 
 0.8243; from the Steam Tables, r, the heat of vaporization 
 = 1007.2; therefore, 
 
 1007.2 x 0.8243 830.23 
 121.09 ( 150 32 ) 3.09 
 
 833.32 Thermal Units. 
 
 remaining in the steam, and given out to the 13 pounds of 
 water during condensation ; this will raise the temperature 
 833.32 -f- 13 = 64.1, or in different form, 
 
 (29) 
 
 so that the sum of 64.1 and 17.83, will be equal to the 
 total rise in the temperature of the feed as the water passes 
 through the injector, or, 
 
 Adding these two values 17.83 -f 64.1 = 81.93 as the total 
 increase in temperature, the delivery temperature will be 
 68.0 + 81.93 = 1W9-93 
 
 which is the same as found experimentally. 
 
 Expressing these results algebraically, by adding (24) (25) 
 and (29), subtracting the work required to force the mixture 
 into the boiler (26), and equating to the increase in the tem- 
 perature of the feed water (30), 
 
 f r + xr )+_?!( + a_;n- 
 
 Reducing and cancelling and changing the form in order to 
 find the value of W, we have 
 
 q, ~ t, ~ -~ 
 
 (32) 
 
 H- 34 - h 
 
 4 ~ /lH " 778 
 
 From this equation have disappeared the terms denoting the 
 
 heat in the steam after expansion, and the important func- 
 
THE ACTION OF THE INJECTOR. 83 
 
 tions remaining, are the total heat in the steam in its initial 
 condition and the temperature of the feed and delivery. Let 
 P denote gauge pressure ; then, discarding unessential terms, 
 
 *-, r v + ft + 32 - /, - (.003) P 
 ~^-^+~(^3}P~ 
 
 This equation enables us to determine the weight of water 
 delivered per pound of steam by any injector by simply ob- 
 serving the temperatures of the feed and delivery and the 
 initial pressure of the steam, and substituting the proper 
 values from any Steam Table ; or, knowing the area of the 
 steam nozzle, the discharge may be calculated by (12), 
 which, multiplied by the ratio W, gives the capacity of the 
 injector. 
 
 As brass is a good conductor of heat, there is always a 
 certain amount of heat transmitted through the walls sepa- 
 rating the steam and feed chambers, which increases the 
 temperature of the delivered water. This heat, being ab- 
 stracted from the steam, renders the discharge through the 
 nozzle more or less wet, and reduces the quantity and velocity 
 at the time of impact ; if excessive, it seriously affects the 
 ability of the injector to receive feed water at a high temper- 
 ature, and also diminishes the range of capacities. This 
 heat appears in the difference (/ 2 O, and this term may be 
 taken as a good test of the practical working qualities of an 
 injector. 
 
 Probably the most satisfactory formula that could be 
 devised to cover the action of the injector, would contain 
 terms denoting the volumes of the steam, feed water and the 
 combined jet, and based upon the exchange of momenta 
 between the moving masses. If limited to the simple case 
 of complete condensation of the actuating steam, this could 
 easily be solved, and this analysis was in fact, followed in' 
 the discussion at the opening of this chapter, and the sizes 
 of the various orifices calculated ; but where only partial 
 condensation is effected an uncertain element is introduced 
 that can only be approximated by a formula based upon the 
 phenomena supposed to take place within the combining 
 tube and then introducing coefficients to make the results 
 
84 THE GIFFARD INJECTOR. 
 
 conform to actual experiments. The percentage of steam 
 mixed with the water at the time of entrance to the delivery 
 tube, and uncondensed, cannot easily be determined, as the 
 temperature of the jet at that time is lower than when 
 measured in the boiler pipe, and both the temperature and. 
 the pressure should be known for its correct determination. 
 The mechanical efficiency of the injector can be found by 
 comparing the results of some of the equations already 
 given ; in the previous example used to illustrate the form- 
 ulae, the work of forcing 14 pounds of water against a head 
 of 141 pounds, amounted to 5.97 thermal units, whereas the 
 total energy expended, which is the sum of (24) and (25) = 
 237.82 thermal units, so that the efficiency is 
 
 But the remainder of the energy is not wasted, but trans- 
 formed to heat and absorbed by the feed water. 
 
 The maximum capacity of an injector is almost entirely a 
 question of proper proportions and efficient action of the 
 impinging jet, so that its value may be determined by equa- 
 tion (21). The minimum however is determined by the 
 admissible temperature of the delivered water that will pass 
 through the delivery tube without overflow ; it should then 
 be determined by (33) substituting for 4 a value correspond- 
 ing to the type of injector. For a single overflow injector, 
 / 2 would be about 190, but for two openings from the combin- 
 ing tube into the same chamber, probably 215, at 60 or 80 
 pounds pressure. If, as an example, the case is taken of an 
 injector receiving steam at 60 pounds, feed 70, and a de- 
 livery temperature of 190, we have from (33), 
 
 *-, n + ft (190 32) 898.8*1- 276.9 - 158 
 -- ~" 
 
 If the delivery temperature at the maximum capacity is 125, 
 the value of W would be 19.69, and the ratio of the mini- 
 mum to the maximum would be 
 
 0>i 8.48 
 
 _J- X ico = - 6 - X ioo = 43 per cent., 
 
THE ACTION OF THE INJECTOR. 85 
 
 which is fairly good, although if the injector were so con- 
 structed that the highest delivery could reach 212, the 
 minimum would be reduced to 7.23, which would be 36 per 
 cent, of the full capacity. 
 
 The highest admissible feed temperature is difficult to cal- 
 culate without determining the special coefficients for each 
 case, which can only be done experimentally. The w T armer 
 the feed water, the greater the weight necessary to condense 
 a given weight of steam and, therefore, the heat equation 
 must be used ; on the other hand, the power of the steam 
 for forcing water through the delivery tube is diminished on 
 account of its reduced terminal velocity owing to the change 
 in pressure w 7 ithin the combining tube, so that its mo- 
 mentum, and the rapidity of its absorption by the water jet 
 is lessened and the density of the mixture reduced. Equa- 
 ting the equation of momentum (21) to (33) and solving for 
 A gives, 
 
 -zO-(* 1 n + ?i + 32-' 2 ) (v KvJi 
 ~ 
 
 By way of illustration, take experiment (</) in Table VII ; 
 under these conditions, the mean vacuum in the combining 
 tube is io|", or 9^- pounds absolute pressure, which gives a 
 terminal velocity to the steam of 3050 feet per second. The 
 water entering the combining tube has a velocity v t due to 
 the low internal pressure which is equal to 
 
 ( 14.69 9.5 ) X 2 31 = 27.76 feet 
 As we are determining the extreme capacity of the instru- 
 ment, it will be assumed that the mixture has just sufficient 
 power to enter the boiler ; its velocity, v t .will therefore be 
 
 * 34.69 9o ) X 2.42 =139-7 feet ; 
 from the table, K = 0.473 an ^ the delivery temperature 
 / 2 = 226 ; substituting these values in (34), 
 
 226 ( 473 x 3 5 ~ ^ 9 - 7 ^ 
 
 o-473 X 3050 139-7 
 
 which agrees very closely with the result found. 
 
86 7*HE GIFFARD INJECTOR. 
 
 The actual cause of the inefficiency of the injector from a 
 mechanical point of view, is due to its dependence upon the 
 principle of impact for the transfer of the actual energy of 
 the steam to the moving water; this induces such an 
 enormous waste of power that its use as a pump for raising 
 water is out of the question except in such cases where 
 economy of steam is no object, or where the heating of the 
 delivered water may be advantageous. The momentum of a 
 body is equal to the product of the mass by the velocity, and 
 the resultant velocity after impact, is equal to the sum of the 
 momenta of the two bodies divided by the sum of their 
 masses ; on the other hand, the actual energy is proportional 
 to the square of the velocities. Kven if there were no loss 
 of momentum in the injector, yet there would be a dissipa- 
 tion of energy, owing to the reduction of velocity ; if one 
 pound of vSteam moving at a rate of 3000 feet per second 
 strikes 9 pounds of water at rest, and imparts all its 
 momentum, the resultant velocity of the mixture would be 
 300 feet per second, yet the energy of the jet of water and 
 steam would only be -^ of that of the initial steam ; for, 
 
 Energy of steam (3ooo) a X I ~ 10 
 
 so that the greater the difference between the velocity of the 
 steam and the final velocity of the delivery, the greater the 
 loss of actual energy ; this also shows that the lower the 
 terminal velocity, the gfeater may be the weight of water 
 delivered, so that the equation simply measures the efficiency 
 as a pump and not as a boiler feeder. 
 
 As an example, the efficiency of the jet will be calculated 
 in the case of experiment (a) Table IV ; the velocity of the 
 steam is 3435 feet, of the entering water, 40 feet, and of the 
 jet, 165.3 feet; the weight of water delivered per pound of 
 steam, 12 pounds; therefore, 
 
 C(3435) 2 ,(40) 2 X 12-1 " (i6S-3J!_ VI 
 t I -/V i /\ !$ 
 2g 2g "2g 
 
 KI 3 per cent, 
 
 and from this value there is not much variation throughout 
 the ordinary range of steam pressures. 
 
THE ACTION OF THE INJECTOR. 87 
 
 Probably the most satisfactory way to compare the in- 
 jector as a boiler feeder with a direct acting steam pump, 
 will be to take an example in which the actual coal con- 
 sumption in the two cases can be compared. No allowance 
 will be made for the advantages which may be claimed for 
 either method in the way of convenience of operation or for 
 economy of maintenance. 
 
 A steam boiler evaporates 10 pounds of water from and at 
 212 per pound of coal, and it is required to deliver 3000 
 pounds of steam per hour at 80 pounds pressure ; water 
 supply at temperature of 63 ; 
 
 (a) The Direct Acting Steam Ptimp. 
 
 As the boiler receives the feed water at 63, the actual 
 evaporation is 
 
 ---- = 8.40 Ibs. water per pound of coal ; 
 1213 63 
 
 where 966 is the latent heat of one pound of water at 212, 
 and 1213 the total heat above o of i pound of steam at 80 
 pounds gauge pressure. The work of forcing 3000 pounds 
 of feed water into the boiler per hour, or 50 pounds per 
 minute, against a head of 80 X 2.31 feet is, 
 
 33000 
 
 where 2.31 is the height of a column of water at a tempera- 
 ture of 63, equal to a pressure of i pound per sq. in. from 
 Table II. According to experiments made by Prof. D. S. 
 Jacobus, of the Stevens Institute of Technology, a direct- 
 acting steam pump requires under these conditions 150 
 pounds of steam per horse power per hour ; accepting this 
 figure, an additional weight of feed water is necessary to 
 supply the steam for the pump equal to 
 
 150 X 0.28 = 42 Ibs. per hour, 
 
 and to force this weight of feed water into the boiler requires 
 0.585 pounds of steam, determined as before, making an 
 approximate total of 42.585 pounds of steam. This can be 
 calculated precisely by the following equation, where y is 
 the total weight of steam required for feeding the boiler : 
 
88 THE GIFFARD INJECTOR. 
 
 50 + -r- ) 80 X 2.31 X 150 
 
 =y = 42.596 Ibs. feed per hour. 
 
 33,o 
 The total weight of coal used by the boiler is 
 
 jooojf J2.596 = 362 . 2I lbs. coal per hour. 
 
 8.4 
 
 () The Injector. 
 
 Feed water is received by the injector at 63 and delivered 
 to the boiler at 123.4. The evaporation is, therefore, 
 
 9 -- - ? = 8.865 Ibs. of water per pound of coal. 
 1213 123.4 
 
 The coal required to supply 3000 pounds of steam per'hour, 
 -||g- = 338.41 Ibs. per hour. 
 
 Work of forcing 50 pounds of water per minute into the 
 boiler, 
 
 8o_x _a.33_X 5? = Q 28 horse 
 33,000 
 
 where 2.33 is the height of an equivalent column of water 
 at the new feed temperature of 123.4. 
 
 As i horse power = 33,000 foot pounds, and i thermal unit 
 = 778 foot pounds, i horse power is equal to 42.41 thermal 
 units ; therefore the coal required per hour for the work of 
 feeding the boiler with the injector is 
 
 0.2824 x 42.41 X 60 
 
 fil -74 lbs - P er hour. 
 
 Coal required to supply the steam that is applied by the 
 injector to warming the feed water, 
 
 3^! = ,8.78 Ito. per hour. 
 
 The total coal used in this case is 
 
 338.41 -j- 0.074 + 18.79 = 357- 2 64 lbs. per hour. 
 Saving of coal when the boiler is fed by the injector, 
 362.21 357.264 = 4.94 lbs. per hour. 
 
 !& per cent. 
 
 This saving of 4.94 pounds of coal per hour corresponds 
 to the heat wasted in the exhaust steam from the pump and 
 which is returned to the boiler by the injector. 
 
UNIVERSITY 
 
 CHAPTER VIII. 
 
 APPLICATIONS OF THE INJECTOR FOREIGN AND AMERICAN 
 
 PRACTICE DESCRIPTION OF VARIOUS PATTERNS 
 
 OF INJECTORS. 
 
 A TECHNICAL classification of the various patterns of 
 injectors would be made according to the essential features 
 embodied in the construction, similar to the method pursued 
 in Chapter II, but as far as practical considerations are con- 
 cerned, the distinctive feature is the presence or absence of a 
 partial vacuum in the suction pipe at the time of starting. This 
 may be produced by an auxiliary jet of steam, or by introduc- 
 ing the proper proportion between the area of the main steam 
 nozzle and the upper overflow in the combining tube ; either 
 of these systems produces a vacuum in the feed pipe prelimin- 
 ary to starting, and constitutes the inherent characteristic of 
 the Lifting Injector ; the other class, requiring a pressure 
 on the water supply, is termed Non- Lifting, or N. L. and 
 permits a simplicity of construction that is seldom equalled 
 in the other form ; either class may be self-adjusting, single 
 or double jet, although the N. L. class is usually made in 
 the more elementary forms. 
 
 The injector as a boiler feeder has been successfully ap- 
 plied to all kinds of service, locomotive, stationary, marine, 
 traction engines, etc., although it is to the first named that 
 it has proved itself most useful, and for locomotive boilers 
 has superceded all other devices. For stationary service, 
 the non-lifting is used whenever a head of water is obtain- 
 able, principally on account of the simplicity of its operation. 
 For large marine boilers, the injector is chiefly used in case 
 of the disability of the pumps, but for small boats, the re- 
 starting pattern has been extensively adopted. This kind 
 
 89 
 
90 THE GIFFARD INJECTOR. 
 
 is also used on traction, logging, and road engines, where its 
 certainty of action and special adaptability, render it invalu- 
 able for the rough work to which such machines are applied. 
 
 The usual division of all the different forms of injectors 
 resolves into a question of location relative to the water 
 supply. A non lifting injector is usually tested under a 
 head of about two feet, and will operate at all higher pres- 
 sures by regulating the feed valve ; this does not, however, 
 give an increased capacity unless the steam pressure is 
 raised. For the locomotive injector, the steam pressure 
 under which it is tested and for which it is proportioned is 
 about 120 pounds, while for stationary boilers, 60 pounds is 
 substituted, although either arrangement will permit a wide 
 range of working pressures. A lifting injector will of course 
 operate with the feed water flowing to it, but the N. L. 
 form will not work under the reverse condition unless it is 
 first primed, and then only with reduced capacity, because 
 the entrance area to the combining tube is too contracted. 
 
 Although there are many different styles specially designed 
 for locomotive service, almost all correspond to one of two 
 standards of capacity and position, of the steam, feed and 
 delivery connections; these are derived from the form of 
 injector manufactured by two of the best known companies, 
 viz., the Sellers' or P. R. R. standard, and the Monitor, or 
 New York standard. This is advantageous in many ways, 
 as it renders injectors interchangeable, and permits a careful 
 and accurate comparison of the relative merits of the various 
 kinds in use. 
 
 Neither in England nor on the continent is there any 
 attempt at uniformity, nor is the construction limited to 
 regular manufacturers ; several railroads have designed and 
 now make injectors for their own use, especially adapted to 
 the requirements of their own motive power ; of the two 
 classes, the N. L. is more generally adopted, and it is 
 usually of clumsy design and differs materally from those 
 in use on American railroads. Fig. 18 is a drawing of the 
 form used by the Great Eastern Railway of England and is 
 applied to both passenger and freight engines. The com- 
 
APPLICATIONS OF THE INJECTOR. 91 
 
 bining tube is heavily ribbed and forms the body, having its 
 upper end spread into a double flange, to which are attached 
 the steam and feed connections ; on one side of the main 
 casting is a lazy cock and on the other a heavy flange, the 
 casing being made either right or left hand depending upon 
 the side of the locomotive for which it is intended as it is 
 bolted in a horizontal position to the frame, below the foot- 
 board. Large numbers of injectors answering to this general 
 description are in use, yet no real standard prevails, and it 
 appears that the English railway managers have not yet 
 determined to their own satisfaction the most advantageous 
 position upon the locomotive. Many manufacturers carry 
 
 Fig. 18 
 
 in stock all conceivable forms of casings for their standard 
 tubes, adapted to be attached to any kind of locomotive in 
 any possible position. Many of the original Giffard in- 
 jectors are yet in use, and are still giving satisfaction ; they 
 are usually placed directly back of the fire box, and part 
 way through the foot plate. 
 
 The most advanced English practice favors bolting the 
 injector on the back plate of the boiler, with self-contained 
 steam and check valves, and with steam and delivery pipes 
 within the boiler. This arrangement is neat and compact. 
 and as all parts are protected against accident, possesses 
 some advantages over the practice in America of outside 
 
THE GIFFARD INJECTOR. 
 
 connections. Objections urged against this system include 
 the heating of the suction pipe and consequent difficulty in 
 starting, and also the rapid closing of the delivery pipe 
 within the boiler when the feed water contains lime or cal- 
 careous material in solution. This is due to the great 
 difference in temperature between the water in the boiler 
 and the water within the pipe, which causes a separation of 
 the solids held in solution and a deposit on the interior sur- 
 
 Fig.19. 
 
 FEtO 
 
 STEAM 
 
 WATER 
 
 OVE.RFLOW 
 
 face of the pipe. With clean water these injectors have 
 given good satifaction, and as they are growing in favor, 
 one made by Gresham & Craven of Manchester is shown in 
 Fig. 19. This form is inverted, the steam passing from the 
 dome through the pipe A' to the chamber A ; steam dis- 
 charging upward from the nozzle a, raises the water without 
 the necessity for a priming spindle, as the sliding combining 
 tube b is forced up, exposing the large upper overflow of the 
 
APPLICATIONS OF THE INJECTOR. 93 
 
 draft tube b' ; the entrance of the feed water and the ensuing 
 partial vacuum draws the tube down against its seat and pre- 
 vents admission of air, while the jet passing through the 
 delivery tube c is discharged at the front end of the boiler 
 by the copper pipe C. This instrument belongs to the 
 re-starting class and is easily primed ; all the tubes may be 
 removed when the boiler is under steam by closing the at- 
 tached valves, and removing the top cap D. 
 
 Another form of injector that is somewhat similarly ar- 
 ranged, is extensively used on the London & Northwest 
 Railway, and was designed by the Superintendent of Motive 
 Power, Mr. F. R. Webb. It consists of an inverted set of 
 tubes connected to a check valve on the back of the boiler, 
 and lying as close to it as possible, but set low down and 
 exposing but a short length of suction pipe to the heating 
 action of the fire-box ; the combining and delivery tubes are 
 rigidly connected and moved to or from the steam nozzle by 
 a convenient handle actuating a quick thread screw. The 
 parts are set so that the water from the tank will flow 
 through the overflow, so that no lifting device is neces- 
 sary. 
 
 In France and in Italy, non-lifting injectors of the Fried- 
 man type are largely used in the latter country, almost 
 exclusively. From information furnished by the Com- 
 pagnie de 1' Est of France fully one-half of the 2441 injectors 
 in use in 1887, upon its lines were of the Friedman pattern, 
 and upon the Chemin de Fer de 1' Etat, over 1000 of the same 
 style were in service in sizes varying from 6 to 10 millimetres. 
 During the last few years, however, many of the French 
 railroads have been experimenting with injectors of Ameri- 
 can manufacture, and this has led to the extensive use by 
 several of the companies of the Sellers' " Self- Acting " In- 
 jector, which is now replacing those of foreign design. This 
 injector is shown in Fig. 26. 
 
 In Germany and Austria, a diversity of styles prevail, the 
 greater number being non-lifting ; even the Korting double 
 jet is there applied beneath the foot board, while in this 
 country it is used almost exclusively as a lifter. Other well 
 
94 THE G1FFARD INJECTOR. 
 
 known patterns are the Friedman, Hasswell, Pohlmeyer, 
 Schaeffer & Budenberg, and Bohler. 
 
 Returning to American practice, it may be said without 
 question, that the lifting injector is the standard class, as 
 few railroads, except in some parts of the West and South, 
 use the other forms. In certain localities where the feed 
 water is impregnated with lime, non-lifters are retained on 
 account of the facility with which they can be cleaned, and 
 they are frequently used in combination with the other form, 
 one being placed on each side of the engine. The chief 
 disadvantage urged against this injector is that it is out of 
 reach of the hand and eye of the engine-driver, and a 
 sudden drop in the steam pressure or a careless opening of 
 the lazy cock when the steam is shut off, may drain all the 
 water out of the tank. This difficulty can be remedied by 
 adding an extension to the overflow nozzle and raising it 
 above the water level, and placing it where it may be con- 
 tinuously observed. This has been applied by the Nathan 
 Manufacturing Co., of New York, to a modified form of 
 their W. F. Injector, shown in Fig. 22. All non-lifting in- 
 jectors should be placed so as to receive the feed water under 
 as high a head as possible, as the action is thus rendered 
 more certain, and the operation of starting, when the pipes 
 are warmed by leaky steam valves, is much facilitated. 
 
 The almost invariable position of the lifting injector upon 
 a locomotive is directly in front of the engine-driver, placed 
 part way through the front cab frame so that the operating 
 handles will be in convenient reach ; the second injector 
 usually occupies a similar position on the other side. On 
 the Philadelphia & Reading R. R., where the camel-back 
 engine is much used, the second injector is set through the 
 back frame of the cab and within easy reach, so that both 
 feeders are under instant command. Both the manner of 
 placing the injectors upon locomotives, and the design of 
 the instruments themselves, seem to be more carefully 
 thought out in this country than elsewhere unless the 
 most recent English practice be an exception and American 
 manufacturers are still striving to bring the injector into 
 
APPLICATIONS OF THE INJECTOR. 
 
 95 
 
 even more convenient shape and to render its action more 
 efficient. 
 
 In order to show clearly the actual construction of the 
 interior of the injector, selections have been made from the 
 various patterns in general use and illustrations will be given 
 of those most widely known. 
 
96 
 
 THE GIFFARD INJECTOR. 
 
 Fig. 20 is a sectional view of the 
 
 SELLERS' INJECTOR OF 1876, 
 
 which contains the ingenious self-adjusting feature described 
 and illustrated on page 15; it is conveniently constructed 
 for operating and for repairing, and is started, regulated and 
 stopped by means of a single lever, requiring no hand ad- 
 justment for any variation in the pressure of the steam, 
 height of lift, or temperature of the feed water. The lifting 
 is effected by drawing the lever H back a short distance 
 until the resistance of the main valve Xis felt; this admits 
 steam to the hollow spindle C, which discharges through 
 the delivery tube D, and produces a partial vacuum in the 
 suction pipe ; when water appears at the overflow, the lever 
 is pulled all the way back, drawing with it the connecting 
 rod Z,, closing the waste valve K and forcing the jet to enter 
 the boiler ; the regulation of the capacity is effected by ad- 
 justing the position of the lever by means of the steel latch 
 V, on the guide rod/. Moving the lever changes the area 
 of the steam nozzle by altering the distance which the taper 
 spindle is inserted, and any variation in the weight of steam 
 discharged, immediately induces an automatic movement 
 of the combining tube which preserves the correct ratio 
 between the water and the steam. An air chamber M, 
 cored in the body, connects with the column of water in 
 the suction pipe, so that all shocks and jars will be ab- 
 sorbed by the elasticity of the air, and will not be trans- 
 mitted to the tube, and cause the injector to '"break" or 
 " fly off." 
 
 The following Table furnished by the makers gives the 
 maximum capacity of this injector in cubic feet per hour: 
 
 
 Size of Injector. 
 
 Steam Pressure. 
 
 No. 3. 
 
 No. 4. 
 
 No. 5. 
 
 No. 6. 
 
 No. 7. 
 
 No. 8. 
 
 No. 9. 
 
 No. 10. 
 
 6olbs. . . . 
 
 30 
 
 55 
 
 86 
 
 118 
 
 161 
 
 210 
 
 269 
 
 332 
 
 120 "... 
 
 40 
 
 70 
 
 112 
 
 1 60 
 
 218 
 
 288 
 
 357 
 
 445 
 
APPLICATIONS OF THE INJECTOR. 
 
 97 
 
 THE MONITOR INJECTOR. 
 
 This is one of the best known of the locomotive injectors, 
 and is a modification of Friedman's arrangement of tubes, 
 with an improved casing. The combining tube is divided 
 into two parts, forming an upper overflow at a point in the 
 tube where the cross-section is about the same as that of the 
 steam nozzle ; this enables the injector to be started with 
 great facility, and the air to be drawn from the suction pipe, 
 by means of the independent ejector, quickly and promptly. 
 
 FIG. 21. " MONITOR.' 
 
 H 
 
 The manipulation is as follows : opening the handle J t 
 admits steam to the lifting jet which discharges through the 
 waste pipe, creating a strong vacuum ; when water appears 
 at the overflow, the handle 5" opens the main steam jet and 
 the water is forced through the delivery tube into the boiler ; 
 with high pressure steam, this draws all the water away 
 from the lifting jet, and clear steam blows from the waste 
 pipe which is checked by closing J\ with low steam the 
 water supply may have to be throttled by the regulating 
 valve W in order to have the injector "run dry.'' This 
 valve is also used to adjust the capacity to suit the require- 
 ments of the boiler, and permits of considerable variation. 
 7 
 
98 
 
 THE GIFFARD INJECTOR. 
 
 H31108 01 
 
APPLICATIONS OF THE INJECTOR. 
 
 99 
 
 A newer form of this injector is that shown in Fig. 22, 
 which can be operated with a lever or a screw handle on the 
 steam spindle, and obviates the necessity for the starting 
 jet /, by the use of the taper lifting-spindle inserted in the 
 steam nozzle which discharges through the combining tube. 
 The steam nozzle is made larger than in the form shown in 
 Fig. 21, which enables it to operate at lower steam press- 
 ures, while a movement of the lever forces the taper spindle 
 into the steam nozzle and reduces the area to a size better 
 suited to high pressures; the range of steam pressures 
 through which the injector will operate satisfactorily is 
 much increased by this change. 
 
 The body is made in three parts, bolted and screwed 
 together, and is heavily and strongly constructed through- 
 out. The following table of capacities is reduced from that 
 given in the catalogue of the Nathan Manufacturing Co., 
 and applies to all the forms of the Monitor Injector, and 
 also to the WF pattern, Fig. 23. This table gives the 
 capacity in cubic feet per hour at a steam pressure of 140 
 pounds per square inch, feed temperature 60 to 70 Fahr., 
 when the injector is placed at the usual height of lift on a 
 locomotive : 
 
 Steam. 
 
 No. 4 . 
 
 No. 5. 
 
 No. 6. ' No. 7. 
 
 No. 8. 
 
 No. 9. 
 
 No. 10. 
 
 No. ii. 
 
 No. 12. 
 
 140 Ibs. . 
 
 73-3 
 
 126.6 
 
 1 66 240 
 
 306 
 
 386 
 
 453 
 
 5 06 
 
 560 
 
 Made by the same company is the 
 
 NATHAN WF INJECTOR, 
 
 designed to be placed below the foot-board with water flow- 
 ing to it. The construction, as shown in Fig. 23, is very 
 simple, and the steam nozzle and combining tube may be 
 easily removed. Steam enters at the top of the casing and 
 the water passes upward from the bottom through a cored 
 passage to the entrance to the combining tube ; the opening 
 to the overflow valve is indicated in the cut by the circle 
 between the combining and the delivery tube, and the 
 
100 
 
 THE GIFFARD INJECTOR. 
 
 FIG. 23. NATHAN W F. 
 
 valve body may be placed on 
 either side of the body, the 
 opposite opening being closed 
 with a screw cap. The pro- 
 portions of the tubes and the 
 arrangement of the overflows 
 are similar to those of the 
 Monitor pattern, and the list 
 of capacities is the same. 
 
 A newer form of the non- 
 lifting injector has been pro- 
 vided with a waste pipe ex- 
 tending above the water level 
 of the tank and into the cab, 
 where it may be under the 
 constant observation of the 
 engine driver, who is thus 
 able to notice and check any 
 tendency to waste. The tubes 
 are arranged more in the form 
 of those shown in Fig. 22, and a spindle is sometimes pro- 
 vided to facilitate starting. 
 
 Fig. 24 is a sectional view of the American form of the 
 " Korting Universal," the 
 
 "SCHUTTE INJECTOR," 
 
 which belongs to the double-jet type. It has no overflows 
 in either the lifting or the forcing combining tubes, as 
 the starting of the jet is effected by direct communica- 
 tion with the air through the compound waste valve 
 placed vertically over the nozzle B. With this instru- 
 ment the manipulation is as follows : Drawing the lever 
 from A to B transmits the pull to *the valve over the 
 lifting steam nozzle by means of the bar held loosely 
 between the caps K K ; as the valve of the forcing 
 set has a larger area, it is held by the steam pressure 
 firmly to its seat, and acts as a fulcrum for the lever. 
 After the appearance of the water at the waste pipe, the 
 
APPLICATIONS OF THE INJECTOR. 
 
 101 
 
 handle is drawn further out, and the vertical overflow 
 valve, connected with the starting lever by means of a rod 
 and bell-crank not shown, cuts off the outlet of the first set 
 and diverts the water through the forcing combining tube. 
 The second steam nozzle is then opened by the continued 
 movement of the starting lever, which, during the latter 
 part of its stroke, closes the waste valve ; the jet then enters 
 the boiler. Although the operation has been divided into 
 several steps, the motion of the lever is practically contin- 
 uous and the starting very simple. An elementary form of 
 
 FIG. 24. "SCHUTTE." 
 
 this injector was shown in Fig. 4, where the principle em- 
 bodied in its construction was explained ; reference has 
 been already made to the peculiarities of construction w^hich 
 enable this type of instrument to receive water at a high 
 feed temperature and operate through a wide range of steam 
 pressures ; at 60 pounds steam, the permissible temperature 
 of the feed water, lifted 5 feet, is 150, and at 125 pounds 
 steam, 135. 
 
 The capacity in cubic feet per hour is shown by the fol- 
 lowing table, based upon the following conditions : Feed, 
 80, lift 4 feet ; at 65, the values will be increased about 5 
 
102 
 
 THE GIFFARD INJECTOR. 
 
 per cent. ; on the other hand, when the height of lift is 10 
 feet, there will be a reduction of the capacity of 10 per cent. ; 
 at 15 feet, 25 per cent. 
 
 Steam. 
 
 No. i. 
 
 No. 2. 
 
 No. 3. 
 
 No 3^ 
 
 No. 4- 
 
 No. 5. 
 
 No. 6. 
 
 No. 7. 
 
 No. 8. 
 
 No. 9. 
 
 No. 10. 
 
 30 
 
 13 
 
 20 
 
 33 
 
 4 6 
 
 60 
 
 75 
 
 9 
 
 128 
 
 165 
 
 195 
 
 240 
 
 60 
 
 16.5 
 
 25 
 
 39 
 
 57 
 
 76 
 
 98 
 
 1 2O 
 
 150 
 
 195 
 
 240 
 
 2*5 
 
 120 
 
 21 
 
 29 
 
 48 
 
 66 
 
 86 
 
 in 
 
 136 
 
 195 
 
 25. s 
 
 300 
 
 345 
 
 150 
 
 24 
 
 3 1 
 
 53 
 
 73 
 
 96 
 
 123 
 
 15 
 
 215 
 
 282 
 
 33 
 
 380 
 
 In the injector illustrated, the capacity is adjusted to the 
 requirements of the boiler by means of a valve placed in the 
 suction pipe, but the more recent improvements include a 
 small taper spindle inserted in the steam nozzle of the lifting 
 set, and adjusted by a handle directly under the starting 
 lever. This varies the flow of steam in the lifting set, 
 changing the amount of water delivered, and materially re- 
 duces the steam consumption at the minimum capacity. 
 It is so arranged, however, that the lifting power of the 
 injector is in no way affected, for the water can be raised 
 and forced into the boiler even when the regulating lever is 
 set for the lowest delivery of the instrument. 
 
 THE BELFIELD INJECTOR 
 
 also belongs to the double-jet type, but differs from the pre- 
 vious form in having all the tubes in the same axial line ; 
 suction is produced by the hollow cylindrical spindle, which 
 is, however, never entirely withdrawn from the first steam 
 nozzle, so that the entering water receives impulses from 
 both the central and the surrounding jet. After passing the 
 first delivery tube, the feed water meets the discharge from 
 the annular second jet, opened by a second movement of the 
 starting lever ; the water then enters the boiler, but requires, 
 in order to obtain more positive action and a fair minimum 
 capacity, the tight closing of the waste valve, effected by the 
 final motion of the lever. Capacities at 125 pounds and 5 
 feet lift are given in the following table ; the size number is 
 arbitrary, and is based upon the capacity. The injector will 
 
APPLICATIONS OF THE INJECTOR. 
 
 103 
 
 operate over a wide range of steam pressures, and gives a 
 fair minimum. 
 
 Steam. 
 
 No. 3. 
 
 No. 4. 
 
 No. 5. 
 
 No. 6. 
 
 No. 7. 
 
 No. 8. 
 
 No. 9. 
 
 No. 10. 
 
 125 Ibs. . . . 
 
 40 
 
 73 
 
 "3 
 
 160 
 
 220 
 
 290 
 
 366 
 
 452 
 
 FIG. 25. " BEHELD INJECTOR." 
 
 A notable characteristic of the 
 
 SELLERS' SELF-ACTING INJECTOR OF 1887 
 
 is the simplicity of its construction ; its special features are 
 obtained by the arrangement of its overflows and the pro- 
 portions of its tubes, instead of by moving parts or special 
 valves. In common with the self-adjusting injectors already 
 described, and double-jet injectors in general, it adjusts and 
 maintains the proper supply of feed water for each pressure 
 of steam, and, in addition, re-starts automatically after a 
 temporary interruption of the water or steam supply. 
 
 As may be seen from Fig. 26, these results are obtained by 
 means of the following arrangement of parts : A small 
 
104 
 
 THE GIFFARD INJECTOR. 
 
 
APPLICATIONS OF THE INJECTOR. 105 
 
 annular steam nozzle discharges through the annular area 
 surrounding a central forcing steam nozzle, and through the 
 overflow spaces in the combining tube ; the proportions of 
 these parts are such that the lifting steam nozzle can always 
 produce a suction in the feed pipe even when there is a 
 discharge from the main steam nozzle, and it is this fact 
 that establishes the re-starting feature. When the feed 
 water rises to the tubes, it meets the steam issuing from the 
 lifting nozzle, and is forced by it in a thin sheet and with a 
 high velocity into the combining tube of the forcing set, 
 without the necessity for the interposition of a divergent 
 delivery tube. The water there comes in contact with the 
 main steam jet, receives an additional impulse, and the mix- 
 ture, reducing in cross-section as the velocity is accelerated, 
 rushes past the overflow openings in the combining tube and 
 passes through the delivery tube into the boiler. 
 
 The self-adjusting action is accomplished by the change of 
 the capacity of the first set with the steam pressure, assisted 
 by the influence of the partial vacuum in the combining 
 tube, which, being communicated to the .surrounding overflow 
 chamber through the apertures in the tubes, assists in draw- 
 ing an additional supply of feed water from the suction pipe 
 whenever the requirements of the jet are increased by a rise 
 of the boiler pressure. This maintains a constantly increas- 
 ing capacity with the pressure, and gives an exceptionally 
 large range of boiler pressures with which the injector will 
 work efficiently. 
 
 On account of the low temperature of the delivered water, 
 the weight of water forced into the boiler per pound of 
 steam is unusually high, especially at locomotive boiler 
 pressures; this small consumption of steam permits -the 
 grading of the capacity through a wide range, and at 120 
 pounds steam, when lifting i foot, the capacity may be re- 
 duced by throttling the feed valve to 36 per cent, of the 
 maximum under the same conditions ; this range of avail- 
 able capacities enables the injector to be used to feed the 
 boiler continuously by simply adjusting the water valve to 
 suit the requirements of the boiler. 
 
106 THE GIFFARD INJECTOR. 
 
 Not only is the range of capacities dependent upon the 
 minimum expenditure of steam to perform the work required, 
 but also the highest admissible temperature of the feed water ; 
 further, as this is dependent also upon the limiting tem- 
 perature of the mixture that will pass through the narrow- 
 est part of the delivery tube without overflow, it follows 
 that an injector that can maintain a continuous jet at a tem- 
 perature of 250 at the minimum capacity without wasting 
 from the overflow nozzle, freely opening to the air, is able to 
 use very hot water. Actual tests when lifting i foot at 1 20 
 pounds steam, show a limiting temperature of 124 for the 
 automatic action, and 136 when the waste valve is held to 
 its seat by throwing over the cam lever ; these results were 
 obtained with fairly dry steam and without special care in 
 handling ; regulation of the feed valve is the only movement 
 required to obtain the minimum capacity. 
 
 The most recent improvement in this form of injector is 
 shown in Fig. 26. A new check valve has been introduced 
 between the water supply and the overflow chamber for the 
 purpose of admitting a supplemental supply of water when- 
 ever there is a partial vacuum therein, caused by the incom- 
 plete condensation of the steam in the upper part of the 
 combining tube. It was found that an insufficient amount 
 of water passed through the annular lifting tube at pressures 
 above 160 Ibs., and produced a high vacuum within the over- 
 flow chamber. This vacuum is now utilized to draw in an 
 additional supply of water, which enters the lower tube and 
 passes with the jet into the boiler, increasing the capacity 
 about 24 per cent, at 1 80 pounds pressure. When the vac- 
 uum in this chamber is less than that in the suction pipe 
 this check valve closes automatically, preventing any steam 
 or hot water entering the suction pipe during the operation 
 of priming. 
 
 That it is very efficient at the high pressure is shown by 
 the accompanying table of capacities, which gives rapidly 
 increasing figures as the pressure is raised. The range is 
 good: 57 per cent, at 200 Ibs., and 63 per cent, at 120 Ibs. 
 
APPLICATION OF THE INJECTOR. 
 
 107 
 
 A full and complete test is given, page 134. This injector is 
 specially designed for locomotive service, and occupies but 
 little room in the cab ; upon removing the overflow sleeve 
 after unscrewing the jam nut, the body may be placed 
 through a round hole in the front cab frame, allowing only 
 the steam and feed pipes and manipulating handles to pro- 
 ject into the cab. Each size of injector is named from the 
 diameter of the delivery tube expressed in millimetres, and 
 the capacity is given in the following table in cubic feet per 
 hour at the ordinary steam pressures employed ; height of 
 lift. 5 feet. 
 
 SIZE. 
 
 30 
 
 6 7 
 
 60 
 
 89 
 
 I2O 
 
 160 
 
 i33 
 
 200 
 
 138 
 
 5& 
 
 121 
 
 ^ 
 
 98 
 
 129 
 
 I 7 6 
 
 i93 
 
 2O2 
 
 1\ 
 
 130 
 
 172 
 
 234 
 
 258 
 
 268 
 
 345 
 
 *i 
 
 167 
 
 221 
 
 301 
 
 33 2 
 
 9^ 
 
 209 
 
 2 7 6 
 
 376 
 
 410 
 
 425 
 
 101 
 
 255 
 
 338 
 
 460 
 
 505 
 
 527 
 
 i 
 
 305 
 
 405 
 
 55i 
 
 607 
 
 630 
 
 THE METROPOLITAN INJECTOR. 
 
 This injector is of the double-tube class. It is provided 
 with a closed overflow, and designed with a view of obtaining 
 a double- tube injector without outside attachments, rendering 
 the construction simple. The lifting apparatus lifts the water, 
 delivers it under pressure to the forcing apparatus, which in 
 turn discharges it into the boiler. There is a double steam 
 valve for successively admitting steam to the lifting appara- 
 tus and to the forcing apparatus ; to the stem of the main 
 steam valve the overflow valve stem is rigidly connected by 
 means of a cross-head. The overflow valve stem passes 
 
108 
 
 THE GIFFARD INJECTOR. 
 
APPLICATION OF THE INJECTOR. 109 
 
 through a cored passage in the injector, which opens directly 
 into the atmosphere through the overflow passage. By the 
 form of valve-stem used, and the means for operating it, an 
 overflow valve of the simplest type can be used. 
 
 The overflow valve seat is made removable from the casing, 
 thus facilitating repairs. The manufacturers state that this 
 injector can be started with 25 Ibs. steam pressure, and with- 
 out any adjustment or regulation of any kind, can be operated 
 at steam pressures up to 275 Ibs. Also, that these injectors 
 will take feed- water at 145 to 150 degrees F., with TOO Ibs. 
 steam pressure ; at 142 degrees F. with 125 Ibs. steam pres- 
 sure ; at 135 degrees F. with 150 Ibs. steam pressure ; at 130 
 degrees F. with 175 Ibs. steam pressure, and at 125 degrees 
 F. with 200 Ibs. steam pressure. The capacity is regulated 
 by increasing or decreasing the amount of steam to the 
 lifting apparatus, by adjusting the small hand wheel below 
 the operating lever 
 
 Its operation, briefly, is as follows : The lever being 
 drawn back slightly opens the steam valve to admit steam 
 to the lifting apparatus ; steam passing through the lifting 
 apparatus raises the automatic valve, which gives instant 
 relief, and thus enables the injector to lift the water promptly. 
 As soon as the water is lifted it takes the same course as the 
 steam, passes out through the overflow, thus filling the 
 water passages of the injector. As soon as water appears at 
 the overflow the lever is drawn back admitting steam to 
 the forcing apparatus, which seats the automatic overflow 
 valve, owing to the increased pressure on top of this valve, 
 and finally closes the overflow valve of the forcing appar- 
 atus, thus turning the water from the overflow into the 
 boiler. 
 
 The injectors already illustrated comprise the most promi- 
 nent in use upon American locomotives, and contain special 
 features, which are in many cases duplicated, though perhaps 
 in a less efficient degree, in other patterns ; special mention, 
 however, should be made of the Hancock Inspirator, Gar- 
 field, Mack, Eynon Evans and Ohio Injectors, but it would 
 
110 THE GIFFARD INJECTOR. 
 
 occupy too much space to illustrate these and all the other 
 patterns in use, and those given embody the most successful 
 application of the essential principles. Many of these in- 
 jectors are also used upon stationary boilers, yet they are 
 more generally applied to railroad service. 
 
 The following have been selected from the great variety 
 of injectors designed for stationary, traction and marine 
 boilers as containing special features of interest ; almost all 
 are well known, and are carefully designed and accurately 
 manufactured. 
 
 FIG. 28. 
 
 Almost all single-jet injectors designed for stationary or 
 marine service, are arranged to start through a lower over- 
 flow or aperture placed between the mouth of the delivery 
 tube and the lower end of the combining tube. An excep- 
 tion, however, is the 
 
 PENBERTHY SPECIAL 
 
 which is shown in Fig. 28, and described by the manufac- 
 turers as an Auto-positive Injector. The usual upper over- 
 flow, characteristic of re-starting injectors is used, but the 
 combining and delivery tubes are made continuous, without 
 spill holes or separating space. The jet, therefore, starts 
 through the delivery tube, as tte final or pressure chamber 
 H has free access to the air when the valve H, is raised from 
 its seat on the bushing M whenever the valve K opens. The 
 
APPLICATION OF THE INJECTOR. Ill 
 
 action of starting is as follows : Steam is admitted through 
 the upper right-hand branch to the nozzle X and discharging 
 through the draft-tube G, produces a vacuum in the suc- 
 tion pipe ; the discharging steam forces the automatic valve 
 K against the end of the pressure valve L,, which is thus 
 held open to permit free outlet of steam and water from the 
 chamber H ; this is made possible by having the area of K 
 about twice that of I,. When the correct proportions of 
 water and steam are admitted and the jet is formed, the pres- 
 sure in the chamber containing the upper overflow is changed 
 to a partial vacuum, which draws the automatic valve K to 
 its seat and allows the pressure valve L, to close, compelling 
 the feed to pass under the check valve P and into the boiler. 
 
 The construction of the injector is very simple and easily 
 understood. As it has no lower overflow, the action is sim- 
 ilar to that of a positive or closed overflow injector, so that 
 the overflowing temperature corresponds to the breaking 
 temperature of the ordinary single jet pattern. The range 
 of steam pressure is said to be good ; the minimum capac- 
 ity is obtained by throttling the water supply, for which 
 purpose a valve must be placed on the suction pipe. 
 
 A sectional view of the 
 
 SELLERS' RESTARTING INJECTOR. 
 
 is given in Fig. 29. The branches for steam, water supply 
 and delivery to the boiler are conveniently arranged, so that 
 all the pipes may be placed close against the boiler wall. 
 The overflow is directly under the water branch and can be 
 provided with a drip funnel and discharge pipe, without 
 bending or springing the other pipe connections. The 
 steam nozzle and delivery tubes are screwed into the body, 
 and do not depend upon the pressure of the steam or of the 
 delivery to hold them in place, so that there is no danger of 
 leakage at these important shoulders. The body and tubes 
 are constructed of the best bronze and are designed to give the 
 longest service with the least amount of attention and repair. 
 The injector is simply constructed, and contains but few 
 
112 THE GIFFARD INJECTOR. 
 
 parts. It is perfectly automatic in its action, restarting 
 instantly after a temporary interruption of the steam or water 
 supply. It raises the feed promptly on long lifts, with hot 
 or cold pipes, and gives a good range of capacities. Steam 
 enters at the top and passing down through the steam nozzle, 
 No. 3, discharges through the draft tube into the overflow 
 chamber and thence to the air, lifting the water to the Injec- 
 tor. The partial vacuum caused by the condensation of the 
 steam within the combining tube raises bushing No. 5 up 
 
 FIG. 29. 
 
 against the draft tube and holds the lower bushing, No. 6, 
 against the delivery tube, thus preventing the admission of air. 
 
 Upon removing the cap at the lower end of the body the 
 end of the delivery tube will be seen projecting below the 
 lower face of the body, so that a monkey wrench may be 
 used to unscrew this tube, drawing out the tubes and the 
 overflow bushings at the same time. 
 
 The size numbers of these injectors are based upon the 
 diameter of the delivery tube expressed in tenths of milli- 
 metres; No. 16, for instance, is IT"? millimetres in diameter. 
 
APPLICATION OF THE INJECTOR. 113 
 
 For all ordinary cases, this injector requires no adjustment 
 and will work satisfactorily from 40 to 100 pounds steam, 
 water flowing from a tank or from the city mains, or when 
 lifting any distance up to 15 feet. If it is necessary to use 
 the injector continuously at 15 or 20 pounds, the small ring 
 placed under the steain nozzle should be removed, and the 
 steam nozzle screwed down closer to the draft tube. For 
 150 pounds steam and very high lifts and feed temperatures, 
 a greater distance is required between the steam nozzle and 
 the draft tube, and a supplemental ring, supplied with the 
 injector, set in place, permitting a wide range of working 
 pressures. The range of capacities is good, and warm water 
 can be fed to the boiler. 
 Fig. 30 is a section of the 
 
 " RUE INJECTOR " (LITTLE GIANT) 
 
 with which the experiments were made which were tabu- 
 lated on page 47. This instrument has a movable combin- 
 ing tube by which the water area may be regulated for any 
 steam pressure or desired change in the capacity, and this is 
 
 FIG. 30. "Rui?." 
 
 effected by moving to the right or left the hand-lever ful- 
 crumed on the body. When arranged as a lifting injector, 
 a separate steam jet is added, either in the form of a central 
 spindle discharging through the combining tube or a special 
 nozzle placed in the waste pipe. The advantage of the 
 
114 
 
 THE GIFFARD INJECTOR. 
 
 adjustable water area has already been described in detail, 
 so that the merit of this feature does not need further re- 
 mark. Outside stuffing boxes permit absolutely tight joints 
 between the water and overflow chambers and the atmos- 
 phere, so that no leakage of either steam or water need 
 occur. 
 
 Another simple form of injector, but which has fixed 
 nozzles, shown in section in Fig. 31, is the 
 
 SELLERS' F. N. INJECTOR OF 1885. 
 
 In this injector the upper and lower overflows are so ar- 
 ranged in the combining tube that it is automatic in its 
 starting action, and gives, in addition, a strong vacuum, 
 suitable for high lifts, when the steam displacing spindle is 
 run slowly out. The form of this spindle also allows of a 
 regulation of the steam supply when the water valve is 
 throttled, lowering the delivery temperature so as to give at 
 
 FIG. 31. SKIVERS' F. N. INJECTOR OF 1885. 
 
 60 pounds steam, a minimum capacity of 25 per cent, of 
 the maximum. The list of capacities conforms closely to 
 that of the Sellers' 1876 pattern, yet the permissible tem- 
 perature of the feed water is higher at the lower steam 
 pressures; at 65 pounds, for instance, feed at 136 maybe 
 used when flowing to the injector under a small head, with- 
 out interfering with the automatic action, and this may be 
 increased to 150 if the waste valve is screwed down. The 
 
APPLICATIONS OF THE INJECTOR. 
 
 115 
 
 tubes are removed from the body for cleaning or repairs by 
 the motion of unscrewing the end cap, without the necessity 
 of breaking any of the pipe connections ; water, steam and 
 delivery check valves are alt contained in the body, so that 
 the injector can be used either as a lifter or a non-lifter. 
 
 An ingenious device by which the area of the water 
 entrance to the combining tube can be varied without the 
 use of a stuffing-box, is illustrated in Fig. 32, a sectional 
 view of the 
 
 METROPOLITAN INJECTOR. 
 
 The operation is as follows : Rotating the handle K opens 
 the steam valve and lifts the feed water by the strong suc- 
 tion produced by the discharge around the plug M passing 
 through the draft tube ; when this plug is entirely with- 
 
 FIG. 32. "METROPOLITAN," 
 
 drawn, the full area of the nozzle is given and the jet is 
 driven into the boiler. If the steam pressure is high or the 
 height of lift is great, the usual supply area of the combin- 
 ing tube will be insufficient, and the spindle must be drawn 
 against the rear stop, pulling the steam nozzle with it until 
 
 OF THB 
 
116 
 
 THE GIFFARD INJECTOR. 
 
 the collar on 5 strikes against the back wall of the water 
 chamber. For low steam, the spindle is screwed until the 
 collar of the steam nozzle reaches the inner stop, when the 
 pressure of the steam holds it firmly in its place. Care must 
 be taken to preserve tight joints between these faces, so that 
 no leakage can occur between the steam and feed chambers, 
 as this will not only reduce the lifting powers, but also the 
 capacity of the injector. 
 
 An English invention that has been used in this country 
 for some years on portable, marine and stationary boilers, 
 is the 
 
 GRESHAM RE-STARTING; 
 
 FlG. 33- 
 
 GRESHAM." 
 
 STEAM 
 
 its interior construction is simi- 
 lar to the locomotive injector 
 shown in Fig. 19. It has the 
 same form of movable combin- 
 ing tube which closes the upper 
 overflow when the injector is in 
 action, and therefore requires no 
 priming device, as the suction 
 produced by the discharge of the 
 steam through the draft tube is 
 exceedingly effective. 
 
 An ingenious invention for 
 permitting a free discharge of 
 steam through the combining 
 tube, yet giving an absolutely 
 continuous tube to the moving 
 water jet, is the flap nozzle sys- 
 tem, invented and patented by 
 Hamer, Metcalf & Davies, of 
 
 Manchester, England, in 1880, and shown in Fig. 34 as a 
 
 feature of the 
 
 MANHATTAN INJECTOR. 
 
 This form of tube is more extensively used in the injectors 
 using exhaust steam than the usual form of overflow open- 
 ings. The tube is cut longitudinally at its middle section up 
 
 DELIVERY 
 
APPLICATIONS OF THE INJECTOR. 
 
 11? 
 
 to a point near the steam nozzle, where the cross section of 
 the tube is sufficient to permit easy exit for water or steam. 
 The force of the steam striking against the loose hinged por- 
 tion, raises it, exposing a large area of discharge ; as soon as 
 the water enters the tube, the partial vacuum caused by the 
 condensation of the steam draws the tube shut and holds it 
 in that position as long as the jet is continuous. This sys- 
 tem is specially advantageous when applied to exhaust 
 injectors, where free discharge is a pre-requisite condition, 
 as there should be no possible danger of the feed water 
 
 FIG. 34. 
 
 MANHATTAN." 
 
 STEAM 
 
 WATER 
 
 rising in the steam nozzle when the jet breaks, as it might 
 work its way up into the cylinder. The illustration shows 
 a live steam injector, re-starting on short lifts, but intended 
 primarity as a non-lifter ; a steam valve and displacing steam 
 plug can be inserted in place of the cap d if desirable. All 
 the tubes are easily removable, and care should be taken to 
 allow no scale or sediment to be deposited upon the surfaces 
 of the closing jaws, as tips is apt to interfere with the tight- 
 ness of the joint and the correct alignment of the tubes. 
 The different parts are clearly shown in the cut a, b, c, 
 being the steam nozzle, combining tube and the delivery 
 
118 
 
 THE GIFFARD INJECTOR. 
 
 FIG. ^5. 
 EXHAUST INJECTOR. 
 
 tube ; / is the overflow valve to prevent air being drawn to 
 the boiler by the suction of the jet at the lower overflow or 
 through any leak between the joints of the tubes. 
 
 The same form of combining tube is applied to the 
 SchaefFer & Budenberg 
 
 EXHAUST STEAM INJECTOR, 
 
 illustrated in Fig. 35. Here the steam 
 nozzle is enlarged to meet the new con- 
 ditions, and a central spindle is added, 
 which, when the pressure of the boiler 
 into which the feed water is delivered 
 exceeds 75 pounds, is made hollow, 
 so as to admit a supplemental jet of 
 live steam from the boiler. The feed 
 water should flow to the instrument 
 and its temperature be as low as pos- 
 sible, never, under any conditions, 
 above 90 ; with the supplemental jet 
 and cold feed water, the injector will 
 feed against pressures up to 150 
 pounds, and without it, against 90 
 pounds; increasing the temperature 
 of the feed diminishes these pressures 
 somewhat. 
 
 THE PENBERTHY INJECTOR, 
 
 which is shown in Fig. 36, also belongs to the re-starting 
 class, as it is arranged with an overflow near the steam 
 nozzle, which is closed when the injector is in action by the 
 loose washer T sliding upon the outside of the combining 
 tube Y. Steam discharges from the nozzle R, flows through 
 the larger opening at the mouth of the suction tube S, and 
 raises the water to the feed chamber, the inlet branch being 
 at the side of the body facing the observer, and therefore 
 not appearing in the drawing ; the partial vacuum caused by 
 the condensation of the steam raises the washer T to its seat 
 in the same manner as the combining tube in the Gresham 
 Injector, so that the two overflows are separated and no air 
 
APPLICATIONS OF THE INJECTOR. 
 
 119 
 
 can enter, even if the waste valve P should leak ; this valve 
 is made large enough to permit free discharge for the steam 
 without producing any back pressure in the overflow cham- 
 ber, as this would interfere with the lifting power of the 
 
 FIG. 36. <l PENBERTHY." 
 
 steam jet. The following tests of an injector of this pattern 
 have been furnished by the manufacturers, but were made 
 by disinterested experts height of lift = 3 feet : 
 
 Limiting feed temperature at 65 Ibs. steam, 128, delivery 200 
 
 75 " 121 " 196 
 
 U 85 II? 
 
 " back pressure without overflow at 65 Ibs. , 92 Ibs. 
 
 196 
 200 
 
 75 
 85 
 
 103 ' 
 
 112 
 
 In conclusion it may be said that the construction of the 
 injector is exceedingly simple, as starting is effected by 
 opening the globe valves placed on the steam and feed 
 pipes. 
 8 
 
120 THE GIFFARD INJECTOR. 
 
 Among other well-known patterns of injectors designed 
 especially for stationary boilers may be mentioned the Han- 
 cock Inspirator, Park, Bberman, Jenks, SherrifT, Buffalo, 
 Globe ; limit of space prevents the insertion of descriptions 
 and illustrations of these also, and a sufficient variety has 
 been given to elucidate the principles upon which all are 
 designed, rendering repetition unnecessary. 
 
 The setting of an injector and attaching the various pipes 
 to a stationary boiler are much more simple than in the case 
 of a locomotive. When required to lift, the chief object is 
 to obtain a position as near the level of the water supply ?s 
 possible, without sacrificing convenience of handling, as the 
 greatest delivery is thus obtained and less care and precision 
 required for starting and regulating. Longer service will 
 also be given before repairs become necessary, because the 
 greater the tax that is placed upon the lifting powers of an 
 injector, the smaller will be the deviation from the correct 
 proportion of the tube which will affect its working. The 
 rate at which the delivery decreases as the lift is increased 
 is shown by Table V, on page 76, and it is only adjustable 
 and self-adjusting injectors, and some forms of the double- 
 jet, that are exempt from this rule. The suction pipe should 
 therefore be as short and direct as possible, and carefully 
 tested under steam pressure to detect any faulty joints or 
 leaks in the pipe ; if very long, it should be made at least 
 one size larger than the nipple of the injector, and should 
 be sloped on long horizontal stretches so as to drain toward 
 the well without air or water pockets. Foot valves are not 
 recommended, but strainers should be used, having entrance 
 holes smaller than the smallest diameter of the delivery 
 tube, with an aggregate area of at least four times that of 
 the suction pipe. 
 
 The steam pipe should be attached to the dome of the 
 boiler, or connected with a main of such size that the pres- 
 sure of the contained steam will be constant and the same as 
 that in the boiler. This pipe should be covered, and sup- 
 plied with a pet-cock for draining accumulated water, espe- 
 
APPLICATIONS OF THE INJECTOR. 121 
 
 ciallyif the conditions under which the injector must operate 
 are at all adverse. Every effort should be made to supply 
 as dry steam as possible, and this can only be obtained by 
 carrying out these recommendations. 
 
 The delivery pipe should of course be short and direct, 
 although this is not as necessary as with the other connec- 
 tions ; it should be supplied with a main check valve, a stop 
 valve and a pet cock for draining. The vertical length of 
 this pipe or the position of the injector relative to the water 
 level of the boiler is comparatively unimportant, as there is 
 always an excess of power in the jet more than sufficient to 
 overcome any ordinary difference in level. In the tests of 
 the Penberthy Injector there is shown to be an excess of 
 counter pressure over the initial of 28 pounds at 75 pounds 
 steam, permitting a difference in level and friction head of 
 about 56 feet, far greater than is likely to occur in practice. 
 
 All steam and water pipes should be carefully blown out 
 with steam to free them from dirt and scale before the in- 
 jector is set in place, and all pipes should be bent so that 
 the unions will fit squarely in their seats, requiring no 
 force to draw them to place when the coupling nuts are 
 screwed up. 
 
 It may appear that needless stress has been laid upon the 
 method of attaching injectors, and that unnecessary refine- 
 ments have been suggested, yet it may be said that the 
 advantages arising from good work and conscientious atten- 
 tion to details of this kind more than compensate for the 
 additional outlay. In this respect, English engineers are 
 much more particular than those in this country ; the gene- 
 ral foreign practice is to substitute pipes bent in easy curves 
 for all cast iron elbows, flanges take the place of malleable 
 iron couplings, while the general solidity and care mani- 
 fested in the attachment of boiler fittings are well w r orthy of 
 
 imitation. 
 
 BOILER TESTERS. 
 
 The principle of the injector has been applied very satis- 
 factorily to the testing of boiler seams, and special instru- 
 ments are made for this purpose. They are so constructed 
 
122 THE GIFFARD INJECTOR. 
 
 that the boiler may be rapidly filled with warm water and 
 high testing pressure then applied, even though the jet is 
 actuated by low pressure steam. 
 
 In this instrument, two sets of tubes are used, one for 
 each operation ; the primary set acts as an ejector, delivering 
 between 600 and 700 cubic feet of water per hour, and fills 
 the boiler up to the safety valve; all openings are then 
 closed, and the smaller set, having a capacity of about 60 
 cubic feet per hour, is started, and the pressure in the boiler 
 raised to the point desired; this secondary set is propor- 
 tioned like an injector, but has a much larger steam nozzle 
 than would be ordinarily employed, as the excess of counter 
 pressure required is much greater than in the case of an 
 injector. If the same relations that obtain in the exhaust 
 injector would hold true at 70 pounds steam, testing pres- 
 sures of 450 pounds would be produced, for with an absolute 
 steam pressure of 14.69 pounds, 105 pounds, or seven times 
 the initial, can be obtained ; but in this case the area of 
 the steam nozzle is nearly sixteen times that of its delivery 
 tube, while at 60 to 80 pounds, 4^ is the greatest ratio that 
 can be used efficiently. The usual pressures are about as 
 follows : 
 
 30 pounds initial will give 135 pounds terminal pressure. 
 
 40 " " 180 
 
 60 " " 260 " 
 
 80 ' " 320 
 
 This method of testing is very convenient for boiler-shop 
 use, as the device is small and compact, easily applied, and 
 requires little skill to operate. Safety valves, which usually 
 form part of the apparatus, may be set for any desired press- 
 ure up to the limits of the instrument, so that there may 
 be no danger of subjecting the boiler to excessive strain. 
 The feed supply should flow under an unvarying head so 
 that the conditions under which the jet operates may remain 
 constant, and rather higher counter pressures are given when 
 the water is received under a head than when it is lifted. 
 
 The suggestion has also been made to apply the power of 
 the injector for increasing available pressure, to use in a 
 
APPLICATIONS OF THE* INJECTOR. 123 
 
 hydraulic press, but it is doubtful if this would ever be of 
 much practical value. By properly proportioning the piston 
 areas of the press, an injector could be substituted for the 
 pumps usually employed, but practical considerations would 
 at once suggest themselves which would render the supe- 
 riority of the injector exceedingly dubious. 
 
 For fire pumps there is a field FrG FJRE 
 
 that is more promising. Small 
 instruments are constructed hav- 
 ing steam nozzles about the same 
 size as the delivery tubes, and are 
 frequently attached to yard and 
 switching locomotives with good 
 results ; one of the_ most promi- 
 nent railroads in this country has 
 supplied its yard engines with 
 pumps of this kind, which are 
 operated by simply opening the 
 tank valve and the steam valve, 
 and coupling up a line of hose. 
 Fig. 37 gives an exterior view of 
 this style of fire pump which 
 will throw 4600 gallons of water 
 per hour through a -J^- fire nozzle 
 a distance of 115 feet. It can be 
 coupled to an ordinary fire hyd- 
 rant in case the supply of water 
 in the tank is insufficient. The 
 capacity and distance given are 
 based upon a steam pressure of 
 125 pounds per square inch. 
 
 For draining or raising large 
 quantities of water a small height, ejectors can be used with 
 a comparatively small expenditure of steam, although with 
 considerable less economy than if a pump w r ere used, and 
 there are but few cases where the increase in temperature 
 the chief source of economy in the injector as a boiler feeder 
 would be advantageous. The advantage of steam con- 
 
124 THE GIFFARD INJECTOR. 
 
 densation is of use indirectly in the case of bilge pumps, 
 which, beside performing the function for which they are 
 designed, are often operated when the engines of steamers 
 are stopped at sea as condensers for the sudden accumula- 
 tion of steam until the fires can be properly checked. These 
 pumps are simply constructed, and, having no valves, are 
 not liable to stoppage. 
 
 The proportions of the parts of an ejector vary somewhat 
 with steam and counter pressure and the purpose for which 
 it is to be used ; the ratio of the area of the steam orifice to 
 the delivery tube may vary from 0.25 to i.oo, to 0.8 to i.oo, 
 the former giving a strong suction in a well-shaped tube at 
 the higher steam pressures. 
 
 As exhaust steam condensers for reducing the back pres- 
 sure upon the piston of an engine, special forms have been 
 designed, which are complete in every detail for regulating 
 the water supply for varying loads, and large numbers are 
 now in use. The arrangement is compact and reliable, re- 
 quires no air-pump, and is usually less expensive than the 
 other systems. 
 
 There are numerous other devices in which jets of steam 
 are used for moving and forcing air or other gases, such as 
 exhausters, blowers, ventilators, etc.; but as they do not 
 depend upon the condensation of the steam which is in 
 reality the inherent characteristic of the injector, they do 
 not come within scope of these pages. 
 
CHAPTER IX. 
 
 DETERMINATION OF SIZE INJECTOR TESTS DIAGRAMS OF 
 
 RESULTS. 
 
 As has already been shown, the maximum capacity of an 
 injector depends upon three factors: Steam pressure, height 
 of lift, and temperature of feed water. It is upon the 
 design of the tubes and the special type of the instrument 
 that the extent of the variation depends when any two con- 
 ditions are given and the third altered. This has been 
 illustrated by the approximate formula, No. 23, on page 
 79, which shows the rise in capacity with the pressure; by 
 Table VII, page 78, where the steam pressure and the lift 
 remained constant, and the feed temperature varied; and 
 Table V, page 76, where the pressure and temperature were 
 constant, and the height of lift increased. Two diagrams, 
 Figs. 39 and 40, indicate graphically the first and second 
 cases more clearly than can be done by tabulation, showing 
 the general rules that hold true, though in a varying degree, 
 for all styles and patterns of injectors. It is thus seen that 
 judgment or, preferably, experiment must be used to de- 
 termine the capacity of an injector for any conditions differ- 
 ing from the usual practice, and a full statement of the con- 
 ditions should be submitted to a competent manufacturer 
 whose wider experience will enable him to give accurate 
 information. 
 
 For each given feed temperature and height of lift there 
 is a limiting steam pressure, beyond which there is no fur- 
 ther increase in the capacity ; this limit of pressure is lower 
 with injectors designed for stationary boilers than those used 
 in locomotive sendee, and it is usually a function of the inlet 
 area to the combining tube, and varies with the pattern of 
 
 125 
 
126 THE GIFFARD INJECTOR. 
 
 injector; considering these things, it is always better before 
 deciding on the size required to submit the following data 
 to the manufacturer : 
 
 (a) Required capacity in cubic feet per hour, 
 
 (b) Steam pressure, 
 
 (c) Height of lift, 
 
 (d ) Temperature of feed water. 
 
 Stating, also, any special conditions, such as length of suc- 
 tion pipe, if unusual, or an excess of back pressure, often 
 required when feeding through a heater. Some forms of 
 injectors are so designed that all ordinary conditions of ser- 
 vice are satisfactorily covered by the usual form supplied; 
 but others require slight modifications of the forms of tubes 
 which enable them to operate more effectively in special 
 cases. 
 
 The first information to obtain is the maximum evapora- 
 tion of the boiler ; formerly the allowance granted per nomi- 
 nal horse power was one cubic foot of water; but with the 
 great advance that has been made in the economy of the 
 modern stationary engine, about one-half that amount, or 
 35 pounds per horse power, is usually regarded as sufficient. 
 It is always more advisable to obtain the actual evaporation, 
 but when unknown or doubtful, the requirements may be 
 obtained from the estimated steam consumption of the 
 engine, calculated from an indicator card, allowing sufficient 
 margin for other drains upon the steam supply. If this can- 
 not be done, the approximate rules here given can be used : 
 
 To FIND WEIGHT OF WATER EVAPORATED PER HOUR <= X] 
 
 FROM AND AT 212. 
 
 When the heating surface (= H) is known : 
 
 For tubular boilers, X = H X 3-4Q 
 " flue " X^HX 6.00 
 
 Water tube " X = H X 3.5. 
 When the grate surface (= G) is known : 
 
 For externally fired boilers, X G X 94.00. 
 " internally " X = G X 180.00. 
 
 When the total weight of combustible burned per hour ( C) is 
 known : 
 
 X= C X 9-00, (Assumed as a fair average.) 
 
SIZES OF INJECTORS. 
 
 127 
 
 When the heating surface (//) and grate surface (<7) are known : 
 For internally fired boilers, X = 36.0 V H X k 1 .' 
 
 (Molesworth Pocket Book.) 
 
 When the heating surface (//") and weight of combustible (C) are 
 known : 
 
 A': 
 
 (12- 4. 
 
 375 X -77 ) X C 
 
 NOTE. To reduce weight of water evaporated from and at 212 to 
 the corresponding evaporation from a feed temperature of 70 and a 
 steam pressure of 70 pounds, divide by 1.183. 
 
 To find weight of combustible, subtract from total weight of coal 
 burned per hour, the weight of the ashes and the contained moisture^ 
 
 In determining the size of injector required for a locomo- 
 tive, the size of the cylinder is usually taken as the stand- 
 ard, although the diameter of the boiler, and the special 
 service for which the locomotive is intended, has a modify- 
 ing influence. The following table embodies the practice of 
 the Baldwin Locomotive Works for all engines equipped 
 with two injectors, one injector and one pump, one injector 
 and two pumps-, 
 
 TABLE IX. 
 
 SIZES OF INJECTORS FOR LOCOMOTIVES. 
 
 Diameter 
 of 
 Cylinders. 
 
 Inches. 
 
 Nominal Size of Injector. 
 
 Diameter 
 of 
 Cylinders. 
 
 Inches. 
 
 Nominal Size of Injector. 
 
 2 Injectors, 
 or 
 i Pump and 
 i Injector. 
 
 2 Pumps 
 and 
 i Injector. 
 
 2 Injectors, 
 or 
 i Pump and 
 i Injector. 
 
 2 Pumps 
 and 
 i Injector. 
 
 6,7,8 
 9, 10 
 
 I 1 , I 2 , 13 
 
 14, 15 
 
 1 6, 17 
 
 18 
 
 3 
 4 
 
 8 
 
 3 
 3 
 4 
 4 
 5 
 6 
 
 ! '9 
 
 20 
 21 
 22 
 
 23 
 2 4 
 
 8* 
 9 
 
 9* 
 
 10 
 10 
 
 II 
 
 6 
 6 
 
 * Use next size larger with specially large boilers. 
 
 Almost all of the best known patterns of locomotive injec- 
 tors of the same nominal size correspond closely in maximum 
 capacity, even though the weight of steam used per hour 
 will vary according to the proportion of the nozzles and the 
 general excellence of the design. With injectors for station- 
 
128 THE GIFFARD INJECTOR. 
 
 ary boilers there is less uniformity, and no general standard 
 prevails. 
 
 With instruments purchased from reputable manufac- 
 turers, there is but little danger of the capacity falling below 
 the figures claimed if the injector is operated under the 
 conditions prescribed in the catalogue ; owing to the present 
 refined methods of duplication of parts, nozzles can be bored, 
 reamed and turned with mathematical accuracy and endless 
 repetition, so that there is no excuse for poor workmanship 
 or any deviation from proper dimensions ; but there may be 
 flaws or spongy places in the main castings, and these defects 
 do not always become apparent until after the injector has 
 been subjected to the vibration of an engine, escaping detec- 
 tion even in the proving room of the manufacturer. If an 
 imperfection of this kind should occur in one of the walls 
 separating the steam from the feed chambers, a useless heat- 
 ing of the feed water would result, and both the maximum 
 capacity and the highest admissible temperature of the feed 
 water would be reduced. Assuming the injector to be in 
 perfect condition, there may be a very serious falling off in 
 the capacity below that claimed, if the steam pressure or 
 the height of lift during the test are greater than those used 
 by the makers ; for instance, if the standard were based upon 
 a pressure of 120 pounds per square inch and a lift of i foot, 
 the capacity (based upon an actual case) might fall off 10 
 per cent, when the lift is increased to 6 feet, and 25 per cent, 
 when the steam pressure is raised to 150 pounds. 
 
 These considerations show the advantage, and, in some 
 cases, the necessity of having a thoroughly-equipped testing 
 department wherever large numbers of injectors are in use, 
 both for the purpose of experiment, but more especially for 
 proving those that have been repaired by the employee in 
 charge of this work. The exact conditions occurring in 
 practice steam pressure, height of lift and length of suction 
 pipe should be imitated as nearly as possible. By this S}^- 
 tem the performance of all injectors submitted may be care- 
 fully determined without premature acceptance of manufac- 
 turers' claims. Many of the large railroad companies have 
 
SIZES OF INJECTORS. 
 
 129 
 
 appreciated these facts, and installed in their laboratories 
 apparatus for testing new forms of injectors, as well as to 
 prove one or two selected from all lots submitted for regular 
 service, which must conform to the standard required by the 
 railroad. 
 
 The standard dimensions to which all injectors used on 
 the Pennsylvania Railroad are required to conform are shown 
 in Fig. 38. and cover chiefly the position of the steam, feed 
 and suction pipes, and the attaching bolt, so that all pat- 
 terns of injectors may be interchangeable. As these pro- 
 portions are known as the P. R. R. Standard, and are 
 used by several of the large companies, a list of the prin- 
 cipal dimensions is given in the table accompanying 
 
 Fig. 38. 
 
 FIG. 38, 
 
 Nominal Size of Injector. 
 
 A 
 
 B 
 
 c 
 
 D 
 
 E 
 
 No. 6 
 
 7* 
 
 Il| 
 
 2 
 
 2 |1 
 
 2 1 
 
 No 7 
 
 7i 
 
 n'4 
 
 2\ 
 
 2U 
 
 2J- 
 
 No 8 . '. . . 
 
 si 
 
 IA.% 
 
 21- 
 
 3-* 
 
 2i 
 
 No 9 ... 
 
 Si 
 
 '48 
 
 id.! 
 
 2l 
 
 s 
 
 2 l 
 
 
 
 
 
 
 
 Regarding the method of testing, it may be said that the 
 simplest way is to empty a tank of known dimensions between 
 given levels, delivering against a check or lever valve loaded 
 
130 THE GIFPARD INJECTOR. 
 
 to the initial steam pressure. A better method is to weigh 
 the feed water, and if possible, the delivery also, so that the 
 weight of steam used per hour can be determined; with 
 small injectors this can almost always be easily done, but 
 with locomotive sizes it is hardly practicable to weigh even 
 the feed. From an experimental point of view it is interest- 
 ing to have both weights, but the proportion of water to 
 steam can be easily determined by observing the feed and 
 delivery temperatures, and applying formula (33) given on 
 page 83, which is sufficiently accurate for all ordinary pur- 
 poses ; the measuring of the temperatures is best done by 
 inserting the thermometer bulbs directly in the flow of the 
 water, screwing the body through a tee having a 3^ gas pipe 
 thread ; special instruments are made for this purpose, and 
 are more satisfactory than the inserted mercury cup. The 
 use of a spring check valve in the delivery pipe is seldom 
 satisfactory, and does not give as accurate results as dis- 
 charging directly into the boiler, or, what is more practica- 
 ble, delivering into a steam trap having a large balanced 
 outlet valve, and connected to the steam area of the boiler. 
 This arrangement is decidedly the best, as it does not alter 
 the level of the water in the boiler, and the pressure against 
 which the injector delivers is maintained constant, giving 
 more precise results when the minimum capacity and the 
 highest admissible temperature of feed water are to be deter- 
 mined. 
 
 The most convenient and instructive method of recording 
 comparative tests is to plot the results in the form of the 
 diagrams given in Figs. 39 and 40. Here are shown tests 
 of three of the best known locomotive injectors in this coun- 
 try, the object being to ascertain the increase in capacity 
 with the steam pressure, and the eifect of high feed temper- 
 atures upon the maximum and the minimum capacities. In 
 both diagrams the lines parallel with the base line represent 
 gallons per hour ; those at right angles indicate steam pres- 
 sure in Fig. 39> and feed temperature in Fig. 40. 
 
 These series of tests were made in the laboratories of two of 
 the large railroad companies at considerable time and expense, 
 
FIG. 39- 
 
 COMPARATIVE TESTS OF THE MAXIMUM AND MINIMUM 
 CAPACITIES OF THREE PATTERNS OF No. 9 INJECTORS 
 AT VARIOUS STEAM PRESSURES. Feed, 74 ; lift, 6 feet. 
 
 2800 
 700 
 2600 
 2500 
 2400 
 
 *2200 
 
 ooo 
 
 1500 
 
 3 1800 
 
 fc , 
 
 1700 
 
 ^1600 
 
 ^1500 
 
 '"1400 
 
 1200 
 
 1100 
 > 
 1000 
 
 900 
 
 800 
 
 * 
 
 40 
 
 60 
 
 80 
 
 re -^ -sure 
 
 100 
 
 120 
 
 140 
 
 MO 
 
 131 
 
.132 THE GIFFARD INJECTOR. 
 
 and are entirely unbiased ; yet, in order to avoid any com- 
 parison of the relative merits of the different patterns, they 
 are here indicated alphabetically, the same letters applying 
 to similar injectors in both diagrams, even though the sizes 
 are different. The maximum capacities are designated by 
 capital letters, and the minima by small letters. The injector 
 marked O is not the same pattern as the C shown in Fig. 
 39; but as it is the product of the same manufacturer, and 
 has similar characteristics, it was added to the diagram. 
 
 The capacities in Fig. 39 were obtained by noting the time 
 required to deliver a measured weight of water against a 
 spring check valve. The lines connecting the observations 
 should form symmetrical curves, and all irregularities indi- 
 cate possible errors of adjustment or of observation which 
 are difficult to detect unless the results are plotted; some 
 patterns of injectors require very careful adjustment of both 
 water and steam, and slight variations in the regulation 
 materially affect the result. The two lines, A and C, keep 
 rather close together, while B delivers more water at the 
 lower pressures, and less beyond 120 pounds, at which pres- 
 sure it reaches its maximum. A apparently starts down- 
 ward between 120 and 140 pounds and then ascends, indi- 
 cating imperfect regulation at the reentrant angle. Referring 
 to the minima, the lines a and b almost coincide, while c 
 gives a much better minimum until 160 pounds is reached, 
 when the sudden rise indicates insufficient throttling of the 
 feed valve. If the steam pressure were raised still higher, a 
 point would be found where the converging lines of the 
 maxima and minima would intersect, which would be at 
 the highest steam pressure with which each injector can be 
 operated under these conditions. 
 
 In the tests given in diagram, Fig. 40, the steam pressure 
 and the height of lift are constant, and the feed temperature 
 raised 5 after each observation until the injector broke. 
 This was done to obtain under each condition the maximum 
 and the minimum capacity of each injector; the latter lines 
 have numerous reentrant angles, due to the fixed notches 
 which the regulating mechanism necessitated. In all cases 
 
TESTS. 
 
 133 
 
134 THE GIFFARD INJECTOR. 
 
 the two capacities should be identical at the breaking and 
 overflowing temperatures ; this is true with lines A and C t 
 and probably with B also, if prolonged to 94. The termi- 
 nation of each line indicates the highest admissible temper- 
 ature of the feed water for each pattern of injector, A 
 breaking at 110, B at 90, and Cat 125. 
 
 A rather curious feature of the tests with the C l c l injector 
 is the crossing of the two lines, showing an apparent para- 
 dox, that the minimum capacity at 115 is higher than the 
 maximum. In this pattern the regulation was obtained by 
 reducing the steam supply, and as it is well known that 
 with high pressures and temperatures a smaller steam nozzle 
 can be used to advantage, it follows that here, where the 
 steam area is reduced, the capacity is larger than when the 
 full steam nozzle is used, but which, when fully opened, 
 gives best results at lower temperatures. 
 
 By contrasting these results, will be seen the difficulty of 
 obtaining a formula which will cover accurately all the con- 
 ditions that enter into the performance of an injector, as the 
 actual construction exercises such a modifying influence 
 upon the results. The equations which were given were 
 based upon the form which seemed to approximate the 
 theoretical results, which would of course be a self-adjusting 
 injector able to preserve the condition of the jet at the 
 instant of approaching the smallest diameter of the delivery 
 tube nearly at a density of unity. In all experimental work 
 upon which theoretical conclusions are to be based, great 
 care should be exercised to secure accurate observations, 
 and these should at once be plotted in the manner shown, 
 so that all doubtful results can be checked, and, if neces- 
 sary, repeated. 
 
 A more recent test than those already described, and one 
 in which special precaution was taken to insure accuracy, 
 will now be given ; this test * was made by disinterested 
 experts with apparatus that can be duplicated in almost any 
 railroad shop with but little* preparation. Any official de- 
 siring to make a comparative test of the injectors in use on 
 his line, can apply the method outlined with satisfactory 
 
 * Reprinted by permission from the Railroad Gazette. 
 
TESTS. 
 
 135 
 
 and decisive results at comparatively small expenditure. 
 The injector used in this test was a No. ioJ/2 of the Im- 
 proved Sellers 1887 pattern, and it was chosen in common with 
 several other well-known injectors for the purpose of making 
 a selection for the equipment of a large number of locomotives, 
 and to ascertain if the performance at certain given steam 
 pressures fulfilled the requirements of the specifications. 
 
 FIG. 42. 
 
 Apparatus. The injector (see page 104) was supplied with 
 dry steam from a 200 H. P. Babcock and Wilcox boiler, 
 through a 3 in. pipe carefully lagged with asbestos covering. 
 The injector was bolted against the side- wall of the boiler 
 with the starting lever and water supply valve within con- 
 venient reach of the operator. The water supply was main- 
 tained at a constant level in a large barrel directly below the 
 injector, into which the suction pipe was extended to within 
 
136 THE GIFFARD INJECTOR. 
 
 one foot of the bottom ; this pipe was 2^ in. diameter up 
 to the nipple of the injector, where it was reduced to 2 in. 
 
 The water supply was weighed and delivered into the suc- 
 tion barrel as follows : Ten feet above the level of the suction 
 barrel were two large tanks, forming a reservoir capable of 
 holding about noo gals. From the 3^ in. flanged pipe 
 bolted to the bottom of each was a vertical 3 in. pipe extending 
 into the suction barrel, 6 in. below the 4 ft. level ; this pipe 
 was made amply large, so that under the head available the 
 tanks would be drained and the pipe emptied before the level 
 of the water in the barrel could be lowered by the injector 
 more than 6 in.; owing to the slope of the bottom of the 
 tanks and the large size of the emptying pipe, this was accom- 
 plished without the slightest trouble. A globe valve was 
 placed in this pipe close to the injector, so that the proper 
 level of the water in the suction barrel could be maintained 
 by the operator. (See Fig. 42.) 
 
 Above and resting on these water tanks was a 500 Ib. plat- 
 form scale carrying a 47 gal. barrel, into the bottom of which 
 was screwed a 2 in. outlet pipe for emptying it quickly into 
 the iron tanks below ; this pipe was kept clear of the sides of 
 the tank and the scale. Another pipe brought the water sup- 
 ply from the city mains above the top of the barrel, and as it 
 was supplied with a quick acting, tight gate valve, the barrel 
 could be quickly weighed empty, filled, re-weighed and 
 emptied into the iron tanks until the quantity of water re- 
 quired for a run of from 15 to 20 minutes was obtained. 
 
 Additional means for supplying the barrel with water during 
 the preliminary run before each experiment was found to be 
 necessary, because when the reservoir was filled with weighed 
 water, none could be withdrawn until the moment the exper- 
 iment commenced. For this purpose a 2 in. hose was run 
 into the barrel and a valve placed near the discharge end, 
 and the water level maintained constant until all prepara- 
 tions were completed ; this valve was closed and hose with- 
 drawn before the valve for the. reservoir tanks was opened. 
 
 Gauges were placed so that the pressure in the steam pipe 
 or in the delivery pipe could be obtained alternately on either 
 
TESTS. 
 
 137 
 
 FIG. 43. ARRANGEMENT OF PIPING AND GAUGES FOR 
 TESTING INJECTORS. 
 
 These sires of ftpes would 
 do for an &' ,."> -9 . mjectar. 
 
138 
 
 THE GIFFARD INJECTOR. 
 
 or both gauges, or simultaneously on separate gauges ; this 
 arrangement worked very satisfactorily, and the admission 
 of any error due to the difference between the steam and de- 
 livery pressures, or to a discrepancy between the two gauges, 
 could be prevented. Fig. 43 shows the arrangement of 
 pipes, valves and gauges by which this was accomplished, 
 and is self-explanatory. 
 
 As the delivery of the injector was too great to be taken 
 into the boiler without affecting the steam pressure carried, 
 it was passed through a special balanced valve (Fig. 44), 
 which maintained a constant pressure equal to that of the 
 boiler. The delivery pipe of the injector was coupled to the 
 under side of a check valve, which was connected to a piston 
 
 FIG. 44. 
 
 of the same area, upon the upper side of which full boiler 
 pressure was obtained by a pipe tapped into the steam supply. 
 
 Measuring Devices. The steam gauges had been sub- 
 jected to careful test and calibration by the makers; at the 
 same pressures the readings of the two gauges agreed 
 exactly. 
 
 The thermometers were tested in oil every five degrees, 
 both up and down, and the corrections noted, from which a 
 table was made and used to obtain actual temperatures. 
 The delivery thermometer was brass cased, and was screwed 
 into the delivery pipe close to the injector, with the bulb 
 well immersed in the passing water ; the scale was divided 
 to single degrees, and could be read easily to half degrees. 
 It was corrected for the error due to the compression of the 
 bulb, as it was subjected to the pressure of the delivery. 
 
TESTS. 139 
 
 The scale used for weighing the feed water was tested by 
 United States standard weights and found to be correct. The 
 barrel resting upon the scale platform was weighed before 
 and after each filling, so that the exact net weight of water 
 passing into the receiving tanks each time could be deter- 
 mined. 
 
 The filling of the water tank reservoir required two ob- 
 servers, one to open and close the inlet valve from the city 
 mains and the valve from the bottom of barrel leading into 
 the reservoir, and the second to shift the tare weight for the 
 empty and the full barrel, and to record the weights upon 
 suitably prepared blanks. As the scale registered to ^ lb., 
 the possible error was very small, for a test by United States 
 standard 100 lb. weights after the experiment was completed, 
 showed no change, 
 
 Method of Testing. The reservoir tank having been filled 
 with the required weight of water, the valves in the steam 
 pipe were opened wide, and the water drained out ; the 
 water regulating valve on the injector was opened and the 
 cam lever over the waste valve was set so as to allow this 
 valve to open freely. Precaution was taken to insure the 
 water supply being free from dirt and chips and the suction 
 barrel clean. The 2 in. hose from the water main was led 
 into the barrel and the injector started against full back 
 pressure. The hose discharged beneath the surface of the 
 water to prevent air being carried down into the water and 
 interfering with the free flow in the suction pipe. Obser- 
 vations were made as to the regularity of the steam pressure, 
 and the readings of the steam gauges and the thermometers 
 in the suction barrel and delivrey pipe were found to be prac- 
 tically constant ; an observer was stationed at the water valve 
 in the 3 in. pipe leading from the overhead reservoirs, 
 another to read the gauges and thermometers, and a third 
 to take and record all readings and note the general perform- 
 ance of the injector. When everything was ready, the barrel 
 was rapidly filled through the hose and then its valve closed 
 and the hose entirely withdrawn ; as soon as the water level 
 in the barrel was drawn down by the injector to the lower 
 
140 THE GIFFARD INJECTOR. 
 
 white line (b) (see Fig. 43), the recorder noted the exact time 
 on a stop watch, the other observers noted the thermometer 
 and gauge readings, while the valve in the 3 in. feed pipe 
 from the reservoirs was quickly opened and the water level 
 raised to the line (a) four feet below the centre of the injector, 
 where it was maintained during the continuation of the ex- 
 periment by careful regulation of the valve. Readings of the 
 thermometers and the gauges were taken every three minutes 
 until the reservoir was empty, which could be immediately 
 noted by the rapid falling of the level of the water in the 
 barrel ; just before the lower level was reached the end of the 
 3 inch pipe from the reservoir was exposed, this construction 
 being insisted upon so that the observer could be certain that 
 all the water in the reservoir had flowed into the barrel ; 
 when the lower level from which the start was made was 
 reached the signal was given to the recorder, and the time 
 again noted, the difference in time being that required to lift 
 and force against initial pressure the total weight of water 
 contained in the reservoir ; from this could be calculated the 
 capacity of the injector in pounds, cubic feet or gallons per 
 hour. 
 
 The method of determining the minimum was the same, 
 except that occasional adjustment of the regulating valve 
 was required during the experiment owing to variations in 
 the pressure of the steam; also, the quantity of water weighed 
 into the reservoir was less than half that used for determin- 
 ing the maximum. Care was taken that the counter pres- 
 sure produced by the back-pressure valve should always be 
 equal to that of the boiler so as to obtain precise results. 
 
 From these two sets of experiments were determined the 
 figures for the ratio of the minimum capacity to the max- 
 imum; subtracting this from 100 gives the "range" in 
 percentage of maximum. 
 
 To determine the relation between the weights of the 
 supply water and of the steam required to force it into the 
 boiler, it is evident that the simplest method would be to 
 subtract the known weight of the supply from the weight of 
 the delivered water and then divide the weight of the supply 
 
TESTS. 141 
 
 by this difference. With small injectors this is often done, 
 as the volume of water to be handled is not large, but with 
 an instrument of the size used, this method becomes imprac- 
 ticable. The method of delivery temperatures was therefore 
 substituted and the same results obtained without the neces- 
 sity for weighing the delivery. The formula used was the 
 following : (See page 83.) 
 
 _ 
 
 W= Weight of water delivered per pound of steam. 
 H = Total heat in one pound of steam (absolute) pressure above 32 
 
 deg. taken from steam tables. 
 T= Temperature of the delivered \vater. 
 t = Temperature of the water supply. 
 /* Steam pressure (gauge). 
 
 Maximum water-supply temperatures were obtained by 
 returning some of the water from the delivery pipe to the 
 barrel or reservoir ; care was taken that the hot and cold 
 water should be thoroughly mixed and that the temperature 
 should not be increased too rapidly. Two sets of results are 
 given : limiting temperatures at each steam pressure for auto- 
 matic restarting without subsequent waste of hot water or 
 steam from the overflow ; also maximum operating tempera- 
 ture at which the injector will run without the jet breaking ; 
 the former were obtained with the waste valve free to rise 
 on its seat; the latter, with the waste valve closed by 
 throwing the cam lever backward ; in this case if the jet 
 breaks, steam will flow back into the suction pipe until the 
 waste valve is allowed to open or the steam supply is -shut off. 
 
 Numerous special tests were also made to determine the 
 action of the injector under conditions frequently occurring 
 in practice, such as variations of the steam pressure, hot 
 water in suction pipe, and the effect of a temporary interrup- 
 tion of the .water supply, such as would occur when the 
 movement of the water in the tank of a locomotive uncovered 
 the end of the suction feed pipe ; also, the amount of water 
 
142 THE GIFFARD INJECTOR. 
 
 wasted during starting and stopping. An account of these 
 tests will be made under the heading of ' ' Results. ' ' 
 
 It should be noted that all the experiments were made 
 without throttling the steam supply ; this was found to be 
 necessary as an early experiment at 150 Ibs. steam showed 
 that the superheating due to wire drawing materially affected 
 the results ; in all subsequent tests the pressure of the boiler 
 was raised or lowered to meet the requirements of the experi- 
 ment. 
 
 The directions given in the catalogue of the manufacturers 
 for stopping and starting the injector were followed: To 
 start : Pull out the lever. To stop : Push in the lever. Regu- 
 late for quantity with water valve. In starting on high lifts 
 and in lifting hot water, it is best to pull the lever slowly. 
 
 Results. To facilitate the comparison, the performance of 
 the injector at different pressures and the results obtained at 
 each set of experiments have been plotted in separate dia- 
 grams, forming curved or broken lines connecting the several 
 observations, so that the results for any intermediate condi- 
 tion can be easily determined ; as the scale of the diagrams is 
 necessarily small, a complete table of results has been given, 
 which contains the actual figures obtained. The results of 
 the tests were remarkably good, for in several cases the 
 claims of the manufacturers were much exceeded. Accom- 
 panying each diagram is a short review of the results. 
 
 As stated above, however, certain tests were made which 
 could not be tabulated, but are almost equally valuable in 
 considering the general performance of a locomotive boiler 
 feeder. 
 
 (A) Variation in steam pressure. The injector was started 
 with the lever-starting valve and the water-regulating valve 
 wide open, and the pressure in the boiler and the back pres- 
 sure were simultaneously lowered from 200 Ibs. to 120 Ibs. 
 and then later, with all valves as before, from 120 Ibs. to 40 
 Ibs. steam, without a drop of water appearing at the overflow; 
 raising the same pressure caused no overflow of steam or 
 water at the waste pipe, and the injector seemed to operate 
 as successfully at one pressure as another, without making 
 
TESTS. 143 
 
 change in the tubes or in the position of its steam or water 
 valve. 
 
 (B) From the fact that this injector worked very satis- 
 factorily with hot supply water, it was evident that its lift- 
 ing power with the suction pipe warm, would also be good ; 
 owing to the provision of large overflows in the forcing com- 
 bining tube, it is not necessary that care should be used in 
 admitting steam to the main jet after priming as is the case 
 with other forms of injectors for even though the feed water 
 be above the limiting temperature as it comes from the lift- 
 ing nozzle, the forcing jet will not break, but will cause an 
 overflow of steam and hot water until the hot water is drawn 
 out, which usually occurs in a few seconds ; in this case the 
 amount of waste was small, but with cold water only a few 
 drops appeared at the waste nozzle at 60, 120, 200 Ibs. of steam 
 or intermediate pressures. The mean of a number of tests, 
 stopping and starting the injector with the supply water at 
 ordinary temperatures, gave one-half pint as the amount 
 wasted each time. 
 
 (C) It has been found by the experimenters that the effect of 
 admission of air to the suction pipe of all injectors which 
 adjust their capacity to suit variations in the steam pressure 
 is to immediately break the jet, and to cause the steam to 
 blow back through the hose into the tank ; but with this 
 instrument any such interference with the normal condition 
 of the jet causes a waste of steam or water at the overflow 
 pipe, which ceases as soon as the disturbing cause is removed. 
 To test this feature, the water in the suction barrel was 
 allowed to fall below the lower end of the suction pipe, so 
 that air would be sucked up into the injector. This caused 
 a discharge of steam and air from the waste pipe, which 
 ceased as soon as the usual level of the water in the barrel 
 was restored. This test was repeated at 200 Ibs. steam pres- 
 sure, when the lifting of the supply water and the forcing 
 it into the boiler occurred the instant the water covered the 
 end of the suction pipe. 
 
 In regard to the injector itself, it may be said that it re- 
 sponded promptly at all times to the movement of the start- 
 
144 
 
 THE GIFFARD INJECTOR. 
 
 ing valve. It is started and stopped by the continuous 
 motion of a single lever, and was regulated by a side motion 
 of the quadrant regulating lever only for the purpose of alter- 
 ing the amount of delivery. Its construction is simple and 
 easily understood ; no outside rods, levers or bell- cranks are 
 used, nor complicated internal valves. When hot water is to 
 be lifted it was found that the strongest suction was obtained 
 when the starting lever was drawn forward about i in. and 
 
 FIG. 45- 
 
 4000 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 y 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 ^ 
 
 
 
 
 3600 
 3*00 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 * 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 
 
 
 
 
 '0 
 
 
 
 
 
 
 
 
 
 
 
 
 s 
 
 
 
 
 
 
 
 
 E 
 
 
 
 
 
 
 
 
 
 
 / 
 
 \ 
 
 
 
 
 
 
 
 q 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N 
 
 
 
 
 
 
 
 
 p 
 
 /- 
 
 -v> 
 
 
 
 
 
 
 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 (T) 
 
 
 
 
 
 
 
 
 
 
 
 
 
 -t\ 
 
 - 
 
 ^2900 
 
 bzeoo 
 *. 
 
 ^2200 
 
 *4 
 
 1800 
 1600 
 I4OO 
 1200 
 /OOO 
 BOO 
 
 
 
 
 
 
 v/ 
 
 
 
 Q 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 
 
 j 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 / 
 
 
 
 
 
 t-v 
 
 
 
 
 
 
 
 
 ^1 
 
 I 
 
 
 
 
 
 / 
 
 
 
 
 
 r-i 
 
 
 
 
 
 
 
 
 Ai 
 
 
 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 k; 
 
 
 
 
 f 
 
 5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^, 
 
 
 
 
 i 
 
 3 
 
 
 
 
 
 
 k, 
 
 
 
 
 
 
 
 
 ^ 
 
 
 / 
 
 
 en 
 
 
 
 
 
 
 > 
 
 
 
 
 
 
 
 
 K 
 
 
 ( 
 
 
 
 
 * 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 ^ 
 
 
 
 
 
 
 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 . 
 
 ^f- 
 f 
 
 
 
 
 
 
 
 
 NJ 
 
 
 
 
 
 
 
 
 j 
 
 
 J, 
 
 "<j 
 
 
 
 
 
 
 5 
 
 
 
 
 
 
 
 ^s 
 
 
 
 
 
 
 ^ 
 
 
 
 
 
 
 k 
 
 
 
 
 
 ^ 
 
 
 
 
 
 <-\ 
 
 
 
 o^ 
 
 
 
 
 
 
 ^ 
 
 
 
 ^ 
 
 
 
 
 
 
 
 
 ^ 
 
 
 
 N 
 
 
 
 
 
 
 
 i 
 
 ** 
 
 
 
 
 
 
 
 
 3 
 
 
 
 3 
 
 
 
 
 a- 
 
 ^ 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 N. 
 
 "V 
 
 I 
 
 u 
 
 
 
 
 
 
 
 
 
 
 
 
 g 
 
 
 
 X 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 rf 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3O 4. 
 
 60 6O 70 & 
 
 , 
 
 09OIOO IZO ffO /6O /SO ZOO 
 5TAM ^/?f33 d//?"S. 
 
 the remainder of the stroke given after water appeared at 
 the waste pipe. 
 
 The Diagrams. Fig. No. 45 is a graphical representation 
 of the capacity tests given in lines I and 5 of the Table of 
 Results. At the top, ranging from left to right, are gauge 
 pressures in pounds per square inch, from 30 to 200 Ibs. ; 
 the horizontal lines indicate gallons of water taken from the 
 supply tank per hour, so that the heavy curved line shows 
 
TESTS. 
 
 145 
 
 the change in the capacity with the steam pressure. The 
 maximum capacity increases as the pressure rises from 1,912 
 gals, at 30 Ibs. to 2,535 at 6o I DS -> then 3,510 at 120, 3,760 at 
 1 50 and 4, ooo at 200 Ibs. steam ; the last capacity is higher than 
 that at any lower steam pressure, and was above that of any 
 other injector of the same size, even though the capacity 
 may be the same at 120 or 150 Ibs. The minimum is shown 
 by the lower heavy line of the diagram, and increases from 
 765 gals, at 30 Ibs. to 1,732 at 200 Ibs. steam ; this possible 
 reduction of the capacity at 200 Ibs., from 4,005 gals, to 
 1,732, is such a great variation in the amount of water deliv- 
 
 FIG. 46. 
 
 62 
 60 
 
 
 
 
 v 
 
 ... 
 
 
 
 
 M, 
 
 
 
 
 
 
 
 1 
 
 
 ?r 
 
 
 
 
 
 
 
 
 ^ 
 
 r- 
 
 ^t. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 *^^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 02 
 90 
 33 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 E 
 
 
 
 "X 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^ 
 
 ^ 
 
 " 
 
 
 
 
 \ 
 
 ^ 
 
 v/, 
 
 v// 
 
 g 
 
 y 
 
 
 ^ 
 
 
 
 I 
 
 $s 
 
 / 
 
 
 
 
 
 
 
 
 
 
 
 
 J 
 
 r j > 
 
 SO 60 7O 80 9O fOO /2O J4O /6O 4. 
 
 ?i? it 
 
 to 
 
 ered that it is very evident that the injector can be used to 
 feed the boiler continuously with either a light or heavy 
 train. 
 
 The ratios between the maximum and the minimum are 
 given by the heavy line in the lower part of Fig. No. 46, 
 and the range in percentages is given in the upper part. 
 These same values are given in lines 8 and 10 of the table. 
 
 Upon the figures obtained during the test to determine the 
 
 maximum capacity, are based the values given in Fig. 47, 
 
 page 146, also line 4 of table, which show the weight of water 
 
 taken from the supply tank per Ib. of steam used by the 
 
 10 
 
146 
 
 THE GIFFARD INJECTOR. 
 
 injector. This represents the actual amount of mechanical 
 work done by the steam, and is a point of special value to 
 practical men ; it is a good gauge of the efficiency of the 
 design of the injector and of its economy as a boiler feeder, 
 as it indicates that the minimum amount of steam is used to 
 perform the work of feeding, and no excess is condensed and 
 utilized only for heating the feed water. Under the same 
 conditions of supply temperature and lift, the weight of water 
 
 FIG. 47. 
 
 26 
 Z 
 
 24 
 23 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Zl 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \l\ 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ] 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 V' 19 
 
 
 
 
 
 
 s 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ^" 
 
 l/ 
 
 / 
 /4 
 /3 
 /Z 
 /f 
 M> 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 v^ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 s 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 s 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 v 
 
 s 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 JO * 
 
 950 60 70 GO 90 /OO i20 /W /6O 180 ZOO 
 STEAM- PfffSSUfffS 
 
 delivered per pound of steam must always decrease as the 
 pressure rises. 
 
 A critical comparison of the results obtained on this experi- 
 ment proves the superior design of this pattern. Referring 
 to the actual figures of Fig. 47 or the table, it is seen that 
 at 120 Ibs. steam, 13.6 Ibs. of water are taken from the 
 tank and forced into the boiler by i Ib. of steam and at 
 200 Ibs. 10.34 Ibs., the latter result being especially remark- 
 able. 
 
TESTS. 
 
 147 
 
 It is very seldom necessary that a locomotive injector is 
 required to feed when the temperature of the supply exceeds 
 loo degs., but when the occasion demands, the action should 
 be certain and permit a fair range of capacities. Many in 
 jectors will not operate at the higher pressures with the sup- 
 ply at this temperature, consequently their action when start- 
 ing with hot feed pipes causes a very large amount of over- 
 flow before the jet enters the boiler. Several temperature 
 tests were made at 30, 60, 120 Ibs. steam, etc., and the results 
 are given in Fig. No. 48 and lines n and 12 of table. The 
 
 FIG. 48. 
 
 ^ 
 
 X5 
 
 
 *l 
 
 30 4O SO 6O TO QO 9O 
 
 limiting temperature at which the injector is re-started at va- 
 rious steam pressures is given by the lower line, and the max- 
 imum temperature to which the supply may be raised before 
 the injector ceases to operate is shown by upper line. The 
 vertical lines of the diagram indicate steam pressures as 
 before and the horizontal lines degrees Fahrenheit. Starting 
 at 30 Ibs. steam, the limiting temperature is 138 degs., which 
 rises to 143 degs. at 60 Ibs. ; at 120 it is 137 deg., and at 150 
 Ibs. pressure 133 degs These were obtained with the waste 
 valve closed to prevent the waste which would occur as 
 these limiting temperatures were approached. 
 
148 
 
 THE GIFFARD INJECTOR. 
 
 It may be stated that the greatest care was taken to pre- 
 vent the possibility of error in observation or in accuracy of 
 apparatus. The operators were all skilled in experimental 
 work and the observation and calculation of each were care- 
 fully checked. The injector was taken directly from stock 
 and without special preparation, and in performance it ex- 
 ceeded in every detail the requirement of the specification. 
 
 TABLE OF RESULTS. 
 
 Test of a Sellers' Improved Injector of 1887 ; Size, 10^ ; Lift, 4 
 feet; Supply Water Weighed. 
 
 MAXIMUM CAPACITY. 
 
 I 
 
 Mean Steam Pressure 
 
 30 
 
 60 
 
 122 
 
 151 
 
 200.5 
 
 Gallons of Water per Hour . . 
 
 1912 
 
 2535 
 
 3517 
 
 37 6 5 
 
 4005 
 50.5 
 
 2 
 
 3 
 4 
 
 Temperature of Supply Water . 
 
 67.0 
 
 67.0 
 
 54-0 
 
 50.0 
 
 Temp, of Delivered Water . . . 
 
 "3-25 
 
 125.0 
 
 133-4 
 
 135-7 
 
 i54.o 
 
 Weight of delivered Water per 
 Pound of Steam Used .... 
 
 25.90 
 
 19.10 
 
 13.60 
 
 12.60 
 
 IQ .34 
 
 MINIMUM CAPACITY. 
 
 5 
 
 Mean Steam Pressure 
 
 30 
 
 765 
 
 60 
 
 1 2O 
 
 I 4 8 
 
 200. 6 
 
 Gallons of Water per Hour . . 
 
 937 
 
 1290 
 
 1432 
 
 1732 
 
 6 
 
 Temperature of Supply Water . 
 
 67.0 
 
 67.0 
 
 212 
 
 54-5 
 
 55.o 
 
 50.0 
 
 7 
 
 Temp, of Delivered Water . . 
 
 171 
 
 238 
 
 250 
 
 263 
 
 RANGE). 
 
 
 Mean Steam Pressure 
 
 30 
 
 60 
 
 121 
 
 '49-5 
 
 
 8 
 
 Per cent. Capacity of Max. . . 
 
 40.0 
 
 37.o 
 
 36.6 
 
 38.0 
 
 9 
 
 10 
 
 Actual Range in Gals, per Hour 
 
 1147 
 
 1598 
 
 2227 
 
 2333 
 
 Per cent. Range of Max. Capac'y 
 
 60.0 
 
 63.0 
 
 63.3 
 
 62.0 
 
 200.5 
 43-2 
 2273 
 56.8 
 
 LIMITING TEMPERATURE OF WATER SUPPLY. DEG. FAHR. 
 
 
 Mean Steam Pressure .... 
 
 3 
 
 60 
 
 1 20 
 
 150 
 
 
 II 
 
 12 
 
 Limiting Re-starting Temp. . . 
 
 130 
 
 135 
 
 122 
 
 120 
 
 Limiting Operating Temp. . . 
 
 139 
 
 144 
 
 137 
 
 133 
 
CHAPTER X. 
 
 REQUIREMENTS OF MODERN RAILROAD PRACTICE REPAIRS 
 
 AND RENEWALS METHODS OF FEEDING 
 
 LOCOMOTIVE BOILERS. 
 
 Owing to the introduction of the compound locomotive or 
 the demand for motive power having increased hauling capa- 
 city, there has been a general tendency during the last few 
 years, to materially increase the pressure of the steam carried 
 on locomotive boilers, and almost all new engines are de- 
 signed to carry from 180 to 205 pounds steam pressure. 
 
 The designing of an injector for these high pressures is a 
 much more difficult problem than was before presented. The 
 higher the operating pressure, the greater the difficulty in 
 fulfilling the requirements of a good locomotive injector. The 
 criticism against most injectors, and a characteristic of the 
 single jet, fixed nozzle type, is that if it is proportioned to 
 operate at 180 or 200 pounds of steam, it cannot be made to 
 work at the low pressures carried in the round house. There 
 are now in the market several special forms of injector, 
 which show commendable effort to meet the new condi- 
 tions, and which operate over a wide range of pressure. All 
 injectors with fixed nozzles are most efficient at the special 
 pressure for w r hich the tubes are designed, and although 
 they admit by hand adjustment of considerable variation of the 
 steam pressure, yet at no other higher pressure will they ope- 
 rate as efficiently. With injectors having two sets of tubes, 
 this permissible range is much more extended, but the 
 mechanical efficiency is seldom as high as that of some of 
 the special types. 
 
 But the principal feature which interests the user or pur- 
 chaser of the injector, is the falling off in the capacity as the 
 steam pressure is raised, which often necessitates the pur- 
 
 149 
 
150 THE GIFFARD INJECTOR. 
 
 chase of a larger size instrument to obtain the required num- 
 ber of gallons per hour. This entails the consumption of 
 more steam, with a heavier drain on the boiler and conse- 
 quent loss of pressure. Further, the action at these pressures 
 is less certain and the jet much more sensitive. 
 
 Considering the problem in detail, it will be seen that 
 there are a number of features which it is advantageous that 
 a modern locomotive injector should possess, and which are 
 found to a greater or less extent in certain of the most 
 improved types. 
 
 One of the chief requirements is simplicity of construction 
 and operation ; special care should not be required to prime, 
 regulate and start ; every unnecessary movement of the engi- 
 neer should be dispensed with; the injector should be ope- 
 rated and adjusted by the movement of a single lever, or at 
 most, by two, one for the steam valve and the other for the 
 feed supply. The action should be positive, so that when 
 once started, it can be depended upon to operate continu- 
 ously without being affected by shock or jar, or change of level 
 of water supply ; it is obvious that every demand upon the 
 attention of the engineer by the lubricator, injector or other 
 device connected with the engine, requires time that should 
 be devoted to the operation of the engine itself, watching 
 the track for passing signals or adjusting the cut-off, and 
 therefore each device should be as nearly self-operative as 
 possible. 
 
 Fully as important as simplicity of construction, is the 
 general action. A locomotive injector should operate equally 
 well throughout the complete range of boiler pressures 
 without hand adjustment of any kind Not only is this 
 necessary on the road for the reasons outlined above, but on 
 account of the use of the injector by the hostlers or men in 
 charge of the engines in the round house. It should give 
 the maximum delivery at 200 pounds steam and work with 
 a wide-open lazy cock at 20 and 30 as well as at 200 pounds, 
 for it is evident from the discussion of the theory of the 
 injector, that the capacity should vary with the steam pres- 
 sure ; now if it is necessary to regulate the water supply 
 
REQUIREMENTS. 151 
 
 when the pressure falls, for instance, from 180 to 160 pounds, 
 due to a special drain upon the steaming capacity of the 
 boiler, the attention of the engineer is taken from his other 
 duties and must regulate and adjust his feed, to prevent all 
 the delivered water from the injector passing out through 
 the overflow into the ash pan. 
 
 It is very desirable also to have a wide range of 'capacities 
 at the usual working pressure. This is difficult to obtain at 
 1 80 and 200 pounds steam, but in order to obtain the best 
 results in feeding, the range should be at least 50 per cent., 
 and more if possible. The special reasons for this will be 
 given in connection with the methods of feeding locomotive 
 boilers. 
 
 On page 65 is given the difference between the amount of 
 steam used by two different styles of injectors ; this is due 
 in great measure to the use of a larger steam nozzle than is 
 necessan r . An injector, is practically a pump in which the 
 actuating steam is condensed. Treat the injector as a me- 
 chanical device ; that device which gives the required 
 delivery with the minimum consumption of steam or energy, 
 is the most efficient. Every pound of steam taken out of the 
 steam space means a subtraction of a pound of steam from 
 use on the pistons, and there are times during a heavy pull, 
 when steam in the cylinders is much more valuable than hot 
 water in the boiler. The author has known of cases where 
 there has been shown a very material improvement in the 
 performance of a badly steaming engine by changing to a 
 more economical form of injector. By an efficient injector is 
 meant one in which the delivery is large per unit weight of 
 steam. What this ratio should be, is difficult to say, but 
 recent improvements have raised the delivery of water per 
 pound of steam, from 8.8 pounds to 11.2 at 180 pounds pres- 
 sure. Besides the advantage of having less tendency to pull 
 down the steam pressure, an efficient injector induces a certain 
 amount of economy in the coal consumption, by increasing 
 somewhat the evaporative capacity of the boiler. The 
 greater the difference between the temperature of the waste 
 gases passing into the smoke-box and the feed water, the 
 
152 THE GIFFARD INJECTOR. 
 
 greater the transfer of heat through the boiler tubes to the 
 water. The temperature of the waste gases varies from 800 
 to 1200 degrees, and if any part of this waste heat can be 
 absorbed by the feed water, it is a clear gain, and this can 
 be obtained by using a lower temperature of the delivery, 
 with smaller steam consumption. 
 
 There is another feature which commends itself to many 
 practical railroad men, not only on account of the added 
 certainty given to the action of a locomotive injector under 
 all conditions, but for the convenience in obtaining the full 
 range of capacities. When an injector is re-starting, it will 
 continue to force water into the boiler as long as it is sup- 
 plied with water and steam, and will not blow steam back into 
 the suction pipe even if the continuity of the jet should be 
 disturbed by an interruption of the water or steam supply; 
 further, the water supply can be reduced below the actual 
 minimum capacity of the injector without breaking the jet, 
 so that close regulation can be obtained without special care. 
 Operating closed overflow injectors at high pressures on fast 
 express trains is avoided by engineers wherever possible, if 
 there be slightest danger of the injector breaking, for the 
 time occupied in repriming and starting an injector with hot 
 pipes may be two or three minutes, sufficient for the train 
 to have covered as many miles, during which time the atten- 
 tion of the engineer must be more fully occupied with the 
 injector than his other duties. The probability is, therefore, 
 that the engineer will run no such chance, and will operate 
 his injector with the least risk of losing time, even though 
 his method of feeding the boiler may not be the best for 
 obtaining economy of fuel. 
 
 The temperature of the supply water does not often enter 
 into the consideration, so long as the range of capacities of 
 a locomotive injector is not affected by summer temperatures. 
 Some of the South American and foreign railroads specify 
 the limiting temperature of the water supply, but these cases 
 are exceptional ; the introduction of blow-back valves or the 
 return of the air-brake exhaust to the tank has required the 
 use of hot-water injectors in a few cases, but at the present 
 
REPAIRS. 153 
 
 time the use is not general, for it is generally accepted that 
 no uncertainty should be allowed in the feed apparatus, and 
 any increase in the temperature of the water supply renders 
 the action of any injector less efficient and consequently 
 reduces the length of life of the tubes. 
 
 REPAIRS. 
 
 Given an injector which has been accepted by the officials 
 of a railroad as the most suitable, the next practical question 
 which arises is that of maintenance and repair. If it fulfill the 
 first-named requirement of simplicity of construction, it will 
 not be difficult for the employee having special care of the 
 boiler feeders to comprehend the ordinary cause of com- 
 plaint on the part of the engineer or fireman who operates it. 
 A few practical hints to the inspector may be of value. 
 
 The first duty on his part will be to examine the connec- 
 tions and joints to discover leaks or stoppage in the strainer, 
 suction and delivery pipe, and the check valves. These 
 proving to be in good condition, the injector should be 
 tested under full working pressure. The methods then to 
 follow depend to a great extent upon the style of injector, 
 and its action during this test. Assuming it to be of the 
 single-jet, fixed-nozzle class, like the Monitor, Ohio, Mack, 
 etc., and that it cannot be made to prime, the fault probably 
 lies with the lifting spindle or priming nozzles. If these 
 nozzles are separate, as in the older form of Monitor, first 
 examine the lifting steam nozzle to see if stopped up, then 
 if tightly screwed to seat, and lastly, joint of body, as air may 
 enter the combining tube at this place and so destroy the 
 suction. The overflow nozzle may be bent, loose or out of 
 line, or the drip pipe so close, or of such small size, that the 
 free discharge of steam is impeded ; or the steam valve to 
 the main steam nozzle may leak to such an extent as to fill 
 the suction pipe and prevent the lifting of the supply water. 
 With other patterns* of injector in which the priming is 
 effected by a central spindle discharging steam through the 
 combining or delivery tube, such as the Sellers' 1876 Injector, 
 
154 THE GIFFARD INJECTOR. 
 
 the Ohio, Monitor 1888, etc., repeat the same method of proce- 
 dure ; examine the lifting jet, unscrewing it if necessary, to 
 remove any obstructing particle of coal or scale. This lift- 
 ing jet should be in exact line with the other tubes, and if 
 bent, discharges the steam against the side of the other 
 nozzle and the desired vacuum is not produced. If the injec- 
 tor is of the double-jet type, the Metropolitan, Hancock, 
 Schutte, for instance, all the applicable tests above described 
 should be used ; also, ascertain if the position of the small 
 spindle which regulates the steam pressure in the lifting noz- 
 zle permits a sufficient amount of steam to enter, and then 
 if the steam valve to the forcer set is closed when the lifter 
 steam valve is wide open ; further, there may be interference 
 with the free discharge of the steam through the body, owing 
 to inaccurate setting of the relief or overflow valves due to 
 wear of operating parts or faulty construction, or to the valves 
 being prevented by some cause from lifting from their seats ; 
 in some cases, where loose valves or soft seats are used, they 
 may have become separated from their attachments and 
 obstructed some of the waste passages. In the Improved 
 Sellers', 1887, and the 1897 Monitor, examine the inlet valve 
 in the passage from the overflow chamber to the suction ; 
 this valve may not seat firmly, due to the bending of the 
 stem or to some obstruction ; this valve should be perfectly 
 tight, or the discharged steam from the lifting nozzles can 
 pass directly into the suction chamber. 
 
 If the preliminary test showed the lifting apparatus to be 
 operative, and that the fault lay with the delivery of the 
 feed to the boiler, the methods to be applied would depend 
 upon the kind' of injector and its action when steam was 
 admitted to the forcing nozzles. If the injector be of the 
 simpler type a heavy waste at overflow indicates a stoppage 
 of the tubes, or that the delivery tube is worn too large for 
 further use ; the latter condition would be still more evident 
 at the lower pressures. Another possible cause of trouble is 
 the interference with the jet due to displacement, breakage 
 of or obstruction by the line-check valve. In some forms 
 
FEEDING LOCOMOTIVE BOILERS. 155 
 
 of the double-jet type, the non closing of the overflow relief 
 valve between the lifter and the forcer sets prevents the pro- 
 per inter-action of the two jets, and the feed water will not 
 enter the boiler ; or, the tubes of the lifter or forcer may be 
 unequally worn, so that the correct ratios are lost, and either 
 too little or too much water may be supplied to the forcer 
 tubes. 
 
 The injector should then be removed from the locomotive, 
 the parts carefully unscrewed and examined ; if the diameters 
 are worn too much, new parts should be subtituted ; exactly 
 how much enlargement is permissible, depends upon the 
 kind of service and the pattern of injector, some special 
 styles permitting more latitude in this respect than others. 
 One of the best practical tests of the superiority of an 
 injector is the length of service without removal from the 
 engine, and the cost of maintenance ; the former may vary 
 from one to six or seven years ; the latter depends upon the 
 design, the cost of repair parts, and the condition of the 
 water and steam supplied to it. It is often possible to replace 
 the old tubes if the diameters are not too much enlarged, after 
 the interior surfaces have been carefully smoothed out with 
 fine emery cloth, and all roughness and abrasion removed, 
 for the slightest impediment to the rapid motion of the moving 
 mass of water and steam interferes with the action of the jet. 
 
 All injectors should be tested before replacing upon the 
 boiler. A simple form of testing plant, such as that shown 
 in Fig. 43, should be part of the equipment of every 
 large railroad shop conducting extensive injector repairs ; it 
 should be connected to a boiler capable of carrying as high 
 pressure as that used on locomotives ; the proving test 
 should be made as complete as possible, and a permanent 
 record preserved for reference. 
 
 FEEDING LOCOMOTIVE BOILERS. 
 
 In the economical handling of the locomotive, the way in 
 which the boiler is fed plays an important part. To lay 
 down a rigid rule is manifestly impossible, but that generally 
 accepted, and adopted where conditions permit, is to main- 
 
156 THE GIFFARD INJECTOR. 
 
 tain a constant water level with a continuous feed. There are 
 a number of reasons for this : the strains on the boiler sheets, 
 due to changes in temperature and unequal expansion, are 
 reduced ; the drain upon the steam supply is more constant, 
 owing to the continuous operation of the injector; less 
 water is wasted in stopping and starting the injector, and 
 less time and attention need be given on the part of the 
 operator; on long level runs, this method can almost 
 always be pursued, if the engineer is provided with an 
 injector having a wide range of capacities ; but there are 
 several special rules to be observed ; in approaching a sta- 
 tion at which a short stop is made, especially between long 
 and fast runs, it is advantageous to stop the injector a short 
 time before the station is reached, to permit a slight checking 
 of the fire, and then, when the station is reached, to feed the 
 boilers quickly with one, or even with both injectors if neces- 
 sary, to prevent blowing off at the safety-valve. Also, the 
 feed should be stopped before starting from a station ; the 
 train can then be drawn out with full boiler pressure, and 
 there need be no drain on the steam supply due to use of 
 the injector, until the train is under good headway and the 
 exhaust has pulled the fire into good working condition. 
 With a badly steaming or overloaded engine drawing a 
 heavy train it is very difficult to raise the pressure to the 
 normal, having a low water level to overcome in addition. 
 
 A curious phenomenon, often observed in connection with 
 the feeding of locomotive boilers, is the variation in the 
 water level after an injector has been started ; when a boiler 
 has not fed for some time and the engine has been working 
 hard with a heavy exhaust, clear fire and waste gases at a 
 high temperature, the entire body of water is filled with 
 bubbles of steam working upward to the surface; this 
 mass of confined air and steam displaces a large volume of 
 water, while the agitation of the surface due to the mo- 
 tion of the engine tends still further to raise the apparent 
 water level. If the boiler feeder is now started, the cooler 
 water condenses the rising steam bubbles and diminishes the 
 volume of water, and lowers the level in the glass. In 
 
FEEDING LOCOMOTIVE BOILERS. 157 
 
 former days, when pumps were used, five to ten minutes 
 were often required before any increase in the height of water 
 would be shown, and although the injector would not pro- 
 duce so marked an effect, yet a deceptive contraction in the 
 volume does occur, which should be carefully watched ; 
 otherwise, when normal conditions of ebullition are estab- 
 lished, it will be found that the water level has risen very 
 suddenly, at the probable risk of water passing over into 
 the cylinder. 
 
 In order to obtain the most economical results, every 
 device used on and about a locomotive should be of the most 
 efficient pattern. No operator, whether on a railroad or 
 elsewhere, can do his best for his employer unless he is sup- 
 plied with suitable appliances or tools, and given the means 
 whereby he can be most economical of his own effort and 
 the material at his disposal. This is especially true of the 
 injector, which should be arranged to give good results with 
 the fewest movements and with the least inconvenience to 
 the operator. This is an important and practical question : 
 if an engineer has several handles to manipulate when one 
 should answer as well, or if fine and careful adjustment is 
 required to give good results, or if there be risk of breakage 
 of the jet when throttling to the minimum, the engineer 
 will operate his injector as best suits his convenience, 
 with little reference to the finer points of economy. On the 
 other hand, all good men are ambitious and anxious to do 
 their best and make records for themselves which may 
 tend to their advancement ; if an engineer or fireman feels 
 that he can obtain good results and will receive material 
 encouragement in his efforts, he will steadily improve. Every- 
 thing about a locomotive should be the best of its kind, 
 whether air-brake, lubricator or boiler-feeder, and should be 
 carefully tested by the motive-power department, and full 
 evidence obtained that it approaches most nearly the standard 
 of modern excellence. Great advancement has already been 
 made, yet further improvements are expected at the hands 
 of those who are working toward the improvement of the 
 
158 THE G1FFARD INJECTOR. 
 
 injector, along the lines laid down at the opening of this 
 chapter. 
 
 In conclusion, a word might be said regarding the careful 
 handling of injectors, which are not delicate instruments, 
 yet require occasional attention in order to give the best 
 satisfaction. A little care in closing the valves and pushing 
 in levers will prolong the life of the parts and reduce the 
 necessity for grinding and repairing. If dirt or scale is 
 caught between a valve and its seat, causing a slight leak, a 
 touch with a seat reamer or a little time taken to grind the 
 surfaces to a bearing, will save much future trouble. All 
 valves should be kept tight, as a leak tends invariably to 
 increase rapidly, and the trouble is magnified if repairs are 
 postponed. The deposit of a hard scale upon the interior 
 surfaces of the tubes and in the overflow vents is a continual 
 source of trouble when the feed water contains lime in solu- 
 tion; a small amount of oil fed into the steam or suction 
 pipes alleviates but does not entirely prevent the formation 
 of the deposit ; it may, however, in most cases be removed 
 by allowing the tubes to remain for eight or ten hours in 
 solution of one part of muriatic acid in ten of water. 
 
 The repairing of injectors should never be intrusted to a 
 novice, and where large numbers are in use, it is customary 
 on large railroads to combine the air brake and injector 
 repairs in the same department, and place the responsibility 
 in the hands of experienced workmen ; every opportunity 
 should be afforded these men for obtaining a thorough 
 knowledge of the theory of action, and the special pecu- 
 liarity of each pattern of injector in use, and it has been the 
 experience of the author that courteous attention is always 
 given by the well-known manufacturing companies to those 
 who seek for information. 
 
INDEX. 
 
 AIR IN STEAM, 72. 
 
 AUTOMATIC DOUBLE JET, 23. 
 
 AUTOMATIC INJECTOR Definition of, 
 27 ; Earliest Forms of, 22 ; Exhaust, 
 113; Gresham, no; Manhattan, 
 III; Metropolitan, 109; Penberthy, 
 110-119; Sellers' 1885, 108 ; Sel- j 
 lers' 1887, 104; Webb, 92. 
 
 BACK PRESSURE, 119; Effect of Feed- 
 Temperature Table, 72 ; Experi- 
 ments, 34, 122. 
 
 BANCROFT, J. S. Improvements by, 22 
 
 BELFIELD INJECTOR, 66 ; Description, 
 
 102. 
 
 BOURDON Inventions by, 3, 
 
 BOILER EVAPORATION, 126. 
 
 BOILER TESTERS, 121. 
 
 BUFFALO INJECTOR, 120. 
 
 CAPACITY See Weight of Water per 
 Pound of Steam. 
 
 COEFFICIENT OF IMPACT, 77, 78. 
 
 COMBINING TUBE, 45, 69; Definition 
 of, 26; Losses in, 74; Vacuum 
 Within, 49 ; Wear of, 51. 
 
 DELIVERY TEMPERATURE, 72, 78, 106. 
 
 DELIVERY TUBE Definition of, 26 ; 
 Pressure in, 33 ; Wear of, 42. 
 
 DENSITY OF JET Effect of Tempera- 
 ture of Feed, 78 ; in Delivery Tube, 
 
 31, 72. 
 
 DESMOND Improvements by, 22. 
 DIAGRAM Maximum and Minimum 
 
 Capacities at different Feed Temper- 
 
 atures, 127; Maximum and Mini- 
 mum Capacities at different Pres- 
 sures, 125, 144; Pressures in De- 
 livery Tube, 40; Pressure within 
 Steam Nozzle, 57, 60; Velocity and 
 Pressure in Delivery Tube, 32. 
 
 DIVERGENT FLARE Delivery Tube, 
 34, 39; Steam Nozzle, 54. 
 
 DIVERGENT STEAM NOZZLE Inven- 
 tion of, 21. 
 
 DOUBLE JET Automatic Form of, 23 ; 
 Definition of, 27 ; Effect of Height 
 of Lift, 77; Pressure between the 
 two sets of tubes, 78 ; Relative Pro- 
 portions to Single Jet, 78. 
 
 DOUBLE JET INJECTOR Belfield, 102 ; 
 Description of, 19; for High Feed 
 Temperature, 50; Schutte, 101. 
 
 EBERMAN INJECTOR, 114. 
 
 EFFICIENCY Cause of loss of, 86; 
 Compared with Pump, 87 ; Defini- 
 tion of, 28; Mechanical, 84, 86; of 
 various types, 66. 
 
 EJECTORS, 123; Definition of, 25. 
 
 ENGLISH INJECTORS, 90. 
 
 ENGLAND Introduction of Injector 
 into, 4. 
 
 EXHAUST INJECTOR, 64; Description 
 of, 118 ; Length of Combining Tube, 
 48; Theory of Action, 70. 
 
 FEED WATER Purity of, 51 ; Tem- 
 perature of, 50, 101, 106, 114, 
 119, 147. 
 
 159 
 
160 
 
 INDEX. 
 
 FEED TEMPERATURE Effect upon Ca- 
 pacity, 78 ; Effect on Maximum and 
 Minimum Capacity (Diagram), 133; 
 Highest Admissible (Formula for), 
 85. 
 
 FEEDING LOCOMOTIVE BOILERS, 155. 
 
 FIRE PUMPS, 123. 
 
 FRANCE Improvement in, 20. 
 
 FRENCH INJECTORS, 93. 
 
 FRIEDMAN Improvements by, 21. 
 
 GARFIELD INJECTOR, 32, 66, 109. 
 
 GIFFARD, H. J., Original inventor, I. 
 
 GIFFARD INJECTOR, 10-32. 
 
 GRESHAM AND CRAVEN, 92. 
 
 GRESHAM, JAMES Improvements by, 
 18. 
 
 GRESHAM RE-STARTING INJECTOR 
 Description of, 116. 
 
 HAMER, METCALF AND DAVIES' EX- 
 HAUST INJECTOR, 21. 
 
 HANCOCK INJECTOR, 109. 
 
 HANCOCK, JOHN Improvements by, 
 18. 
 
 HOLDEN AND BROOKE Improvements 
 by, 23. 
 
 HORSE POWER Required to Feed 
 Boiler, 65. 
 
 IMPACT Coefficient of (Table), 72; 
 Losses during, 86. 
 
 INJECTOR Definition of Term, 25. 
 
 IRWIN INJECTOR, 48. 
 
 ITALIAN INJECTORS, 93. 
 
 JENKS INJECTOR, 120. 
 
 KNEASS, STRICKLAND L. Improve- 
 ments by, 23. 
 
 KORTING, ERNEST Improvements by, 
 1 8. 
 
 LIFT Effect upon Capacity (Table), 
 76. 
 
 LITTLE GIANT INJECTOR (RUE), 66; 
 Description of, 113; Experiments 
 with, 46. 
 
 LOCOMOTIVE INJECTORS Sizes Re- 
 quired (Table), 121. 
 
 LOFTUS, JOHN Improvements by, 22. 
 
 LOSSES BY CONDUCTION OF HEAT, 83. 
 
 LOSSES IN COMBINING TUBE, 69, 74. 
 
 MACK INJECTOR, 32, 66, 107. 
 
 MANHATTAN INJECTOR Description 
 of, 117. 
 
 MAXIMUM CAPACITY Conditions af- 
 fecting, 119. 
 I MECHANICAL EFFICIENCY, 64. 
 
 METROPOLITAN INJECTOR, 32, 66, 108, 
 
 121. 
 
 MILLHOLLAND Improvements by, 13. 
 MINIMUM CAPACITY (Formula), 84. 
 MONITOR INJECTOR, 32, 36 ; Descrip- 
 | tion of, 97. 
 MONITOR STANDARD, 90. 
 NATHAN MFG. Co., 94. 
 NON-LIFTING INJECTOR, 89. 
 OHIO INJECTOR, 109. 
 OIL Use of to Prevent Scale, 129. 
 PARK INJECTOR, 120. 
 PENBERTHY, WILLIAM Improvements 
 
 by, 22. 
 PENBERTHY INJECTOR Description of, 
 
 Il8; "SPECIAL," IIO. 
 
 PENNA. R. R. STANDARD, 90, 129. 
 
 PRESSURES Within Delivery Tube 
 (Diagram of), 32; Within Steam 
 Nozzle, 57. 
 
 PUMP Compared with Injector, 87. 
 
 RANGE OF CAPACITY Definition of, 28. 
 
 RE-STARTING Sellers', in. 
 
 REPAIRS, 153. 
 
 ROBINSON AND GRESHAM Improve- 
 ments by, 13. 
 
 RUE INJECTOR See " LITTLE GIANT." 
 
 RUE, SAMUEL Improvements by, 13. 
 
 SCALE Prevention of, 158. 
 
 ScHAU Improvements by, 21. 
 
 SCHUTTE (KORTING) INJECTOR De- 
 
 scription of, 100. 
 
 SELF-ADJUSTING INJECTOR Action of 
 Combining Tube, 48; Advantages 
 of, 47 ; Definition of, 27 ; Descrip- 
 tion of, 15, 96. 
 
 SELLERS, WM. Inventor of Self-Ad- 
 justing Principle, 14. 
 
 SELLERS' 1876 INJECTOR, 32. 
 
INDEX. 
 
 161 
 
 SELLERS' 1885 INJECTOR Description 
 of, ic8 ; High Feed Temperature, 51. 
 
 SELLERS' RE-STARTING, in. 
 
 SELLERS' 1887 INJECTOR, 66; De- 
 scription of, 103. 
 
 SETTING INJECTORS, 120. 
 
 STEAM DISCHARGE, 58 ; Phenomena 
 of, 54; Rate of, 61, 63; Formulae, 
 61, 63. 
 
 STEAM JETS Air mixed with, 71. 
 
 STEAM NOZZLE Definition of, 26 ; 
 Ratios Throat to Initial Pressures, 
 58, 60, 61. 
 
 STEAM PRESSURE Effect on Maxi- 
 mum and Minimum Capacity (Dia 
 gram), 124, 144. 
 
 STEWART, ROBINSON AND GRESHAM 
 Improvements by, 20. 
 
 TABLE I. Delivery of various Injec- 
 tors per Unit Area of Delivery Tube, 
 
 32. 
 TABLE II. Pressure and Weight of 
 
 Water, 43. 
 
 TABLE III. Values of 0, 65. 
 TABLE IV Density of Jet in Delivery 
 
 Tube, 72. 
 TABLE V. Variation of Weight of 
 
 Water per Pound of Steam with Dif- 
 ferent Heights of Lift, 76. 
 TABLE VI. Tension of Vapor of| 
 
 Water, 77. 
 TABLE VII. Variation of Capacity 
 
 with Feed Temperature, 78. 
 
 TABLE VIII. Weight of Water per 
 Pound of Steam at Different Steam 
 Pressures, 79. 
 
 TESTS OF INJECTORS, 129. 
 
 THEORY OF INJECTOR, 67. 
 
 VABE Early Form of Automatic In- 
 jector, 22. 
 
 VARIATION OF DENSITY OF JET WITH 
 FEED TEMPERATURE (Table), 72, 
 78. 
 
 VELOCITY OF STEAM DISCHARGE, 55; 
 At various pressures, 61 ; Formulae, 
 61 ; At Impact, 72. 
 
 WATER Weight of i Cubic Foot of, 
 at various Temperatures, 43. 
 
 WATER ENTRANCE TO COMBINING 
 TUBE, 46. 
 
 WATER IN STEAM, 64. 
 
 WEAR Combining Tube, 51. 
 
 WEBB'S INJECTOR, 93. 
 
 W. F. INJECTOR Description of, 
 
 99- 
 
 WEIGHT OF STEAM DISCHARGE, 61 ; 
 Experiments and Formulae, 63. 
 
 WEIGHT OF WATER Delivered per 
 Sq. Mm. of Delivery Tube, 66 ; per 
 Pound of Steam (Formula for), 76, 
 83 ; per Pound of Steam at different 
 Steam Pressures, 79 ; per Pound of 
 Steam, 72, 146; Variation with Feed 
 Temperature, 78. 
 
 WILLIAMS, G. C. Automatic Injector, 
 22. 
 
UNIVERSITY OF CALIFORNIA LIBRARY 
 BERKELEY 
 
 Return to desk from which borrowed. 
 This book is DUE on the last date stamped below. 
 
 DEC 4 1947 
 
 LD 21-100m-9,'47(A5702sl6)476 
 
YC 12874