MINUTES OF PROCEEDINGS OF THE INSTITU- TION OF CIVIL ENGINEERS, LONDON, ENGLAND, SESSION OF 1862-63. AMERICAN IRON BRIDGES. ABSTRACT OF THE DISCUSSION UPON* A PAPER SUBMITTED By ZERAH COLBURN. EDITED BY CHARLES MANBY, F. R. S., M. Inst. C. E„ HONORARY SECRETARY. AND JAMES FORREST, Assoc. Inst. C. E., SECRETARY. NEW YORK: D. VAN NOSTRAND, 192 BROADWAY, 1 8 6 7 . ^ Digitized by the Internet Archive in 2017 with funding from University of Illinois Urbana-Champaign Alternates https://archive.org/details/americanironbridOOcolb INSTITUTION OF CIVIL ENGINEEES. May .5, 1863. JOHN HAWKSHAW, President, in the Chair. Mr. Zerah Colburn submitted a Paper describing the several plans of Iron Bridges in use upon the Bailways of the United States, illustrated by a series of diagrams. No. 1,091. — “ American Iron Bridges.” By Zerah Colburn. After reading the Paper, Mr. Colburn said, he had re- frained from entering more into detail to avoid making it tedious, but he should be happy to reply to any questions. No doubt, in looking at the diagrams, it would be considered that American engineers had practised great economy in de- signing these bridges. The fact was, that if bridges could not be constructed of iron at a moderate cost, the Bailway Companies in the States would not adopt them, but would continue to use timber bridges, which could be built at from £5 to £7 per lineal foot. It was on the score of cost, alone, he believed, that American engineers had adopted cast iron for all part^ in compression ; so many square inches of sec- tion could be put into the tubular form for one-third the cost in cast-iron that would be incurred with wrought-iron plates. In the case of the Green Biver Bridge, the top chords and the vertical posts were of cast-iron pipes. Care was requi- site in casting the pipes. Mr. Fink tested the iron for all the pipes, and holes were drilled in the side that lay uppermost in the sand, to ascertain that the cores had not floated, and thus that the metal was of uniform thickness. The fact of cast-iron being used in the top chords was a sufficient expla- nation why bridges were not made continuous over two, or more, spans in America. In nearly all these bridges almost the whole of the iron was made to do work in carrying the load. They, no doubt, differed from what English engineers were accustomed to, and he feared the diagrams would pre- sent an extraordinary appearance to English eyes ; but they were faithful representations. Mr. Colburn then called attention to the drawing of a bridge with truss rods to every upright post. The span was ^ 45939 4 AMERICAN IRON BRIDGES. 125 feet, and tlie depth was 23 feet from the centre of the top chord to the centre of the bottom chord ; yet all the iron posts, the truss rods, the diagonal tension rods, and the lean- ing end columns, were not together equal to a quarter of an inch in thickness of iron, if spread out as a plate, oyer the whole side, and half of that thickness was cast-iron ; still no portion of the iron was subjected to a greater working strain than 4 tons to the square inch. Under ordinary circum- stances a working strain of 3 tons to the square inch was the utmost that existed. He contrasted with that design some plate girder bridges, erected three years ago, on the Boston and Worcester Bailway. They were 87 feet span and 7 feet 6 inches deep. The plates were 6 feet 3 inches wide and 7 feet 6 inches high. At every vertical joint there was a pair of butt straps, 8 inches wide, double riveted on each side of the web, and over these a pair of angle irons, 3 inches by 6 inches. Midway between the vertical joints, there were two angle irons, 6 inches by 3 inches, the longer sides being turned over at right angles, to form a knee by which they were riveted to the top and bottom chords. Mr Philbrick, the engineer, adopted that form of stiffening the webs of girder bridges, because, as he said, he had found that in England and in France, the T irons used for stiffening had begun to split in some cases. He had only called attention to the butt straps and angle irons used* by Mr. Philbrick, be- cause these, spread out as plates over the side of the web, were alone equal to one-quarter of an inch in thickness, while in the Murphy- Whipple Bridge the whole of the mem- bers forming the web amounted to less than a quarter of an inch. He might add, that none of the bridges shown in the drawings had broken down, and he had never heard that they had exhibited any signs of weakness. They were con- sidered in America, and he believed them to be, very strong, and good bridges. If they had any peculiar merit, it might be in the fact that they were not overloaded with ballast. The trusses were deeper, and, therefore, there was not so much iron put into the sides as might be expected to keep the strains down. In Bollman’s Bridge the test load was 1^ ton per lineal foot, for a single line. In America the assumed load was arrived at, by supposing that there was on the bridge a train of loaded goods wagons, with two engines, and a snow- plough weighing about 15 tons ; and, in addition, when the bridge had a floor, that this was covered with snow, equal to a weight of 30 lbs. per square foot of floor ; and that a side AMERICAN IRON BRIDGES. 5 wind, equal to 30 lbs. per square foot, was blowing. It was further assumed that, as every wheel had a break, all the breaks were put on at once, which would throw the strain upon the top chords of the bridge. That was the assumed load of the bridge of 200 feet span, and it was adopted by Mr. Murphy, Mr. Whipple, and Mr. Bollman. In answer to inquiries from several members, Mr. Colburn said that the highest breaking strain of the cast-iron alluded to in the Paper was 20J tons per square inch ; ‘but it was certainly never subjected to that strain in a bridge. It was boiled cast-iron, that was melted and kept in fusion for sev- eral hours, and partially decarbonized. Previous to the present war in America, every cast gun was made of iron melted two, three, or four times. Mr. F. J. Bramwell regretted to hear that the arrange- ments for the reading of other communications were such, that it was desirable to limit the present discussion ; as he feared there would be great difficulty in investigating the principles of construction of the Trusses described in the Paper, unless some considerable time were devoted to that purpose. With regard to the principles involved in the construction of the Trusses of American Iron Bridges, he referred to three diagrams he had caused to be prepared, showing, re- spectively, the Fink truss (Fig. 1), an ordinary diagonal truss which was so well known in this country (Fig. 2), and the Bollman, or Harper’s Ferry Bridge, truss (Fig. 3). In these diagrams, one-half of each figure was drawn merely in lines, running along the centres of the various members of the truss ; while in the other half of each figure the members were indicated either by dotted fines simply, or by dotted fines and shading on each side, so as to show, by the width of the shading or of the dotted lines, the relative amounts of iron required jn each member ; and by the different kinds of dotted lines and the different kinds of shading, to distinguish between the parts in compression and those in tension ; dis- carding, for the sake of simplicity, all consideration of rivet holes, joints, and such matters, and assuming the whole of the iron to be brought to an uniform thickness, and to be equally valuable for compression and for extension. Thus, if the shading along one member of the truss were twice as wide as along another, then it indicated that there must be double the sectional area of metal in the first of these two members, that there was required to be in the second of such members. In this way the diagrams would at once instruct the eye, as to the consumption of metal in each part of each 6 AMERICAN IRON BRIDGES. truss, while the difference of the shading would show which part was in compression and which in tension. Further, at the end of each figure a parallelogram had been drawn, which indicated the relative amount of each kind of metal required in the three different trusses. Assuming that there would be no special circumstances of convenience, or incon- venience, attending the manufacture of any particular truss, there could be no question that was the best truss which consumed 'the least metal in its construction. And if the decision, as to the merits of the Fink truss (Fig. 1), and of the Bolhnan truss (Fig. 3), were to be arrived at by the con- sideration of this test, Mr. Bramwell believed he was cor- rect in saying, that they were both of them inferior to the ordinary diagonal truss (Fig. 2), and in the following ratio : — Diagonal truss required 262 parts of metal. Fink “ “ 896 Bollruan “ “ 370 The distribution of the metal would be as follows : — Diagonal Truss : Compression 136 Tension 126 262 Fink Truss : Compression 200 Tension 196 396 Bollman Truss : Compression 185 Tension 185 370 These results were arrived at by a calculation which he believed to have been based on correct principles, and to have been accurately worked out. He would, however, en- deavor to explain the causes of the extra consumption of iron in the Fink and the Bollman trusses. First, as to the Fink truss : — The diagram was made on the assumption, that the depth of this truss, and of course of the other two trusses also, was one-eighth of the span, and that the truss was loaded with fifteen units of load, one over each of the vertical struts. The stmts on the left-hand side were lettered from A to H, the latter letter being over the central strut. The effect of these units of load would be, that unit A would, by its vertical A /, be carried one-half by the tie-rod / a, and one-half by / B ; and that the unit B would be carried on its vertical B b } which would also have AMERICAN IRON BRIDGES. 7 to carry the half of unit A, brought on by the tie-rod f B, and also one-half of the unit C, brought on, in a similar manner, by the tie-rod g B. In this way the vertical B b would get a load of two units, one-half of which, or one unit, would have to be upheld through the tie-rod b a, but as the upper part,/ a, of this tie was already strained by the weight due to one-half of the unit A, and as the angle of the tie was 45°, the tie b a would bring a compression of exactly a unit and one-half on the top member of the girder. It was clear that a downward load of one unit and a half must also be brought, by the tie b g D, on to the point D, and that a similar load of a unit and a half must be brought on to the same point by the tie d h D, making together a load of three units. This, added to the unit D itself, would cause a load of four units to have to be carried by the vertical D c. Of these four units two would be carried by the tie c a, and the other two by the tie c H. The tie c a would bring the com- pression on to the top member of the girder due to these two units. If the tie c a had been at the same angle as the tie bf a, this compression would only have been two units ; but as the tie c a was at an angle only half as favorable as the tie bf a, the compression brought on to the top member of the truss by the tie c a would be double that of the verti- cal load carried by it, or a compression of four. The last source from which compression would come upon the top member of the truss would be from the strain of the tie e a. This tie, it would be seen, had to bear one-lialf the duty of upholding the central strut H e. This strut H e would have to support, first, two units of load brought on it by the tie c H ; secondly, one unit of load brought on by the tie d i H, owing to that tie upholding one-half of the load of two on F d, and, thirdly, one-half of a unit of load brought on by the upper part i H of the same tie d i H, which upper part would have to support the half of the unit of load on G i. These three sources of strain would amount together to three and a half units. There would, of course, be an equal load brought on by the corresponding ties to the right hand of the strut H e, which strut would thus be loaded with seven units, „ to which must be added the unit on H itself, making eight as the total load on the strut H e. Of this eight, one-half, or four, would be upheld by the tie e a, and by this tie would exert a compressive force on the top member of the girder. It would be seen, that the angle at which this tie e a lay was twice as unfavorable as that of the tie last considered, and four times as unfavorable as that of the tie bf a, the angle of which was 45°. Owing to this unfavorable angle of the 8 AMERICAN IRON BRIDGES. tie e a, the load supported by it would exert a compression of four times its own amount, or sixteen. The sum of the compressions, therefore, on the top member of the girder would be : — For the tie/ a £ unit x by the effect of the angle 1 = \ “ “ ha 1 “ “ “ “ 1 = 1 “ “ c a 2 “ “ “ “ 2— 4 “ « ea4 t( “ “ “ 4 = 16 / It would be seen, that this compression extended from end to end of the top member of the truss, so that the top mem- ber would require to have as great a sectional area at the extremities as in the middle. Assuming that one part of iron was required to resist one unit of load, then the sectional area would be 21J, and the length of the truss being taken as 8, the amount of iron consumed in the compression bar would be equal to 21J x 8, or 172 parts of iron. Before considering the amount of iron in the other mem- bers of the Fink truss, Mr. Bramwell would compare the compression bar, or top member, of this truss, with that of an ordinary diagonal truss (Fig 2). In the latter case it was well understood that, as fifteen units of load, distributed as shown, were equal to a .central load of eight, and as the girder was eight times as long as it was deep, the compres- sion in the centre would, of course, be sixteen units of load. This, then, would represent the maximum sectional area, which would exist only at the very middle of the compres- sion member, as it would become less and less towards the ends of the girder ; the diminution being in such a ratio, if the number of points of attachment of the ties and struts were infinite, as would give a parabolic curve, the area of which would, <5f course, be two-thirds of the parallelogram containing it. Therefore 16 x § = 10§ would represent the average depth of the compression member, which multiplied by the length 8, would give 85^ as the amount of iron in that member, or rather less than one-half of the iron in the sijni- lar member of the Fink truss. In the diagonal truss, with the number of ties and struts as drawn, this theoretic amount was diminished only to the extent of 1^ ; that was to say, the whole amount of iron was 84, arrived at thus : — AMERICAN IRON BRIDGES. 9 Pressure Between the points a A the compression would be 3£ units, which plied by the effect of the angle , “ * AB “ 10£ “ B C “ 16i “ CD “ 21i “ DE “ 25£ “ E F “ 283 “ PG “ 30£ “ GH “ 31 k , multi- t II 54 84 104 124 144 154 154 Giving for the sum of the compressions 84 This divided bj the number of spaces 8, made 10| as the average depth. It was, he thought, not improbable that a cursory view of the drawing of the Fink truss might lead to the idea that it was one that must be economical ; inasmuch as it appeared to have a compression member only, and to be without any corresponding formal tension member, the diagonal ties be- ing made to serve the purpose of a regular tension member. But even if this were so, and if the ties were made at no greater consumption of metal than the ties and struts of a diagonal girder, which he would prove was not the case, even then the Fink truss would have no advantage ; inasmuch as its one compression member contained slightly more iron than the compression member and tension member together of the diagonal truss. He had stated that the compression member of this latter girder contained 84 parts. The ten- sion member contained rather more, thus : — Tension. Between the points b c the tension would be 4 units, which, multiplied by the effect of the angle = 2 a c d 66 11 66 5 h n d e 66 17 66 H a ef 66 22 11 cc f g 66 26 6 6 13 a gh 66 29 66 14 h a h i 66 31 15 h 66 i k 66 32 16* Giving for the sum of the tensions 86 This, divided by their number, 8, made 10| as the average depth. The amount of metal, 86, in that member, added to that in the compression member, gave 170 parts for the metal in those two members, as compared with 172 parts in the compression member alone of the Fink truss. He would now revert to the consideration of the Fink truss : — 10 AMERICAN IRON BRIDGES. It would be remembered that the main tie bar e a had to transmit four units of load ; but these became multiplied in their power of strain on the tie, by the number of times the length of that tie exceeded the depth of the girder. This depth being 1, and the half length of the girder being 4, the length of the tie would of course be equal to Vl 2 + 4 2 ='v/17. The strain, therefore, would be the load y Vl7. But as the length of the tie was also Vl7, it followed that the amount of metal must be Vl7 x -\/l7 x the load of 4 ; in other words, the amount of metal increased as the square Parts of of the length of any sloping tie. In this case the Metai. amount was 2 ties each 17x4 = 136 Similarly, it could be shown, that the ties ca , cH, and their Counterparts on the other side of the truss, would require V5 x V5 x 2 load x 4 of them = . 40 That ba, bT>, d D, (7H, and their counterparts on the other side of the truss, would require V2 x V2 x 1 load x 8 of them = ..... 16 and that the upper parts of these ties, fa, gD, hD, ^'H, and their counterparts on the other side of the truss, with the corresponding short ties, / B, gB, hF, iF, and their counterparts on the other side of the truss, would require Vj x Vj x J load x 16 of them = 4 Making the total of iron in tension, as already stated 196 Further, the compression member, it had been P Metai! f proved, would contain 172 The strut He would contain 1x8 units = 8 The strut D c, and the corresponding one on the other side of the truss, would contain 1x4 units x 2 of them = 8 The struts B b, F d, and their corresponding ones on the other side of the truss, would contain 1x2 units x 4 of them = 8 The struts A/, Cg, E li, Gf, and their corresponding ones on the other side of the truss, would contain ^ long x 1 unit x 8 of them == 4 Making the total of iron in compression, as already stated 200 AMERICAN IRON BRIDGES. 11 In order to complete the comparison between the Fink truss and the diagonal truss, it would be necessary to ascer- tain the amount of metal in the struts and ties of this latter truss. The lengths of the various struts and ties would be uniformly Vl*, the multiple of strain would therefore be Vl j, and the amount of iron in each of them 1J x the units of load : — On the strut A b there would be 4 units. tt B c (t 3J a it C d a 3 a it D e (( 2J a tt E/ tt 2 a tt E 9 tt li a tt Gh • tt l a tt Hi a 1 2 a Making a total of 18 This sum multiplied by the 1 j, would give 22 To which must be added the end upright ah = 3 Making 26 Or, for the two sides of the truss 52 Adding the amount of metal in the compression mem- ber before given 84 The total of iron in compression would then be, as already stated . 136 Next, as regarded the parts in tension : — On the tie a c there would be 34 units. vyii me tiu a Uj V lilfcJI Ad tJ WUUJLU ft 3 uiiil». tt a B c • tt 2i tt a c/ it 2 tt a Vg tt li it a E h tt i tt a F i tt 1 2 tt it Gh tt 0 tt Making a total of This, multiplied by 1|, would give. . . , 14 12 AMERICAN IRON BRIDGES. Making for the two sides of the truss 35 Adding the amount of metal in the tension member be- fore given 86 The total of metal in tension was thus found to be ... . 121 4 But to this should be added the ties G 1c, which, although not called into action, when an uniform load was on the girder, would have to bear a strain if the load were a passing one. They might then be required to support a single unit of load. Under similar circumstances two ties on each side might require to be strengthened, so as to bear a further half unit each ; then 4 ties 1 unit x thelj = 5 • Giving as the total of iron in tension, as already stated . 126 He would not occupy any further time, by entering into the details of the calculation of the quantity of metal in the Bollman Truss ; but would simply refer to the figures already given, by which it would be seen, that while it was more economical than the Fink Truss, in the proportion of 37 to 39.6, it was less economical than the ordinary diagonal truss, in the proportion of 37 to 26.2. The cause of the waste of metal in the Fink and in the Bollman trusses appeared to him, on a little consideration, to be sufficiently obvious. The strains which related to the centre part only of a truss, and which might be got rid of in a comparatively short distance near that centre, were, in the Fink and in the Bollman trusses, carried to the very ends of the compression member, so that the whole length of that member sustained a pressure that need only be borne near the centre. Further, to carry out this bad arrangement, the tie-rods were, of necessity, placed at most unfavorable angles, by which not only was an unnecessary amount of metal consumed, but the already useless strain on the compression member was aggravated. It was unnecessary to investigate the construction of the Murphy- Whipple truss, as the drawing showed that truss to be free from the radical error, of carrying the whole of the compression to the very ends of the top member. It was, however, clear, that the use of struts in a vertical position was not so economic as the use of struts in a diagonal posi- tion, inasmuch as the struts when placed vertically did not assist in the progression (if such a term might be allowed j AMERICAN IRON BRIDGES. 13 from end to end of the truss. The ties, also, were not dis- posed at the most economic angle ; but with those excep- tions, he had little doubt the Murphy- Whipple truss would, on investigation, prove to be one of good construction, and equal to such as were in ordinary use in England. He regretted having to pronounce so unfavorable an opin- ion on these trusses ; but he was glad to be enabled to do so, without the fear of his criticism being displeasing to the Author, who, it was understood, had merely submitted these particular trusses, as matters of interest in the history of, railway works in America, without expressing approval of their construction, and for the very purpose of having their merits, or demerits, fully discussed. Mr. Phipps said, the Paper was for the most part limited to several varieties of trussed girders. On these he would not at present offer any observations, but would confine his remarks to one or two iron arched constructions, also referred to in the Paper, as he had paid particular attention to the subject of cast-iron arches. It had been truly remarked, that whenever these arched constructions were loaded irreg- ularly, it became difficult to calculate the strain to which the material then became subject. For instance, iron arches, particularly on railways, might be loaded upon one-half of the span only. In such a case the curve of equilibrium would often shift so much from the middle of the arched rib, as almost to touch the extrados and the intrados of the arch on opposite sides of the centre. Having obtained, however, the position of the curve of equilibrium, it then became a question, how to estimate the effect of the pressure in its detrimental action on the outer ‘fibres of the rib. This was a point upon which, he agreed, that no practical, or reliable, information was to be obtained from books, and, in conse- quence, he had invented a method which he had found both simple and accurate in its application. To illustrate the de- gree to which the shifting of the line of pressure might affect the strain upon an iron arch, when placed under compression, he stated that in a prismatic bar of iron, in shape like a 3-inch plank, the pressure on the outer fibres on one side would be doubled, and on the other side be reduced to noth- ing, by the removal of the line of pressure only the one-sixth part of the whole width away from the centre of gravity. Now in this, as in every other case, whenever a piece of metal, or other elastic material, was compressed by a force on any other line than that of its centre of gravity, the action upon the outer fibres might be obtained, by conceiving the material, first of all to be compressed squarely throughout, 14 AMERICAN IRON BRIDGES. as if the strain were applied over the line of centre of grav- ity, and then, finding the strain due toHhe angular motion, by the same process that would be used for obtaining the strain on that portion of a cantilever contiguous to where it was attached to a wall ; using the total compressive force applied as the weight, and the distance of that pressure from the line of centre of gravity as the leverage. Then the strain on the outer fibre thus obtained, added to the former square-on pressure, would give the whole detrimental action upon the outer fibre of the material. Thus, as he had pre- viously explained in reference to the Charing Cross Bridge, when a pressure of 475 tons was applied, square-on, to a piece of iron of 161.25 square inches sectional area, all the fibres would be strained uniformly up to 2.94 tons on the square inch ; but by removing the line of pressure 3.6 inches away from the centre of gravity, the pressure on the outer fibres would be increased up to 6.86 tons on the square inch. Mr. Phipps added that, in cases where the arched rib proper was well connected with the roadway bearer above, by an efficient system of diagonal bracing, the centre of gravity of the whole section of the arched rib and roadway bearer to- gether must be taken, as that around which the previously- named angular motion must be ’computed. Mr. F. W. Sheilds could not agree, that the object of the bracing in a girder bridge was merely to keep the top and bottom members asunder, and to enable them to do their work. In his opinion, the chief object of bracing in all parallel girders, was to carry the vertical weight of the bridge, with the extraneous load upon it, to the points of support at the piers ; and that duty was just as important as, and quite distinct from, the work to be performed by the top and bottom flanges. The bracing, ought to be specially designed for that duty, and in that view he thought the American bridges brought under their notice, even those of the simplest form, showed a want of scientific knowledge in their construction. Thus, if a girder bridg# be supposed to consist of twelve bays, and the driving-wheels of a locomo- tive loaded with twelve tons rested over the third bay, that load would be transmitted to each pier, by the bracing, in the inverse proportion of the distance of the loaded point from the piers, being nine tons to the nearer, and three tons to the further pier. One of the diagrams exhibited was just in such a case, and he observed that at the' third bay, there was no diagonal tie from the lower flange which could carry the load to the further pier. Consequently the girder would have to depend merely upon the rigidity of the top flange, AMERICAN IRON BRIDGES. 15 considered as an independent beam of perhaps a few inches deep, for the transmission of that part of the load which reached the further pier, and which should be sustained by the whole girder of several feet deep. Diagonals of great inclination were not advisable, independent of all increase of strain arising from their inclination, as they were free to de- flect on a curve struck from one of their ends as a centre and the other end |ts a radius, and that curve would coincide in practice, for a considerable distance, with a vertical line, so that there was little or no resistance to deflection. There were some bridges in the Regent’s Park upon that principle, which fully bore out his foregone conclusions, for he found that by jumping upon those bridges, the whole could be set in motion by his own weight alone. Mr. Zerah Colburn remarked, with reference to the com- parison which had been made between the various trusses, that it was well understood in the States, as applying to the Fink and B oilman Bridges. The Murphy- Whipple Bridge was believed to be the best class known in the States ; and if there was not much difference between that and many bridges in this country, it was, no doubt, owing to the fact, that when the principles were fixed, there could not be much room for difference between one good bridge and another. Mr. Hawkshaw, President, expressed his regret that, owing to the late period of the Session, and the necessity for read- ing two other Papers before its close, the discussion upon a subject of so much general interest as American Iron Bridges should have been necessarily restricted. The Insti- tution was under great obligations to Mr. Colburn for the able and elaborate manner in which the communication had been made ; and he hoped that the Author would, on other occasions, contribute to, and take part in, the proceedings. A I ■ . I £ X A f