On Hollow Railway Axles. By J. E. M'CONNELL, of Wolverton.* [Paper read at the Institution of Mechanical Engineers.]

The subject of railway axles was brought before this Institution on a former occasion by the writer, when he gave the result of various experiments, showing the form and dimensions most economical of material, with a proportionate and proper strength of the several parts, and also the changes in the structure of the iron which appeared to have taken place from various causes during the course of working. Since that period the writer's attention has been constantly directed to the subject, and the opinion he then expressed respecting the fractures of axles arising from changes from the fibrous structure of the iron, to a brittle, shortgrained, or crystalline condition, has been confirmed by repeated instances which have come under his knowledge.

With the view of improving the strength and durability of railway axles-the two most important points for insuring the safety and security of railway traveling-the writer, after repeated experiments, and obtaining all the experience and information he could collect on the subject, arrived at the conclusion that the hollow or tubular axle combined in itself, if properly manufactured, all the properties necessary to secure the best form for lightness, strength, uniformity of structure in the material, elasticity to neutralize the injurious effects of blows and concussions, and consequent durability, from having a greater freedom from deteriorating effects.

The selection of the tubular form of axle originated from the knowledge, that with a considerably less weight of material in the form of the tube, *From the Civil Engineer and Architect's Journal, October, 1853.

VOL. XXVI.-THIRD SERIES.-No. 6.-DECEMBER, 1853.

31

a much greater strength can be obtained to resist torsion, deflexion by pressure or weight, or concussion from blows. The resistance of a solid system to deflexion and torsion, increasing in proportion to the fourth power of the diameter (or the square of the square), but the weight increasing only as the square of the diameter, two solid cylinders, having the respective diameters of 4 and 5 inches, or 1 to 14, will have a proportionate weight of 16 to 25, or 1 to 14, but a resistance of 256 to 625, or 1 to 21. Then if a hollow of two-thirds the diameter be made in the larger axle, its weight will be diminished (or nearly), and its resistance only th (XXX=1, or th nearly), and the comparison with the smaller solid axle will then be 1 to 1 in diameter, 1 to in weight, and 1 to 2 in resistance, being double the resistance, with th less weight.

The use of hollow axles was tried some years ago, but was not continued, the main objection being, that there appeared a great difficulty of insuring, by the particular mode of manufacture adopted at that time, a sufficient uniformity of thickness of the sides of the tube throughout, and also of the soundness of material. The mode adopted consisted of rolling two or three bars of a semicircular cross section, which were welded together with butt-joints, but with no internal pressure, and with solid ends where the bearings came. These axles, having no mandril or internal pressure during the process of welding, were found to be of a very uncertain strength throughout the axle, and the weakest point might be close to that part where the greatest force or strain would be exerted.

To overcome these objections, a mode of manufacturing railway axles has been introduced by the writer, which, it is believed, effectually accomplishes the object in view, securing the utmost strength with the least possible amount of material, uniformity of structure of the iron, perfect equality of thickness of material, and soundness of manufacture.

The plan adopted is as follows:-A number of segmental bars of the best quality of iron are rolled to a section, so as to form, when put together ready for welding, a complete cylinder, fig. 1, about 1 times the diameter of the axle when finished, the bars fitting correctly together, so as to

have no interstices, and overlapping in such a manner as to insure a perfect and sound weld when completed, as shown in fig. 1.

This cylinder of loose segmental bars is temporarily held together by a screw-clip, and each end being put into the furnace until a welding heat is produced, the bars are then partially welded together, and the clip removed. The whole cylinder is then placed in the furnace, and brought to a proper welding heat; it is then passed through a series of rollers, B, B, fig. 2, which have each a mandril of an egg-form, A, in the centre of the circular openings, which are attached and supported on the end of a fixed bar, the bar being firmly secured at the opposite end, to resist the end pressure or strain during the process of rolling. The mandrils are made of cast iron, chilled, fitting on like a socket on the end of the bar to a shoulder, and they are secured by a screw-nut, so that they are easily removed when required.

The motion of the rolls is so arranged, by a reversing-clutch on the shaft, that as soon as the axle-cylinder has been drawn clear through, the motion is reversed, and the axle, which has been drawn on to the mandril-rod, is again drawn back through the same opening in the rolls; it is immediately passed through the next smaller groove of the roll, with a decreased size of mandril, and again reversed back through the same groove in a similar manner, and so on through a series of grooves in quick succession, each decreasing in size, and consequently increasing the compression and strength of the iron of which the axle is formed, and by the last groove it is passed through it is reduced to the proper diameter. At each time it is changed from one groove to another, the axle-cylinder is turned by the workman a quarter round, so as to equalize the pressure on every part of its surface, to insure uniformity of the compression of the iron, and thoroughly complete a sound welding throughout every part of the axle.

The specimens before the meeting showed the soundness and perfection of the manufacture, as a proof of which, in every test applied, either by blows on the outer surface or by an immense splitting pressure, by driving a mandril in the interior, there has never been found in any one instance a failure of the weld, although the test has been applied to pieces cut off the extreme end, where it might be supposed the welding of the cylinder of the axle, from various causes, would not have so good a chance of being perfect.

The axle at this stage, after being welded and drawn down in the rolls to the proper size, is taken at once to a hammer, where it is planished between semicircular swages over its entire surface. A small jet of water plays upon it during this process, which enables the workman to detect at once, by the inequality of color, any unsoundness in the welding. From the hammer it is taken to the circular saws, where it is cut accurately to the length required, and ready to have the bearings formed upon it.

On coming from the hammer, the axle is found to be perfectly clean both inside and outside, the scale being entirely removed. The ends are then re-heated, and gradually drawn down by a hammer to the proper dimensions and form of the journals, a mandril being inserted in the end of the tube during the process of hammering.

The formation of the journals can also be produced by a rolling machine, constructed of tables the entire length of the axle, rolling transversely, each table being a duplicate of the other, and matrices of the axle when finished. Or in another way, by two sets of rollers, each set consisting of three rollers running vertically, being of the same diameter, and driven at the same velocity, formed exactly to the shape of the bearing, and set the proper distance apart from shoulder to shoulder of the journals.

The manufacture of these axles has been intrusted to the Patent Shaft Company, and a great amount of credit is due to Mr. Walker, the managing partner of that firm, for the very excellent system he has adopted and carried out in the process of manufacture.

As an illustration of the saving in dead weight, take, for instance, a railway employing 15,000 wagons and carriages, and assume each of these vehicles to run on an average 10,000 miles per annum. The weight of two axles of the solid description finished, say 5 cwt., and if replaced with hollow axles of equal strength, the weight per vehicle may be reduced 1 cwt.; this taken over one mile of the above stock per annum will be 11,250,000 tons, and assuming the cost of traction for locomotive power at d. per ton per mile, the saving will amount to 11,7007. per annum, without taking into account the other advantages, and also the saving to the permanent way, &c.

In the samples of axles submitted to the meeting, two different kinds of bearings were shown, the parallel bearing with the rounded shoulder, and also the double conical bearing, such as is used on the Great Northern, Great Western, Bristol and Exeter, South Wales, and South Devon Railways. In either description of bearing, the hollow axle is good, although it is believed that the conical bearing for either the solid or hollow axle has less tendency to injure the texture of the iron during the formation of the journal than the parallel shouldered axle, and it appears a matter well deserving the consideration of the Institution, to ascertain what, under all conditions, is the best form of axle bearing.

Experiments, conducted by Mr. Marshall, the Secretary of the Institution, have been tried for the purpose of ascertaining the comparative strength of the hollow and solid axles to resist a transverse strain. Each axle was supported on massive cast iron blocks, fixed at a distance of 4 feet 11 in. apart, to represent the support given by the rails to the axle. A cast iron block weighing 18 cwt. was then let fall on the centre of the axle from a height of 12 feet, and the extent of bending was measured. The axle was then turned half round, and another similar blow given on the opposite side, bending it in the opposite direction. This proceeding was repeated until the axle was broken. The general results of these experiments are as follows:

An Old Solid Axle, 32 inches diameter in centre, 4 inches at ends, which had been at work three years, was bent 83 inches by the first blow; it was nearly straightened by the second blow in the opposite direction, then bent 10 inches by the third blow, and with the sixth blow it was broken in the centre square across.

A New Solid Axle, of the same dimensions, was bent 9 inches by the first blow, then nearly straightened by the second blow, and bent 93 inches

by the third blow, and by the fourth blow 2 inches, and by the fifth blow it was broken 2-inch from the centre. The appearance of the fracture was crystalline over three-fourths of the section, the remaining part tough fibre.

A New Hollow Axle, 43 inches diameter throughout, was bent 5 inches by the first blow, nearly straightened by the second blow, and bent again 5 inches by the third blow. The ninth blow bent it 4 inches, and the tenth blow 13 inch. Up to the fifteenth blow it was bent alternately, the bends varying from 2 to 3 inches. There was no appearance of failure or cracking, but a slight rising of the surface at the fifteenth blow. The blows were continued to the twenty-seventh, the bends varying from 2 to 3 inches, and at this blow a fracture took place across the middle of the axle 1 inch long. The twenty-eighth blow bent it 3-inch, and closed the fracture on the opposite side made by the preceding blow. By the twenty-ninth blow it was fractured two-thirds through, and bent 9 inches, the appearance of the fracture being very fibrous.

A second series of experiments was made, to ascertain the comparative strength of the journals of the hollow and solid axles to resist breaking. Each axle was supported on an anvil, with the inner shoulder of the journal projecting 14 inch beyond the edge of the anvil, to represent the support of the axle in the nave of the wheel; 100 blows with two 24 lb. sledge-hammers were then struck upon the upper side of the outer end of the journal, the men being changed after striking each twelve or thirteen blows alternately. The amount of bending of the journal was then measured, and the axle turned half over, and another 100 blows similarly given on the opposite side of the journal; the same proceeding being then further repeated. The general results of these experiments are as follows:

An Old Solid Axle, with 3 by 5 inch journals, that had been at work three years, had one journal broken off with 205 blows, and the other with 53 blows; both fractures were square across the journal at the shoulder.

A New Solid Axle, with 3 by 6 inch journals, had the journal broken off with 570 blows, the fracture being irregular in form, and fibrous.

A New Hollow Axle, with 3 by 5 inch journals, had 400 blows on the journal, which bent down the end §-inch, and produced a longitudinal split on the under side 33 inches long, but no transverse fracture.

A New Hollow Axle, same size journals, received 800 blows on the end of the journal, which bent it down 4-inch, and split the journal longitudinally on both sides, but caused only a slight transverse crack near the shoulder, 2-inch long.

The experiments on transverse strength by a heavy weight falling on the centre of the axle, and giving the blow on opposite sides alternately, show that the hollow axle is nearly double the strength in that respect of the corresponding solid axle, the amount of bending being only 5 inches instead of 93 inches; and the number of blows required to break the hollow axle being 29, whilst the solid axle broke at the fifth blow, shows the hollow axle to be greatly stronger in resistance to fracture.

The hollow axle became -inch oval in the centre after receiving the seventh blow, and it was only -inch oval after receiving the twenty