Historical Notes: Universal Joints

When Louis Renault first donned his bowler hat and went motoring, it is common knowledge that the power from his small de Dion power unit was transmitted through an open propeller-shaft and two universal joints. Unfortunately, Mr. Hooke, who gave his name to these joints, was not alive to see this new application of his principle, but perhaps this was just as well, as they revolved at a speed which might well have caused him to ponder deeply, more especially as he first applied the idea to sundials. As a matter of fact Hooke’s main work was far more that of a scientist than that of a mechanical engineer, but in spite of this he prepared some memoirs which dealt with the elliptical motion of the universal joint, which he read before the Royal Society. He lived from 1635 till 1703. The expression “Cardan shaft” is sometimes used to describe this type of drive, but it would appear that Cardan, who lived from 1501 till 1575, merely saw a diagram of three hoops joined together at a friend’s house and was unable to put the idea to practical use. Presumably the first introduction of the average mechanically-minded boy to the principle that Hooke enunciated is normally through his Meccano set, but the application to motor cars was not quite so easy as might have been supposed from the splendid coloured picture of a motor-car chassis which used to be portrayed on the lid of Meccano boxes.

To start off with, the drawing-room floor is usually somewhat cleaner than the standard macadam roads of 1900 or thereabouts and the first joints were extremely prone to damage on this score, situated as they were, in the best possible place to pick up the grit and water that was invariably flung up at them during the actual motoring process. It is not proposed to deal in this article with the chains that preceded the propeller-shaft, nor with the whole transmission system, as, if opportunity offers, these will no doubt form the subject of future articles, but it was obviously the limitations of the chain that caused designers to persevere with the shaft arrangement. It was those who made the lighter cars that were first converted. Without going deeply into the mechanics of the Hooke joint, it is not easy to describe the difficulties that faced these designers, but some of the characteristics may be noted. Firstly, the loads imposed upon the pins are normally high, as for a given torque their effective radius is small. Secondly, the wear that takes place is normally through a small angle, and this makes the provision of adequate bearing-area difficult, and further, other things being equal, the wear is proportional to the diameter of the pin. It is, of course, common knowledge that the second shaft suffers a variation in angular velocity when the joint is deflected, but this may be put right if the second joint is introduced in proper relation to the first, provided the shafts remain reasonably in line viewed from the top of the chassis. The general suspension and rear-axle geometry is quite important, but a number of early designers depended upon the shock-absorbing qualities of pneumatic tyres to overcome slight geometric errors, and this was a justifiable procedure provided the errors were not fundamental. Even this, however, meant that the joint had to bear much heavier loads than those dictated by pure torque and propeller-shaft weight. Without detailed knowledge of the actual spring deflections to be anticipated on each design, it is not fair to criticise individual designs on this particular point, but some, such as Decauville, certainly ran into trouble through faulty layout. Early converts from the orthodox chain layout in the light-car field were Decauville, Ariel, Darracq, New Orleans and, of course, de Dion, but the latter falls rather outside the scope or the present article. It is only fair, too, to point out that Lanchester used a shaft drive from the earliest days. With the exception of de Dion and Lanchester, all these makes laid themselves open to the criticism of inadequate bearing areas, their joints usually consisting of malleable iron or, at best, steel forgings which rapidly became sloppy and had to be renewed, as in very few cases was provision made in the design for separate bearing bushes, which could be replaced. It is rather surprising that few of the early designers used a “long and thin” pin which would have appeared more logical. However, running experiences soon showed the limitations of the early joints and efforts were speedily made to enclose them in leather or metal casings. Mr. Renault certainly encased his properly from a very early date, about 1902, and no doubt reaped the benefit of satisfied customers a year or so after he first sold them their motor cars. A great improvement in the “lot” of universal joints came when engines could he properly throttled instead of merely governed on a “hit-or-miss” basis, and also when clutches became softer and less inclined to be of the “in or out” variety.

For a long time, however, the transmission brake carried on to plague the universals, and in badly designed examples which snatched, or in cases of indifferent driving skill, severe additional loads must have been imposed. In order to deal with fore-and-aft movement, some manufacturers resorted to the slipper-block type of joint, usually leather enclosed, which was very satisfactory provided it was well-made and properly maintained. The writer recalls examining carefully such a set of joints from a chauffeur-maintained 12/16 Sunbeam which had done some 85,000 miles, and these showed no perceptible wear at all. They could not have been cheap to make, however, especially as grinding limits had to be called for to ensure that there was no play in the assembled joint. Needless to say, during the pre-war period, the “lot” of universal joints varied with the type of drive arrangements fitted, and perhaps the Automotor Journal in 1911 summed up the position as well as any, in spite of the somewhat unwieldy wording of its prose. Reviewing the position at Olympia in that year, this journal says “from the gearbox to the back-axle, the shaft drive is, of course universal; very few firms indeed remain who even provide the alternative of chain-drive on their most powerful cars. The problem of tying the back axle casing to the frame is still one of the dividing questions of the day, but there is undoubtedly a tendency on the part of the manufacturers to give more and more support to the practice of utilising the tubular propeller-shaft casing for this purpose entirely. For a long time it has, of course, been used partially as a torque stay, or rather as a means of stiffening the propeller-shaft so that member can act in this capacity.

“The practice is not to be commended, for it necessarily throws additional strain upon the universal joints, which already have quite enough work to do, and it is now comparatively uncommon to find the tubular casing riding on the propeller-shaft in this way. Either the casing is anchored to the frame independently or it is not employed at all, the torque being taken by an independent stay situated alongside the propeller-shaft.

“Many firms now consider that the propeller-shaft casing, if used at all, might just as well be used in the dual role of torque-and-thrust stay, that is to say, it might just as well serve the purpose of pushing the car. This, of course, is otherwise taken by the springs or by separate radius rods, in the few instances where such members are fitted to a live-axle chassis. It is this practice of utilising the tubular propeller-shaft casing in the dual capacity of torque-and-thrust stay that is distinctly gaining ground among automobile designers. Like everything else, it is a half-measure, and theoretically it can be shown to be incorrect. Thus, for example, in such a system, when the wheels rise towards the axle, the radius action of a propeller-shaft casing either forces the axle to move bodily backwards or accelerates the car bodily forwards, for it is impossible that the car and the wheels should both avoid longitudinal movement if there is any vertical displacement. In all probability it is the wheels that slip, but it is a moot point whether the amount is seriously detrimental to the life of the tyres. At any rate, it is very difficult to see how this particular problem is to be overcome, for there is bound to be the equivalent of the radius rod, and as often as not the practice is to use the rear springs in this capacity. The length of the radius rod in this case is only half the length of the spring, and the consequence is that the relative longitudinal movement for a given vertical movement is more severe than when the radius is struck from some point well up in the centre of the chassis, as is the case when the front end of the propeller-shaft casing is supported in the frame. There is at least this to be said for the practice of using a propeller-shaft casing in this dual capacity, a practice which, by the way, was first introduced for pleasure-car work in the Sheffield-Simplex cars, that it makes an extremely neat job of the design.”

Thus somewhat heavily, does the Automotor Journal describe the agonies of the joints of those far-off days and unconsciously re-demonstrates the argument for what has now become the standard practice, that of having the two joints plus a sliding sleeve coupling combined in the open propeller-shaft. Before leaving the pre-needle-roller era and reconsidering briefly the fabric types, it is interesting to recall that early transmission efficiency tests cast doubts upon the mechanical efficiency of universal joints, and it was not until it was firmly pointed out that the horse-power losses, if they really existed, must serve to heat up the universals to an intolerable degree, that the blame was put where it belonged, at the door of the bevel-gear makers of the day!

The first glimpse of the next step forward in universal joints, that of providing a disc which required no lubrication and was theoretically everlasting, came when Mr. Weller [later the A.C. designer,—Ed.] proposed such a joint for use between clutch and gearbox, where misalignment amounting to more than about two degrees could not arise. This joint (previously used in similar form in electric motor practice) consisted of a star-shaped diaphragm of thin sheet steel attached to one member at its periphery and to the other at its centre and was actually fitted to the Weller car, but it would appear that Mr. L. A. Legros and Mr. G. Knowles were the first to apply the new idea so far as propeller-shafts were concerned. Some very successful runs were eventually made (presumably on an Iris car), after early trials had shown failures after 400 miles, but the steel diaphragm rusted and became weakened unless stiffened to such a degree that the expense of making them exceeded the production cost of the standard Hooke joint.

The next step was to use leather or fibre, and in spite of the advent of the Hardy-Spicer needle joint, that company continued to sell the disc-type well into the 1930s. It had the obvious merit of simplicity, and was impervious to weather conditions. It was, of course, also capable of absorbing longitudinal movement. An early form, not of the “spider” type, was first applied by Isotta-Fraschini in 1912, but it was during the ’20s and early ’30s that the type really flourished, particularly for the lighter cars. Naturally, its construction had to be properly carried out, the warp and weft of the separate fibre plies being displaced by an angle suitable to the number of plies used–60 deg. for three plies, for example. One disadvantage of the early applications of this type was that the two joints had to bear the weight of the propeller-shaft, and were thus liable to the whirling stresses and this led to the practice of spigotting one spider into the other, the location being on a hardened-steel ball-joint.

As has been said, its use tended to be restricted to the lighter type of car, the standard Hooke joints (in various forms), or the sliding “pot”-type, continuing in favour for the heavier and more luxurious cars, but now, of course, of adequate proportions, properly protected and with adequate provision made for periodic lubrication. The transmission brake, too, had disappeared, which helped to preserve these components. These joints, however, were still expensive to make properly and to replace when wear did take place.

But speeds were going up, and the coming of the “tin-can” car in the early 1930s, demanded a standard universal joint that lent itself to mass production and got over the out-of-balance snags of the fabric type. The Hardy-Spicer people answered that demand with the altogether better needle-type, which has remained the standard-type since. It is difficult to imagine an improvement on this, which must be too familiar to need description here, as it is either illustrated or described in every modern car handbook.

Today, if the modern racing-car engine demands that the first stage of the transmission turns over at 10,000 r.p.m. or if race conditions demand that the second stages between de Dion axle casing and road wheel have to accommodate severe angular displacements with really high torques, it is surely adequate praise for this component that nobody loses any sleep over the problem?—” A.B.C.”