Big I-4:nd Bearing Design
W. S. BRAIDWOOD, B. A. TT is interesting to observe that throughout the develop’ ment of the racing motor-cycle engine, designers
keep on reaching a stage in. the obtaining of increased power, where one particular component of the engine appears to set a limit to the power and speed possible with a given cubic capacity. Eventually each source of trouble is eradicated by experiment and research, only to reveal that when one component is made to stand up to the work, some other part, not previously considered a difficult point in design, becomes in its turn the limiting factor to power output.
At one time pistons were a constant source of trouble, at another exhaust valves were thought to be the greatest trouble in the design of a fast engine and so on, but throughout the evolution of the high speed engine, especially in the motor-cycle world, there has been one thing which has always been a potential source of trouble, and that is the big-end bearing. This has always been a source of anxiety when the othei troubles have been solved, and at the present time we may safely say that this component is causing more engine failures in racing than any other, and it is therefore a good opportunity to consider the chief causes of big-end failures and to see what steps can be taken to reduce their possibility.
The big-end bearing is unlike any other in an engine in that it combines an extremely high rate of revolutions and heavy alternating loads with a rapidly varying angular velocity. Also its actual motion is not only that of a journal rotating in a bearing, but the bearing itself, instead of being stationary as in a main bearing, is describing a circle with a diameter equal to the stroke of the engine, and therefore has to cope with an extra load due to centrifugal force.
It is therefore this point and its varying angular velocity that are the cause of its being a problem all on its own in the bearing line, and they are the cause of nearly all the big-end trouble incurred.
The actual size of the bearing is limited chiefly by the fact that the crankshaft must be compact and as rigid as possible, and it is for this reason that nearly every motor-cycle engine of to-day has a roller big end, in view of the very high loads which this type of bearing can carry on a small area. However, the fact that trouble is still experienced in many designs goes to prove that the mere employment of a roller bearing of sufficient size to deal with the maximum loads involved does not constitute a way out of the difficulty, and we will therefore consider what happens to a simple roller bearing at high speed. It will be simpler to take the two causes of trouble separately, so we will first take the effect of the centrifugal force. The first effect of this is of course the obvious increase in the bearing load due to the centrifugal force on the lower end of the connecting rod, and for this reason this part must be kept as light as possible consistent with rigidity. The second and more important effect is that of the centrifugal force on the rollers them selves. This force tends to pack the rollers together on the outside of the bearing, and the force is such at high speeds as to break down the oil film between the rollers and the consequent rise in temperature softens the case hardening of the rollers and causes failure of the bearing. The usual course now adopted to prevent this rubbing is to carry the rollers in a cage of steel or bronze so that they are kept separate and can only rub against the cage in. which they are a running fit and therefore avoid the intensity of pressure resulting from line contact with the roller. adjacent. Also, if the cage is constructed of some material which makes a good bearing for hardened steel, wear of the rollers should be practically confined to that caused by the actual bearing load on the big end. Of course, the cage itself is subject to the same centrifugal force as the rollers, and therefore it must be sufficiently strong and rigid to avoid being broken up by the corn bined effect of its own mass and that of the rollers it contains. Although breakage of the cages is by no means unknown, in the more successful designs it is
extremely rare, and we must look further for the cause of the continued trouble which is still being experienced with big-ends. The caging of the rollers in a bearing, however, has no effect on the unfortunate circumstance that a big-end bearing has a continually varying angular velocity. In other words, although the engine may be maintaining a constant speed, and the average angular velocity or revs, per minute may not change, such happy state of affairs does not obtain in the big-end, whose actual angular velocity varies from a maximum to a minimum and again to its maximum every revolution of the engine, owing to the varying angularity of the connect ing rod. If the con-rod was of the infinitely long variety so often postulated in text-books this would not be so, as there would never be any angularity, but owing to
considerations of space the con-rod of the motor-cycle engine is not only not infinitely long, but is actually shorter than in almost any othez form of reciprocating engine.
At T.D.C. the angular velocity of the bearing about its journal is equal to the angular velocity of the crankshaft plus the angular velocity of the conrod, and at B.D.C. it equals the difference of these angular velocities.
Let us consider the case of an engine in which the length of the con-rod is 3 times the radius of the crank, a case which is much nearer the truth in many designs than it ought to be. If the engine is running at a constant speed the linear velocity of the crank-pin is a constant which we will call V feet per second, and the radius of the crank R. Then the length of the con-rod. will be 3R.
Thus the angular ve1ocit31 of the crankshaft is V/R radians per second, and that of the con-rod V/3R radians per second.
At T.D.C. the angular velocity of the bearing therefore equals V/R plus V/3R or 1.33 V/R radians per second.
At B.D.C. it is V/R minus V/3R or 0.66 V/R radians per second. From this it will be seen that the angular velocity of the bearing at T.D.C. is twice that at B.D.C. This means that if the rollers are to roll on the journal and the outer race they have got to be accelerated to twice their angular velocity in half a revolution of the engine, or in one two-hundredth of a second at 6,000 r.p.m. However, owing to the moment of inertia of the roller it refuses to do so, and the result is that it skids instead of rolling, and eventually causes failure.
This trouble has been partially met in various designs by decreasing the moment of inertia of the roller by decreasing its diameter, this trend having led to the “bundle of needles” type of big-end, either caged or not, which is used in a number of designs at present.
However, all these ingenious adaptions of the roller bearing are merely admissions that the roller bearing is inherently unsuitable for a big-end, and it becomes increasingly obvious that a plain bearing fulfils the requirements to a far greater degree than any other type. The reasons advanced against the use of plain big-ends in motor-cycles is that they need more thorough lubrication than rollers, and that there is not room in a motor-cycle engine for a large enough plain bearing to carry the loads incurred at racing speeds. The former reason is a relic of the past, as now most engines have dry sump lubrication in which the whole of the oil supply is fed through the big-end under pressure to keep the skidding rollers cool enough to avoid failure, and a momentary stoppage of the oil supply is quite sufficient to wreck the bearing, and therefore the plain bearing is no longer at a disadvantage on this score. The second reason carries more weight, and the writer
agrees that the ordinary car type of white metal big-end would not be suitable owing to its liability to failure by crushing and also to its considerable weight. However, there is a good choice of stronger materials, and an aluminium bronze, consisting of 313 per cent. aluminium and 12 per cent. copper makes an excellent bearing for hardened steel shafts, being very light and a very good conductor of heat.
In high speed bearings the actual maximum pressure is not of great moment as long as it is not so great as to crush the material, and the important thing is the load factor, that is the product of load and speed.
Under good conditions of forced lubrication, and with good facilities for getting the heat away from the bearing, load factors as high as 20,000 lb. ft. per second can be, and have been used.
Furthermore, in the case of very high rubbing speeds load factors can be increased by as much as 50 per cent. where floating bushes are employed between the members. In this case the rubbing speed between each face is halved and a larger quantity of oil can be circulated through the bearing. This gives an allowable load factor of 30,000 lb. ft. per second, and there is certainly room to design a bearing well within this figure without appreciably changing any other part of the engine as now used in motor-cycles. The bearing would consist in its essentials of the present type of crankpin and connecting rod with a plain bush substituted for the present assembly of rollers.
It would suffer from none of the disadvantages of the roller big-end which we discussed at the outset, and it would have the added advantages of simplicity, silence, and extreme cheapness compared with the present type of bearing, which last item at least ought to appeal to the manufacturer, and it will be interesting to see who will be the first with sufficient courage to flout the present fashion and take what appears the obvious and correct course.