US2142478A - Bearing device - Google Patents

Description

Jan. 3, 1939. w. T. MURDEN 2,142,478 v BEARING DEVICE Filed Sept. 16, 1955 INVE NTO R: \/\J1 LLIAM TMURDEN H15 ATTORNEY Patented Jan. 3, 1939 UNITED STATES PATENT oFElcE l EAmNG DEVICE William T. Mnrden, Forestville, Conn., assigner,

by mesne assignments, to General Motors Corporation, Detroit, Mich., a corporation of Del- Application september 1c, 1933, serial No. 689,733

1o claims. (ci. sos-193) over the area of engagement betweenthe con-- v tacting parts, and 'another object is to greatly reduce shifting of the parts from desired mutual relationship, as by imparting increased rigidity, especially in the line of load. To these ends, and also to improve generally upon devices of the character indicated, my invention consists in the various matters hereinafter described and claimed. l l

Stated'generally (but not by Way of limitationhaccording to my invention, the opposing surfaces of the co-operating load-supporting members (as a ball and its co-operating racemember) are so shaped that when the device is free from load those members are, at alll vpoints beyond the desired length of their vcontact under load, spaced apart so `far that the rated load will not cause them to interengage, but, when still free from load, those surfaces are, throughout that length of contact under load, as much in coincidence as it is possible to make them and yet permit the use of only a.smooth` line throughout such length of contact and for juncture with the above-mentioned portions that are not interengaged under'lo'ad. One eicient and commercially practicable Way of accomplishing this is by shaping the surface of one of the contacting parts as the curve of a circle, and shaping the co-operating surface of the other such part as an elliptical curve that is very nearly the curve of the said circle throughout substan-` tially the ultimate arcof contact, has its central part tangential to said circular curve in the line of the `load, throughout several degrees from the point of tangency progressively (but only most minutely) departsfrom the circularcurve as that arc proceeds from the tangent point, then appreciably accelerates its rate of such departure as the' extremities of such ultimate `arc of contact-'are being reached., and then rapidly further increases such rate of departure substantially "at and beyond the extremities of that arc,

As will more fully hereinafter appear, this results in substantially even distribution of stress over the WholeI length of contact under even the lightest loads, and deflection of material ofthe,

contacting elements isv greatly reduced in the contact area, especially at its center, whereby the endurance properties of the device are greatly enhanced ias that it has longer life under a given load at a given speed) and the parts of the device do not materially shift from their initial, and desired, mutual relationship.

In theaccompanying drawing, Figure 1x is an elevation, partly in section, of a portion of a ball bearing exemplifying my invention, the view showing the bearing when not under load; Figure 2 is a plan view of a fragment of the outer race-ring of such bearing, the contact area between ball and raceway whenv under loadbeinlg indicated by broken lines; Figure 3 is a fragmentary, diagrammatic view corresponding to Figure 1 but enlarged and with the curvature relationship necessarily exaggerated; and Figure 4 is a table of approximate radial distances between the ball surface and race surfaces of different illustrative kinds, such distances being taken at points transverse the bearing.

vIn the illustrated embodiment of my invention the co-operating load-sustaining parts are shown as the inner (or cone) race-ring l of a'ball bearing, the outer (or cup) race-ring 2 thereof, and the circular series or set of spherical balls 3. Allof these load-sustaining parts are shown as commonly heretofore constructed except that, the rolling-elements 3 having circular curvature transverse the device when not under load, the curvature of each of the raceways, 4 and 5, tangent, as at 6,*to the illustrated ball at the line of vload (which is that of the diameter 1) is, when the device is free from load, that of an arc of an ellipse Whose curvature at such ,point of tangency is practically that of the ball curvature and whose axis (the major axis in the illustration here given) is in the line of that be. fore-mentioned diameter, 1, ofthe ball that includes the points of tangency between ball and raceways, the center of the aforesaid elliptical arc being at the before-mentioned point of tangency. One satisfactory Vsuch ellipse for a raceway to be used with a ball one-half inch in 'diameter is indicated at 8, Figure 1, and dened by the equation x2 cna (.346o2)2+(.29412)2 in which :c indicates the distance along that major axis from any point, a, in that axis to distance, perpendicular to that axis, from that point a to the periphery of the ellipse. Many advantages, hereinafter appearing, are secured by relatively spacing the surface of the raceway and the co-operating surface of the rolling element as just above described.

Such spacing is so minute in devices of the present character that no accurate graphic illustration can be given within the limits ofthe accompanying drawing, but the clearance is indicated diagrammatically (and necessarily in an exaggerated way) in Figure 3, where the ellipticai race-way is indicated by the surface 5, and the broken line 9 indicates the circular curve commonly heretofore given the raceway; and the clearance is also stated approximately in Figure 4, where distances are given in units of one onewhen using such elliptical curves as above indicated. But even the figures stated are necessarily only approximate.

In ball bearings and other devices of the present character it is, of course, advantageous that when under load the co-operating surfaces of the bearing members (as the ball 3 and its racemember 2) contact with each other over an appreciable length (as the arc extending to each side of the point of tangency 6), so that adequate mass of material is presented to sustain the load. In ball bearings such arc of eventual contact is commonly of total extent of from about 40 degrees to about 50 degrees of the ball circumference. But any circular transverse racecurve, even though it have a radius only barely greater than that of the co-operating ball, necessarily results in a relatively rapid divergence between ball-surface and race-surface from the yinstant that the curves proceed transversely away from the point of tangency, the rate of Vsuch divergence rapidly accelerating 'as such point is proceeded from. For example, as shown in Figure` 4, line 11, when a ball of one-half inch diameter is used with a circular race-curve whose radius is but 51.6% of the ball diameter, at only three degrees from the tangent point'the clearance between balland race is well over 1 onehundred-thousandth of an inch, at ten degrees it is almost 12 one-hundred-thousandths, and at ,twenty degrees (a common minimum for onehalf of the arc engaged under rated load) ,the separation ,is about 461/2 one-hundred-thousandths, which is almost one-half of a thousandth. The result of such co-operating circular curves is that, as indicated in Figure 3, the ball is, transversely, in the nature of a rounded wedge against the co-operating raceway surface and, therefore, not only must considerable relative approximately-radial movement (as that of compression or indentation) 4occur between ball and race-ring in. order to obtain load-supporting contact between those members at only a few degrees away from the point of tangency, but also that relative movement must be almost onehalf of a Athousaridth of an inch to permit such load-supporting contact over even, the customary the center, o, .of the ellipse, and y indicates the minimum arc. To permit such relative movement the material between the limits of such contact arc is necessarily correspondingly compressed or displaced and the resultant stress varies with each unit of length of the contact and is most concentrated at theiinitial point of tangency. One result of this is that undue stress is placed upon the intermediate portions of the engaged arc in order to permit any contact at all between the arcs terminal portions, and, consequently, they life of the dcvice is limited by the endurance capacity of that small, central portion of the arc that is subjected to the greatest strain. Were it attempted to overcome this difficulty by simply making the whole surface, as 5, of the race-member to the same circular curve as that of the co-operating ball, that part of that race surface engaging the ball beyond the above-indicated arc of Contact would create so much friction as the rolling ball rubs against it that even more detriment would occur to the bearing, it being detrimental to a bearing, and uneconomical from the standpoint of eiiiciency, to have the contact arc too long, the economical arc being that which includes only those bearing-'member portions whose advantageous properties (as that of sustaining load) are not outweighed by their disadvantageous properties (as that of producing. detrimental friction); and-if such circular race-curve were abruptly stopped at the ends of that economical arc of contact, detrimental stress would occur at those ends. A second diiiiculty resulting from. lthe above-mentioned relatively great compression or displacement is that the `4 approximately states those clearances when employing a one-half inch ball and the aboveindicated arc of the ellipse whose equation is given above. As shown by that table, there is little or no appreciable clearance for a distance that extends over several degrees of the contact arc and at even ten degrees from the center the clearance is but slightly more than one one-hundred-thousandth of an incha distance that is not'one-tenth of that occurring at the same point when the typical circular raceway is employed, and that is no greater than the distance obtaining at less than a third of that arc when employing the circular curvewhile at twenty degrees from the'center (which is not infrequently the limit of the contact arc) the clearance of the elliptical curve is only about one-fifth of that occurring with the circular race, being less than one ten-thousandth of an inch, and at even twenty-five degrees (about the maximum for contact arcs heretofore employed) the clearance is barely over two ten-thousandths' and is little more than one-fourth of that which results with the circular curve. Therefore, when a device employing the above-stated elliptical curve is free from load, the opposing surfaces of the co-operating Aload-supporting members are but very slightly spaced apart at any point in the whole arc of inter-engagement under load, and over la large proportion of that arc (about half of it) they are in actual or substantial coincidence, their separation there being so slight that even a negligible load (as one applied by the mere hand) will cause interengagement, so that in obtaining a typical total contact arc of about fortyfour degrees (twenty-two degrees upon each side of the center) although with the previously-employed circularly curved raceway that center must compress or bedeflected more than 56 onelhundred-thousandths (more than half of a thousandth), with my elliptically curved raceway the deflection is but 13 one-hundred-thousandths (about one and a third ten-thousandths) Moreover, it is known that when opposed circular curves are used, the shape of the whole contact area when under load is that of an ellipse whose center` is the point of tangency of those curves (thus showing markedly unequal stress per unit of length of contact) but with my device the gasdscoloration process shows that, with the exception of advantageous very short taperingoif extremities. I4, Figure 2, the whole contact area under load is substantially rectangular, as indicated at I5, Figure 2, thus confirming the mathematical calculation that throughout all but Athese slight extremities the stress under load issubstantially uniform per unit of length. Thus,V

the result of the new structure is that when load is applied no deflection or compression of material to any substantially appreciable amount is necessary to afford the length of contact requisite for sufficient support, and, also,V each unit of that length of contact is subjected to substantially the s ame amount of stress. Consequently, amount of displacement and ilexure of material of such contact area is greatly reduced. Furthermore, not only is the stress more evenly distributed throughout. each contact area,y but also the total load is better distributed among those several balls of the bearing that are in the loaded zone. As a result the device is given'great endurance (especially against failure because of constant flexing of material), and there is greatly reduced approach between the supporting and. supported parts (as the race-rings I and 2) when load is applied. And this is true'under even the lightest loads as well as under the maximum for which the device is rated.

In devices of this character it is highly advantageous, as previously intimated, to avoid abrupt ending of the line of contact and stress (so that detrimental stress forces may not concentrate at such an abrupt end), and my above-described relationship between the opposing faces of the load-supporting members (as the race-rings I and 2 and the balls 3) eifectually accomplishes tls. Again referring to line I2 of Figure 4 it will be seen thatup to about t'en degrees from the center, the rate of departure of the race curve from the curve of the co-operating ball is almost nothing and almost constant, the resultant clearances increasing only about one-eighth of one one-hundred-thousandth of an inch per degree; then for about the next eleven degrees the rate of such departure slightly accelerates; and beyond that there is an increasingly great acceleration. The result of this is that although the transverse-contour of the race-member throughout its whole contacting portion, and even for a safe distance beyond 'that portion, is a smooth, regular, and gradual curve and thereby avoids abrupt breaks with their consequent stress-concentration, and although that race-member lies extremely close to its co-operating rolling-member across the desirable arc for eillcient load support, nevertheless that race surface is safely and relatively-quickly carried beyond reach of the co-operating` ball at the limits of the useful and economical arc of contact, thus effecting the desirable tapering-olf of stress, as indicatedat I4 in Figure 2, and also avoidingany -side contact between raceway` andl ball that would provduce detrimental friction due to rubbing between those parts.

In such devices as ball bearings constructed in accordance with my invention the increased resistance that the above-mentioned stress-distribution presents to compression or deiiection of material is such that the extent of arc of contact under a given load remains about the same as that of the profitable or economical arc here- ,tofore used,` for the same load, in devices having circular curves in opposition. But I find that in my device I can increase such economical arc, becausethe substantially even distribution of stress over such arc (with the consequent elimination of concentration of penetrating force in the intermediate portion of the arc) so greatly increases the xity of the parts as respects each other (the rigidity of the device) as to more than oH-set the effect of any rubbing friction resulting at the extended arc portions. Moreover, this gain in length of economical arc enables a lesser number of balls, or other rolling-elements, to support a given load (I have found that while still coniining the contact arc to an economical length, ten balls in my present device will support at least twice as much load as will fourteen ballsin a similar device having the circular curve'transverse the raceway) and this is advantageous, not only in that it reduces cost by economizing as to material, but also vin that for a given load the total amount of fside friction (for the bearing as a whole) due to the lesser number -of balls rubbing upon the aggregate amount of longer contact arc is no greater than such total amount produced by the larger number of balls when acting upon the shorter arc; also, as I prefer to assemble the lesser number of balls of a given what into the total area of contact under maximum rated load, thereis actual orv substantial conformity between the periphery of thel rolling member and the co-operating face of the racemember over a large middleportion of that area and no pressure above the average is necessary in the neighborhood of the lling slot to bring about the Vabove-described rigidity and uniformity of stress.

But also, if desired, devices according to my invention can have their economical lengths of contact less than those above mentioned, as, for example, to provide such a restricted arc of contact in an antlfriction bearing that under maximum rated load the rolling-element will roll clear of a filling-slot end. One ellipse to whose curve the above-described race curve can mission, it lis advantageous to make the racemember flat or of a circular curve and to employ a co-operating rolling-element that is a relatively narrow disc whose transversely curved periphery engages the race-surface, and in such instances I prefer to construct such transverse curve of the rolling-element upon that elliptical arc whose center is in the ellipses minor axis rather than in the major axis.

It will be apparent that my characteristic inter-relationship and clearances between the loadsupporting elements of my device can be obtained by many other curves than those of the circles and the ellipses whose diameters or equations are above given, and that those characteristics can also be obtained by many other lines than those of circular curves opposing elliptical curves. For example, the closely-conforming portion of the race (say, for an included angle of 45 degrees) can be the actual circular curvature of the cooperating ball, and beyond that arc the race surface can be so far removed from the ball that no contact will occur under maximum rated load, but in such event I prefer to join the contact arc and the outlying non-contacting portions by smooth and gradual curves, so that no abrupt changes of line will be presented. One method of making such a circular-curve contact area is by producing an elliptical curve as previously described and then using the designed co-operating ball, or its equivalent, to lap the contact area to the desired arc of the ball circle.

Thus, although my present device aords desired economical extent of contact between the load-supporting elements and avoids both detrimental engagement beyond such contact area and stress-concentration at the ends of such area, it also has substantially even distribution of stress per unit of length over substantially the whole contact area'even when subjected to the highest loads, and possesses great endurance properties and rigidity.

I claim:

1. In a device of the character indicated, cooperating load-supporting-members, the transverse surface of one of which when the device is free from load is a curve whose central portion engages the other said member in substantially the line of load, said curve as it progresses from said point of engagement nrst minutely increasingly departing from the direction of the surface of the Vsaid other member over a substantial proportion of the length of contact under normal load and then relatively rapidly increasing its rate of such departure until as said curve passes beyond said length of contact it is spaced from said other member by a distance greater than that obliterated by application of the load; substantially as described.

2. In a device of the character indicated, cooperating tangent load-supporting members that are curved transversely, one of said members being non-circular, and said members having substantially the same radius of curvature at said point of tangency, and over a substantial proportion of the arc oi contact under normal load minutely separating from each other as said point is proceeded from and then over the remainder of said arc more rapidly so separating; substantially as described.

3. In a device of the character indicated, cooperating load-supporting members whose engaging surfaces are related to each other substantially as are a circular arc and an elliptical arc which has its central point in an axis of the ellipse and tangent to said circular arc in substantially the line of load, such central portion of said elliptical arc being substantially the same curvature as is said circular arc; substantially as described.

4. In a device of the character indicated, cooperating load-supporting members whose engaging surfaces are related to each other substantially as are a circular arc and an elliptical arc which has its center tangent to said circular arc in substantially the line of load, said elliptical arc being very nearly the curve of such circle throughout substantially the arc of contact under load and rst only minutely progressively departing from said circular curve as said elliptical curve proceeds from said point of tangency, then appreciably accelerating its rate of such departure as the extremities of said arc of contact are being reached, and then reiatively rapidly further increasing such rate for the portions that pass beyond said extremities; substantially as described.

5. In a device of the character indicated, cooperating bearing-members which are transversely curved in lines that are related to each other substantially as the circumference of a circle of one-half inch diameter is related to an ellipse defined by the equation x2 yz '(.s' 46oz)2+(.29412 2 1 substantially as described.

6. In a device of the character indicated, cooperating load-supporting members Whose surfaces in the region of contact under load are spaced from `each other substantially as are a circular arc and a cooperating elliptical arc of the region of the vertex on the major axis of the ellipse, said arcs being tangent at the said ver- `tex; substantially as described.

'1. In a bearing, cooperating bearing elements having curved opposing surfaces, the curve of one of said elements being unitary and having a vertex, and having, at said vertex, a radius substantially the same as that of the other said element and also having, at points upon opposite sides of said vertex, a radius greater than that of said other element; substantially as described.

8. In a device of the character indicated, cooperating load-supporting members having engaging surfaces of which one is a circular curve also being initially in contact with said cooperating member at the centra. portion of said arc; substantially as described.

10. In a. device of the character indicated, co-

operating tangent load -supporting members,

which are curved transversely and one of such curves being non-circular, said members initially WILLIAM T. MURDEN.

cERTIFicATE `or CORRECTION.

Patent No 2,114.2, 14.78

January 5 195 9 i INIIL'LHIAFI T. MURDEN. It is hereby certified that error appears in the printed specification of the above numbered patent requiring correction as follows: Page 2, first column, line 52, for "in" read by; page LL, first column, line 51, for the word "highest" read lightest; and that the said Letters Patent should be read with this correction therein that the same may conform to the record of the case' lin the Patent office.

Signed and sealed this 21st day of March, A. D. 1959.

(Seal) Henry Van Arsdale.

Acting Commissionerjof Patents.

Images