£ E I £ I _> F I_ C A T I_ O N

This invention relates to method and apparatus for improving, the retention of oil in the environment of rotating shafts. Background of the Invention

For the purposes of discussion it will be assumed ^' that the term "rotating shaft" or just "shaft" includes wear sleeves providing the actual rotating surface in contact with an oil seal, as well as those shafts with ^• which the oil seal makes actual contact.

In spite of the many years and great efforts that have been spent on the arts relating to lubrication, perfect understanding is still unattained. The present invention deals with the relationship between a rotating shaft, a radial-lip type of oil seal, and a housing. The retention of the oil seal in the housing and the assurance of leak-tightness there is not pertinent to the present invention and will be assumed to be present. It is readily obtainable since there is no movement between the installed oil seal case and the housing.

For the retention of oil inside a housing for a rotating shaft, it is common practice to rely on a radial- lip type of oil seal having a casing held in a leak-tight fit with the housing. Such an oil seal has an elastomeric lip for leak-tight engagement with the surface of a rela¬ tively rotating shaft. At least three elements are necessary

SUBSTITUTE SHEET O PI

for the retention of oil: 1) the elastomeric ring or shaft seal, 2) the shaft, and 3) the lubricating oil or other fluid which is to be sealed. The lubricating oil provides a very narrow film separating the elastomer from the rotating shaft and thereby preventing the destruction of the elastomer.

Heretofore, it has generally beenassumed that a shaft or wear sleeve which has been ground and polished provides an absolutely smooth surface which has nothing ^to o with the problem of lubricating oil retention unless it is marred by defects such as scratches which are quite visible to the naked eye. It has also been assumed that a shaft which feels smooth and looks smooth is smooth, and that any problem that may arise leading to loss of some of the oil is therefore due to and the fault of the elastometic sealing element.

The present invention is based upon the discovery that the assumption that a smooth-appearing and ^" smooth- feeling shaft plays no part in the possible leakage of oil-is not strictly true. The grinding of the shaft to provide the "smooth" cylindrical surface that is necessary, produces microscopically observable irregularities which have heretofore been largely neglected but which, it has been discovered, can result in oil leakage. The problem has been that, during grinding, by either the conventional technique of plunge grinding or centerless grinding, the longitudinal axis of both the shaft being ground and the grinder, while they have been assumed to be parallel, have in fact been askew, due to tolerances generally permitted such parallelism. Until only recently (as mentioned below relative to an earlier patent application filed by co-workers) the importance of eliminating all such tolerances was not recognized.

This aske ness during grinding creates a direction- ally biased series of non-smooth microabrasions on the

SUBSTITUTE SHE

shaft, resulting in something analogous to a screw thread, and imparting a liquid-pumping effect.with similarities to the effect of an Archimedes* screw. It will be recalled that this screw, whether invented by 5 Archimedes or by others before his time, comprises a rather broad threaded screw encased by a cylinder and was long used in Egypt and elsewhere to raise water. In the present case, the cylinder may be considered as analogous to the radial lip shaft seal, while the screw 0 is provided by the shaft turning within the seal, and the liquid is typically oil. Thus, depending on the direction of shaft rotation, the oil may be pumped away from the lip, or the oil may be pumped in the opposite direction so that the oil flows beyond the seal, pro- 5 ducing leakage. In the latter case, the user has usually believed that he has a faulty seal, and he may have changed the seal, only to discover that the prob¬ lem remained, without anybody knowing why.

The ignorance of* this problem is such that neither ^* 0 the-ultimate users of shafts, nor the shaft manufacturers, nor the oil seal manufacturers have heretofore appreciated it. As a result, when new designs or new types of oil seals were tested, leakage sometimes occurred for no apparent reason that could be determined. On retrospect 5 it appears that the reason was this Archimedes' screw effect due to the directional effect of the grinding of the shaft—even though the shaft seemed to be perfectly smooth and met all the established surface smoothness specifications. 0 The grinding of a shaft, just discussed, does not result in what is actually a screw thread, but it is similar in that it creates microscopic troughs, all ex¬ tending in the same general direction, namely, that governed by the degree of askewness, and each individually 5 .and collectively tending to give the effect of a screw thread pattern. The angle is so small that it cannot be

^S U ^BS TI ^T UT ^"

seen even with a 250x power microscope, but its existence can be verified by a variety of tests. On the other hand, many instances of dynamic oil leakage, i.e., during relative rotation with a shaft, can actually be traced to this basic phenomenon.

A typical example of a shaft intended for use in the automotive or light truck vehicular market is given below in the detailed description of my invention. However, suffice it to say that literally hundreds, if not thousands, of these microscopic troughs on the shaft surface are within the seal lip contact area with the shaft and each imparts an axial force on the thin film of lubricant, commonly oil, generally present between the seal and shaft within this limited area. As a result, the lubricant is literally pumped in one axial direction or the other dependent on relative direction of shaft rotation, out from the seal-shaft contact area. In one such direction of shaft rotation, the seal will be considered a "leaker". Although the effect is not gross, the effect over a long period of time, and in some parti¬ cular environments, will be serious ^" enough that the user of the device will often believe that his oil seal is leaking and will replace it under the faulty assumption that the oil seal is defective or worn out, when that is not the case at all.

The dynamic oil leakage problem may or may not be severe, but the effect on the user, as already noted, may be severe because no one wants to see oil leaking in any of the situations in which oil seals are used, and if the wrong remedy is attempted, considerable expense can result without any beneficial result being obtained.

An earlier patent application by co-workers and subject to the same ownership as this application has been filed, which clearly recognizes the problem as being caused by askewness between grinder and shaft.

E SHEET

as aforesaid, and purposes as a solution to either (i) eliminate virtually all tolerance heretofore permitted in establishing parallelism between grinder and shaft so as to produce a no-lead shaft or (ii) grind the shaft at a known degree of non-parallelism with the grinder so that the direction of shaft lead could be accurately controlled during production, identified, and taken advantage of during installation, to provide a shaft seal combination which would pump the thin oil film in the direction of the oil reservoir rather than to the air side of the seal. This, however, does not solve the problem that occurs when the shaft is intended to rotate in both directions, such as a transmission drive or drive axle, or where for production reason a compatible direction of shaft rotation can not be assured, and it is to this specific problem that the present invention is directed. Summary of the Invention

The present invention in its method aspect relates to "the making of a cylindrical surface which is sub¬ stantially neutral, hydrodynamically speaking, when it is used as the contacting surface in engagement with a radial-lip oil seal during both static and rotational conditions. The invention calls for balancing the directionally biased microabrasions produced during preparation of that cylindrical surface between clock¬ wise inclined microabrasions and counter-clockwise in¬ clined microabrasions. In other words, these directionally biased microabrasions are put into equilibrium by adopting a shaft-surface finishing procedure which assures that approximately the same quality and quantity of directional microabrasions tend toward one direction as tend in the opposite directions. The effect is to cancel the screw- thread effect and hold the oil in status quo. The sur- face formed is, of course, smooth to the touch and to unmagnified sight.

OMPI ET ™

The balance provided by this invention can be achieved in various ways. It can be done by such expedi¬ ents as a traverse grind in which the shaft is sent through twice, one time with a left-hand grinder feed and the second time with a right-hand grinder feed. Or the part itself may be sent through the same grinder twice, in these two different directions. Or it may be sent in and then retracted. Other ways of achieving it are by double angle plunge grinding in which the angle of cutting is reciprocated, either "waggling" the wheel or the work piece, or by a combination of the principles of dress, feed, angle, and so on. Each of these will be discussed in detail below, and other objects and advan¬ tages will appear from the illustration and description of some preferred embodiments.

Hydrodynamic flutes on a seal and hydrodynamic threads or microabrasions on a shaft are quite different i-n relationship and physical position with respect to the oil film. The hydrodynamic flutes of a seal are on €he- "air side" of the seal, so that the flutes are active only when oil gets through the seal to the air side. Shaft microabrasions may be on both sides of the seal, but only the area in contact with the seal can be functional to cause leakage of oil past the seal. A bi-directional hydrodynamic seal is supposed to keep a seal from net leakage no matter in which dir¬ ection the shaft is turning, principally by return of oil from the air side to the oil side. On the other hand, a microscopically directionally balanced shaft as described in the present invention, establishes a substantially stable oil film between the seal and the shaft.

Brief Description of the Drawings

Fig. 1 is a view in elevation and in section of a portion of a typical oil retention system, embodying an oil seal in a housing with its radial lip bearing on

OMPΓ

the surface of a shaft. Only the upper half of this combination is shown. On the scale shown, the shaft appears perfectly smooth, and the lip appears to be nearly a knife edge, although wear soon blunts that edge. This view represents both the prior art and the present invention, for at this scale they look alike.

Fig. 2 is a greatly enlarged fragmentary view also in accordance with the prior art, showing the Archimedes' screw effect caused by grinding practices which theoretically are sound enough, but which lead to a directional flow effect.

Fig. 3 is a greatly enlarged, and possibly somewhat exaggerated, fragmentary view of a portion of the shaft of Fig. 2, showing how the supposedly smooth shaft contains elongated microabrasions with a marked tendency toward screw-thread or directional structure resulting from the way in which the grinding is done.

Fig. 4 is .a photomicrograph showing a shaft surface as ground pursuant to prior art shown in Figs. 2 ahd .3.

Fig. 5 is a photomicrograph showing an actual shaft surface as ground pursuant to the present invention, but greatly exaggerated as to angle.

Fig. 6 is a diagrammatic view of a grinding system according to the present invention, employing grinding with relative longitudinal motion between the grinding wheel and the shaft being ground, first in one direction and then in the opposite direction.

Fig. 7 is a diagrammatic view of a modified form grinding system also used to produce equal and opposite angles of microabrasions, balancing skewness of the grinding wheel with its relative longitudinal movement.

Fig. 8 is a diagrammatic view of a modified form of grinding system balancing dress against longitudinal movement for producing equal and opposite angles in a

§yiδTϊTUT

single one-directional pass.

Fig. 9 is an enlargement of a portion of the shaft or wear sleeve after grinding according to Fig. 8. Fig. 10 is a diagrammatic view of a centerless grinding operation in accordance-with the present invention, using skewness to balance longitudinal bias.

Fig. 11 is a view in perspective of a wear sleeve ground by the system of Fig. 10.

Fig. 12 is a view in side elevation of an on-center grinding operation illustrating application of the present invention, with relative tilting used. Description of Preferred Forms of the Invention Preliminary considerations

Fig. 1 shows a simplified and typical oil seal installation. A housing 20 surrounds a shaft 21. Between the shaft 21 and the housing 20 is interposed an oil seal 22 comprising a supporting annular case 23 having an outer periphery -24 engaging a housing bore 25 in a stationary, leak-tight fit. An elastomeric member 26, molded. and adhered to the case 23, has a lip area 27 which engages the shaft 21. A garter spring 28 urges the lip 27 against the shaft 21 with a desired amount of pressure. The left hand side of the figure is the air side, and the right hand side is the oil side, where the oil is to be retained. The structure shown in Fig. 1 is merely exemplary, for the principles of the invention apply to a very wide variety of lip-type seals. It will be under¬ stood that Fig. 1 shows only substantially the top half and that the bottom half is substantially a mirror image along the axis of rotation.

In Fig. -1 and in actual installations, the shaft 21 appears to have a very smooth, polished surface 29. The shaft 21 may be either a complete shaft, or it may be only a wear sleeve on a shaft, but in either event, its surface appears to be perfectly smooth. However, as a

SUBSTITUTE SHEET OMPI

matter of fact, if one were to take a very great enlarge¬ ment, it would be evident that the shaft 21 is not com¬ pletely smooth.

Fig. 2 shows a great enlargement of a small portion of Fig. 1 as applied to a prior-art shaft or wear sleeve. The lip 27, which in Fig. 1 appeared to be provided with a sharp edge, actually has a flat portion 30"facing the shaft 21, and the shaft 21 is seen not to be smooth at all, but to have troughs 31 between lands 32 and acting substantially like a screw thread. This results from the very slight imperfections which are always found on ground surfaces.

As a shaft (or wear -sleeve) is put through a grinder in one direction, or as the grinding device is moved along a stationary shaft, the grinding device, which is, of course, not smooth and must not be com¬ pletely smooth, includes irregularities. It is supposed to grind the shaft so that the shaft will be very smooth, and to the naked eye it appears to do so, but the smooth- nes ^' s is only relative, and roughness shows up dramatically when one approaches molecular dimensions. In fact, the surface is not smooth even when seen at enlargements sub¬ stantially above molecular dimensions. The result of this type of imperfection, which has apparently been over- looked or disregarded in the prior art, is that when the shaft 21 rotates in one direction, it will tend to return oil to the oil side, whereas when it rotates in the opposite direction it will tend to carry oil out of the oil side and into the air side, somewhat like an Archimedes' screw. The observer who believes the shaft to be perfectly smooth may well blame this ejection of oil on the oil seal or on something else which is not concerned at all.

Fig. 3 is a very much larger scale enlargement of a very small portion of the prior-art shaft of Fig. 2,

SUBSTITUTE SHEET O PI

showing microabrasions 33 having directional inclination, substantially as are produced by conventional grinding. There are recesses 34 and projections 35, all inclined and tending therefore to act like a screw-thread struc- ture and result in directional flow. This feature (the microabrasion) can be described as the generally parallel, interrupted-helical microabrasions that are generated on a cylindrical surface of a rotating shaft by the inter¬ action of obliquely rotating, cylindrically-oriented abrasive matrix on the working surface of an abrasive wheel.

The microabrasions 33 often look like elongated four-sided pyramids in relief. The sizes and angles of these pyramids are controlled by (1) the coarseness of the abrasive on the grinding wheel, (2) the strength of the adhesive holding the grinding wheel together, ^• (3) the differential speed between the rotational speed of the shaft and that of the grinding wheel, ^" (4) the angle of skew (if any) between the neutral axis and the grinding axis, and (5) the direction of rotation of the work and grinding wheel. If no care is taken or if the grinding is done in the typical manner, the abrasive action removes material, leaving the pyramids, and also tends to provide a definite direction of Archimedes screw effect, as in Figs. 2 and 3.

By way of example, an automotive or light truck vehicular application commonly requires a 1.875 inch diameter shaft having a surface finish of 15 RMS. Such a shaft would typically be ground-with an 11 inch diameter grinding wheel, which will produce on the shaft pyramids averaging in size from about 0.0003 inch wide, 0.00015 inches deep and 0.015 inch long. Naturally, these dimensions will vary with other shaft diameters, specified surface finishes, and grinding wheel diameters, among other things. Nevertheless the principles involved.

SUBSTITUTE SHEET ^~ W ^B ^X]

OMPI

specifically the oil film dynamics, remain the same.

The photomicrograph of Fig. 4 was made as follows: A prior-art shaft substantially as described above was coated with collodion. When the collodion hardened, it was stripped from the shaft and light was shined through it. The original picture was taken on color slide film through a microscope. On the 35-millimeter film, the magnification was approximately 50x; from this was made, via an inter-negative, a lOx enlargement (8" x 10" paper, with margin) ; this was reduced when making the patent drawing; the resultant enlargement is approximately 250x relative to the original. The microabrasions do not appear to have any directional inclinations—but they do, as just described above and as demonstrated both by oil leakage and by sensitive string tests. Fig. 4 shows how it appears to be impossible to ascertain the effects of shaft microabrasions visually even at great magnifi¬ cation. A substantially neutral hydrodynamically compensating shaft •

By practicing the present invention, a hydro¬ dynamically directionally balanced structure as shown in Fig. 5 can be obtained. Given a 1.875 inch diameter shaft as described above, the seal-shaft contact area will be approximately 0.010 to 0.020 inch wide along the axial direction and within this span literally hund¬ reds of the aforesaid pyramid shaped troughs will be exposed to the thin oil film.

Fig. 5 shows an actual enlargement of a shaft which has been treated according to the present invention and is substantially hydrodynamically microscopically balanced. As shown, there are a number of pyramid-like microabrasions on the shaft surface. Microabrasions cannot be avoided. The depth of what appear to be scratches range up to about 0.00025 inch deep. Again,

SUBSTITUTE SHEET ( ■. ^OMPI .

these microabrasions are very small, but each one tends to move the oil in a direction depending upon the dir¬ ection of shaft rotation. The angles of inclination of each relative to a plane perpendicular to the shaft 5 axis would be in the order of about 0.5 to 3° with most probably averaging less than 1.5°. Likewise many, but not all, of the scratches may cross one another. Fig. 5 shows an actual photomicrograph at 15Ox magnification. To reiterate, in order of grasp the concept of

10 cumulative effect of the micro interrupted-helices that give the shaft a surface texture.

The Fig. 5 photomicrograph was taken generally like that of Fig. 4 but was taken on black-and-white 35 mm. film. On the film the enlargement is almost

15 exactly 50x, and from the film the 8" x 10" print repre¬ sented a somewhat greater than lOx further enlargement, followed by reduction when the photolithographic repro¬ duction was made or the formal ^" Patent Office drawings. For Fig. 5, a special shaft was subjected to micro-

^" 2~0 abrasions at an exaggerated angle of about 15 to a radial plane and this was followed by imposing micro¬ abrasions of the same magnitude and same angle in the opposite direction. The shaft was then coated with collodion, and when the solvent evaporated, the collodion

25 was removed and light sent through the collodion, the angle being adjusted to show the microabrasions clearly. The exaggerated angles of the micro-abrasions for the purpose of showing them clearly, and would not be used in practice, where the angles would, as in Fig. 4, be

30 practically unobservable, even at this magnification. An analysis of the Fig. 5 photograph of highly magnified shaft portion shows that it takes about 5 microabrasions to equal one cut that is 0.002 inch wide. That means it would take 25 microabrasion side-by-side 35 to span the 0.010 inch width of the static band. If the

SUBSTITUTE SHEET

O PI

microabrasion average 0.012 inch in length, that means it would take 490.8 end-to-end to go around the 5.89 inch circumference of a 1.875 inch diameter shaft.

Therefore there would be 490.8 x 25 = 12,271 5 microabrasions, all acting on the oil seal where it touches the shaft.

If each micro-abrasion is biased or skewed only 0.000001 inch (a millionth of an inch), in one revolu¬ tion it could move oil axially 0.0005 inch (one-half of a 10 thousandth of an inch) . This means that in 20 revolu¬ tions, one micro-groove could move the axial equivalent of one static lip width.

One micro-groove does not hold much oil, but with 12,271 of them all working together it is no wonder 15 that a seal will leak several drops a minute when the shaft is running at 5000 r.p.m.

Carrying this thought on a little farther, a skew of a millionth of an inch in a space of twelve ^' • thousandths of an inch is the equivalent of one inch ^' 20 in a thousand feet. You can hardly see an inch if it is three football fields away. You can see a telephone cable a block away if it is against a clear blue sky. You can't, however, hold two strings in your hand and detect the angle between them if one goes to each side 25 of the phone cable. This is the "micro-lead" of the interrupted helices that make up the surface texture with which we are concerned. Collectively they cause problems even though individually they fade into insig¬ nificance. 30 The invention broadly

The present invention provides ^' three different approaches to arriving at a more neutral shaft:

1. Produce two sets of microabrasions each being at relatively equal and oppositely inclined angles, 35 relative to the axis of rotation of the shaft, so that

SUBSTITUTE SHEET MPI

the effects are off-setting to the fluid flow (Figs. 6, 7 and 12) .

2. Produce microabrasions by diamond-dress lead that are opposite and off-setting to thread-like micro- abrasions produced by the normal grinding (see Figs. 8 and 9) .

3. Transfer two relatively equal and opposite diamond dress leads to the shaft by grinding (Figs. 10 and 11) . Producing microabrasions at angles to the axis of rotation of the shaft, but with equal and opposite angles (Figs. 6 and 7)

Fig. 6 shows diagrammatically a grinding wheel 40 with circumferential abrasive particles which tends to produce a right-hand effect on a shaft 41 when the shaft 41 and wheel 40 are moved in one direction relative to each other, and a left-hand effect when moved in the opposite direction. There may be two ^' -such grinding wheels that are substantially identical, or a single grinding wheel 40 which is moved relative to the shaft 41, first in one direction to produce microabrasions

42 and then in the other to produce microabrasions 43. In both instances the microabrasions 42 are at an angle to the axis 44 of rotation of the shaft, but if properly made with proper velocities, contact, and parallelism, the angles of inclination of the microabrasions 42 and

43 can be made substantially equal and opposite, and an approximately neutral shaft thereby made.

In Fig. 7 a similar effect is obtained by using a grinding wheel 45 which is somewhat skewed relative to a shaft 46 and thenmoving it in such a direction and at such a rate that instead of producing an effect on the shaft 46 in the nature of a screw thread, or with directional tendencies, tends to produce a neutral series of scratches without inclination.

- UHX

SUBSTITUTE SH EX OMPI

A similar system might use the skewed wheel 45 in plunge grinders, and then plunge-grinding the same shaft with a different wheel of opposite skew. Another alternate system would be to follow the plunge-grinding 5 with the skewed wheel 45 and then using longitudinal movement with a wheel like the wheel 40 to produce micro¬ abrasions in the opposite direction and crossing at about the same angle. Thus, there will be a combination of left-hand and right-hand directional tendencies, which 10 amount to. a substantially neutral tendency in the end. Adding balancing threads against microabrasions (Figs. 8 and 9)

Fig. 8 shows an abrasive wheel 50 which has been skewed so that it produces microabrasions 51 in one 15 direction, on a shaft 52 that is moved longitudinally. This grinding wheel 50 is then trimmed, as illustrated in Fig. 6, by a diamond dress device 53, which advances about.0.008 of an inch per revolution of the wheel, thereby producing something ^' like a screw thread 54 on •2*0 the- shaft or wear sleeve 52. This is shown in Fig. 9 where the microabrasions 51 and screw thread 54 are shown enlarged and crossing each other.

The grinding may be plunge grinding, if desired, or longitudinal, if desired. The combination of skewing 25 the abrasive wheel 50 slightly to produce the micro¬ abrasions 51 and the threads 54 produced as a result of the diamond dress 53 results in the structure shown in Fig. 7 in which threads 54 are inclined in one direction, and the microabrasions 51 are inclined in the other. 30 By trial and error, or by careful design, the micro¬ abrasions may be balanced against the screw thread to assure a substantially neutral action. Thus, a right hand lead from a diamond dressing may result in a thread which would be clockwise in the direction of rotation 35 shown, whereas the abrasion scratches -would result in

SUBSTITUTE SHEET ^~ ^ π }

O PI b _* . ^WIPO

a left hand lead. The scratches 51 are superimposed on the thread 54, which was generated by the diamond dress 53.

A centerless grinding operation (Figs. 10 and 11) Fig. 10 shows an operation in which a grinding wheel 55 and a rubber driving wheel 56 are paired with their axes slightly skewed so as to achieve the driving effect. A blade 57 supports a series of rings 58 to be ground into wear sleeves. By having the blade 57 tipped to lie askew relative to the centerline 59 of the grinding wheel 55, the microabrasions below the centerline 59 give a clockwise orientation to the micro-structure, and those above give the opposite orientation. By raising or lowering one end of the support blade 57, the centerline of the rings 58 being ground will cross the centerline 59. The support blade 59 can be adjusted accurately so as to balance the scratches in one direction with those of the other, and thereby produce a wear sleeve 60 like that shown in FigI 11 with crossing troughs.

Thus, raising or lowering one end of the support blade 54 so that the center lines being ground tend to cross, produces something like the thread structure. The centerline of the grinding wheel 55, if raised in the opposite direction, changes the direction of the micro-abrasion lead. Thus microabrasions below the centerline give a clockwise orientation to the micro- structure, and those above the centerline tend to give a counter-clockwise orientation. The result when this is done in both directions, again, will produce a sub¬ stantially hydrodynamically balanced series of rings which later become wear sleeves, and which will prevent the kind of leakage of oil which has been described above, A centered grinding operation (Fig. 12) Fig. 12 shows an on-center grinding set up which

EET

will produce a structure having the same effect as that shown in Fig. 5, which is substantially hydrodynamically balanced so that the oil will not be worked out from the oil side to the air side no matter which direction the shaft rotates in, and in fact oil that might tend for other causes to leak out can be returned to the oil side of the seal.

In Fig. 12 a work piece 64 is mounted on a drive shaft 65 by a suitable chuck 66 with a suitable driving mechanisms 67 to rotate it. The outboard end 69 of the work piece 64 is retained in place by a suitable end support member 69. A grinding wheel 70 is then applied to the surface on one side of the shaft, or wear sleeve 64 to be ground and it is to be plunge-ground or moved along the work piece 64. Conventionally, this has been done in only one direction, and the traversing of the grinding wheel 70 along the work in either direction produces a ground thread on the work, .as demonstrated earlier. However, by traversing ^' in both directions, the thread, tends to be broken up, for reversing the rotation also breaks up the thread. Also, wobbling the axis of rotation of the abrasive wheel 70 (as shown in the broken lines 71) or of the work piece 64 (as shown by the broken lines 72) gives microabrasions in both directions which tend to produce a neutral shaft.

Wobbling implies repeatedly changing the axis of rotation, while tipping implies a single change in position which lasts for many revolutions. Thus, the microabrasions may be changed by wobbling or by tipping the grinding wheel either way, or by raising or lowering the end of the workpiece which tips the work in either way. In each instance, there is a movement through the center line of the other part. In contrast, making the axes nonparallel makes the microabrasions have a definite direction. Therefore, by either tipping or controlled

_S U _BST ITUTE SHEET / *!_ *

wobbling, the result of the present invention can be obtained.

To those skilled in the art to which this invention relates, many changes in construction and widely differeing embodiments and applications of the invention will suggest themselves without departing from the spirit and scope of the invention. The disclosures and the descriptions herein are purely illustrative and are not intended to be in any sense limiting.

^■ w ^β mυrέ _SHEET