on the other hand, had a radius of  r  8  7  mm and face width of 6。3 mm。    The disk had a circular lead crown of radius r            7  mm。 This was to ensure an elliptical

contact pattern with no edge loading between the disks。 Figures 2。4 and 2。5 show the engineering drawings of the roller and disk specimens, respectively。  Figure 2。6(a) shows the variation of the half-track width b (half-width of the contact ellipse in the axial direction) and the half-width a in the circumferential direction of rolling and sliding as a function of the normal load        acting on the rollers。   Similarly, Figure

2。6(b) shows the variation of the maximum contact stress o        [32] as a function of     。

This indicates that the desired contact stress levels can be obtained with this disk pair

design with a reasonable contact size。

Two different gear steels were used: (i) AISI 5120 steel intended to represent a typical automotive gear steel, and (ii) AISI 9310 intended to represent a common aerospace gear steel。 During the final finishing of the outside  surfaces,  a special process developed earlier [17, 23] was applied to induce grinding marks in axial direction to simulate the contact conditions of gears。

Once the specimens were evaluated for surface roughness, they were ranked by roughness values and sent out for further processing。   A batch of ground specimens

with roughness within Rq C [   ,   7] m were kept as the baseline ground sets (Figure

Figure 2。4  Engineering drawing of the roller specimens。

Figure 2。5   Engineering drawing of the disk specimens。

Figure 2。6。 Variation of (a) half-lengths a and b of the contact ellipse and (b) maximum contact stress with the normal force。

2。7(a))。 A batch of ground roller and disk specimens were processed using  a proprietary commercial polishing process to achieve ultra-smooth contact surfaces with

roughnesses of Rq  =             m as illustrated in Figure 2。7(b)。   Similarly, a   commonly

used chemical polishing process was applied to a third batch of ground specimens to achieve smoother isotropic surfaces at Rq  =        m as illustrated in Figure 2。7(c)。

One last batch of specimens were chemically polished and then coated with a commercial nickel-boron (NiB) coating。 This NiB coating selected by the Sponsor was dip-applied to all external surfaces of the specimens。   It  had  thickness   controllability

within [     ,       ] m ([              ,       ] inches)。  In this study, a recommended coating

thickness  of  25  m  (0。001  inches)  was  used。    The  specimens  used  for  this  were

intentionally undersized by the coating thickness as to allow for direct accommodation into the test machine with no hardware changes。 This particular coating had a hardness of 63-65 HRC if not heat treated after its application, and up to 74 HRC if heat treated。 Due to concern that the case hardened substrate could be tempered and lose much of its hardness during the heat treatment of the coating, the coated specimens were not put through further heat treatment。 Figure 2。7(d) shows a pair of rollers that are nickel- boron coated。

Figure 2。8 shows measured initial surface profile traces (in the direction of rolling) of representative rollers from each of these four surface variations。   In    Figure