In a study by Krantz and Kahraman [3], the effect of lubricant viscosity on gear wear rates and pitting lives were investigated through examining spur gear specimens tested with seven different lubricants。 It was found that the higher viscosity lubricants result in longer pitting lives as well as reduced wear coefficients。 Spur and helical gear efficiency studies by Xu and Kahraman [4], Petry-Jonhnson et al [5], Moorhead [6] and Vaidyanathan [7], Li and Kahraman [8], Li et al [9] and Britton et al [10] all showed experimentally that an increase in oil viscosity through a high viscosity fluid or a reduction in operating temperature reduces the load-dependent (mechanical) power losses。 This was confirmed by Talbot and Kahraman [11] through planetary gear set power loss experiments as well。 Some of these studies [4,8,9] also indicate that increased viscosity causes load independent power losses (due to churning and gear mesh pocketing), in the process wiping out some or most of the reductions  in mechanical losses。 The windage power loss studies of Seethraman and Kahraman [12], Seetharaman et al [13], Talbot [14] and Hilty [15] point to such increases in spin losses with increased lubricant viscosity as well。

The second method of increasing ratio through reductions in surface roughness have been investigated in various studies。 There are a number of commercially marketed processes that can polish machined surfaces either chemically or mechanically to reduce ground or shaved surface roughness amplitudes by 4 to 20 times。  Gear experiments in

references  [4,5,8,14]  all  show  sizable  reductions  in  spur  or  planetary  gear       mesh

mechanical power losses of shaved or ground gears when the surfaces chemically polished。 For instance, Petry-Johnson et al [5] report a 19% decrease in mechanical power loss of a spur gear boxs when surface roughnesses are reduced from 0。32 m for ground gears to 0。08 m through chemical polishing of  gears。  These  studies also showed that the surface roughness changes do not influence spin losses at all。

Several other studies preferred to use simpler contact arrangements such as a two- disk set-up to study roughness effects。 Diab et al [16] used a two-disk machine to compare traction behavior of polished and ground surfaces within a narrow slide-to-roll

ratio  (ratio  of sliding  velocity  to  rolling  velocity)  range  of  0。2 。 Similar two-disk

experiments [17] or ball-on-disk experiments [4,18] show that the friction coefficient of a contact interface can be reduced by reducing surface roughness。

A number of contact fatigue investigations studied the influence of surface roughness of gear durability。 Among them, Bluestein [19] and Klein [20] performed shaved and chemically polished spur gear pitting tests using FZG type machines to show 3 to 5 times increases in pitting lives when surfaces a polished。 Klein’s data was later correlated to the pitting model of Li et al [21] who showed the same improvements by smoothening the surfaces。 In addition, Li [22] performed two  disk  pitting  tests on surface roughness effects which were correlated to a point contact pitting model of Li and Kahraman [23]。 Gear fatigue tests by Krantz et al [24] and several other studies [25-27] confirm the same effects of surface finish。

There are several other studies concerned temperature induced failures such as scuffing due to heat generated at the contact interfaces。  These studies looked into the

impact of surface roughness as well as methods to reduce boundary friction coefficient at asperity contacts through oil additives that provide tribo-films or surface coatings。 Popgoshev and Valori [28] studied the impact of materials and lubricant types  on scuffing failures。 Lee and Chang [29] focused on correlating asperity contacts and the lubricant film thickness to resultant scuffing performance。 Liou [17] used a heat balance model and two-disk experiments to search for a scuffing limit for an automotive gear steel-lubricant combination。 In Alanou et al [30], the effect of  different surface treatments and coatings on scuffing was investigated。 In these tests, ground, superfinished, and superfinished plus coated surfaces were tested at a given constant sliding speed value with the contact pressure increased incrementally up to 1。7 GPa to determine the load at which scuffing occurs。 These tests indicated that superfinished surfaces scuffed at higher loads。 Results of this study were not conclusive as flaking of the coatings in certain conditions was observed。 A recent theoretical and experimental study by Li et al [31] indicates the same effects on a ball-on-disk contact。