Bowden and Tabor^11 have shown that the coefficient of friction for one lubricated surface moving relative to another will depend upon the material used in the two surfaces, the type and velocity of motion between the surfaces, and the composition of the lubricant. The first two factors are fixed by machine design, but the gear lubricant can be altered so as to aid in providing desirable friction characteristics. The contribution of this last factor is evident if we consider that the coefficient of friction of lubrication metal surfaces is about 1.0 while in the case of such surfaces, lubricated with a boundary film, the value is about 0.05 to 0.15.
Any explanation of friction is based on the fact that there are irregularities in surfaces. As one such surface moves relative to the other there is an interlocking of the asperities of the two areas. The role of the lubricant in overcoming boundary friction may be partly by filling in the low spots and in the case of a thick film, of lifting one surface over the other. However, under the best conditions, there will be some interlocking of the high spots. If, in so doing, some asperities plow through those of the other surface this will cause one type of friction. The remainder of the friction may be a result of welding and shearing of minute junctions between the two surfaces.
We are concerned, in the main, with two steel surfaces in rolling or sliding contact. While the thought has been advanced that rolling friction is essentially independent of the lubricant used, gears in operation, particularly hypoid or worm gears, have both a rolling and sliding motion. An example of the variation of friction with the type of motion concerns first a steel rider sliding across a lubricated plate of the same steel to give a coefficient of friction of about 0.14. With the same combination of materials, the rolling friction coefficient was about 0.00015.
Other operating variables to be considered are load, speed, and temperature. Rounds^45 using a test machine in which ball bearings had both a rolling and sliding action and, thus, simulated the action of gear teeth, concluded that, in general, the coefficient of friction decreases as either the oil temperature, the ball velocity, or the load increases. However, the magnitude of the oil temperature and ball velocity effects tend to decrease as the load increases.
In choosing a gear oil one is faced with a compromise because the lowest viscosity oils have the least internal friction and, yet, high viscosity oils will maintain the most satisfactory film to prevent metal to metal contact, particularly at low speeds. This is true when hydrodynamic or thick film lubrication conditions prevail and where viscosity is the most important property affecting friction. While a condition of low friction is desirable in a lubricated gear set, it must be kept in mind that there is not necessarily a correlation between friction and wear. Beeck^3 has suggested that this may be due to the fact that while wear takes place momentarily at isolated spots, friction is ordinarily measured as an average of a large area and a longer time interval.
Irrespective of viscosity, various types of mineral lubricating oils may give different friction values. Rounds^45 found that naphthenic base oils gave higher friction values than paraffinic base oils, even after limited attempts to change the friction properties by fractionation or additional refining. In this investigation, the most commonly used synthetic fluids gave friction values similar to straight mineral oils. Since paraffinic oils show less increase in viscosity with pressure than do naphthenic type oils, this may account for the lower friction values of the former type of oil when used in lubricants. The discussion above has been concerned with kinetic friction of fluids containing no additives. When boundary lubrication conditions exist, nonreactive gear oils will not suffice; therefore, consideration must be given to lubricants containing additives, Rounds,^45 using a naphthenic type oil as a base, found that different additives would lower kinetic friction and/or raise or lower static friction. Fatty acids and related compounds were the most effective of the agents investigated for lowering static friction. To lower kinetic friction at velocities above 100 fpm, reactive chlorine and sulfur compounds were most effective.
However, the type of service which demands the use of EP gear lubricants nullifies reduction in friction. For example, Bisson et al.,^5 when investigating the effect of chemical reactivity of lubricant additives on friction and surface welding at high sliding velocities, found some decrease in friction up to about 1300 fpm, when using p-dichlorobenzene, followed by a very abrupt increase in the coefficient of kinetic friction. Thus it was found that :For all additives, the existence of a critical sliding velocity where the friction coefficient increased and surface welding occurred was verified.
Kinetic friction in fluids causes heating and thus power losses. The friction due to gear oils themselves may be film friction, which has been considered, or churning friction which occurs as the gear teeth rotate in the oil bath. Other things being equal, gear oils which will have the lowest internal friction will be the most desirable from a power efficiency standpoint.
In connection with friction, consideration might also be given to the oiliness characteristics of gear lubricants. The following definition was adopted by the SAE Crankcase oil Oiliness Committee in 1937: Oiliness is a term signifying differences in friction greater than can be accounted for on a basis of viscosity when comparing different lubricants under identical test conditions.
This quality is desirable in boundary lubrication and is no doubt due to adsorbed layers on the metal contributed by polar compounds such as fatty acids, esters of fatty acids, metallic soaps, some organic compounds containing chlorine, nitrogen, phosphorus, or sulfur, etc. Unfortunately there is no standard method for measurement of oiliness.
