The dilution of crankcase oils with the heavy ends of gasoline constitutes the outstanding present-day problem in the lubrication of automobile engines. This paper first presents the results of extensive tests designed to determine the rate and extent of dilution for various types of car under various operating-conditions. From these results it appears that, in the average car in winter service in the northern half of the Country, the viscosity of the original oil drops to about one-third of its original value in the first 150 to 180 miles of intermittent operation. Beyond this the viscosity, contrary to general opinion, remains fairly constant at an equilibrium value corresponding to about 15 per cent dilution, where, on the average, the rate at which fresh diluent enters the oil is practically balanced by the rate at which it is eliminated therefrom. Individual cars, of course, vary rather widely from this average figure, depending partly upon the design and condition of the car, but to a greater extent upon the operating conditions. Certain theoretical equations with empirical constants are shown to correspond closely to the observed curves giving the change in dilution with mileage.
With the average oil losing one-third of its viscosity in the first 150 miles of winter operation, to prescribe any ordinary type of oil which will have a really proper viscosity throughout any reasonable period of service is obviously impossible. In general, the widely used medium grades of oil, having around 325-sec. viscosity, are probably the best compromise available; but, for the first 50 miles of operation, their viscosity is considerably too high for easy starting when cold or for effective cold-lubrication in most cars. After 150 miles, the amount of dilution is, in general, so great as to reduce the viscosity to a point too low for safety, especially as road dust and grit accumulate. The use of a heavier oil accentuates the first set of difficulties and undoubtedly causes much of the scoring and wear of cylinders and rings, while the use of a much lighter oil gives an average viscosity after about 150 miles, which is far too low for safety.
As a remedy, this paper suggests the use of a fairly heavy oil, of from 500 to 575-sec. viscosity at 100 deg. fahr., blended with from 10 to 12 per cent of a distillate having a boiling range substantially identical with that found in the average crankcase oil at equilibrium. By this means it is possible to produce an oil with an initial viscosity around 220 sec, which, being thin enough, gives easy starting and good cold-lubrication and yet is so near the average equilibrium-dilution that, in general, it maintains a viscosity in this optimum range throughout its entire period of service. In warm weather, or in cars in which the average running-temperature is high, some of this diluent will shortly be driven-off to give a somewhat higher viscosity, which is desirable for such vehicles, while cars kept in cold garages and frequently started and stopped will have a somewhat lower equilibrium-viscosity, as is desirable for such vehicles.
These theoretical considerations have been fully confirmed by extensive comparative tests in two fleets of cars and trucks operated both summer and winter on the old and new types of oil. Changes in dilution and viscosity were followed closely, and the results for different types of car are presented graphically. If a car dilutes its oil to more than 20 or 25 per cent, the viscosity of even the new type of oil enters a range where the factor of safety is low but, in any case, the new oil is better than the old. If car builders can, by design and by education, keep dilution below 20 or 25 per cent, this oil should solve the remaining dilution problem in a very satisfactory manner. The factors and limitations that determine the optimum composition of the new type of oil are also discussed. Present specifications are, of course, not applicable to these oils.