The history of the Stribeck curve and ball bearing steels: The role of Adolf Martens

Abstract

The historical development on the discoveries of lubrication regimes for lubricated sliding contacts has been reviewed from the very beginning. It was found that the functional relationship between the coefficient of friction and the product of sliding speed and viscosity divided by the normal load well known as the Stribeck curve has been experimentally explored much earlier by Adolf Martens in 1888 long before Richard Stribeck did his pioneering measurements in 1902. However, basic work in this field of experimental hydrodynamics was done and results were published even earlier than that back in 1854 by Gustave A. Hirn, in 1879 by Robert H. Thurston and in 1883 by Beauchamp Tower.

Introduction

Long before the term “tribology” was coined in 1966, phenomena in the fields of friction and lubrication were named after relevant researchers. Recent research in the archives of BAM, the German Federal Institute for Materials Research and Testing (founded in 1871 in Berlin as the Royal Prussian Technical Testing Institute, MPA), has re-examined the work of Adolf Martens (1850–1914) and Richard Stribeck (1861–1950), both of whom worked in that institute.

One of the first lubricant testing machines, which give directly and precisely the value of the coefficient of friction, were designed in the mid 1870s by Thurston [1], [2] at the Stevens Institute of Technology in the United States. His machines were made commercially available. The work was focused on railroad purposes. The liquid-lubricated journals of one of his machines had a diameter of Ø ∼9.500 mm with a length of 17.800 mm. The pulley driven shaft could realize sliding speeds between 0.6 m/s and 1.5 m/s. The published data showed the variation of the coefficients of friction by load (contact pressure) and temperature at a given speed and not as function of speed.

This testing philosophy was consistent with the understanding of Thurston, as he wrote on p. 186 of Ref. [1]:

“…, it is readily seen, that the change in value of the coëfficient of friction with change of velocity is not great in machinery in which the velocity remains within usual limits,…”

and on p. 274 in Ref. [2]:

“The conditions which produce most serious differences in ordinary work are the nature of the unguent, the pressure, and the temperature. Velocity of rubbing determines a limit beyond which the intensity of pressure cannot be carried without danger of heating; but the effect of its variation upon the coefficient of friction is usually less considerable than is that of either of the other conditions specified”.

On the other hand he concluded on p. 209 in Ref. [1]:

“The resistance due to friction varies with velocity, decreasing with increasing velocity rapidly at very low speeds, as from 1 to 10 feed per second, and slowly as higher speeds are reached, until the law changes and increase at ordinary temperatures takes place, and at a very low rate throughout the whole range of usual velocities of rubbing in machineries”,

followed by the statement on p. 210 about

“The temperature of minimum friction under the conditions of these experiments is about T = 15√v for a pressure of about 200 psi”.

Thurston was therefore close to establish the “Stribeck” curve, but he presented no “Stribeck”-like graphs. Yet, the data provided by him in Table 71 on p. 185 of reference [1] showed clearly the minimum now known as the transition between full hydrodynamic lubrication and some asperity interaction. However, the data would have allowed this interpretation. The table presents coefficients of friction data determined each at different oil temperatures, speeds and contact pressures. Fig. 1 shows graphs re-constructed from the data published by Thurston.

It might be of interest that the very beginning on research on hydrodynamic lubrication is most probably associated with G.A. Hirn in France. He built and developed a “balance de frottement” (friction balance) in the mid of the nineteenth century. This test rig mated a polished grey cast iron drum (Ø = 230 mm, L = 220 mm) to a half-shell in bronze (Cu:Sn = 8:1) and was capable to run between 45 and 100 rpm. He tested different vegetable and natural base oils. He concluded from his intensive tests with this lubricated journal bearing, that the friction is proportional to the square root of the contact pressure (top of p. 217 in Ref. [3]) and not linearly proportional with it, which was contradictory to the teaching from Leonardo da Vinci for dry friction.

He also found, that by doubling the speed from 51 to 92 rpm, the friction of the lubricated journal bearing was reduced (see pp. 208–215 [3], data in “tableau E”), which was contradictory to the teaching from C.-A. Coulomb for dry friction. He concluded also (bottom of p. 209 in Ref. [3]), that with increasing speed the load carrying capacity of his lubricated journal bearing is also increased. However, Hirn did not perform frictional experiments by systematically varying the sliding speed in a wider range or with continuously recording of the resulting friction [3].

This was done and made public some decades later in 1883 [4] and 1885 [5] as Beauchamp Tower published his widely discussed reports on friction measurements of oil lubricated journal bearings. He conducted this work on behalf of the friction committee of the Institution of Mechanical Engineers in London. His data contained also the information that with increasing speed and load the friction from low to higher angular velocities is going through a minimum. By these experiments Tower discovered the effect of hydrodynamic lubrication and could explain the load bearing capacity of the oil film in the journal bearing but could not further interpret the reasons for the minimum in the friction data. This is best described by Towers own words, as follows [4]:

“A very interesting discovery was made when the oil-bath experiments were at the point of completion…. While the brass was out, the opportunity was taken to drill a ½-in. hole for an ordinary lubricator through the cast-iron cap and the brass. On the machine being put together again and started with the oil in the bath, oil was observed to rise in the hole which had been drilled for the lubricator. The oil flowing over the top of the cap made a mess, and an attempt was made to plug up the hole, first with a cork and then with a wooden plug. When the machine was started the plug was slowly forced out by the oil in a way which showed that it was acted on by a considerable pressure. A pressure gauge was then screwed into the hole, and on the machine being started the pressure, as indicated by the gauge, gradually rose to above 200 lbs. per sq.in. The gauge was only graduated up to 200 lbs., and the pointer went beyond the highest graduation…. This experiment showed conclusively that the brass was actually floating on a film of oil, subject to the pressure due to the load….”