by Kevin Cameron

In this little book is an excellent brief history of the plain bearing, beginning on page 69. Rolling-element bearings support their load in an obvious way, and need lubrication only for cooling and to discourage the grosser forms of metal-to-metal contact. Plain bearings developed from the simplest of models -- the tallow-lubricated wheel of an oxcart, rotating on its fixed axle. Steel shafting running in bronze shells supported much of the industrial revolution, but where there was contamination in the oil supply -- as in internal combustion engines whose bearings shared their lubricant with piston and cylinder -- it was necessary to use bearing surfaces softer than bronze. Tiny particles carried by the oil into the running clearance of a bearing were liable to go around and around until they had completed an entire career of destruction, cutting up both surfaces as they rolled, lodged, gouged, welded, and broke away again. If the shell were of a soft material, the first pass through the loaded zone would push the particles down into the bearing surface, and they would be gradually hammered down flush into harmlessness.

At first, cavities were provided in bearing saddles, intended to be poured full of the soft babbitt bearing metal, against a bar that stood in for the crank journals. The resulting rough surface would then be laboriously hand-scraped into reasonable contact with the actual crankshaft, using the time-honored method of bluing the journals with Prussian blue, setting the crank in place, and then scraping off all the now-blue points of contact.

Many aircraft engines of WW I used replaceable bronze shells carrying a considerable thickness (almost a millimeter) of this babbitt metal. It was reasoned that, should the babbitt wear away or crumble off, the journal might look benevolently upon the bronze as a substitute partner.

The famous V12 Liberty aircraft engine went to France in 1918, says Heron, and promptly suffered big-end failures from fatigue of the weak babbitt. The problem was discussed with Ettore Bugatti, who had already encountered this kind of failure and had overcome it by reducing the thickness of the babbitt; the more of it there was, the more it could flex or yield under load, and the sooner it would fatigue and begin to break up. Bugatti's shells were steel-backed, with much thinner babbitt, and they worked well.

Later, in the early 1920s, there were more Liberty big-end failures, but in the US. Norman Gilman of Allison Engineering (of Indianapolis, later builders of their own famous V12 engines) also made up steel shells with thin babbitt. These were used in fork-and-blade rod pairs in wich the blade rod rode, not on the crankpin, but on the outside of the bearing shells in the fork rod. Gilman put copper-lead material on the outside, and it worked so well that Army engineers wanted to test shells with copper-lead on the inside as well. The resulting bearings worked so well that they were applied as main and con-rod bearings on many subsequent aircraft engines -- and the technology continues today.

An interesting aside on the development of plain bearings is that, in tests run by Rolls-Royce in the 1930s to determine the possible cause(s) of Merlin V12 main bearing failures, it was found that these bearings could run happily and without detectable temperature rise for some seconds after their oil supply was shut off.

If, however, air were to get into the oil supply, so that a jet of air at oil-system pressure was forced into the bearing, it would seize almost instantly. This, I think, is a very important point with respect to plain bearing failures in car and motorcycle racing; keeping the flow going is not enough -- what is flowing must be oil, for air will blow the oil out of the bearing and fail it at once. No wonder fancy racing engines include air/oil separators, rather than leaving matters to chance.

The obvious advantage of the inserted shell-type plain bearing over the earlier, poured-in-place babbitt metal is that bearings can quickly be replaced with identical parts, making unnecessary the old business of laboriously hand-scraping the poured babbitt metal into fair contact with the journals that were to run against it.

As S.D. Heron notes in his review of plain bearing history, the Allison Co. developed technology leading to the widespread adoption of steel shell inserts surfaced with the more-durable copper-lead material. Copper is better able to bear load without cracking than is babbitt, while the presence of the soft lead provides the ability to embed foreign particles, preventing them from causing further damage. Lead has only limited solubility in copper as the material cools from melt temperature, forming a sort of lead-filled matrix of spongy copper.

Unfortunately, there were problems bonding this useful stuff to the steel shells. There were attempts to use solid shells of copper-lead (parallelling the earlier use of bronze shells with thin babbitt lining them) but this didn't work.

Ultimately, modifications to the copper-lead alloy - additions of tin or silver - improved the bond to steel -- and steel-backed, thin copper-lead plain bearings were widely adopted.

Heron notes that such copper-lead bearings were inadequate for one model of P & W radial, leading to attempts to use silver in place of copper-lead - presumably because of its greater ability to resist fatigue cracking and/or detachment from the steel backing. Silver, like copper, work-hardens strongly. Anyone who has hand-forged silver has seen it harden, then split as the extent of the cold-work increases. Possibly for this reason, silver was a tricky bearing material; if it ran-in OK, it would give good service, but then again, it might suddenly seize.

P & W solved this split-personality by lead-plating the silver, forming a forgiving break-in surface, but then the lead was attacked by certain oils, requiring that the lead in turn be plated with indium.

In the meantime, aircraft engine oil had been changing. Castor oil had been the preferred lubricant for aviation and racing engines for many years, but its limited supply and tendency to gum made it unacceptable for widespread use. One valuable feature of castor oil, however, was its polar nature, bonding strongly to metal surfaces to form a strongly rust-resisting molecular barrier. Early petroleum-based aviation and motor oils had no comparable quality, and plain bearings run with mineral oils were subject to rapid corrosion -- often producing curious meandering patterns called "hen-tracking". In time, anticorrosion additives that either neutralized acids or mimicked the polar behavior of castor oil were developed.

The use of indium as an anticorrsion plated layer for bearings was pioneered by General Motors in their plain bearing metallurgy experiments.

Heron tells of an airline's returning a con-rod big-end shell to P & W for lead/indium replating after 7000 hours of service. It was decided that this bearing could continue as it was.

Pratt & Whitney's engineer Hobbs said that the silver/lead/indium bearing's surface consisted of physically strong points of silver in a sea of lead. This is analogous to the surface of a Nikasil cylinder bore of today, which consists of small hard regions of silicon carbide, slightly proud from a surface that is mainly softer nickel.

General Motors looked for another way to make such bearings for its radial tank engines. Their manufacturing method was to surface the steel shell with silver, then knurl the silver. This knurled surface was then plated with soft babbitt and the whole bearing bored to size. The result was a grid of silver with its spaces filled with the softer material -- in effect just what Hobbs had described.

Most of us are familiar with the widespread practice of slightly abrading the surfaces of plain bearings with ScotchBrite before installation. This has its believers and its scoffers, but I have no personal experience with it. Heron notes that in the final development of P & W's silver/lead/indium bearing, the silver was sandblasted before final surfacing with lead and indium;

"The sandblasting provided a slightly relieved bearing surface which increased the load bearing capacity under very severe load conditions which would rupture (or smear) a smooth lead-indium film."

Hobbs' belief was that this process was what created the local regions of silver, whose strength would support the load (with lead present to act as a solid lube where needed) where the above-mentioned smooth film of lead and indium would not, but would be deformed enough to breach the oil film. According to Heron, when the Rolls-Royce mission saw these rough-looking bearing shells being installed in the Merlin V12s being built by Packard, they were horrified.

I wonder if there can be any possible relation between P & W's process, and the folk-lore business of abrading new shells slightly with ScotchBrite?

Heron's section on plain bearing development ends with a note that first-class plain bearings "have, in general, proved more reliable than ball and roller bearings". It is also noted that "the plain bearing usually saves weight."

In the current era, auto makers are adopting rolling-element bearings in valve trains as replacements to the formerly universal plain bearings. This is because of the peculiar nature of valve train friction; it has a friction minimum at some moderate speed, with the friction level rising above that because of increasing viscous oil shear losses, and rising below it because of the onset of boundary (as opposed to hydrodynamic) lubrication, as full oil films are largely squeezed out from between tappets and camlobes. Since any scrap of added fuel economy is gold in the car biz, and since auto engines spend most of their lives at low RPM, roller cams and tappets are likewise gold.