Re: Roller Bearing advantage



First, full disclosure. While I am a mechanical engineer, I am not a bearing design engineer. Engineers that go into that field are making a life-time professional commitment. It is a very sophisticated field.

So I am not an expert, but the theory behind roller bearings, and the reason they are referred to as "anti-friction" bearings, is that in the ideal design there are NO sliding surfaces anywhere in the bearing. ALL loaded surfaces involve rolling "point" or "line" contact with NO slippage (just as the "ideal" line contact for a steel wheel on steel rail should not have any slippage). If there is no slippage, then there is no "friction" (but there are other losses - fortunately they are much lower than sliding friction).

I believe that the "ideal" is VERY difficult to attain in nearly all bearing applications - but I think you can come pretty close in some applications, and freight car wheel bearings may actually be one of them.

What surprised me last night going through the 1940 CBC was how "primitive" (by today's standards), the roller bearings were for the STMFC era (20 years prior to the end of the STMFC era). The SKF bearings were somewhere between spherical (ball bearing) and taper. As a result, while portions of the contact area were truly rolling, near the edges of the contact area there was slipping under load. And if there was slipping, it must be lubricated. So it looks like they immersed the roller bearings in oil.

But with all of that oil being squeezed and pushed around, the roller bearing is beginning to look like the automotive equivalent of a gear driven engine lube oil pump - but without a discharge. So a lot of work is done to move that oil around - but only as speed increases (which explains why the rolling resistance of a 1942 era roller bearing was comparable to a journal bearing at higher speeds.)

In this era I am sure there was also considerable conservatism from a business standpoint. I can imagine a tapered roller bearing engineer telling management that these bearings really DO NOT require oil immersion - and management unwilling to take what they perceived as significant financial risk and/or having to convince every railroad mechanical engineering head that lubrication was not required. I am 99% sure the engineer would lose that argument. So in the 1940 CBC the roller bearing ads focused on "Journal boxes" that would maintain lubrication of the roller bearing while keeping out the dirt.

In a quick scan of the 1940 CBC, it did seem that Timken was a lot closer to the "ideal" tapered roller bearing than the others, which is likely why they were using "Timken Bearings" as a trademark of sorts.

Fast forward 70 years to today's modern tapered roller bearing. In these bearings, machined to mirror finishes, the "line" contact areas that carry/transmit the "principle" loads (car weight and thrust loads in curves) are truly rolling along the entire contact area. So no sliding friction from the "primary" loads. But there are locations in the bearing where secondary loads involve some sliding (e.g the forces that may keep the rollers in the proper location), so some lubricant is required. The bearing designer's challenge is to minimize these forces at sliding points of contact. In the case of modern freight car bearings, the lubrication is a "lifetime" surface lubricant (I think defined as 6 years - but I am not sure), and the bearings do NOT have lubrication fittings (introducing grease would likely ruin the bearing).

In fact, grease can be viewed as a "dirt" magnet in these bearings. The real design challenge today is making seals, which DO include sliding surfaces, that last 6 years and keep dirt out of the bearing's internals.

As for modern road trucks and their roller bearing use - I really do not know why they are using oil immersion. Bearing design is very complicated, and often involves taking many other factors into consideration - including the nature of the equipment it is being installed into. But an oil filled roller bearing will have higher losses than a dry roller bearing. Whether that dry roller bearing survives in a truck is another matter.

Dave Evans

PS - for more info, Google Timken and check out their web site - they have a number of interesting pdf downloads for railroad wheel bearings.

--- In STMFC@..., "John Hagen" <sprinthag@...> wrote:


I don't see how the immersion in oil could have much of affect on increased
resistance. Oil is less viscous than grease at all temperatures so I don't
understand how they would add resistance. Trucks use oil immersed bearing
for reliability and less resistance. I always thought (when I thought about
it which is all that often) that the reason railroads don't use oil bath
rollers bearings the susceptibility of the housing/reservoir to damage from
all the crap that ends up on the right of ways.

John Hagen

--- In STMFC@... <> , "Mike
Brock" <brockm@> wrote:

Well...what the hell...everyone else seems to have commented on some
part of
this subject so why not me?

From The Steam Locomotive by Ralph Johnson, Chief Engineer of Baldwin
Locomotive Works, 1942, "Roughly it may be said , that the ratio of
resistance in starting a car equipped with roller bearings, as compared
one equipped with solid bearings is eight to one in favor of the roller
bearing". "As the speed of the train increases the difference in
resistance between the two types of bearings decreases rapidly, and
above 10
mph the difference becomes very small. In cold weather the starting
resistance of a train with solid bearings is quite high and therefore
acceleration of a train equipped with roller bearings is aided very
materially". " As a general propostion it may be stated that, by the use
roller bearings, the journal friction will be reduced 50% at starting,
approximately 10% from 5 to 35 mph and nothing above 35 mph".

Reckon he knows anything?

What do I call non roller bearings? Depends upon how many beers I have

Mike Brock

I hope my latest post does not drive you beyond a six pack....

This seems like a good starting point for 1942. What surprised me from
reviewing the car builder's cyclopedia was the immersion of roller bearings
in oil. This would no doubt significantly increase the friction of roller

There has been significant progress in reducing roller bearing friction
since then - not sure where things stood at the end of this group's era
(1960), but roller bearing friction today, at speed, is much less than the
JOURNAL bearings of old. Friction that results from trucks tracking through
curves is also reduced (by a combination of better bearings and better truck
geometries), such that the old curve compensation standard for laying out
right of way may be out of date (note - this is a theory based on some
analysis of a very complicated subject a few years ago that I am not going
to get into again - not worth my time.)

One 2003 engineering report I have from a prototype truck manufacturer
indicates that rolling resistance of modern freight car trucks on tangent
track is only 1.8 pounds per ton of car weight - without any speed
dependency. The increased car resistance with speed is now mostly
aerodynamic. Tapered roller bearings have come a long way.

Dave Evans

Join to automatically receive all group messages.