The roller bearings you are referring to are very different from the modern tapered roller bearings in freight cars. Each RR roller bearing is often carrying a load of over 30,000 pounds!
These loads, and speeds, are very different from engine bearing requirements. The lubricant used in RR RB's is for the dust seals and the roller cages (now plastic to further reduce parasitic friction and the need for grease.) RR RB's do NOT want grease on the surfaces of the rollers - it is not necessary and increases friction. There are advisories in the RR industry specifically to not over-lubricate RB's.
Adding oil into a roller bearing will significantly increase friction - effectively the rollers are pumping oil around the races, much as a gear driven pump is used to pump the oil out of the crankcase. So I can understand why RB's in a crankcase, or a machine tool spindle, might have higher drag than journal bearings, and even require cooling.
But that is an entirely different environment from freight car axles.
There are piles of railroad engineering documents about the MUCH lower starting friction for RB equipped freight cars vs. plain journal bearing equipped cars, and for most of the era of this list, the at speed frictions were similar, with a slight advantage to the RB's (the RB's were oil immersed in this era). But this was all before the later Timken tapered roller bearing designs (AP) that have rolling "friction" as low as one pound per ton of freight car - at all speeds, including starting. During the steam era (mandatory STFMC content), the minimum for plain/friction bearings was about 3 pounds per ton - typically at a speed between 15 and 20 mph. Starting friction could be as high as 25 pounds per ton - 35 if it was below freezing (source is 1999 AREMA manual on railway engineering.) Today the RB friction can be under 1 pound per ton of freight car. For a loaded car, friction from wheel flange rubbing the rail head, and poor roadbed (flexing of rails and absorption of energy by the truck's suspension) can be larger than the RB's. Today aero losses are by far the biggest consumer of fuel, especially for double stacks in cross winds.
This is a totally different set of challenges from motorcycle engines and high speed tools. Ball bearings have much higher friction than tapered roller bearings, since in a ball bearing some of the contact surface is actually sliding, unlike a tapered roller bearing where ALL load bearing surfaces are in rolling, not sliding, contact (hence the term "anti-friction" bearing.) There were some early roller and spherical RB's for railroad axles, but they have faded away - the opposed tapered roller bearings (Timken's patent, licensed to other manufacturers) are far superior, since they also handle the lateral thrust loads when going through curves.
---In STMFC@..., <toddchorton@...> wrote :
Dave, all those Kawasaki KZ 900/1000/1100's & Suzuki GS 1000/1100/1150's all used roller bearing cranks and rods. The Japanese started phasing them out in the late 80's. None of the new bikes today use them. Solid bearings make more power and they are cheaper to manufacture.
In Nascar and Foumula 1 racing, money is not really a consideration, if roller bearings made more power they would use them in these applications.
If you take a sealed grease packed roller bearing there's a lot of drag from grease, rollers and seals just to turn one. None of these will free spin unless you remove all the seals and grease. A dry bearing will spin freely but it won't last long. Roller bearing do generate considerable amounts friction even if being used in an application with a light air/ oil mist. In the last 10 yrs or so ceramic balls and sometimes ceramic races are used in high speed applications where normal steel balls and races generate too much heat to survive. I'm in the machine tool industry and usually anything with over a 10,000 rpm spindle range requires ceramic balls to survive for extended periods of time. Machine tool builders use bearings in spindles to keep from having to having to add a closed loop oiling system. Most high speed spindles use a spindle chiller to keep from overheating the bearings.
Roller bearings absolutely have their place as we all know. Can you imagine going to get in your car and having to lube the wheels before you take a trip? This, In my opinion, was the main selling feature of roller bearings for freight car wheels. It eliminated a LOT of maintenance.
I'm not 100% convinced that a car with roller bearings takes less effort to start moving it but I can see a good argument for it. At speed? Well, that's another story.............
From: "devans1@... [STMFC]" <STMFC@...>
Sent: Thursday, March 23, 2017 5:15 PM
Subject: Re: [STMFC] Solid Bearings vs. Roller Bearings
"Solid bearings have less drag than roller bearings at speed."
Not true for freight cars. I can't comment on the HP difference for motorcycle engines, but the cost of a RB equipped motorcycle engine would go through the roof if all the connecting rod and crankshaft bearings were RB. I am skeptical HP goes up using journal bearings in place of RB's if the RB's could be kept DRY. But with the need for a motorcycle engine to oil lubricate the cylinder walls, I suspect it would be just about impossible to keep a motorcycle crankshaft RB dry - and oil in the RB's would increase there "friction" considerably.
Fortunately freight car wheel bearings are not inside a crankcase filled with oil and combustion products, so their RB's can remain dry.
I hate to fan this fire again, but roller bearings are properly called "anti-friction" bearings, because all load bearing surfaces are rolling contact surfaces, not sliding surfaces. (So the "friction" in the bearing is no different than the "friction" of the wheel rolling on the rail, when the wheel flange is not contacting the rail head. RB friction is LESS than the wheel friction if the wheel flange is sliding against the rail head (as happens when trucks are hunting and cars are in curves.))
Today, wheel flange friction and aerodynamic drag are bigger sources of friction/drag, compared to the RB's.
---In STMFC@..., <toddchorton@...> wrote :
Solid bearings have less drag than roller bearings at speed. The initial start up from a dead stop might favor roller bearings but the mechanical drag of metal against metal in a roller bearing causes more resistance at higher speeds,
Roller bearing are superior when it comes to minimal amounts of lubrication where as a solid bearing will quickly fail under the same conditions.
Internal combustion engines are a prefect example of this. In 70's and 80's a lot of motorcycle engines were built using roller bearings on the crankshafts and connecting rods, As the power in these engines continued to elevate to higher levels these engines went to solid bearings for cranks and rods because of mechanical drag and horsepower loss.
A somewhat educated guess, but a 1000cc 4 cylinder engine at 10,000 rpms will probably make about 10 hp over a similar engine with roller bearings.
Two cycle engines, with roller bearing cranks and rods with a slight amount of oil mixed in with the gas will live a long time. They could make more power if they used a closed loop pressurized oiling system but that presents another set of problems not for discussion on a freight car related message board.
From: "thecitrusbelt@... [STMFC]" <STMFC@...>
Sent: Thursday, March 23, 2017 3:37 PM
Subject: [STMFC] Solid Bearings vs. Roller Bearings
I found this statement in Volume 72 of the Proceedings Of The American Railway Engineering Association:
"There is no appreciable difference between solid bearings and roller bearings above 35 mph, but other mechanical considerations heavily favor the use of all roller bearings."
I believe I have seen this same statement elsewhere.
So what other mechanical considerations favor the use of all roller bearings?