Railway Preservation News

roller bearing conversions
Page 1 of 2

Author:  dr dodge [ Wed Nov 07, 2007 10:56 am ]
Post subject:  roller bearing conversions

Anyone ever convert a locomotive from friction to rollers?
I assume there is a trick piece that is used instead of pressing the wheels off of the axles.
this is research for the porter overhaul, 0-4-0T
(search engines have yielded no usable info)



Author:  Wolf [ Wed Nov 07, 2007 12:48 pm ]
Post subject:  Re: roller bearing conversions

I believe the Great Smokey Mountain RR tried it on one of their locomotives. They apparently had some problems and went back to friction bearings. I have not heard any details, but one would suspect that perhaps the bearings were not properly matched to the load, or they steam clean their running gear and their bearing box design allowed water to sneek into the bearings or ???

While I suppose you could custom design some sort of split roller bearing, there are reasons that is not standard practice in industry. If you want to make the conversion, I suspect you will be stuck with having to press the wheels on & off.

Hopefully, someone with more experience in this area will chime in...

Author:  Dave [ Wed Nov 07, 2007 5:27 pm ]
Post subject:  Re: roller bearing conversions

GSMRR retrofit used roller bearings applied to axles with wheel centers off. Once the wheels were presed on again, bearings were locked on as well. The bearings were held into the boxes with a fabricated holder, which allowed the boxes to travel within the pedestal independently, unlike the "standard" steam driver roller bearing design, which held the bearings sealed in a hollow trunk cast integral with the boxes, in a fixed relationship relative to each other and the axle. Jones and Laughlin had some 23" gage Porter 0-4-0's built for hauling ingots around the mills with this arrangement, which is the only 0-4-0 roller bearing application I know of. I never heard what the critical problem was in the GSMRR experiment - but i applaud them (and the late Greg Dodd) for trying something different. I think it was probably the diagonal forces caused by the bearings floating seperately, personally - but I wasn't directly involved.


Author:  dr dodge [ Wed Nov 07, 2007 7:58 pm ]
Post subject:  Re: roller bearing conversions

I found these guys

Haven't done any calcs yet, but they make some pretty beefy split race bearings

pressing the wheels is not an option,
they need to be easily replacable

the loco has stephenson (sp?) valve gear, so no hollow tube issues
any help is appreciated, so thanks


Author:  whodom [ Wed Nov 07, 2007 10:46 pm ]
Post subject:  Re: roller bearing conversions

It would certainly be worth contacting that manufacturer and asking for their help with the application. While I've never personally run the numbers, IIRC the loads from the piston thrust are MUCH higher than the static loads the bearings support from the weight of the engine. Make sure they take this into account when sizing the bearings.

A couple of other notes regarding the problems with the GSMR roller bearing conversion: Most steam locomotives with rollers on the driving axles were built this way from new. This generally required larger pedestal openings in the frames to accommodate roller bearings of sufficient size for the loads along with the 2-piece cast bearing housings (sometimes called cannon boxes) as described by Dave. GSMR attempted to use off-the-shelf sealed freight car type roller bearings which would fit within the existing pedestal openings on their engine. I've read speculation that the bearings really weren't adequate for the loads encountered and that the sealed grease lubrication may have been part of the problem. The thought being that oil lubrication, with periodic changes of the lubricant, would have flushed contaminants from the bearings. This may be pure speculation, but it does make some sense. Like Dave said, you have to applaud them for trying.

One way around the limited pedestal openings is the bearing arrangement used on the early N&W A's. They used a "floating driving axle" arrangement with a single, very large diameter roller bearing fit into the hub of each driver. The inner race of this bearing rode on a cannon box that went from pedestal to pedestal. The bearings did not contact the axle; the axle just held the wheels in gauge and transferred the torque from one side to the other. A very interesting arrangement, but it supposedly had its own problems and would likely require new drivers to be cast to retrofit it to an existing engine.

Please keep us posted on the split bearing application attempt.

Author:  jasonsobczynski [ Thu Nov 08, 2007 1:15 am ]
Post subject:  Re: roller bearing conversions

To convert a locomotive to roller-bearing with a long term success would require either Timken or SKF to fill a custom order for steam era bearings. MODERN BEARINGS WILL NOT HOLD UP! Timken can do it, I am not sure if SKF retained the necessary documentation. For the application you speak of, the SKF spherical steam locomotive driving box bearing would be the proper selection(if available or affordable). This design allows for the boxes to not be rigidly connected versus that of timken.
I can speak with you more off line if you like.

Author:  whodom [ Thu Nov 08, 2007 9:27 am ]
Post subject:  Re: roller bearing conversions


I don't completely buy this "modern bearings won't hold up" argument. If the loads and conditions the bearings have to operate in are properly considered and the bearings are selected accordingly, they should work regardless of when they were made. Of course, the roadblock may be that no currently produced roller bearings meet all the required design parameters.

I think the main problem with GSMR's application was they did their best to adapt a modern existing railway bearing to an application it really wasn't intended for and it didn't work out.

In his book "The Red Devil", Wardale relates how he did a complete detailed design for full roller bearings on an improved Chinese QJ locomotive (which was unfortunately never built) using SKF bearings. The odd thing he ran into was the SKF's design guidance for steam locomotives (which they still had on file) recommended designing the bearings based on axle loads alone, which should have provided significantly under-designed bearings. SKF claimed they had never run into problems using this criteria, but Wardale selected the bearings based on the higher loads that result from the piston forces anyway.

Of course, China canceled this project after the detailed design for the improved QJ was 100% complete, and as a result, a full compliment of SKF roller bearings (for axles, rods, valve gear, etc.) sat unused on the Datong shop floor when he left China.

Based on this info, I'd bet SKF still makes suitable bearings for steam locomotive applications.

Author:  jasonsobczynski [ Thu Nov 08, 2007 9:37 am ]
Post subject:  Re: roller bearing conversions

The statement of "Of course, the roadblock may be that no currently produced roller bearings meet all the required design parameters" is correct. Modern bearing geometry does not allow for A: shock load from piston thrust and B: all aspects of lateral/thrust loading experienced on a locomotive.
This information is from the Timken Corporation

Author:  dr dodge [ Thu Nov 08, 2007 11:56 am ]
Post subject:  Re: roller bearing conversions

but, lets remain in perspective

this is not a massive engine dragging a long passenger train in everyday use
it is an 0-4-0 HK Porter tank and will pull only 1-2 cars ocassionally

the cooper site has some massive units made for heavy industrial use
anyone know what the original material the timkin bearings were made with? how they differ?
I just suspect that the rolling resistance will decrease massively when converted thus increasing efficiency



Author:  Sully [ Thu Nov 08, 2007 3:26 pm ]
Post subject:  Re: roller bearing conversions

I believe that in practice (read: Testing) has shown that rolling resistance is almost the same or less? for 'friction' or journal bearings in full-film hydrodynamic lubication mode. Roller bearings primary advantage is that they have lower friction during starting and very slow movement, where a plain journal bearing is in transition stage for those modes of operation. You can roughly calculate the mode of operation given the speed of your loco if you know the axle's journal dia, and your lubrication oil's properties. Check out plain journal bearings in a Fundamentals of Machine Elements book.
To me, the pros of rollers are that today's locomotives in recreational service do a lot of starting and stopping, not much long steady running.
The con, that they are less forgiving with the dirt, stresses, and the availability and price for our application is a roadblock.


Author:  Dave [ Thu Nov 08, 2007 3:31 pm ]
Post subject:  Re: roller bearing conversions

Doc, I don't think the 4 bearings on your engine will impart anywhere near the rolling resistance difference (once turned enough to bring lube up from the cellar) to be more than marginal. Probably piston drag and valve gear powering, including shoving the valve back and forth even if it is balanced, will make any potential bearing resistance marginal at best, inconsequential at most likely. What roller bearings can do is reduce maintenence associated with lubrication, and with lateral wear issues. It's more about long-term savings than daily efficiency gains.

I'd be looking at more efficient combustion engineering like GPCS and Lempor / Koopmans front ends to get more efficiency related savings if that's your goal. You might also check out what Sulzer has developed with boiler insulation and light oil firing.

BTW, I heard from a pretty knowledgable guy that I ought not to have applauded Greg Dodd for his experiment. FWIW, I can't see any point in discouraging innovation. The fact that it didn't work simply educated us all about how not to do it, which costs us nothing and might save us some trouble down the line. GSMRR might have had to devote some resources to putting it back, but that doesn't diminish the value of the lesson from my POV.


Author:  dr dodge [ Thu Nov 08, 2007 8:18 pm ]
Post subject:  Re: roller bearing conversions

in my opinion, between the ease of maintainence, frictional (=heat) losses,
it is definately worth exploring
especially if the sidegear is done, too

as far as combustion efficiency I have some even more bizzarre ideas
(alternate fuels, pressurized fireboxes,preheaters, turbochargers, and more)
small scale experiments are currently underway, but this thread is not the time or place.

when we raced dirt cars, the last one we built, we went weight and rolling resistance reduction crazy. when we were done, the car (missouri street stock) weighed 2300# and won track champion (by a bunch).
I say that because it taught me that every little gain is worth the effort to try it.
(plus then the loco would be FRA legal on a main line)


Author:  David Farlow [ Thu Nov 08, 2007 8:42 pm ]
Post subject:  Re: roller bearing conversions

It would be FRA legal with friction bearings, just like Flagg Coal #75. As for the mainline and an 0-4-0 ....

David Farlow
Whitewater Valley Railroad

Author:  whodom [ Fri Nov 09, 2007 7:51 am ]
Post subject:  Re: roller bearing conversions

dr dodge wrote:
in my opinion, between the ease of maintainence, frictional (=heat) losses,
it is definately worth exploring
especially if the sidegear is done, too

I strongly suggest you start off talking to the bearing manufacturer. They have application engineers who do nothing but figure out how to use their bearings and they should be able to tell you if they'll work and which ones to use.

Author:  dr dodge [ Mon Nov 12, 2007 4:45 am ]
Post subject:  Re: roller bearing conversions

does anyone know the specs weight wise of a fully loaded 0-4-0 tank?
length (wheel base and knuckle to knuckle)?
when you calculate the side loads on the wheels do you count the mass of the cars pulled, or just base it off the tractive effort available?
I figure its the same as calculating the seesaw thing as far as side loads based upon your worst radius/angle (which, I am guessing, is the switch frogs)



Page 1 of 2 All times are UTC - 5 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group