Robert, on the topic of temperature monitoring of babbitt bearings that I would agree with you is far cheaper and wouldn't take much to actually do. Are any groups out there doing any sort of real time hotbox detection such as you speak? It would actually be pretty neat to devise a little system like this, better yet make every sensor wirelessly transmit their temperatures to a central panel perhaps even an IPAD? I am not quite sure how historically accurate say having a turn of the century train instrumented with little arduinos and a IPAD to display real time bearing health on all bearings would actually be but it would be rather cool if someone's is actually doing this. I have no idea what the savings would be but I could see it making sense especially for groups like the Durango and Silverton, Cumbres and Toltec who have long rides though the middle of no where where help could be hard to come by and keeping the line open is a huge requirement. Are these groups already incorporating such technology?

As far as the roller bearing vibration based health monitoring was this comment meant to be a joke, or are railroads today actually taking advantage of vibration monitoring of bearings to watch for impending failures? Some how I doubt that when Timken outfitted the 4Aces to go on the road and sell their roller bearings they had 2012 state of the art vibration monitoring systems so I tend to think that would be a little overkill for this hypothetical project?

In my experience doing this for a living on helicopters you need to actively watch and trend the vibrations of each bearing over a period of time to look for issues. I could see it being done on a locomotive but how do the railroads handle monitoring the health of the bearings in every freight car that is currently in service on the line? I highly doubt they have a system that actively monitors and trends vibration on every freight car so what do they do out there or are bearings just a life limited part to hours of service or miles traveled?

Has vibration monitoring worked its way into rail cars and or even the locomotives? I know GE is really advanced in aircraft bearing and health monitoring so I doubt they should have any lack of ability to transfer that knowledge in order to create a smart locomotive that tells the operators when and what needs to be repaired hundreds of hours before the engine must go in for service. I could easily see major savings from having real time condition monitoring on modern locomotives. Imagine a locomotive that calls up the headquarters automatically and via software says "I need XYZ bearing removed sometime with in the next 1-2months." I envision how lots of time spent inspecting perfectly good parts could be skipped while at the same time maintenance could be moved to scheduled and reliable intervals. I realize this site being mostly being based on the discussion of historic preservation the knowledge may be limited but if anyone works for the modern class 1's or GE/EMD I am really curious.

As far as if one were to outfit roller bearings to a steam locomotive the better thing to do if space allows is to over design the bearings and put in as large of a margin of safety to get to get as close to the infinite life of the bearings. Unlike the helicopters I deal with with under designed bearings to save weight you don't care of weight on a locomotive, only so much as space to squeeze it into. Crankshaft type loads are pretty well understood in the world of bearing design so I would think the design of something with a good enough margin of safety to toss it into the next major overhaul period where you will be dropping the drivers anyhow should allow enough life to check the bearings and forget about them until then.

My guess is much of what the original poster would like to know has already been heavily documented and written about by Timken back in the 1930's both when they tried to sell roller bearings on new design to the railroads. I know I read elsewhere once that they were retrofitting roller bearings in a lot of places too so a lot of the research, knowledge may already be other there.

If anyone on the other hand has a serious need for either vibration and or active temperature monitoring of historic machinery let me know I do this for a living so depending on the need it could be fun to get to do a little side job integrating it with the older machinery that I really love.

The dynamics of the loadings experienced on a long rigid wheel base, coupled with rather impressive impact loadings, are greatly different from those incurred by a truck that swivels.



First, you sound to be assuming that no matter the engineering such a modernization will fail....that's to bad. The engineering methodology behind applying rollers to an existent plain bearing locomotive does exist, it comes down to utilizing paid professionals possessing said information whom are capable of enacting such a modernization. Having said that, by determining what is required for conversion of a particular locomotive it may well be determined that locomotive x,y or z may not be able to be converted for one dimensionally limiting factor or another.

20 years down the road the bearings originally designed for steam locomotive use back in the 30's, which are still made today per original design for alternative uses, will likely still be available..... but why would we need them? Properly maintained they will last notably longer.

Roller bearing failures are generally detected, PREDICTED, via having oil samples analyzed. I get what you are saying about all of the sensors...but.....

Just pointing out a few things.....

Cheers, Jason

This has morphed into some interesting trains of thought in the past couple days......

Onboard bearing temp monitoring has been done on at least one mainline steamer. I'm unaware of any with a constant feed of vibration sensor feedback......which doesn't mean it hasn't been done. It isn't hard to manually monitor a roller bearing in a machine that sits in one place, but to do so on a locomotive at speed would require something bolted in to the bearing housing, perhaps borrowing something from the world of helicopters as an experiment would be a nice thing to try. We'd no doubt learn a lot about what those bearings actually go through in service that may simply be theoretical now.

From the perspective of the original question, I think he's no longer contemplating a conversion of something old school with rollers so let's assume that any theorizing about it now is more academic than practical. Horse dead, no beating required.

dave

The BNSF has wayside bearing failure detectors, apparently a failing roller bearing has a distinctive sound that is easy to detect.



-Hudson

With journal (plain) bearings, would it be possible to force oil in at high pressure to keep the metal separated even when stopped? I realize it would require better seals. The oil coming out from under would not be under pressure at that point but it would be a lot to contain it. It would have to be recycled and filtered---a complex system. Maybe the answer is yes but too little gain.

Steve

It would require the bearing box to be sealed....which is not SOP. It operates at a total loss system, where used lube gets flung off and new lube sticks onto the bottom of the axle from either grease packing or an oil soaked waste in the celler and the rotoation caries it up into the journal. I have seen the top reservoir of the boxes (sort of a secondary source for the bearing and primary source for the shoe and wedge lube) fitted with a top and mechanical lubricator fed, but not pressurized.

So, this is interesting......can a robust seal be developed that would not only maintain oil inside but keep crap outside of a steam locomotive driver box? They do jump around a bit as well as slide back and forth and move diagonally....forcing lube down through the top is not difficult - just thread a teminal check into the top of the box crown brass feed. I can see a gravity fed cellar drain leading to an undermounted reservoir but you hit one high rock in the ballast......

Bob's a pretty original engineering thinker. I'd be interested in his take on seal design.

dave

Idea found cooking dinner tonight.....what if......we fabricate a sort of cannon box like split tube oil sump that seals against both inside surfaces of left and right boxes on each axle? We could possibly also install impellers on the axle that would force oil from the sump up and to the top feed of the crown brass as the wheels rotate...... from whence it would flow back down to the sump again...except I can't figure out how to deal with damn hub thrust bearing outside surface seals.

Well, maybe it's a start for some more visualization. This would help us halve the quantity of lost lube, though if something like it can work, and help limit the flow of crud in from the inside. More lube flow and less crud in seems like a good thing to me.

dave

None I'm aware of. But the cost of this technology is in freefall... both in terms of the silicon bits and the learning curve needed to develop the software. I mean, Adruino is for children!


Yes, wireless is the key to dealing with bearings that are thrashing around somewhere you can't run a cable, e.g. rod bearings.


Well, vibration monitoring is becoming more achievable at an exponential rate, because of the increased processing power of computers, the widespread publishing of pre-existing open-source software, and the proliferation of technologies like Arduino. It would still be a virtual "manhattan project" of code, though infinitely reusable on anything that rotates or reciprocates.


No, I'm assuming that it's likely to fail. And paid professionals are not even slightly magic. Paying someone does not make them smart or competent, and the steam locomotive business is pockmarked with proof of that.


Don't confuse current practice with ironclad protection, or even with what is achievable or modern. There are a lot of people out there stuck in their ways who get very butthurt about anything modern. Obviously, they prioritize "entrenching in the old ways" over "how best to make my locomotive reliable." Why should we care what these naysayers think?

At least 3 people have alluded to the notion that a roller bearing will put out signs of failure months before it fails. That may be so for certain failures. However I believe there are a great many other failures which are only detectable minutes in advance. The railroads had many incidents of a roller bearing heating, seizing, and dumping the train on the ground IN BETWEEN hotbox detectors.


Sure you could do that, internal-combustion engines already do that, just modify the journal so it has an oil passage connecting to the top of the bearing. Have a small electric motor that delivers low volume at high enough pressure to float the car on the journal. Mount the motor internal to the journal box, so the oil never leaves the space it's already in, so no new seals are needed. Outside the box you have an Arduino monitoring current flow to the motor to detect failures or dry oil (overspeed motor). However, I do not quite see the value in doing all that.

When it comes to monitoring temperature giving this one a little more thought it could be done and very cool for that matter. Say you have the Durango and Silverton with 3 trains a day. If you made provisions to measure temperature of each plain bearing on every car, have the cars communicate wirelessly with a host computer on the back of the locomotive. This computer would be connected to a cellphone link. Temperature gets sampled at each bearing 1 time per minute on the trip, compared to ambient air temperature. The system would have a hotbox detection mode built to actually alarm the crew, but otherwise when the train gets into the station the data file would be sent out over a cellphone and offer a trend plot of all the bearing temperatures. Assuming that a hot box develops over a long period of time say 20hours or so you should easily see where you are having bearing problems. After observing how a few faulty bearings rise in temperature you could automate this entirely the train arrives in the station and the computer would check the health of all bearings on the train. Should one need extra care, it would be able to divert that car to overnight maintenance where the bearing could be worked on prior to the car entering service the next day. Now the key question is would this actually be worth it and how does an organization like the Cumbres and Toltec or Durango and Silverton who have respectable distance train rides for extended periods of time actually deal with the threat of hot boxes. As much as they want to recreate the scenes of yesterday for their paying customers, I doubt that a hotbox is one of those experiences that they would really need to recreate?

My suspicion is that they are already treating such an issue through more diligent lubrication and regular maintenance of the pads, such that this is likely no more an issue? Is this right? What really caused the hot boxes back in the day? Was it more an issue for train cars that would say sit out in the open and not get oil on a regular basis or is it still a random occurrence that happens even on cars seeing regular lubrication. If they are having trouble with hot boxes a system like the one described above would be really cool, but I bet unless someone wanted to volunteer their time to set up the above it wouldn't be cheap, probably a few hundred grand as described above, for the full blown automated system.


Now to vibration monitoring if you want something relatively shake the box I would investigate something along the following; http://www.ni.com/mcm/ Maybe there are some sudden failures that occur in bearings especially if something else breaks to alter the load path but for the most part they are a long term deterioration that leads to a slow and predictable failure, especially if the whole system is properly designed. If you really cared enough to monitor bearing health on a locomotive I thought they were all sealed greased bearings so monitoring oil quality would be difficult. Vibration would be the way to go and a system like the National Instruments ones would come with enough precanned industry accepted algorithms to get the job done. Any engineering firm you contracted to design or help specify the bearings should be able to set you up with the right people to configure this system. The real question would still be if this was even warranted to begin with. The Timken 1111 locomotive was able to travel over 100,000 maintenance free hours back in the 1930's with roller bearings and the super power generation of steam had enough experience running roller bearings previbration monitoring that I just can't see it being much of either a concern in a properly designed system or something to have. Where the vibration stuff really helps a ton is when you have a whole fleet of equipment that is more or less identical to compare trends across. Then when your first locomotive fails a particular bearing you look at the past vibration energies in the weeks leading up to the failure, and develop an algorithm that looks for that in the future elsewhere in the fleet. So for a one off conversion you would likely either have a few false alarms that would create expensive tear downs or you would catch a few failures after the horse left the barn. All and all I just don't see it as needed. Fun to discuss, temperature monitoring of plain bearings is something I could see as practical, vibration monitoring though very feasible and mature technology in other areas would have marginal benefit here and you would be best putting your money into a better bearing design to ensure that the bearings had large enough factors of safety built in to get you to the next major overhaul.

I'm bumping this very old thread after finding it in a search. I think I have some thoughts to contribute, being as I am a Carman at the Durango & Silverton.

We seem to have themed seasons regarding recurring problems with the rolling stock. In 2023 it was broken leaf springs. This season has been hotboxes. I suspect the reason may be our emphasis during Annuals, last season we replaced a lot of springs, this past winter we repoured a whole bunch of bearing brasses. Some of the hotboxes this season have been tinning failures -- the pour wasn't done quite right (theres a knack to pouring bearings). However, that only accounts for a few of them. Running 140+-year-old cars with plain bearings 90 miles every day, a few hotboxes are inevitable.

The idea of real-time temperature sensing is an interesting one. However, it seems pretty costly. For about the same cost, we can convert to roller bearings, and are gradually doing so. We have developed an adaptor that will allow us to convert plain bearings to a standard roller. We just bolt on new pedestals, set the adaptor on a standard roller and wedge, and lower the pedestal onto the adaptor. Why spend money on developing a brand-new idea when converting to a mature, proven technology costs about the same?

And the thing is, plain bearings DO have indicators of their temperature. If the box is smoking, it's too hot! ;-) Yes, the temperature chart over time might be interesting data, but it's really not necessary. Taking the bearing apart and reading the brass and journal condition may not tell you exactly WHEN it got hot, but you don't really need that information to fix the problem.

Finally, speaking as one of the poor sots who has to oil the trains... rollers on the D&S can't come too soon! Rear seals leak no matter what. Box covers either get stiff and hard to open, or the spring goes bad and they won't stay open. Oil cans develop leaky spouts. And the overhang of narrow gauge equipment makes me routinely glad for my hard hat, especially with the passenger gons that are standard gauge width and the ridiculously low 3000-series boxcars.

I love working on a railroad that feels a lot like a 75-year time warp. But that doesn't mean I'm against embracing new tech! There is a good reason K valves and plain bearings are obsolete. As one of the guys who maintains that stuff, AB valves and roller bearings are VERY attractive!

Dave, I'm not familiar with some of your terminology but you seem to be barking up the wrong tree with tapered roller bearings held rigidly in a tube. Spherical roller bearings would allow the needed angular movement. They are attached by a tapered mount. The biggest issue that I can see is that they would require a larger box than tapered roller bearings.
They stand up to shock loads well. At railroad speeds I would expect a long life. They are not cheap. Installation must be done per manufacturers instructions.

https://maedlernorthamerica.com/product ... QYsAAlR-CR

I wouldn't think that converting to rollers would be that difficult provided your cars had open pedestals and that bearings, adaptors and wheel sets are available in the sizes you need. Don't forget that back when things were transitioning from brass (actually bronze) to rollers, there were kits to modify journal boxes to accept rollers. With some ingenuity, one could recreate something similar and retain the look of plain bearing boxes.