Bearing End Play: Reasons for No Concern

[brewnog]

Can anyone think of any reasons why I should not be concerned about a seemingly large amount of end play in a bearing?

The bearing is a twin taper roller bearing with a split inner race. It's carrying a mild, pretty steady axial load, and a moderate, pretty steady radial load, at a moderate, steady speed. The end float is small and nothing for concern, but I'm a bit worried about the amount of end play, - you can see the attached wheel wobbling when you grasp and 'rock' it from side to side. Surely this is causing edge loading on the rollers? Any situations where it might be expected, or no cause for concern?

 

[Danger]

You're the engineer, not me, but I'd be concerned about anything that wasn't within the design specs. If there's more end play than there's supposed to be, something's wrong. I also don't like the idea of a wheel 'wobbling', regardless of what its purpose is. That has to be feeding excess load forces back into the bearing. Seems like a degenerative feedback loop to me.

 

[Michael_E]

Tapered roller bearings are designed to operate with a specific preload that keeps them from having any end play at all. If you do have end play, then you are correct in that the rollers will end up deteriorating much faster than their design life would indicate, usually on their large diameter where the loads and speeds are greatest. Doesn't the assembly have some method to take-up and preload the bearings ?

Michael E.

 

[Shawnzyoo]

if you can i suggest taking a vibration reading of the bearing assembly in the axial direction
you can then accurately gauge what is failing within the bearing
be it a race, cage, or ball itself
is it a drive end or off end?
direct drive or attached to a sheave of sorts?

 

[brewnog]

Thanks for the insight boys and girls.

I was wondering whether the clearance was to allow an oil film to be established when moving, hydrodynamic lubrication stylee.

Michael - yeah, there's quite a neat method of preload. The inner race is in two self-opposing halves, with a clearance between them. The inner race is loaded by an end plate, which pushes the two halves together, eliminating the clearance, which sets up the preloading condition. However, all our measurements show that the cast mounting is to drawing, and that the clearance is as it should be. It's a puzzle.

Shawn, I might give the vibration measurements a go, how would different failures manifest themselves in axial measurements? This play is present on brand new components, it's not a failed bearing as such. It's an off end, as you put it, and direct.

 

[FredGarvin]

Thanks for the insight boys and girls.
I was wondering whether the clearance was to allow an oil film to be established when moving, hydrodynamic lubrication stylee.
Michael - yeah, there's quite a neat method of preload. The inner race is in two self-opposing halves, with a clearance between them. The inner race is loaded by an end plate, which pushes the two halves together, eliminating the clearance, which sets up the preloading condition. However, all our measurements show that the cast mounting is to drawing, and that the clearance is as it should be. It's a puzzle.
Shawn, I might give the vibration measurements a go, how would different failures manifest themselves in axial measurements? This play is present on brand new components, it's not a failed bearing as such. It's an off end, as you put it, and direct.

Brews,
You're using a lot of "relative" terms there...My idea of low speed is probably not the same as yours. A large amount of endplay (whatever that amount is) could be a design requirement. How much endplay are we talking here? .005", .010", .100"? Oh crud, you're in metric...;-)

Personally I don't like throwing ideas out about bearing configurations without knowing all of the parameters. Does the bearing have to experience a larger thermal gradient by chance?

I would check to make sure you have the right tolerance grade bearing. The tolerances will open up for cheaper grades. Also, is the bearing new or being reused? Was it inspected prior to use? Have you inspected it yet? You've checked the housing, but how about the shaft fits? Also, there is ALWAYS the possibility that the print is wrong. It wouldn't be the first time, especially if it is a prototype or low production run component.

The problem with vibration analyzing is that you have to have an idea as to what you are looking for. What good will knowing that you are seeing at 2 ips at ball pass frequency unless you have some kind of baseline to compare it to? None (unless you have a good feel for things in this area). Once you have that base line, you can pick up quite a lot of problems, especially in bearings.

One thing you could look for, which sometimes manifests itself (more for spherical roller bearings) is the presence of sidebands about ball pass frequency when the bearing is getting ready to fail. Here's a pretty good article about looking at the vibration analysis side of a bearing failure:

http://www.maintenanceworld.com/Articles/DLIEngineering/BearingWearExample1.pdf

Make sure you check out the easy things first. Make sure you have the right bearing set. Make sure it's installed properly (i.e. preloaded properly) and make sure someone didn't pull a bonehead mistake anywhere else. Depending on what's going on, it may be easiest to just get a new one, install it and see if the problem goes away.

 

[brewnog]

Brews,
You're using a lot of "relative" terms there...

Ha, I knew someone would say that, and I was almost certain it'd be you! Cheers anyway lad!

I can't give too many specifics, but the speeds are 1500-2000 rpm. I haven't measured the endplay myself, but it's somewhere between 0.1 and 0.25mm (depending on which sample is used). In an identical application on a different component (different bearing arrangement), there's no noticable end play at all. There's no thermal gradient at all. This isn't just on one set of bearings, it's on lots, across different batches. The tolerance is as it should be, the bearings are brand new, and it's being seen on lots of samples, it's not just one batch. I just can't seem to find any justification (in terms of design intent) why there should be any play, that's what's bugging me, and I just want to be able to rule it out as a potential cause of failure. Anyway, it's on hold for the time being...!

 

[Shawnzyoo]

be warned i am a senior in mech. engineering
and i have been interning for 9 months now doing vibration analysis on rotating machinery
so you were warned...student and intern
but from what i have picked up most of the rolling element bearings i have seen how up at the following frequencies with sidebands sometimes
Inner Race >4x motor
Outer Race around 3x motor
Ball Spin around 2x motor
Cage .3 to .4 x motor
and we classify bearing alert levels around .005 ips depending on the machine size and design.

are they pressure fitted bearings?
hope that could be of some help

 

[Danger]

Sorry to complicate the issue here, Brewski, but I don't quite understand your description of the set-up. The term 'twin taper' is new to me. Does this mean that there are 2 sets of rollers at different orientations, or that the rollers themselves are tapered as well as not being parallel, or...?
Also, I've been assuming that your relating of the split race means that it's in 2 semi-troughs like a pea pod (but circular), but then I got to wondering if you meant split longitudinally like a crank bushing.
In any event, what crossed my mind was the possibility that there might be some mismatch between the 2 halves, or between one or both of the halves and the rollers. If that were the case, your preload might not be having the desired effect even if it's properly calibrated. A long-shot, I know, but... possible?

 

[brewnog]

Sorry to complicate the issue here, Brewski, but I don't quite understand your description of the set-up. The term 'twin taper' is new to me. Does this mean that there are 2 sets of rollers at different orientations, or that the rollers themselves are tapered as well as not being parallel, or...?

Two sets of tapered rollers, sitting at an angle to the axis. The sets face each other in opposing directions. The outer race is one ring, the inner race is two separate rings, which are pushed together in order to close the gap and set up the preload.

There's no mismatch between the two halves, they're part of one bearing unit. As far as I can tell, this play is meant to be there, and it's not overly worrying anyone, but I want to know what purpose it serves, and if it really is supposed to be there!

 

[Michael_E]

Danger,
Here's a typical picture of a split inner race tapered roller bearing from Timken's site. The inner races are tightened with a preload in order to remove all play from this assembly. If the assembly will receive significant thermal expansion, then the lubricant will need to be of a particular type, (oil and not grease), and in the case of a high speed spindle the lubricant will need to be circulated and cooled by some means. Generally the preload is such that moderate thermal expansion just increases the preload slightly.

In this case the problem seems to be that the assembly has no preload and I would suspect that the end plate mentioned as a method of preloading the assembly is missing a part. These assemblies often use a wave like washer, belleville spring or even just a flat precision washer to give the proper preload. If it's missing, then you get axial play and "wobble" which will cause premature failure. The failure shows up as initial pitting at the most highly loaded spots followed by rapid failure as the vibrations become more severe and the bearing finally gives up the ghost and seizes or leaves entirely. This can be spectacular if it happens in the vehicle that you're riding in, ( I've had this happen and it sure does stop you fast !)

 

[Danger]

Okay, I've got it now. Thanks muchly for the picture, Michael; it clarified everything. And yeah, a buddy of mine burned out a wheel bearing several years ago. It heated up so badly that the spindle sheared off.

 

[FredGarvin]

Two sets of tapered rollers, sitting at an angle to the axis. The sets face each other in opposing directions. The outer race is one ring, the inner race is two separate rings, which are pushed together in order to close the gap and set up the preload.

I can't say that I have ever seen a bearing that uses the inner race to set up the preload like you are mentioning. I have to agree with Michael and say that it sounds first and foremost that there is nothing keeping this entire assembly preloaded when there are no operational loads imposed. Is there any way you can give us some more details on the actual installation?

Ha, I knew someone would say that, and I was almost certain it'd be you! Cheers anyway lad!

You're a lot smarter than Danger says you are, you know! :tongue2:

 

[brewnog]

Michael, that is exactly like the bearing I'm talking about, thanks!

The problem is, (and I'm looking into this further now!), the inner races are being pressed up against each other, and the play still seems to be present. There's no thermal element to this problem.

Might post more when I find out more, but are we agreed that there isn't a widely known reason why this play is intentionally there?

 

[Danger]

I'm agreed, but that's just 'cause we're buds. I don't really know what I'm talking about.

 

[FredGarvin]

I want a cross section...pronto.

 

[Danger]

Just one more wee question: why on Earth would anybody bother making something like that anyhow? What's the point of having tapered bearings if the shaft that you're rotating is straight?

 

[FredGarvin]

Just one more wee question: why on Earth would anybody bother making something like that anyhow? What's the point of having tapered bearings if the shaft that you're rotating is straight?

In a nutshell, they are a great comprimise for carrying a large radial and axial load in one bearing.

Straightness of a shaft relates to vibration, which will lead to increased loads, but the straightness is relatively small in comparrison to the operating loads some installations will see due to other forces present.

Really, all shafts are "straight" but not perfectly so.

 

[Danger]

I should have been a bit more specific in the wording of my question. While I realize that tapered rollers in general are used for combined thrust/rotary applications, having two opposing ones seems weird. If there's a preload, then why not just one angle of taper? Two would suggest to me that the shaft is expected to oscillate longitudinally.
Since Brewnog said that there was only mild axial force, I figured that ball or straight rollers would be sufficient.
The other application that I know of for them is when the shaft is conical rather than cylindrical, which is what I meant by 'straight'. I wasn't referring to the 'trueness' of the shaft. Sorry for the confusion.

 

[brewnog]

I was wondering why they'd bothered to use tapers too. The same part on a much beefier engine has one set of ball bearings and one set of rollers, but they're spread over a larger distance so there's no play at all.


I did you a lovely cross section Fred. You'd never guess I've been sat in front of Pro E for the last 5 hours would you?!

http://personalpages.manchester.ac.uk/student/E.Smith-2/bearing.jpg

Arrows are where you squish the inner races together by a ring or something. Note that it's not just the races that are tapered, the rollers are too. It's pretty neat really, I'd not seen it before.

 

[Danger]

Why'd you bother with Engineering, Brewski? You should have had a scholarship to an art college. :tongue:

 

[Danger]

Hey, guys... a buddy here at my little house party just added a perspective that I hadn't thought of. Is there a chance that lubricant pressure or viscosity has an effect?

 

[FredGarvin]

I was wondering why they'd bothered to use tapers too. The same part on a much beefier engine has one set of ball bearings and one set of rollers, but they're spread over a larger distance so there's no play at all.

Where exactly is this bearing located? The other application could possibly see less axial loads (and in one direction) so it can get away with using an angular contact and a roller configuration. Who knows. I have gotten myself into snags before by assuming that a scaled up or larger version of something should have the same bearing configuration.

I did you a lovely cross section Fred. You'd never guess I've been sat in front of Pro E for the last 5 hours would you?!

And they say that the art of engineering illustration is a dead one...When I saw that I wept. Well, actually I laughed pretty hard (coffe may have come out my nose).

Arrows are where you squish the inner races together by a ring or something. Note that it's not just the races that are tapered, the rollers are too. It's pretty neat really, I'd not seen it before.

You should have a spring or some kind of device to keep that left side loaded. That would be my first concern. I think I am just going to have to hop on a plane and come out there. I'll also run it by my bearing guru on Monday to see if he has anything to say about it.

 

[FredGarvin]

Hey, guys... a buddy here at my little house party just added a perspective that I hadn't thought of. Is there a chance that lubricant pressure or viscosity has an effect?

Usually those aspects will manifest themselves in heat build up and decresed life in a rolling element bearing. They really should not have anything to do with the mechanical fits. Especially since the problem is checked statically with no power or active lubrication happening.

 

[Danger]

Actually, Fred, it's because the thing is checked at rest that we wondered about it. Our thinking is that if normal operating oil pressure and viscosity are high enough, that fluid might effectively eliminate the end play when the thing is running. Once again, we're grasping at straws here. (Well... he was grasping a straw; I was drinking from the bottle.)

 

[FredGarvin]

Actually, Fred, it's because the thing is checked at rest that we wondered about it. Our thinking is that if normal operating oil pressure and viscosity are high enough, that fluid might effectively eliminate the end play when the thing is running. Once again, we're grasping at straws here. (Well... he was grasping a straw; I was drinking from the bottle.)

Ahhh. I see what you're getting at. With rolling element bearings, in my experience, you don't rely on hydrodynamic effects like you do in a plain bearing. The reason being is that you do not, under any circumstances, want the rolling elements to slip along the perimeter at any time. An excess of lubricant will cause that. The lubrication is only there to remove heat from the bearing, not to contribute to the load support.

I'm not saying that your idea is not possible, but I can't say I have ever heard that in these kinds of bearings. I don't have a whole lot of experience in tapered roller bearings because their speed ranges are too low, so I may be learning something here as well.

 

[brewnog]

I did think about the possibility of a hydrodynamic effect being to blame, I think I mentioned it in an earlier post. However, the bearing is essentially a sealed unit, and has minimal grease in it. I don't know enough about tribology or, urrm, lubrication to know what kind of conditions grease would break down into a thin oil film under, but isn't high-temperature grease supposed to stop this happening anyway? Thanks for that insight anyway Fred, didn't realize that lubricant was only for heat transfer purposes in roller bearings.

The two applications I mentioned are identical, one's just a bit bigger and had (out of tradition) been much more beefily designed, - and hence has two widely-spaced sets of bearings. The one I have the query about has been designed down to a tight budget, we think. Interesting about the scaling-up assumptions though, I'll be sure to be a bit more careful when thinking about that.

Heh, the diagram was done at home, and I'd been sat on ProE all day (I've moved to Design now for a few months, woop!). There's no spring or device to maintain compression of the inner races, just a ring pressed up against it with a couple of capscrews holding the ring into the shaft.

 

[Danger]

I did think about the possibility of a hydrodynamic effect being to blame, I think I mentioned it in an earlier post.

So you did, Mr. Nog; I'd forgotten about that because I was thinking a bit differently about it than you were. Rather than an established film between the faces of the rollers and the races, I envisioned a pressurization that would dampen longitudinal movement of the rollers like little hydraulic struts. As I said, pretty far-fetched.
Thanks for all of the info, Fred. I never knew that about bearings.