A question about bearings.

[Flynth]

A bearing should never be spun without a load on it, the rolling elements can skid and damage the race surfaces, no I have never done that with the Kluber greases.

This, to me, sounds like an extraordinary claim.

Where did this idea come from? I never heard about bearings requiring a minimal load. And getting damaged because of no load being spun by hand? No way. Perhaps if one tried to spin one up to very high speed with compressed air, but not by hand.

Both the rolling elements and the races are hardened steel designed to withstand much higher loads placed on the bearing during use. There is rolling and sliding of surfaces under much higher loads than those ever attainable by spinning a bearing it in one's hand...

What if the installed bearing runs an extremely low load? Let's say a very light, air driven plastic shaft with just a mirror that weights few grams at most. Will the bearing get damaged too?

The drag from the lubricant comes from it’s viscosity, and the amount of lubricant in the bearing, NBU15 is an ISO 15 grease, so very low viscosity and should have

I see. Thank you for mentioning it. rrently using an iso vg 10 oil and it is just barely ok. I think iso 15 will cause too much drag, but now knowing the viscosity of the base oil for this grease I can test with oil instead of buying an expensive grease and finding it doesn't work. Is there grease that uses iso 10 base oil? Or even less?

That is meant for high speed applications and to determine if it would be suitable for long life in your application, you would need to run the calculations. This document from FAG runs you though the calculations the application engineers do and if you have a good grasp of the loads on the bearing, you should be able to fairly accurately determine what viscosity is suitable.

https://www.schaeffler.com/remotemedien/media/_shared_media/08_media_library/01_publications/schaeffler_2/publication/downloads_18/wl_81115_4_de_en.pdf

Thank you for the link. They have a chapter on calculating the frictional moment at page 14. They only provide numbers for oil bath lubrication. Regarding greases they provide this very confusing statement:

The f0 values of freshly greased bearings resemble, in the starting phase, those of bearings with oil bath lubrication. After the grease is distributed within the bearing, half the f0 value from the table
(fig. 16) has to be assumed. Then it is as low as that obtained with oil throwaway lubrication. If the bearing is lubricated with a grease which is appropriate for the application, the frictional moment M0 is obtained mainly from the internal frictional resistance of the base oil.

There are no values given for the "oil throwaway lubrication" and there is no definition of it. Oil bath is defined as level of oil that reaches half the lowest rolling element. One can perhaps infer "oil throwaway lubrication" is basically some oil in the bearing, but no oil level and that type of lubrication is half that of oil bath (at the same viscosity).

Furthermore, they claim grease lubrication will provide half of the drag of an oil bath (providing the viscosities of the bath and grease base oil are the same). This is very interesting. On one hand it claims oil bath (only reaching half way the height of the lowest ball) will be double the drag of my bearings now (as they are used with oil throwaway lubrication). On the other if I can find or make a grease with base oil with iso vg 10 viscosity it should have exactly the same drag as it does now.

Shielded bearings are not as good as sealed, but they will work fairly well at keeping contaminates out provided it is not a harsh environment.

Unfortunately, It is a harsh environment with grinding dust and coolant everywhere...

 

[Ischgl99]

I used to work at FAG Bearings, it doesn’t take much abuse to damage the bearing races and rolling elements and any skidding is bad for the bearing and will reduce its life. While doing it by hand might not do much damage, there is enough of a risk that it should never be done. The problem with spinning it by hand is you are rapidly accelerating it without lubrication to keep the running surfaces separated.

What if the installed bearing runs an extremely low load? Let's say a very light, air driven plastic shaft with just a mirror that weights few grams at most. Will the bearing get damaged too?

Absolutely! The bearing needs to be chosen so that the loads on it keep the rolling elements from skidding. Typically the bearing needs at least 2% of the dynamic load rating to properly load the bearing.

Is there grease that uses iso 10 base oil? Or even less?

I’m not aware of any, the only ISO10 lubricant I know of are oils.

Regarding greases they provide this very confusing statement:

When a greased bearing is installed, it is packed with grease. That causes a lot of friction and resistance until the rolling elements push the excess grease out of the way. Once the grease is pushed off to the sides of the bearing, you have a thin oil film on the races that provides the lubrication. The way a grease works is the oil contained in the soap base slowly leaches out of the grease to the sides of the path of the rolling elements and provides continuous lubrication just like a throw away oil system that sends a small quantity of oil into the bearing during operation. So, you would treat grease lubrication like throw away lubrication at the viscosity of the oil in the grease since they are basically the same.

With an oil bath, since the oil is usually at about half the ball height, the rolling elements are constantly running through a pool of oil resulting in higher frictional resistance. An oil mist or throw away lubrication system only has a thin film of oil on the races/rolling elements, so the resistance is much less. Think of running in the water along the beach, if you run in the section of the beach where the wave just receded (throw away lubrication), your feet are getting wet, but you have only the resistance of the sand. If you go into the water and try running in water say 200mm deep (oil bath), you will have a lot more resistance to running since you have the sand and the water to run through. I hope that analogy makes sense.

Unfortunately, It is a harsh environment with grinding dust and coolant everywhere...

That’s too bad, some sort of seal is likely needed. Can you explain why you need low resistance for an extended spin down time? I don’t see how that would matter in a grinding application.

 

[Flynth]

I used to work at FAG Bearings, it doesn’t take much abuse to damage the bearing races and rolling elements and any skidding is bad for the bearing and will reduce its life. While doing it by hand might not do much damage, there is enough of a risk that it should never be done. The problem with spinning it by hand is you are rapidly accelerating it without lubrication to keep the running surfaces separated.

No one said anything about "without lubrication". The whole point of spinning it by hand is to check the drag caused by the lubricant used, but fine, if there is a rule at FAG Bearings not to spin them by hand when not lubricated that's understandable. However, when they are lubricated? IMO that's an overabundance of caution. If I had an expensive class 4 bearings I probably wouldn't spin them too out of fear of dropping them . A $1 deep groove bearing is another matter...

Absolutely! The bearing needs to be chosen so that the loads on it keep the rolling elements from skidding. Typically the bearing needs at least 2% of the dynamic load rating to properly load the bearing.


I’m not aware of any, the only ISO10 lubricant I know of are oils.


When a greased bearing is installed, it is packed with grease. That causes a lot of friction and resistance until the rolling elements push the excess grease out of the way. Once the grease is pushed off to the sides of the bearing, you have a thin oil film on the races that provides the lubrication. The way a grease works is the oil contained in the soap base slowly leaches out of the grease to the sides of the path of the rolling elements and provides continuous lubrication just like a throw away oil system that sends a small quantity of oil into the bearing during operation. So, you would treat grease lubrication like throw away lubrication at the viscosity of the oil in the grease since they are basically the same.

Is the only way to implement oil throwaway lubrication to basically drip oil through the bearing (of use mist), or are there some other ways? I wonder if I put a small quantity of oil in the space enclosed by my bearings (a lot smaller than the oil bath amount). The oil will tend to fling away from the bearings, but the turbulent air should redistribute it hopefully preventing lack of lubrication.

An alternative is to make my own grease, but I'd rather avoid that particular rabbit hole at the moment.

With an oil bath, since the oil is usually at about half the ball height, the rolling elements are constantly running through a pool of oil resulting in higher frictional resistance. An oil mist or throw away lubrication system only has a thin film of oil on the races/rolling elements, so the resistance is much less. Think of running in the water along the beach, if you run in the section of the beach where the wave just receded (throw away lubrication), your feet are getting wet, but you have only the resistance of the sand. If you go into the water and try running in water say 200mm deep (oil bath), you will have a lot more resistance to running since you have the sand and the water to run through. I hope that analogy makes sense.



That’s too bad, some sort of seal is likely needed. Can you explain why you need low resistance for an extended spin down time? I don’t see how that would matter in a grinding application.

Its not a grinding spindle, but the application is in a grinding environment. I can't say more about it at the moment. I'd like to finish my video about it first. Then I'll post my thread in projects etc.

 

[Ischgl99]

The only other method I can think of is oil-air which is a larger oil particle than oil mist, but a drip feed system would be easier to do. The simplest method is going to be grease lubrication, if you combine that with a labyrinth seal instead of a radial contact seal, it might do what you want with an ISO15 grease.

 

[Lo-Fi]

The point about bearings needing load is covered by the term "preload" if you want to research. Properly configured bearings need to be preloaded so that the balls or rollers are forced into contact with the races and cannot slide. This obviously introduces some friction, and, indeed, the preload is sometimes measured in torque it takes to overcome that friction. Yes, they'll work with no preload, but they won't last long.

Labyrinth seals are usually made for application, so often not available off the shelf. They're very easy to make, and can be radial, axial or both.

 

[Flynth]

The point about bearings needing load is covered by the term "preload" if you want to research. Properly configured bearings need to be preloaded so that the balls or rollers are forced into contact with the races and cannot slide. This obviously introduces some friction, and, indeed, the preload is sometimes measured in torque it takes to overcome that friction. Yes, they'll work with no preload, but they won't last long.

Labyrinth seals are usually made for application, so often not available off the shelf. They're very easy to make, and can be radial, axial or both.

Isn't preload a completely different thing? Is it not used to increase stiffness and lower runout rather than make sure rolling elements don't skid and damage races? I'm not arguing there is no rule like this... I'm just puzzled it exists. I would love to see some document that talks about bearing damage when they run with too little load (and good lubrication).

I know about preload in context of high precision bearing pairs, but to avoid damage? Let me provide one personal anecdote about this.

I've recently refurbished a grinding spindle that uses a hydrodynamic bearing in the front and a bearing pair that should've been preloaded in the back(small preload of 30N-50N). It is a very old machine. The spindle hasn't been touched in decades, but it was being used. During the rebuild I discovered someone replaced class 5 angular contact bearings in the back with tapered roller bearings. The outside race spacer that was previously setting the preload of old bearings was hitting the cage of this new bearing so they left the screw pushing two bearing outside races together loose - 0.3mm (12 thou) of clearance. It has been running like that (in iso10 oil bath) for years. I don't know how long exactly, but a long time. How do I know that? Because the front hydrodynamic bearing had the bronze bushing wear the hardened shaft 0.02mm (7 tenths in) since it was adjusted last (likely by the person who "fixed" the rear bearings). Those rear bearings that run so long with no preload have no signs of wear at all. Furthermore, in terms of runout and dimensional accuracy they measure only slightly outside class 5 precision requirements. So not wanting to spend money on the machine now I kept them, but I set preload correctly this time. If running under loaded was damaging they should've been damaged, right?

Coming back to the subject. Does anyone know (or has written materials) about grease chemistry? I know they contain base oils and soaps, but what kind of soap? I read Msds of one of those ultra low drag greases and all it said was "barium soap". Hmm, barium? Ok, one can relatively easily obtain barium hydroxide to make soap, but what fat to use?

 

[Ischgl99]

I would love to see some document that talks about bearing damage when they run with too little load (and good lubrication).

https://www.schaeffler.us/remotemedien/media/_shared_media/08_media_library/01_publications/schaeffler_2/poster/downloads_19/dws_de_en.pdf#:~:text=Discoloration%20of%20the%20rings%2C%20rolling%20elements%20and%20cages,the%20bearing%20components%20will%20undergo%20deforma-tion%20%28hot%20run%29.

Good lubrication will help prevent that, and you might get away with under loading the bearing, but in general it is poor practice since lubrication is rarely ideal all the time.

Following is an excerpt from the book Ball and Roller Bearing Theory, Design and Application.



Each section of a bearing catalog also has design guides that will state this as well, at least with the Schaeffler bearing catalog.

 

[Lo-Fi]

You won't really "see" damage from skidding rolling elements until it's really bad and tapered rollers are particularly robust. There's a reason almost every car on the road uses them for wheel bearings, along with excellent load bearing... It's something that accumulates over time and might eventually manifest in out of round, out of spec, noise, rough running and suchlike. I thought my Bridgeport spindle bearings were fine until I measured the runout. Smooth, quiet, no detectable play, but 15 thou runout! What you've got are two very hard surfaces potentially rubbing over each other, so some grinding is inevitable, depending on load, speed and application, of course.

If you really need ultra low friction, rolling elements of any kind are maybe not be best option. Is increasing the rotating mass to offset the friction an option? Do you actually need bearing accuracy and longevity?

To take a step back: My experience when problems involve heavy compromise in one aspect or another is that I'm inevitably looking at the issue the wrong way and need to rethink the requirements. Rule number 1 (and I don't mean this to sound rude, merely pragmatic): "Make your requirements less dumb". It's always good to question your requirements when you end up looking for esoteric solutions to what should be a simple problem.

My Quorn tool and cutter grinder spindle I designed showing simple labyrinth seals either end:


The single rear bearing is lightly preloaded by the end cap, hence the tiny relief just ahead of the outer race to the right. The dual main bearings are a Nachi matched pair, where the races are ground by the manufacturer to give the correct preload "back to back" with inner and out races clamped firmly together and sold as a pair. Lubrication is thin spindle oil, same as specified for the Bridgeport (I can't remember exactly what offhand). Turning as fast as you can by hand, freewheel is about 1/2 a turn.

 

[Flynth]

https://www.schaeffler.us/remotemedien/media/_shared_media/08_media_library/01_publications/schaeffler_2/poster/downloads_19/dws_de_en.pdf#:~:text=Discoloration%20of%20the%20rings%2C%20rolling%20elements%20and%20cages,the%20bearing%20components%20will%20undergo%20deforma-tion%20%28hot%20run%29.

Good lubrication will help prevent that, and you might get away with under loading the bearing, but in general it is poor practice since lubrication is rarely ideal all the time.

Following is an excerpt from the book Ball and Roller Bearing Theory, Design and Application.

View attachment 447232View attachment 447233

Each section of a bearing catalog also has design guides that will state this as well, at least with the Schaeffler bearing catalog.

Well, this thread has certainly been very educational. Thank you for those references.

You won't really "see" damage from skidding rolling elements until it's really bad and tapered rollers are particularly robust. There's a reason almost every car on the road uses them for wheel bearings, along with excellent load bearing... It's something that accumulates over time and might eventually manifest in out of round, out of spec, noise, rough running and suchlike. I thought my Bridgeport spindle bearings were fine until I measured the runout. Smooth, quiet, no detectable play, but 15 thou runout! What you've got are two very hard surfaces potentially rubbing over each other, so some grinding is inevitable, depending on load, speed and application, of course.

Interesting. I've measured the runout too. I don't remember the exact value, but it was within spec.

However it is possible it would be better when they were new. Unfortunately I haven't thought about aligning them to cancel out the tolerances on the race widths (they were about 8 microns out if I remember correctly). So when putting together I had to shoot for that very light preload when warm. This results in slight clearance when cold. And those bearings do get warm sooner than the hydrodynamic bearing in the front. So perhaps there is some invisible wear.

If you really need ultra low friction, rolling elements of any kind are maybe not be best option. Is increasing the rotating mass to offset the friction an option? Do you actually need bearing accuracy and longevity?

To take a step back: My experience when problems involve heavy compromise in one aspect or another is that I'm inevitably looking at the issue the wrong way and need to rethink the requirements. Rule number 1 (and I don't mean this to sound rude, merely pragmatic): "Make your requirements less dumb". It's always good to question your requirements when you end up looking for esoteric solutions to what should be a simple problem.

Yes, the most logical thing to do would be to either make the whole thing bigger to have more rotating mass. Or use air bearings. That's clear to me now. However, when I was designing it (big word for a pencil drawing in a notebook) I didn't even know about the "freewheeling" requirement.

When this was in concept stage I thought, "you know what I really would love a small one of these, but no one makes them small. So I built one. Then while using it I thought "it would work so much better if it could freewheel for longer". So I "resolved" that by removing one seal each of the bearings and putting in that very low viscosity oil. Now it works exactly as I want, but I wonder about bearing longevity. Is it important? Probably not so much. It is fairly easy to replace them, but I'm planning to make a video about it. Maybe a set of plans for other people. And if that's the case I'd rather bearings do last. As for precision, it doesn't have to be super precise, but there can be no slop (so runout has to be repeatable like in a grinding spindle).

What I'll do is I'll probably add an oil port to the housing tube between the bearings. I'll add a little more oil (definitely less than half a bottom ball height, but enough for the rolling elements to pick it up). Then I'll set it up so it's driven at let's say double the normal speed and I'll leave it running for two weeks. If the bearings last this test I'll probably accept it.

Otherwise I'll try the grease option.

And the third is redesigning for air bearings.

My Quorn tool and cutter grinder spindle I designed showing simple labyrinth seals either end:

View attachment 447234
The single rear bearing is lightly preloaded by the end cap, hence the tiny relief just ahead of the outer race to the right. The dual main bearings are a Nachi matched pair, where the races are ground by the manufacturer to give the correct preload "back to back" with inner and out races clamped firmly together and sold as a pair. Lubrication is thin spindle oil, same as specified for the Bridgeport (I can't remember exactly what offhand). Turning as fast as you can by hand, freewheel is about 1/2 a turn.

What are those seals made of? Did you make them yourself or are they an off the shelf item? If you made them could you say more about what clearance they are running with and any other important details, please?

 

[Lo-Fi]

It sounds like your compromise here is probably bearing longevity... If you're not using expensive bearings and they're easy to replace, it seems pretty reasonable, even if you treat them like a service item.

The labyrinth seals are essentially more of a deflector than a seal, all machined from steel for this project. Brass would work well, though. They won't do much to stop oil getting out; more to stop stuff getting in. There's no contact between the rotating and stationary, though the clearance is very fine.