A conundrum about ball and roller bearings............ (1 Viewer)

[Spriddler]

For years I've always considered ball and roller bearings to be near the peak of mechanical bearing efficiency.
However, it has occurred to me that since the diameter of the inner race is smaller than the diameter of the outer race and one of them is always fixed i.e. only one of the races rotates, the balls or rollers must have to skid during their rotation in order to cover the different surface distances over which they have to travel. Maybe they do skid but that seems to be a very disappointing/inefficient feature.
Here's a simplistic graphic of a ball bearing which may help visualise my confusing assumption:

 

[TheTwoOfUs]

The balls are travelling at a slower speed than the shaft so scuffing doesn't occur.

 

[Stealaway]

They had major problems in the early days with big end bearing skidding, especially in large single cylinder racing motorcycles
Much research was necessary to overcome it but it no longer happens.

^{_{Subscribers  do not see these advertisements}}

 

[jockaneezer]

Years ago, I made a simple reduction drive for a model boat using an electric motor to drive the inner race, the propshaft was attached to the ball race cage by a crude spider and the outer race was fixed solid.

 

[Jake vE]

If the correct bearings are selected and more importantly loaded properly then there should be no skidding. I'm involved with vibration analysis through work and you would normally be able to see it

 

[tonyidle]

Take all the balls out except one. Keep either inner or outer stationary and rotate the other. Watch the ball - no scuffing. Any number of balls will do the same as that one.

Big end roller scuffing was caused by uncaged rollers either bunching or, more often, not tracking at right angles to the crank pin. Crowding the rollers or fitting a cage usually solved it.

^{_{Subscribers  do not see these advertisements}}

 

[68c]

It's a bit like those diagrams showing the path of a fixed point on a wheel as it rolls along. I have included the math to explain how it works.

 

[Spriddler]

The balls are travelling at a slower speed than the shaft so scuffing doesn't occur.

How can that be?
E.g. An inner race of 10 mm dia will have a circumference of 31.42mm. (2πr x 5).
Say the balls are of 5mm dia (circumference 15.71mm) then the outer race (on which balls run) must have an inner dia. of 15mm and a circumference of 47.12mm. (2πr x 7.5).
The difference between the circumference of the inner and outer race surfaces is 47.12 - 31.42 = 15.70mm.
Therefore without any skidding each ball will rotate exactly twice (31.42 divided by 15.70) to pass around the inner race.
However, the outer race has a circumference of 47.12mm which means that to pass around it without skidding the ball has to rotate 2.99 times.
It is physically impossible for the same ball to cover 31.42mm and 47.12mm without skidding since one 'side' of a ball cannot travel at a speed different from the other.


Edit - Hmm, thanks to everyone for all the information which was posted whilst I was still struggling with this post.
I'll digest it all over my lunch, much delayed due to mumbling, maths and typing at the keyboard.

 

[Jake vE]

Reading your latest musings I now see what you mean, I was thinking mostly about skidding that causes damage....There will be some sliding of the balls towards the outer race but they are sliding within the film thickness of the lubricant (< 1 micron). But then with the friction of the lubricant when the ball is sliding across it at the outer race and the friction of the lubricant as the ball counter rotates to the inner race you will lose some efficiency. I still think off the top of my head from what some of the manufacturers state you are still getting 96-99% efficiency

^{_{Subscribers  do not see these advertisements}}

 

[Spriddler]

Big end roller scuffing was caused by uncaged rollers either bunching or, more often, not tracking at right angles to the crank pin. Crowding the rollers or fitting a cage usually solved it.

Coincidentally the big end crankpin with three rows of uncaged rollers on my 1954 Matchless motorbike suffered exactly that.

 

[68c]

Not only is the ball rotating as it runs along its track, it is also rotating around the axle. Imagine the roller having a much smaller inner race, as it rotates around the centre it is not hard to envisage the face against the inner race moves slower than the on the other side of the ball. The ball is not actually rotating about its centre, rather around a point closer to the inner race when considered from a global perspective. The red dot in diagram.

 

[Spriddler]

....There will be some sliding of the balls towards the outer race but they are sliding within the film thickness of the lubricant (< 1 micron).

I've been restoring classic British motorbike engines of the 30's, 40's and 50's for myself and others for more than 45 years. Ball and roller bearing lubricant 'shearing' may be an issue. In their wisdom many newcomers to old bikes use modern multigrade oils in the belief that they 'must be better' than the originally specified 'old fashioned' monograde oils.
Multigrades are essential in high speed modern engines having shell type big end and/or main bearings with continuous high pressure lubrication. In my experience (and in discussions with various oil company engineers) multigrade oil is not ideally suited to earlier engines which typically have a max useable rpm of between at between 2,500 and 4,500 and which have total loss, splash, or very low pressure lubrication delivered in pulses from a reciprocating plunger type of oil pump to the ball/roller big end and main bearings. According to those oil engineers the long chain molecules in multigrade oils and their friction-reducing additives are prone to "shearing" when used inappropriately for highly stressed lightly lubricated or poorly designed ball and roller bearings when (if) the balls or rollers skid cause rupturing of the long chain molecules.
Well that's the theory.

^{_{Subscribers  do not see these advertisements}}

 

[Steve and Denise]

Not only is the ball rotating as it runs along its track, it is also rotating around the axle. Imagine the roller having a much smaller inner race, as it rotates around the centre it is not hard to envisage the face against the inner race moves slower than the on the other side of the ball. The ball is not actually rotating about its centre, rather around a point closer to the inner race when considered from a global perspective. The red dot in diagram.

So really speaking the outter of the ball covers more distance due to the center moving hence not slipping on the larger circumference, if I’ve got that correct ?

 

[Steve and Denise]

I've been restoring classic British motorbike engines of the 30's, 40's and 50's for myself and others for more than 45 years. Ball and roller bearing lubricant 'shearing' may be an issue. In their wisdom many newcomers to old bikes use modern multigrade oils in the belief that they 'must be better' than the originally specified 'old fashioned' monograde oils.
Multigrades are essential in high speed modern engines having shell type big end and/or main bearings with continuous high pressure lubrication. In my experience (and in discussions with various oil company engineers) multigrade oil is not ideally suited to earlier engines which typically have a max useable rpm of between at between 2,500 and 4,500 and which have total loss, splash, or very low pressure lubrication delivered in pulses from a reciprocating plunger type of oil pump to the ball/roller big end and main bearings. According to those oil engineers the long chain molecules in multigrade oils and their friction-reducing additives are prone to "shearing" when used inappropriately for highly stressed lightly lubricated or poorly designed ball and roller bearings when (if) the balls or rollers skid cause rupturing of the long chain molecules.
Well that's the theory.

The pit head one was much simpler

 

[stevec]

It's a bit like those diagrams showing the path of a fixed point on a wheel as it rolls along. I have included the math to explain how it works.

View attachment 472005

Where do I send my answers?!

^{_{Subscribers  do not see these advertisements}}

 

[Carpmart]

The balls in a ball bearing are not rotatating around a common central axis point, so it’s all academic

 

[pappajohn]

Theoretically, if there were no skidding the bearing wouldn't need any grease, the balls would rotate perfectly in the inner and outer race.

 

[Spriddler]

The balls in a ball bearing are not rotatating around a common central axis point, so it’s all academic

and a bit eccentric as well, too, also.

^{_{Subscribers  do not see these advertisements}}

 

[Steve and Denise]

and a bit eccentric as well, too, also.

Come on Spriddler try to think of a tricky one

 

[Spriddler]

Come on Spriddler try to think of a tricky one

O.K. I'll post a question in a new thread about a squashed wasp.

As you will realise, I'm lockdown bored, having finished picking out the fluff from my jacket Velcro, washed and polished the wheelie bins and been online and checked the tax, MOT and insurance status of nearby vehicles............

 

[funflair]

So if this works and it does it suggests that there would be no skidding with the balls in a ball bearing

Excuse the drawing, and I couldn't be bothered to draw any more "balls" in, the shaft turns the "balls" and the "balls" walk around the outer race turning the cage but at a different speed to the shaft, just the basis of an epicyclic gear box as in Sturmey Archer.

^{_{Subscribers  do not see these advertisements}}

 

[Jake Royd]

I've though about this for a moment.
I'm an ex automotive transmissions design and development manager so hopefully I know something.
Essentially the 'top of the ball is rotating in the direction of travel of the outer race, so it speeds up (compared to the angular velocity of the ball centre) so it matches the surface velocity of the outer race.
The 'bottom' of the ball is slowing down relative to the same centre (note, it's 'bottom' contact point is going backwards relative to the centre) , so It's surface speed matches the surface speed of the inner race.
Therefore there is no skidding.
If as a design you alter the diameter of the ball and change the diameters of the two races the same holds true.

Sounds plausible , but i don't know it's true.

The balls are an interference fit between the two races , so if there was skidding it would very quickly wear out the bearing.
Also the contact area of ball on race must have an area (ie. it's not line or point contact) otherwise the bearing could not support the radial load it's designed to withstand.

 

[funflair]

I've though about this for a moment.
I'm an ex automotive transmissions design and development manager so hopefully I know something.
Essentially the 'top of the ball is rotating in the direction of travel of the outer race, so it speeds up (compared to the angular velocity of the ball centre) so it matches the surface velocity of the outer race.
The 'bottom' of the ball is slowing down relative to the same centre (note, it's 'bottom' contact point is going backwards relative to the centre) , so It's surface speed matches the surface speed of the inner race.
Therefore there is no skidding.
If as a design you alter the diameter of the ball and change the diameters of the two races the same holds true.

Sounds plausible , but i don't know it's true.

The balls are an interference fit between the two races , so if there was skidding it would very quickly wear out the bearing.
Also the contact area of ball on race must have an area (ie. it's not line or point contact) otherwise the bearing could not support the radial load it's designed to withstand.

If you alter the diameter of the balls and the races everything will still work as you say and the only thing that will change will be the speed that the cage of the bearing travels at.

 

[Spriddler]

So if this works and it does it suggests that there would be no skidding with the balls in a ball bearing

View attachment 472067

Excuse the drawing, and I couldn't be bothered to draw any more "balls" in, the shaft turns the "balls" and the "balls" walk around the outer race turning the cage but at a different speed to the shaft, just the basis of an epicyclic gear box as in Sturmey Archer.

That seems to work but when trying to visualise it in motion my brain locks up and resets to stationary. I'll keep at it though.
It's rather like when I've tried to imagine absolutely 'nothing' (i.e. the time before the universe and the gas cloud etc. existed) since if it's something I can imagine (visualise) it it won't be 'nothing'.

I've though about this for a moment.
I'm an ex automotive transmissions design and development manager so hopefully I know something.
Essentially the 'top of the ball is rotating in the direction of travel of the outer race, so it speeds up (compared to the angular velocity of the ball centre) so it matches the surface velocity of the outer race.
The 'bottom' of the ball is slowing down relative to the same centre (note, it's 'bottom' contact point is going backwards relative to the centre) , so It's surface speed matches the surface speed of the inner race.
Therefore there is no skidding.
If as a design you alter the diameter of the ball and change the diameters of the two races the same holds true.

Sounds plausible , but i don't know it's true.

The balls are an interference fit between the two races , so if there was skidding it would very quickly wear out the bearing.
Also the contact area of ball on race must have an area (ie. it's not line or point contact) otherwise the bearing could not support the radial load it's designed to withstand.

I think I've got my head around the rationale but it still puzzles me how the surface of the same ball can move at two different speeds at the same time..............
I'll ponder it more deeply whilst I'm preparing my veg for dinner. (I seem to remember reading that those were the times when Leonardo had his inspirations ).

^{_{Subscribers  do not see these advertisements}}

 

[tonyidle]

How can that be?
E.g. An inner race of 10 mm dia will have a circumference of 31.42mm. (2πr x 5).
Say the balls are of 5mm dia (circumference 15.71mm) then the outer race (on which balls run) must have an inner dia. of 15mm and a circumference of 47.12mm. (2πr x 7.5).
The difference between the circumference of the inner and outer race surfaces is 47.12 - 31.42 = 15.70mm.
Therefore without any skidding each ball will rotate exactly twice (31.42 divided by 15.70) to pass around the inner race.
However, the outer race has a circumference of 47.12mm which means that to pass around it without skidding the ball has to rotate 2.99 times.
It is physically impossible for the same ball to cover 31.42mm and 47.12mm without skidding since one 'side' of a ball cannot travel at a speed different from the other.

It's a good job nobody mentioned that to the inventor of the epicyclic gearbox. There really isn't any slipping or scuffing.

 

[bigtwin]

This thread is the best forum example of the difference between the nature of what an Engineer does and what people who do things to/with pieces of metal do!

Thanks Spriddler.

Ian