Working principle of a ball bearing

[Weber308]

Hi, I am currently a jnr,mechanical engineer in training. I started 3 weeks ago. My employer has given me a challenge. He asked me to explain to him how a ball bearing works. He said to me that mathematically a ball bearing is not suppose to work. I have done a lot of research on the internet and cannot find any ball bearing design links concerning the calculation of the speed of the ball itself with respect to the inner race and outer race. I understand that the design of ball bearings are a specialized field and my text books and scope of learning only covers plain bearings.

Could anyone pls help me in understanding how the basic working principle of a ball bearing works. Wikipedia is of no help. Am I maybe over thinking this or is there really a mathematical problem in how bearings work.

Your help will be greatly appreciated.

 

[DRWeig]

I'm not a bearing guy, but did you find this? Principles and Use of Ball and Roller Bearings



Best to you,

Goober Dave

Haven't see the forum policies? Do so now: Forum Policies

 

[electricpete]

He asked me to explain to him how a ball bearing works

too broad a question. Narrow down what aspect you're interested in. Else do a google search to find out everything about how a ball bearing works.

He said to me that mathematically a ball bearing is not suppose to work

In what sense? Please elaborate

I have done a lot of research on the internet and cannot find any ball bearing design links concerning the calculation of the speed of the ball itself with respect to the inner race and outer race.

Page 1 here gives frequency of balls passing by a point on inner ring (bpfi) or a given point on the outer ring (bpfo) for rotating inner ring configuration under assumption no sliding (kinematic frequencies). These are "defect frequencies" useful in vibration monitoring

http://www.ntnamericas.com/en/websi...e/tech-sheets-and-supplements/frequencies.pdf

Derivation is not given but not too hard to figure out. If you get stuck, see if you can solve the simpler case of contact angle =0.





=====================================
(2B)+(2B)' ?

 

[Skogsgurra]

Your boss is smarter than most bosses.

He is right, in a way. The ball bearing has a problem with the math. As you probably know, or soon will learn, the contact between ball and raceways is not a point contact. Instead, it is a line contact and that implies that there are different speeds involved, dependent on the different radii working in the raceway.

I do not want to help you too much. But that is the mathematical paradox of the ball bearing. A roller bearing doesn't have that problem. At least not as pronounced as a ball bearing has it.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[electricpete]

The ball bearing obeys the kinematically derived equations quite well in most circumstances.
The miniscule deformation has insignificant effect on the frequencies.
The predictability of the frequency helps with rolling bearing diagonistics.
The deformation (related to hertzian contact theory) is relevant to load transmission and friction and the mysterious phenomenon called "elastohydrodynamic" lubrication (EHD).
Google EHD lubrication, you will find plenty if that's what you're interested in.


=====================================
(2B)+(2B)' ?

 

[Weber308]

Thank you very much for the help and the links. I'm compiling a report on what I've learned till thus far and understand about the workings of a bearing. I will present it to my boss tomorrow and hear what he says. I do feel that the question is a bit vague although learning as much as I can about bearings can only be beneficial.

Skogsgurra, I'm not sure what you mean by line contact. Do you mean that when a load is applied radially and axially that it is uniformly distributed?

I sincerely appreciate the help!
Thank you guys very much.

 

[Skogsgurra]

Yes, Weber.
The contact surface is circular only when there is no load on the bearing. When radial load increases, the surface area gets more and more elliptical and will, at higher loads, cover a substantial part of the width of the raceway. Since the radial distance from the center of rotation differs while the ball's rotational speed is the same at all distances, it follows that there must be a sliding action between ball and raceway. This sliding action produces the "fish bone" patterns that can sometimes be seen in the raceway.

Please note that it has nothing to do with wash-board patterns - they are something entirely different.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[electricpete]

One view of the sliding pattern Gunnar is talking about is shown in Figure 1 at the bottom of the 2nd page (page 282) here:

http://www.skf.com/binary/tcm:12-97854/520034.pdf

During forensic analysis, SFK did recently tell me they could distinguish the zones by direction of sliding apparent in individual marks under a microscope. From my experience (could be wrong), I didn't think you can see anything of this pattern by the naked eye. Often there is an obvious ball path which is shinier or duller than the rest of the race (especially if the bearing has been running in distress for awhile). Sometimes the eye can see some variations within the ball path (from lubricant overheating comes to mind during a recent inspection). I personally have never seen anything visually that I recognized as related to the three distinct sliding zones.

=====================================
(2B)+(2B)' ?

 

[electricpete]

The link seems flakey. Here's an excerpt in case it doesn't work

=====================================
(2B)+(2B)' ?

 

[Skogsgurra]

Worked well for me. That is a pdf that I have been looking for for a long time! Thanks.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.