Thrust bearing design

[xlxc34]

Hi Everyone,

I'm working on a BMX bicycle hub, and space is at a premium to say the least. This is because the design of the hub means that the bearings must be fitted underneath the driving sprocket, which has 9 teeth and is designed for a chain with a 1/2" pitch. This effectively places an upper limit on the bearing diameter of 24mm once the chain dimensions are taken into account, and the axle on which it runs is 15mm in diameter. The old design of the hub (which I am updating) had 2 single row deep groove bearings next to each other, but this is not up to the job and we're getting a lot of returns. This is partly because they are seeing radial loads that are too high and also because there are axial loads which weren't accounted for by my predecessor. These are pretty high, and my new design is going to incorporate a needle roller bearing for the radial loads and a loose ball thrust bearing for the axial loads (there is no way I'm fitting an off the shelf thrust bearing into this part, so I'm going to have to use the axle and housing as raceways for both the needle roller bearing and the loose ball thrust bearing.

I have a number of questions about the calcs I should be doing on the thrust bearing. I'm guessing they involve Hertzian contact stresses. However, I have a lot of unknowns that I was hoping some of you guys could help me with:

- I can't find info on allowable stresses for Hertzian contacts anywhere. Can anyone give me a ballpark for a steel part machined from a high strength steel?
- What sort of hardness should I be specifying for the raceways?
- How do I make them hard, or do I just machine the whole part from hard stuff?
- How are the raceways machined at this sort of scale?

Thanks, any help will be much appreciated!

 

[caoimhin1]

- For a bicycle, fatigue life is not going to be the issue. Work out the maximum shock load that you expect to see in operation and compare to the static load ratings in the bearing catalogues. I'd aim for 10-20% max of the catalogue static load ratings.
- 62 Rockwell hardness is typical for bearings
- Usually hardened before final machining
- Good luck, this is the black art of bearing manufacture. I'd recommend you use off the shelf ball bearings to be honest. Figure out why the current design is failing and work on a solution to that.
For example, unless they are a matched pair, two single row deep groove ball bearings beside each other are unlikely to work. Consider a pair of universal angular contact ball bearings which will take the thrust load and minimise clearance.

 

[electricpete]

Just curious....where does the axial load come from?

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

 

[xlxc34]

Thanks caoimhin1, that's very useful. Do you have a very rough idea what kind of Hertzian contact stress would be allowable? This is just for future reference, as I think I'm going to take your advice and try to have a more major design rework to eliminate the need for a thrust bearing in the space constrained part (so I can use an off the shelf bearing).

And electricpete, the axial load occurs because I am designing a type of hub called a freecoaster. Instead of having a ratchet mechanism like most hubs, There is a clutch inside the hub which has some kind of drag (usually provided by a spring) on the axle. The clutch is threaded onto the sprocket, so that when you pedal it moves axially along the axle until it engages with the inside of the hub shell. This means that the pedals don't rotate when you're travelling backwards, which can be useful! This guy runs a freecoaster:

http://www.youtube.com/watch?v=LuLyDJCZqi8

 

[dvd]

Can you re-design the clutch to double-engage and eliminate the axial load? Right- and left-hand threads cause both sides to move outward/inward and thrust forces cancel.

Cool video... but put on a helmet? I love somebody who should have.

Brain Injury Association of America

http://www.biausa.org/

 

[AshleyK]

I don't understand where the axial loads are coming from, aside from the shock load when somebody drops the bike on its side.

Deep groove ball bearings can take a fair amount of axial load as well. Have you considered a damping washer in between, or on the outside of the bearings to save a total re-design? We used to use lamiflex washers sandwiched between the outer bearing race and the mouting plate of a driven high-speed gearbox shaft with great success.