Timken static stress and bearing life Query 1

[morris9791]

Dear Folks,

I have 2 queries regarding the output results of the Timken bearing calculation program.

1) Static Rating
Based on Timken recommendations for calculating the static rating of bearings - one can compare the static radial rating Co against the Po the static equivalent radial load to obtain a factor of safety etc.
Using manual calcs (based on Timken method) we can get sufficient FoS e.g. 1.4 etc but when the Timken program is used we get the following kind of result.

“Condition 1 - right position: caution bearing has stress greater than 350ksi (2400MPa)”
“Condition 2 – left position: life prediction invalid, raceway exceeds 580ksi (4000MPa)

Note that we get the same radial ad axial forces as in program etc.

Does any one know what the ‘built in FoS’ is in this Timken program? It would be useful to know the minimum FoS is so that we can avoid this message in future.
We tend to compare this program against the manual calcs for verification reasons etc.

I could back calculate this value as a last resource I guess. I was hoping some-one has a more detailed understanding of what is happening.

2) Bearing Life
The Timken bearing life

in revolutions is given as L10 = (C90/P)^10/3 * 90e6 revs and
in Hours L10 = (C90/P)^10/3 * 1.5e6/n where n is in rpm.

Since the Timken program calculates in hours, I am wondering is it safe / correct to convert to revs since we can obtain L10 and n from program and work out the (C90/P)^10/3 term and hence use this to back calculate the life in revolutions?

Any information will be appreciated. Best Regards

Steady Eddie

 

[dinjin]

It is safe to convert to revs. You should be able to
backtrack and find the Fos. If not call Timken. I assume it is 2 or greater.

 

[morris9791]

Thanks Dinjin,

I will probably do that; it would be interesting to see what FoS they use.

Another query I have.
Which of the following equations do people tend to use? It appears that using the first one assumes that n is very low (of the order 0.016rpm)
I think i would be more inclined to use the second one as it is more realistic since it includes rpm etc.
Of course it depends on the application I guess.

1) L10 = (C90/P)^10/3 * 90e6 revs and in Hours
2) L10 = (C90/P)^10/3 * 1.5e6/n where n is in rpm.


What are you opinions on this?

Thanks
Ed

 

[dinjin]

90 over 60 equals 1.5

No mystery. Depends whether you
want revs per minute or revs per hour.
Am I missing something?

 

[morris9791]

dinjin,

my bad, i should have looked more closely. the 1st equation assumes an rpm of 500.

All makes sense now.

Cheers
Ed

 

[Eratosthenes]

Ed,
the bearing steel has a strength of around 300,000 psi. If the contact stress roller/raceway is higher, you start getting plastic deformations, although at 350,000 psi they will be very small. I assume this is where you first warning is coming from.
The static capacity of a roller bearing is defined at the maximum theoretical contact stress of 4000 MPa (580,000 psi). Theoretical, because it actually won't happen. At this stress level the deformation is equal to 1/10,000 of the roller diameter, still pretty small, but per your second warning, the standard life calculation no longer applies. Personally I cut off fatigue life calculations before that depending mainly on speed.
Calculated fatigue life (=dynamic) is a statistical measure and there is no FoS. But there is a static FoS. It can be as low as 0.78 even 0.5 in automotive transmissions. Again, it depends on the application.
You are using tapered roller bearings with case carburized inner and outer rings. You may want to stay from a static FoS less than one to avoid crushing the core. 1.4 sounds like a good number to me.

 

[morris9791]

Dear Eratosthenes,

Thank you for the interesting information.
Just out of curiosity I compared the static radial rating against the resulting static equivalent for this warning and the FoS works out to be 2.8.
This seems to be the limit, any lower forces and the warning subsides.
However, I am guessing that the stress calcs are non linear and hence the resulting stress FoS would be a lot lower?

Thanks again
Ed