non matched bearings working in parallel

[sancat]

Dear friends,
We have a design problem, where we have to fit in a constrained space an approximate load capacity of 12800 N.
One of the alternatives will be to use a pair of 6201 series bearing working side by side. Yes, we know we can use a paired match, and that the dynamic load would be around 162% higher, but for a number of reasons we would prefer to use off the shelf regular bearings.
We would like to know if you know about a paper or a calculus method discussing life rating for this alternative:
off the shelf (non paired) deep groove ball bearings, working in parallel.
Thanks in advance
sancat

 

[diamondjim]

I am confused in what you are asking?
Is the load radial or axial? I assume
you are stating each can take approximately
80 percent of the rated load when paired.
That sounds correct if you are talking
about radial load.
What is a paired and non paired bearing
arrangement? Are you saying two different
bearing models in a paired arrangement?
By paired arrangement, do you mean identical
clearance or preload in the two bearings?
I would submit your application to the
manufacturer of the bearings for a rating
and dynamic life calculation with your
given loads. Sounds like an interesting
problem. AFBMA has some definite guidelines
for different bearing arrangements.
You might want to consider a two row ball
bearing.

 

[Hippo41]

The point of using matched bearings is so that you get equal load distribution over the two bearings,otherwise it's gong to be a bit of a lottery.

Useful to know the radial and axial loads you need, maybe some angular contacts or small cylindrical roller bearing may be nore suitable.

There are ways of determining the statistical life of bearings working in sets (e.g. if you have one that gives you 30 000 hours life, one that gives you 20 000 hours life, the combination may offer 15 000 hours life).

Can you e-mail me your fax number or e-mail and I will send you the appropriate catalogue sheets?

Lester Milton
NBC Group Ltd, Telford, Shropshire, UK

 

[sancat]

Thanks for your answers,
Matched or paired bearings are those manufactured toghether and "paired" at the factory so each would take an equal amount of load. They are suppossed to work as one, and the set is supposedly able to carry a 162% higher load.
We will use the bearings to replace some very small bearings in a series of conveyors, in a cookie factory. The bearings, thus, are mainly loaded radially.
They work non stop (24/7), at a speed of 300rpm. They should last 2 years. Using the basic load equation, I have determined that I need 12800 newts so a 90% of them should last 24 months.
What we intend is to place a pair of bearings on each side of the roller. An additional comment, due to the conveyor special use, one side should be electrically isolated from the other, so, one bearing (in my new configuration one pair of bearings) are installed over a nylon sleeve, which isolates the inner ring from the shaft.
We dont want to use actual paired bearings for stocking and logistic problems, and also because we dont trust the client will know/respect this special indication.
sancat

 

[sancat]

Hippo, Thanks!, What's your email?
sancat

 

[diamondjim]

Sancat,
Hippo's analysis seems strange
If bearing number one gave you 30000 hrs life
and the second gave you 20000 hrs life, I would
expect the pair to give you something like
115740 hrs life if the input loads are the same.
Do you know how clean the 6201 series is as far
as cleanliness level and grain size? This is one of
the most important factors. Vacuum degassed material
can raise the life 5 times the rated life of regular
steel bearings.

My thinking is that since life is a cubic function
and one bearing may see 60 percent of the load
part of the time and the other may see 60 percent
of the load at other times. Life would not be simply
8 times or (1/0.5)cubed times the rated life.
The life would be 4.63 times (25000 average value).
4.63 equals (1/0.6) cubed.

Keep us posted with your results.

 

[Hippo41]

Hi to both Sancat and DiamondJim!

My e-mail technical@nbcgroup.co.uk

Let me know your contact details and I'll send you both the life formulae that I'm talking about.



Lester Milton
NBC Group Ltd, Telford, Shropshire, UK

 

[Hippo41]

Reading DiamondJim's last post again, I will try and put the formula up here, but it may not be clear.

Assume Bearing A has a life La of 20 000 hours
Assume Bearing B has a life Lb of 30 000 hours

The life of the system Lab is given by the formula

1/Lab^^0.9 = 1/La^^0.9 + 1/Lb^^0.9

The numbers above result in a system life Lab of 11 132.9 hours.

Most small bearings (from the reputable suppliers anyway) are made from vacuum degassed steel now. In fact, as the designs have all just about reached the 'optimum', improvements in bearing life are now mainly down to metallurgy improvements (e.g. SKF Explorer series, NSK Super-Tuf and Hi-Tuf steels). Admittedly there are some interior changes (control of surface finishes etc), but it's more down to the control of impurities, crystal structure in the metals etc..

However, I usually caution machine designers that using these 'new' super bearings is dangerous if those who maintain these machines are going to simply throw in whatever they can buy (or are told to buy!) locally - as usually happens in my experience. Hence I appreciate Sancat's desire to stay away from 'specials'.

Lester Milton
NBC Group Ltd, Telford, Shropshire, UK

 

[diamondjim]

Sancat and Lester,
I do not know where the formula came
from but is makes no sense that if you
saddle two bearings that the life would
decrease. I think for the formula
to make any sense L would have to be .50
times L for the combination.
I know that each bearing would not see half
the load because of inaccuracies, but if this
were true, your could expect 8 times the life
of the single bearing. So I think my analysis
may be nearer to the life expectancy i.e.
from 4.63 to 8 times the life for the combined
pair.
I did not use the weighted average life of each
bearing to get a true answer. I would have had to
use the load and assume that the first bearing
would see 60 percent of the load for some
percentage of the time and 40 percent the
remaining percent of the time. Then do the
same weighted average for the other.
Does this make sense?

 

[Hippo41]

The data comes from the current NSK Technical Handbook. From memory (as above) I used a factor of 0.9, having delved into the cupboard to retrieve the book I see that it should be 1.1 - Sorry guys!

That means the combined life of the two bearings as above would be 12754 hours.

Essentially it's probablility theory - if any one bearing fails then the machine becomes unserviceable.

Post your e-mail addresses and I'll scan the pages in and send to you for reference.


Lester Milton
NBC Group Ltd, Telford, Shropshire, UK

 

[diamondjim]

Les, Sancat,
I have no problem with the statistical
theory part, but neither bearing will see
the full load. So there must be an increase
in life if two bearing are used assuming
each will see less than the full load until
one of them fails. I think my analysis is
correct that the life will increase in the
magnitude of 4.63 times the life or greater.
I would go back to NSK for clarification.
You are assuming no load sharing and that
is not correct.

 

[electricpete]

I haven't read all the replies and I don't know the answer to original question.

One comment is that closer tolerance on dimension of the bearings makes matching less important. Many manufacturers offer variety of tolerance classes. Don't know if cost makes this impractical... was cost the reason for avoiding matched pair to begin with?

 

[Hippo41]

DiamondJim

I think we're both debating the same point - but from different directions! I agree with your thinking as well!

The ultimate system life will depend on the load split. However, if Sancat uses 'known brand' bearings they will probably be fairly well matched anyway because of modern production controls.

The other problem is that modern bearing steels are so 'clean' that the life figures achieved in practice won't bear out any amount of calculation anyway! Under certain conditions (unfortunately not cookie conveyors!) some manufacturers are now claiming infinite life is a possibility!

Two bearings will be better than one, simple deep groove ball bearings will be cheap and easy to fit/ maintain/ replace. I guess that's what's required at the end of the day.

Specify a decent grease if you have high temps or a lot of washdown water involved, and watch your sealing arrangement.



Lester Milton
NBC Group Ltd, Telford, Shropshire, UK