Custom Design Ball Thrust Bearing - Where to start? 2

[NuclearNerd]

I need to spec a rolling-element thrust bearing in a fairly corrosive environment. Unfortunately I can't seem to find anything commercially available (400 series stainless won't work).

I am considering making our own thrust bearing. It should be easy to manufacture a cage from 304SS (just a thick washer with holes), thrust washers from 17-4 ph, ground flat, and COTS ceramic balls. Now I just need to figure out how to do the analysis. I can calculate the hertz contact stress for a given load and ball arrangement, but I'm not sure where to set the contact stress limit to guarantee a long service life. I'm hoping I can come up with a conservative number analytically, because I don't have the budget for an extensive test program.

Where do I begin?

 

[dinjin]

You are really limiting the function by using
flat washers as raceways. Contact several
bearing manufacturers with your request as to
ball and race material. If you can buy it, so
can they. Give the them the loads involved and how
you plan on shielding the bearing. What type of
grease is allowed. ON and On.

 

[MikeHalloran]

Just pick up the phone.

Call someone who does know where to begin, e.g. a sales engineer for a rolling element bearing company. You don't need to reinvent the bearing, and climb a learning curve that's taken hundreds of years to surmount. You just need a custom bearing.



Mike Halloran
Pembroke Pines, FL, USA

 

[Tmoose]

How long is a "long" service life?

If the loads and required life are low enough some bearing manufacturers or specialists will offer plating of fairly normal bearings to buy chemical resistance when sealing to prevent lube contamination etc is impossible.

http://www.kaydonbearings.com/Endurakote_bearings.htm

FWIW FAG considers 2000 MPa the "contact stress" endurance limit of 100Cr6 bearing steel, so "infinite" life is attainable if EHD lubrication, kappa > 2, and excellent cleanliness are maintained. Your description of the bearing environment as "fairly corrosive" sounds like none of the 3 requirements are met

FAG and SKF publish (differing) maximum loads relative to endurance limits for each of their ball and roller bearings.

Some Individuals within the bearing fields question whether an endurance limit actually exists.

https://www.stle.org/assets/document/8-10_tlt_commentary_ISO_Standard_281_2007_Part_II.pdf

 

[NuclearNerd]

Just pick up the phone

I wish it were that simple. Bitter experience tells me that most companies are not interested in engaging on a small custom job, even if I'm willing to pay a premium. If you know otherwise, maybe you can point me to a willing company?

You are really limiting the function by using
flat washers as raceways

I know, but that is the easiest way to go for designing these in-house since I won't have to do any precision grinding. I can always increase the ball size (up to 1/2") & number of balls to limit stress. I might change my mind though after I run the contact stress numbers.

some bearing manufacturers or specialists will offer plating of fairly normal bearings to buy chemical resistance

In this case I know from experience that plating is insufficient protection unfortunately. Thanks for the stress numbers from FAG. I have a document from Koyo that suggests a limit of 4200 MPa for ceramic bearings, so it's in the same ballpark. That same Koyo document suggests loading ceramic bearings no more than 10% of the dynamic load rating to protect from shock, so I might try to use the contact stress limit you gave, calculate a maximum load, and then take 10% of it.

 

[Tmoose]

depending on the operating conditions rolling element bearings can need some "minimum" load to reduce skidding which can really cause problems.
It is usually a number like 5% of basic dynamic load.

So you have the possibility of shock in your application too?

Modern bearing life estimates rely on beautifully pure steel and exquisitely gentle machining to create ideal surface conditions free of stress raisers.
I think commercial washers and commercial grinding may not qualify as either

 

[Tmoose]

A Google search for tpi_64_de_en.pdf should find FAG's Corrossion resistant products catalog. It includes some discussions about coatings and lots of other stuff.

 

[electricpete]

It should be easy to manufacture…. Now I just need to figure out how to do the analysis. …I'm not sure where to set the contact stress limit to guarantee a long service life. I'm hoping I can come up with a conservative number analytically, because I don't have the budget for an extensive test program.

With due respect, it seems to me that you are grossly underestimating the engineering,experience and precision that goes manufacture of into modern rolling bearings. Good advice from the other responders.


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

 

[NuclearNerd]

With due respect, it seems to me that you are grossly underestimating the engineering,experience and precision that goes manufacture of into modern rolling bearings.

I don't mean to minimize. I would dearly love if there were a commercially available solution for my application. Unfortunately the best I can do is bound my uncertainties and greatly over-design.

Unfortunately, a first run of the calculation, with 1/2" diameter balls, flat washers, and a 200MPa contact stress limit gives me a capacity of only 0.1N per ball, which is FAR too small. Back to the drawing board...

 

[NuclearNerd]

The numbers are looking a lot better using rollers rather than balls (lower Hertz stress for a given diameter). Keeping the contact stress below 200 MPa, I figure a 1/2" diameter x 1/2" long dowel pin could hold ~ 90 N (20 lbf). I can build an arbitrarily large diameter bearing, so squeezing in 25 rollers gives me a 500 lbf capacity (less if I don't assume the rollers share the load equally).

Unfortunately I can't get ceramic rollers. I can make both the washers and rollers from 17-4ph, but I'll probably need to chrome plate one or both to prevent galling. I'm closer. Thoughts?

 

[coolbreeze]

If the idea of using a roller bearing instead of a ball bearing is novel to you, than it should be getting more obvious that their are commercially available solutions which you have not considered yet in your application.

While Koyo, SKF, Timken and the like will provide basic engineering information that you can use to "design" a custom bearing yourself, they all have application engineers that will help you get it right with a phone call.

Have you considered fluid film bearings?

http://www.ussbearing.com

Creative Motion Control also has a unique solution which I am sure they will customize for you at a price.

 

[Tmoose]

What are the rpm you have and life you need?

To survive at anywhere near the FAG contact stress the rollers MUST be lubricated and operate with "utmost cleanliness.
===========
Is the 4200 MPa TOYO info you referenced here ? -

http://www.koyousa.com/brochures/pdfs/cat208ex.pdf.

Look at page 41. The life when lubed with various coatings lubes is 1 million revolutions only when loads are tiny. Like 0.02 of the basic rating.
=============
Tapered roller bearing roller and race geometry create uniform rolling contact.
Straight rollers in a radial array must slide like crazy. Under significant loading (even with pretty good lubrication) sliding will tear up the rollers and races.
Fig 74 here-

http://www.schaeffler.com/remotemed...publication/downloads_18/wl_82102_3_de_en.pdf

 

[NuclearNerd]

To survive at anywhere near the FAG contact stress the rollers MUST be lubricated and operate with "utmost cleanliness.
===========
Is the 4200 MPa TOYO info you referenced here ? -

http://www.koyousa.com/brochures/pdfs/cat208ex.pdf.

Look at page 41.
The life when lubed with various coatings lubes is 1 million revolutions only when loads are tiny. Like 0.02 of the basic rating.

That's the source, yes. Page 18 talks about 4200 MPa as the static load limit from ISO 76-1987. The limit for austenitic stainless would of course be much less. The plan *is* to limit the load to something like 0.05 of the basic rating of an equivalent carbon steel bearing. I have somewhat arbitrarily chosen 200 MPa as my design limit for contact stress. That should be low enough to prevent static deformation, and it's below the fatigue limit for 304SS iirc.

Tapered roller bearing roller and race geometry create uniform rolling contact.
Straight rollers in a radial array must slide like crazy. Under significant loading (even with pretty good lubrication) sliding will tear up the rollers and races.

I can see that. By keeping the loads down, I hope to minimize the wear, but it is inevitable with a cylindrical roller design. Testing may be required at this point.

Have you considered fluid film bearings?

http://www.ussbearings.com

Creative Motion Control also has a unique solution which I am sure they will customize for you at a price.

Thanks for the links. My application is too slow to support a fluid film (and I'm not interested in distributing pressurized water to all of the bearings). USSbearings solution looks interesting though, because of the low unlubricated coefficient of friction. I'll look them up. Creative Motion Control's bearing looks interesting, and I'll call them too, but I'm not holding my breath that they can make the same design work with austentic stainless.

 

[sreid]

Consider using air bearings from New Way Air Bearings.

 

[NathanSM]

We, at PdM Specialists, can help you find out what type of bearing you are looking for based on your needs. Look us up!

http://www.pdmspecialists.com/