Bearing Preload 1

[janderson2]

Is there a simple rule of thumb for determining preload. I have checked the mfr's pages and they just say heavy enough to do the job and light enough not to cause excess heating.

 

[MASSEY]

I'm not sure what 'mfrs pages' you checked but I have never experienced that problem.

SKF, Timken, Torrington all have very specific pre-load information.

What specific bearing manufacturer/bearing are you working with?

The companies mentioned above also have engineers on staff to answer these type of questions.

 

[janderson2]

I've checked with NMB and NTN.

I'll follow-up with the ones you mentioned.
Thanks

 

[swertel]

In an effort to help you through this site, what kind of bearing are you using?

For example, as MASSEY stated, the catalogs for Timken and Torrington have preload values, but I'm only aware of seeing those values for duplex or triplex bearings (2 or three bearings back to back). They need a certain amount of preload in order to run true.

I haven't ever seen preload data in a catalog for single or mounted bearings. These typically don't require it. Then again, I've focused mainly on radial ball bearings and not needle bearings or thrust bearings in my designs.

--Scott

 

[willdo]

NTN has an excellent reference card showing preload values for sphericals and they also print a very in depth technical catalog. I use them all of the time. They also have good application engineers who have always been very helpful to me. You have to be able to supply them with enough information for them to help you (type of bearing, clearance, fit, application, etc.)

 

[janderson2]

Thanks for all the suggestions.

I am using a fairly standard NMB 608 with double shield.
I will have an axial load of 200 lbs and a radial load of about 50 lbs.

ressed on a shaft with an interference of a couple of tenths. We're using a tolerance ring to help center the bearing in a loose nest.

 

[willdo]

Just guessing, but I think what you're looking for is just proper shaft and housing fits. These help determine your final internal clearance of the bearing. We don't usually talk "preload" when we're talking about single row radial ball bearings. If you need shaft and housing fits you'll need to provide the clearance of your bearing (if the bearing is marked just 608ZZ it is a C0 clearance; if the bearing is marked 608ZZC3 it is a C3 clearance)and whether the inner race or outer race rotates. NMB can help you if you call. I would ask them very specifically about having a 200# axial load on this bearing.

 

[MASSEY]

I concur with willdo, the pre-load is determined by shaft and housing fits. That is why I say that I have never encountered not enough information because the suppliers I mentioned above are emphatically exhaustive concerning shaft and housing fits and tolerances.

In addition to shaft and housing specs. there are specs. for end play in the bearing assemblies. These are built into the assembly by the manufacturer.

When you speak with the bearing suppliers they are going to talk in terms of a resultant 'thrust' load which is the resultant load vector of your combined axial and radial loads.

Your application looks a little unusual in that your have more axial load than radial. You may find your suppliers recommending a thrust bearing incorportated into your design. Or at least an angular contact bearing in lieu of a deep groove ball bearing.

Tapered roller bearings are good for combined axial and radial loads. Deep groove ball bearings are designed primarily for radial loads and can accept some axial loading. Angular contact bearings are better for axial loading than deep groove ball bearings. Thrust bearings are designed to be loaded axially. Whereas needle style bearings are designed for radial loading.

To answer your original question: there is no standard rule of thumb for determining pre-load for 'bearings' because as you see, bearing designs are application specific. And each design has their specific shaft and housing bore specs. and tolerances as well as pre-determined built in end play.

Talk to your bearing manufacturer, they answer these questions every day.

 

[electricpete]

OK, so angular contact ball bearings have a well-defined and published minimum axial pre-load, and deep-groove ball bearings and roller bearings do not have any published minimum preload.

But radial bearings, particularly roller bearings CAN skid without sufficient radial load. Toshiba recommends that a bearing always be loaded to 4% of it's static load rating to avoid this problem. Also the type of lub makes a difference.

Skidding seems to be a particular problem for uncoupled run of motors with roller bearings on the drive end.... Not that coupling up to a pump adds any radial load, but presumably it stabilizes the bearing somehow.

 

[blfarrar]

Sorry to contribute to this discussion at the end...

To answer Janderson2 question - there is no rule of thumb all assemblies to designs of machines are build & application specific.

for the record - I spent time in my apprenticeship building precision spindle with either Timken taper rollers (prec 3 - copper dots etc) or with angular contact pairs with specified amounts of preload.

This preload amount was from the designer of the machine (metal cutting machine spindles)which were designed to handle specific cutting forces both radial & axial.

Bearing preload is there to eliminate any freeplay in the assembly that would compromise the function of the rotating element. It must also be considered as the bearings run the assembly must accomodate any additional loading caused by expansion through heat. Certainly not guesswork or trial & error.

The company I worked for had experimented with preload amounts that worked for the general purpose machines they built & the multitude of uses they encountered. Generally an unknown.

Only the very specific use machines (high speed spindles for machining metals to fine tolerances had values that were very specific.

As for the tolerances of the bearings (diameters internal & external) that create in some cases the interferance fits that mean the bearings have to be shrunk or heated for assembly - The amount of size change in the actual bearing element is negligable (or should be) extreme force being used to get any bearing onto a shaft or into a housing bore should not be necessary.
It also means there is very little scope for check the bearing is "home" to a shoulder or correctly fitted. The company I worked for just didnt design assemblies that used heat or freezing to assemble bearings for this very reason - as a company philosphy. It must have worked as the company name was revered for quality machines (William Asquith - Halifax.WestYorks).

Good reference sources for bearing technology are BARDEN - they make bearings for "problem" designs - high speed spindles & rotors. There technical reps are first class & will help - also very good products.
I have found that most of the good bearing makers will assist with tech data if they get a specific input of info.
If the assembly is a critical piece of kit it is worth doing a bit of research - prevention better than cure.

Note !
Keep away from obscure or low quality makes of bearings - any $s saved will be lost in the short life of the bearing.

Also
If the bearings arent sealed with lube - be cautious of grease overload (look at the reliability centre website)
Kluber make some very good speciality greases that can lube for life assemblies that are both high speed & high load.

 

[electricpete]

where is the "reliability centre website"?
thx

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[blfarrar]

electric pete

My apologies - next time I use a reference I will give the web address

the relibility centre website is at

http://www.reliability.com

some of their papers are really good.
they had a series that went into bearing lubrication & the faults that happen when this is over done or hardened grease is pumped into races.

also some good data on vee belt tensioning.

This was very applicable to a sawmill I worked in that they destroyed a couple of really large motors, (before I got there might I add) but part of the folklore of the site in the catolog if disasters that went unexplained until we root caused the problems.

 

[JeanMicheling]

Can someone tell me the best preload to use for taper roller bearing? I don't have no data exept the axial clearance of duplex taper roller bearing.

 

[Tmoose]

Those Wavy preload spring washers generally apply axial preload ~ "light" for ball bearings. One of the bearing outer races has to slide in order for them to work. Some loading needs clamped outer races to prevent creep, thus foiling easy preload OR floating schemes.

Tapered roller bearings are mighty danged stiff, and by necessity have widely varying dimensions, so the TIMKEN software and worksheets uses actual dimensional inputs to try and hit ~ zero enplay at operating conditions.

 

[JeanMicheling]

Is it possible to get this timken software and worksheets?

 

[Tmoose]

For a while Timken was sending out almost as many CDs as AOL.

Mine says "Tapered Roler Bearing Selection guide 1.0"