bearing bore positional allignment

[ozzkoz]

I'm fairly new to using ball bearings and I was hoping for some advice on what I think should be a very common topic but I can't actually find any info on.

When I have a simple shaft system with a bearing on each end and mounted in a 2-part housing. There is going to be position of the bearing bore in the housing relative to a tight fit pilot, then the position tolerance of the other bore relative to the pilot in the mating part. There will also be some runout of the bearing lands of the shaft relative to each other. I would typically want to keep enough clearance in the system to account for the positional tolerances but bearing manufactures recommend very tight fits (press on shaft and a few tenths clearance on the OD).

It seems like I will have to hold 3 position tolerances down to the knats rear end in order to ensure my system assembles. Is this typical or am I missing something? I figure I can consider the radial clearance in the bearings as helpful but this still only adds a few tenths. Is there a better way to control this? Should I be considering the angular misalignment of my bearing bores instead of the radial position?

 

[MikeHalloran]

There's nothing inherently wrong with 2-part housings, but if you're going to put a bearing in each part, you may have to machine both bearing bores in the same setup, in the assembled housing, or provide equivalent tight tolerances between each bearing bore and the housing split.

It may be easier to put both bearing bores in one half, and have the other half envelope the first, e.g. move the bearing support functions into the same part.

Or go to flanged cartridge bearings with spherical self-aligning features built in.

Or use a self-aligning roller bearing in one half.

Or use bearings with extra clearance.

Or accept reduced life, or use larger bearings. The thick bearing catalogs can help with tradeoffs like this. Bearing company sales engineers can help more if they smell any sales volume.



Mike Halloran
Pembroke Pines, FL, USA

 

[Tmoose]

http://www.skf.com/skf/productcatal...tableName=1_0_t21.tab&maincatalogue=1&lang=en

ignore the recommendations at your peril.

flanged pillow blocks include (static) misalignment capabilies because they need to.

http://www.skf.com/skf/productcatal...tableName=1_0_t21.tab&maincatalogue=1&lang=en

 

[electricpete]

Did you mean to post the same link twice?

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

 

[Fugeguy]

Can you upload a cross section?

Thanks

 

[caoimhin1]

All bearing manufacturers recommend the same positional tolerances, as given in Tmooses link to the SKF ones. But I've found that, apart from machine tool manufacturers, in reality nobody can actually meet any of the tolerances, they are very very tight.

 

[unclesyd]

I concur with colmhin1's post. The prometric centers between two brands are usually off, not as much now as it 15 years ago. A lot of the center of rotation was an orbit, not much but this plus the tolerance of the on the geometric centers could throw you and require a lot measuring.
The above conditions were sometimes pretty bad on bearings with bores 2" and above.
We had very long 2" jack shafts with a gear every 8" and it to one or two shifts to get the TIR down to acceptable levels by shimming the bearings.

 

[ozzkoz]

Thanks for the table, that is the kind of thing I was looking for, I just have to figure out what the fits are (must be metric equivalents to LCx). It doesn't sound like I I'll be able to meet these fits, is it maybe ok to have a little bit of radial interference loading the balls?

 

[Tmoose]

What shaft or housing misalignment would create radial interference loading on a two bearing shaft system?

 

[unclesyd]

Tmoose,

We have one particular application where the rolls used in the
production of nonwoven fabrics where the web flow can create
radial interference on the bearings. This is especially true on the wider webs 6' or better wide. The reasoning is that the web
directional properties are not omnidirectional at this point.
Contributing to is the we are running the smallest possible shafts to keep the momentum down for quick stops.
This problem was overcome by the use of Slimb Bearings. These
were originally controlled by the operator but I understand that now some are being controlled by a Laser Displacement Device to
operate the Slimb Bearing. An added benefit was they were also
able to narrow the width needed for the selvaged edge of the web from 4" to 2".
We have some equipment used in the production of polymer that has bearing 18' c/c and we get radial interference due to the a mass build up off the center of rotation. As the component is so big we can't swing it to get a TIR of the 2 journals so we straightened the shaft on each end to less than 0.005" runout. We can do this because the sections between bearing will act like a crankshaft. While running we found that we will get radial displacement of the outboard bearing.

http://host.web-print-design.com/yorkshire industries/slimb_ap

plications.htm

PS:
The Slimb Bearing was one of my Attaboys. Didn't keep it too long though.

 

[Tmoose]

Unclesyd wrote:
"...........to get the TIR down to acceptable levels by shimming the bearings."

I'm assuming you were "shimming" the inner race ID. I have no love for most set screw or eccentric lock bearings, which jack the race off center and make the shaft rotate eccentrically ( and tip the race, too) but a big part of the problem is commercial shafting is pretty small. Several thousandths under nominal. Makes life easy for the folks assembling the parts, but if mechanical runout is important, I'm doomed.

 

[Tmoose]

Unclesyd wrote - "We have one particular application where the rolls used in the production of nonwoven fabrics where the web flow can create radial interference on the bearings. "

Hi syd,

i'm having difficulty picturing that application. Could you post pictures or links?

thanks

Dan T

 

[unclesyd]

I think this link will work.
Look at the applications page and other where it shows rolls handling a web. The mechanics of our web formation caused numerous problems with trying to achieve a uniform product. In correcting some of the problems we would impose radial loads on the bearings.

http://host.web-print-design.com/yorkshire industries/index.html

In respect to shimming the bearings were a pillow block type and the center of rotation height was off between sets of bearings and between manufacturers. I hate to quote an absolute number but I do recall 0.005" in one incidence. Bearing life was important but keeping the noise down around these machines was the big picture. We coupled work on the bearings with other things, like using a flexible disk coupling to effectively shorten the shaft, more adjustment in the many cross shafts.
Physical conditions around the bearing would tell us if it was under duress by an increase in high pitch noise, by meter, rapid heating as the bearings were nowhere near loaded in design sense, and vibration levels at certain points on the machine.

Sorry I can't get a picture of the web being processed. This operation is now an independent satellite plant and based on work done for them I don't thick I'd be welcomed with a camera.

The machines the bearing work was done on were designed in 1935 built in 1947. From new they had undergone an evolutionary upgrade in and design process with accompinining problems, the bearings being one of them. The load on these machines was not obvious just looking at them, but they were drawing/ stretching Nylon fibers with a ratio of 1.3 to 4.5 to 1.

PS;
If you need to take pictures and people get a little apprehensive take them with a film camera and allow them to develop the film and cull the trade secrets. I have done this twice recently.

 

[Tmoose]

Ozzkoz wrote - "It doesn't sound like I I'll be able to meet these fits, is it maybe ok to have a little bit of radial interference loading the balls? "

When I read that, I interpreted it as a internal bearing clearance reduction resulting from tight shaft or housing fits, NOT a radial load from the process, or coaxial misalignment of multiple shafts, or <2 bearings on one shaft.

Dan T

 

[unclesyd]

oes it really matter how the bearing is loaded radially in a practical sense.

Again if I recall the bearings I referenced all were the 200 and 300 series bearings. One our problems was we had 8 bearings on a single shaft. This was a design condition imposed on us due to having to power the cross shafts or jack shafts at some multiple of the original speed and load.

All around our plant e had to use bearings with increased clearance to accomodate thermal expansion. At the time you usually couldn't get the figure bearing we need for increase clearance. Our biggest problem with bearings with high clearances was the mechanics getting carried away with the preload.

You have to give me a little latitude as most of these events happened thirty or more years ago. If I could get to my old files or to the closing reports of some projects I could paint a better picture with numbers.