Foo - Aaah, trying to preload these ___ bearings

I have no idea if I am doing this right, there seems to be little to no documentation that describes the steps involved in properly preloading angular contact bearings. I understand I should use a small spacer between the inner races. Can somebody confirm that I am doing this right?

Firstly, these are going on a pretty nice ground ballscrew for my CNC mill. So there be bidirectional thrust loads for sure. Max RPM will be ~700.

Oh, the bearings are 7200's. That is, 10mm ID, 30OD, and 9mm thick.
They have a 30mm diameter bore that is 19mm deep, the fit is perfect. I made them deeper than need be just in case, since I can always remove material later.


****The important part here****
So, first I measured the axial "slop" of the inner races. I supported each bearing only by the outer race, and applied force to the inner race. The inner race of course is pushed downward until the balls hit the retainer and it can no longer moved. Using a .0001" drop indicator, I measured just how much movement there was. I tested four bearings, my findings were as follows:

.004"
.0041"
.005"
.0042"

I was warned that this dimension will not always be consistent, so I am not startled by the range.

Now, I believe I need to insert a spacer between the inner races, and then clamp on the outer races. I already have an outer race clamp all set up and ready to go.

Spacer thickness....
Do I pick two bearings that go together and add the two dimensions together? So one spacer would be .009" thick exactly, eliminating all room for axial play (assuming my measurements are correct!). But then I would need to add some more thickness to actually preload it, correct? How much extra thickness are we talking about here? Given that steel in this situation won't want to compress much, I would imagine that .001" would be more than adequate, is this correct?

Is my method correct? Any verification is well appreciated! I can make the spacer to an accuracy of .0001", but I don't want to deal with this until I KNOW this is the right way, because making things that thin on a grinder is a PITA since there is not a lot of magnetic force holding it down.

Thanks

Sounds like a problem with the Fetzer valve to me. You are gonna need to get a new Fetzer valve.

Hey! It's all ball bearings nowadays. Now you prepare that Fetzer valve with some 3-in-1 oil and some gauze pads. And I'm gonna need 'bout ten quarts of anti-freeze, preferably Prestone. No, no make that Quaker State.

I think the Flange gasket might be throwing your readings off. Maybe try re-grouting your Anderson hinge, then reset the Flange then take another reading.

Remember to shimstock the bimmelfarger as well!

http://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gifhttp://ganjataz.com/01smileys/images/smileys/geek-speakinleet.gif

As long as the rotational thrust dampener is properly syncronized to the directional output transducer and the oscilating torque multiplier is isolated from the centrifugal clutch mechanisim, you should be ok.

phantomcow2 - prennial emo kid

Any drawings of your arrangement? Typically these bearings are installed in pairs so that they can take load in opposite axial directions. Preload is then adjusted based upon the thickness of the spacers clamping the inner-races together vs. the spacing of the outer-races/bearing-carrier. The inner races can then be floating. If you're holding a ballscrew, you're looking at a pair of bearings on each end or four total.

Correct, they are placed together back to back (DB) to take axial loads in both directions.
I've actually got this angular contact setup on the screw coupling end, and a radial bearing for support on the other end. I don't need to fix it on the other end, because of the minimal speeds this ballscrew is rotating at and the diameter (20mm) will not whip.

Reread my first post though. I need to know how to determine the proper inner race spacer thickness.

And, how much extra thickness should be added for preload. Is .001" considered astronomical? Or is it considered just right for this type of application?

I don't think it's that easy as preload is a force that's not in a direct relation to displacement of the inner-race. Here's a way to calculate the preload forces:

http://i42.photobucket.com/albums/e346/DannoXYZ/AutoTech/AngularBearingPreload0.jpg
http://i42.photobucket.com/albums/e346/DannoXYZ/AutoTech/AngularBearingPreload1.jpg
http://i42.photobucket.com/albums/e346/DannoXYZ/AutoTech/AngularBearingPreload2.jpg

The bearing-manufacturer should give you an Fa preload-value for the bearings. Personally I would use a spring between the bearings to achieve this. You can then calculate the force from the spring-constant K. Or maybe even use some spring-washers:

http://i42.photobucket.com/albums/e346/DannoXYZ/AutoTech/SpringWashers.gif

The other thing you have to account for is the axial-loads of the screw itself as it will overcome the preload on one bearing.

What is the application for those bearings?

For MY bearings? They will go for a ballscrew, which is for my soon-to-be CNC mill.
For angular contact bearings in general? Applications where you will see a radial and axial load, angular contact's can cope with both fine. They only handle axial loads one direction, but you can mount two of them opposite each other to cope with axial loads both directions. That is what I am doing. You can mount two oriented hte same way for more thrust load capacity. You can mount them in slightly different arrangements, each with their own Pro's.
Very useful :).

Also, I believe I had a solution to this all. See, I have bearing blocks which have been bored 1mm too deep. So how about I make a spacer of some arbitrary thickness, say .03". Then, grind the top of my bearing block (this also reduces the depth of my bore) until just .001" (or whatever extra thickness I would have added to my spacer) of the bearing's outer race is sticking out. I would have had to have a clamping mechanism to clamp the outer races together anyways. Now that clamp can clamp the outer races together, but all the while it squeezes that .001" of preload.
It is much easier to precisely control the depth of that bore than the thickness of a thin spacer

Your preload isn't a distance, it's a force that corresponds to the distance.

Do a stack of everything in the system from top to bottom, figure out the surface elastic modulus of the bearings themselves, then the thru-part modulus of the races and the bearing block, from there, find out the correct preload force and you'll get your preload distance.

Any chance I could just go by feel? :D
One guy told me that a general rule of thumkb to follow if in doubt would be that the force required to rotate the shaft on a preloaded bearing should be about 4:1. That is, 4 times the effort needed when the bearings are not preloaded. I don't even know the preload force required, not sure what amount of thrust loads I will see are. I don't think the bearing manufacturer will give me any sort of modulus information. The most I get as far as information is "ISO 9001 certified". Tells me a lot, I bet theres only one factory out there with that certification :rolleyes:

Try blowing on it.

Errr...bearings aren't my specialty, but I've got a pretty good suspicion that the worst that could happen if you go by feel (and don't flat out crush the bearing race by clamping down way too hard) is you shorten the life of the bearing by a little bit and maybe add a little bit of extra friction in. With a homemade CNC, unless these are flat out tiny rollers compared to the application, I can't imagine the durability of the bearings being a major problem. Even if you do wear them out, it sounds like you can unbolt your clamp and replace them pretty easily. Any friction should be totally negligible compared to the force required from your steppers to move the mill bit into the workpiece.

Oh, and 0.001" sounds like a fairly small compression fit. I can compress the jaws of my calipers by that much with my thumb no problem.

However, I will definitely lay rest my opinion if svloid disagrees, since he seems to be more familiar with bearing applications.

By the way, I never asked before. How many axis is this mill going to be?

I can check some tech pubs at work tomorrow for you. I remember reading a lot about duplex pairs in the past.

<edit> They do have a few applications in submarine equipment.

10 axis.

Why blow it when I could use my air compressor:D

I know it won't wreak havic on the world if my bearings are just preloaded by feel and with aluminum foil as shims. But, I really want to have it as perfect as I can get it so I can say "I did it right".

I am not saying that .001" is how much of a difference there will be between outer race and bore diameter, that would actually make a decent interference fit. .0005" is what I aim for when I make a bearing block for fun. I am saying that it will compress the setup .001" to preload it. I have no idea how many pounds that actually is though.

As for axes...
I am going to start at 3 and work my way up :D . I've got the bulk of the hardware needed all produced for 3 axes, just need to buy one more helical beam coupling. But I have a few goodies given to me waiting to become a 4th.

As for axes...
I am going to start at 3 and work my way up :D . I've got the bulk of the hardware needed all produced for 3 axes, just need to buy one more helical beam coupling. But I have a few goodies given to me waiting to become a 4th.
If you can get it to the Fifth Dimension (http://www.youtube.com/watch?v=53hQAPBuW5o), it will really rock!