Shim thickness for bearing preload.

[Flynth]

I'm rebuilding an old surface grinder's spindle and there is a set of two tapered roller bearings with preload spacers/shims between them. One of the shims got loose and became deformed/a lot thinner.

Those are drive side bearings so they don't need a large preload. I estimate 50N up to 100N tops based on what I read.

Unfortunately, I can't find how to set the axial distance (shim/spacer difference for inner-outer race) to achieve a given preload.

If anyone knows, please let me know a good starting point for very light preload. The bearings are 30206.

 

[Firstram]

The bearings in the drawing (first thread) look like angular contact bearings, not tapered roller. Is it possible the acb's were replaced with rollers by Bozo? The races should be ground to provide preload with equal height spacers. Just throwing it out there.

 

[Flynth]

The bearings in the drawing (first thread) look like angular contact bearings, not tapered roller. Is it possible the acb's were replaced with rollers by Bozo? The races should be ground to provide preload with equal height spacers. Just throwing it out there.

It's not impossible, thank you for mentioning it. It didn't even cross my mind they can be non original. Checking in the manual it specifies 7206 class 5 (angular contact ball bearings) as a recommended replacement. It is anyones best guess what they have used in the factory 70+ years ago, but whatever it was it couldn't be worse. So if they used different bearings they should have been selected for better tolerances.

IMO The difference between normal and class 5 where it matters most is in maximum radial runout of races. Normal tapered roller bearings allow for 18/25 microns of inner/outer race runout while class 5 is only 8/8. I'll have to measure the actual runout of those bearings before I decide if I want to spend $140 for a set of 7206 class 5 (my local shop has them in stock if I wanted to buy them). The tolerance class specifies the worst case. I'm hoping those bearings are not that bad. I plan to install the shaft with just those two bearings preloaded as planned and measure the runout without the front bearing. The bearings are about an inch apart while the measurement will be 12in away so I expect any runout to be magnified 12 times. If I read the bearings tolerance specs correctly class 5 should have at worst 16microns of total runout, if both bearings are arranged in worse possible way that may be 32 microns total, that becomes 380 microns (1.5 thou) 12 in away! If I measure more I'll replace the bearings. I also might try "clocking" them to lowering the runout of one by the other.

Coming back to preload.

Searching for information about 7206 high precision bearings I found "class A" (light) preload is 98N and SKF also specified axial displacement for this preload class at 47N/micron So I guess I have my answer. I should have 1 micron difference between the spacers to achieve ~50N of preload.

I don't feel it is particularly achievable in a home shop.

If I choose to replace the bearings with let's say FAG 7206-B-XL-TVP-P5-UL I assume they will already have the axial displacement ground into their races. Then I'll have a difficult task of making two spacers exactly the same in width up to a micron (with no surface grinder)...

 

[Flynth]

I've measured the bearings (races width, OD and ID plus runout when installed - held by a 4 jaw rather than on the shaft to exclude influence of the spacers, the journal etc) and they are good enough.

Specifically the races width is within 7 microns(3 ten thousands in) . ID and ODs are within 5~6 microns, runout is about 6~9 microns. So they wouldn't pass class 5, but only by few microns. The problem now is the spacers. The inside race spacer must be original because after stoning off all the burrs its width is within 1 micron. The outer.... Is horrible. I have to make a new one, it will not be easy to hit 1 micron of tolerance with surface grinder down... That's one of the reasons I'm just t that worried about the bearing runout and races width exceeding class 5 tolerances. My spacer will exceed them too. If I'm lucky by similar amount and it they will cancel out (one can hope right)?

Pity they don't sell those spacers in variety of sizes like the bearings. Oh well, anything I'll manage will likely be better than before.

BTW, if anyone is interested how I measured those tolerances down to a micron... That's how:

That's a microkator and one division is 0.2microns(8 millionths inch). I used it for races widths and ODs. For runout and IDs I used a 1 micron resolution test indicator on a lathe.