Bearing position tolerances 1

[tlicense]

Hello,

I have been looking at the bearing installation for a roll and was intending to use FAG super precision bearings. However upon looking at the tolerances required for the housings (screen grab taken from the catalog attached), I'm doubting the ability of any of our machinists to be able to achieve them.

If I'm reading this correctly, it's saying the two bearing bores should be concentric to datum axis A-B to within the diameter given as t5 divided by 300.
For a 100mm bearing, that means that the coaxiality tolerance is 6/300 um = 0.02um!
If the roll was say, 1000mm long, what machining process would you suggest to hit a 0.00002mm coaxiality tolerance? I've taken a look at SKF's tolerances and they're much the same.
I don't know of anything that is going to hit that tolerance. From my brief research, CNC honing should be able to maybe get 0.25um at best.
Anyone have any idea what process you would use to get the rest of the way there? Even if someone did take the job on and machined it, I've no idea how we would inspect it, maybe an air gage of some kind?

Many thanks,

-Tony

 

[MikeHalloran]

Super precision bearings are normally used in, e.g., machining spindles, where both bearing bores are located in the same casting, and probably machined in the same setup.

Rolls are normally supported by separate frames of some sort, where the bearings may be one or several meters apart. Corresponding bores are typically not in the same part.
I think that situation is why self-aligning bearings exist.

A talented machinist might be able to meet the spec you quoted with a line-boring rig, but if the entire machine is not in precisely conditioned space, the bearing bores will still move around because of temperature differences.




Mike Halloran
Pembroke Pines, FL, USA

 

[Tmoose]

".... the bearing installation for a roll and was intending to use FAG super precision bearings. "

What is the benefit you are expecting from superprecision bearings?

 

[tlicense]

Sorry gents maybe I've not framed my question correctly. The specifics of the roll I was looking into when I came across this is a inconsequential to what I was asking about.

Regardless of what it's for, what machining process would you use that could reliably hit a concentricity less then 0.1um? I've experienced *precision* machinists struggle to hit less than 10um.

Even if it were machined in the same setup you would have to be going some to achieve anything like that.

 

[BrianE22]

I'd start with jig grinding. If the housing is not too long, I think you could do it.

 

[imcjoek]

The divide by 300 bit looks weird to me. If it's a constant value, why not just bake that into the chart?
More likely the tolerance is just the "T5" value. That puts you at 6um [~0.0002" for our imperial viewers] which is more in line with what I'd expect to see for a bearing bore tolerance.
Bearings (Super precision or not?) are not mirrors or lenses. I don't think wave-lengths-of-light is an applicable measurement ballpark.

(Call manufacturer to verify).

 

[tlicense]

I don't think wave-lengths-of-light is an applicable measurement ballpark.

That's a very good point!