Shaft won't fit bearing

[Richard Hed]

 

[user 284]

It's got nothing to do with making the shaft fit as is. Obviously it's the method of measurement and measuring tool used that's the root cause. And at the small sizes of shafts, bushings or even rolling element bearings most of us would be using, trying that temperature differential trick isn't ever going to work well. Even less so with any ball or roller bearing. Check the thermal expansion rates for something like mild steel. 100 degrees change from room temperature and that shaft or bore is going to enlarge or shrink by about .00072". And that's for something at a full 1" diameter. Under that diameter you'd need way more difference, to get the few thou reported in the OP's initial thread post. You'd need a soak in liquid nitrogen and heating the bushing way up there. Even then I'm still not sure that would be enough. And commercial ball or roller bearings would get destroyed using that much shrink fit. The better bearing manufacturers all have charts with maximum allowable limits for press fits. And for the general sizes we might be using, that's down to very few + - 10ths at most.

Even for sleeve bearings or bushings that might be used on something like a model engine, your still requiring fairly tight tolerances. A thou or two up or down between both parts and it's going to be either way too tight or too loose. I've got a few different pieces of metrology equipment that could get me very close. But if I didn't and for anything small and with that critical fit. I think I'd invest in a pin gauge at the correct size I wanted and then a few up and down in size as go / no go gauges. Below about 1/2" in diameter, there pretty reasonable for cost. Or if you've got decent reamers, George Thomas mentioned he'd ream the hole and since pretty much all reamers will cut oversize to there marked diameter, then make the shaft to that nominal size and then try it for fit to the shaft. That of course doesn't work for something like a ball or roller bearing.

 

[Steamchick]

Manufacturers' limits for fits should be followed - if you can measure the related shafts and housings that accurately. This guarantees the durability of the bearing in service.
Unfortunately, so few model engineers can actually measure to the required accuracy - Or manage to achieve anything that accurate with their machines. Despite having the "knowledge" of what I need to achieve - I can't do it. (Sometimes hit the mark, but mostly fail!). I am lucky on any job to hit within half a thou of size. Usually happy at a thou under or over size - then bodge it! So I can't machine to clearances or interferences "from the book". - In a way, that is a part of the fun, and helps me enjoy and respect those that can achieve models with the right tolerances and fits so they run like smooth watches! I only get close with machining, then linish a bit closer - and hope I don't go too far! If clearances/fits are too large my friend "Loctite" comes to the rescue! - But actually, you are losing alignment accuracy then.. or relying on "the job" to self align before the loctite cures....
A comment on Rolling Ball burnishing: In industry, this is usually used so a "cheaper" steel can be used as the base material, but the burnished surface has an improved finish, but more importantly is has been locally compressed so is harder. The harder surface is required for some special applications and occurs naturally in other designs - such as on cams (with roller followers), cam followers (flat or curved faces), valve rocker ends that impinge upon a valve end, and actually slide across the surface of the end of the valve as the valve is opened (the spring compressed), eccentrics, etc. that need a very durable surface.
In fact, the same action also occurs naturally on steel tyres and rails on railways, ball and roller races, engaging gear teeth, etc.
You can see it as the witness mark where the surface looks smoother - more reflective, or "shiny" - as the surface imperfections have been "cold-forged" into a smother surface. But usually it is too small for the home workshop to measure.
All good stuff, and FUN!
K2

 

[L98fiero]

Manufacturers' limits for fits should be followed - if you can measure the related shafts and housings that accurately. This guarantees the durability of the bearing in service.
Unfortunately, so few model engineers can actually measure to the required accuracy - Or manage to achieve anything that accurate with their machines. Despite having the "knowledge" of what I need to achieve - I can't do it. (Sometimes hit the mark, but mostly fail!). I am lucky on any job to hit within half a thou of size. Usually happy at a thou under or over size - then bodge it! So I can't machine to clearances or interferences "from the book". - In a way, that is a part of the fun, and helps me enjoy and respect those that can achieve models with the right tolerances and fits so they run like smooth watches! I only get close with machining, then linish a bit closer - and hope I don't go too far! If clearances/fits are too large my friend "Loctite" comes to the rescue! - But actually, you are losing alignment accuracy then.. or relying on "the job" to self align before the loctite cures....
A comment on Rolling Ball burnishing: In industry, this is usually used so a "cheaper" steel can be used as the base material, but the burnished surface has an improved finish, but more importantly is has been locally compressed so is harder. The harder surface is required for some special applications and occurs naturally in other designs - such as on cams (with roller followers), cam followers (flat or curved faces), valve rocker ends that impinge upon a valve end, and actually slide across the surface of the end of the valve as the valve is opened (the spring compressed), eccentrics, etc. that need a very durable surface.
In fact, the same action also occurs naturally on steel tyres and rails on railways, ball and roller races, engaging gear teeth, etc.
You can see it as the witness mark where the surface looks smoother - more reflective, or "shiny" - as the surface imperfections have been "cold-forged" into a smother surface. But usually it is too small for the home workshop to measure.
All good stuff, and FUN!
K2

As I said earlier, manufacturers specs are how the bearings are designed to work, with that said, maybe, apart from miniature turbines for model airplanes and grinding spindles, there aren't many places where amateurs need to be that concerned, too tight can be a problem but Loctited bearings will last much longer than most model engines will ever run. Also, from personal experience, a mining company I once did work for wanted all electric motor bearing fits 0.0005 oversize and Loctited as the bearings lasted longer that way and the Timken axle bearings on their 85 ton heavy haulage trucks were fitted with Loctite as well.