Bearing Interference fit problem on shaft in wheel - how to solve using only temperature differential if possible - already tried and failed.

I have replaced many 6003 bearings, they are really common in Japanese transmissions and axles. They are always press-fit to the shaft, always. If a body has the right tools to accompany a hydraulic press, it's a safe and routine process. You must support the bearing's inner race, and if you can't, try making a better set of press tools, because it's fundamental to the process. Good (not 3rd world garbage) bearing press plates are worth their weight in gold. Press arbors and shells either purchased or made from pipe/tubing material will do the job even better. You need something close fitting to support with, and something well-suited to drive with. Without that stuff, a press is just... well, it's just another hammer to ruin your parts with.
 
The general rule for fitting bearings is the point load gets a slip fit and a circumferential load gets a transitional to interference fit. In your application, the inner ring is the point load since that shaft is stationary, so that should be a slip fit. The wheel rotates making the load rotational on the outer ring, so to keep the outer ring from spinning in the wheel as the load rotates with the wheel, it needs at least a transitional fit. The exception to this is cylindrical roller and tapered roller bearings since those are separable bearings and you typically use a transitional fit on each ring to keep them in place, and they are designed to have the proper clearance with those fits.

Typical fits for this would be for the shaft h6 (+0 -0.011mm) and the wheel would be something like a K6 or even an M6 fit, +0.002/-0.009 to -0.004/-0.015, these are all millimeters. K6 would be a transitional hole fit, and M6 would be more interference. Your 17.05mm measurement is WAY oversized even if the shaft was rotating, so it is no surprise you had trouble installing them.

You want to make sure you do not go too tight with the fits, or have tight fits on both the shaft and the hole, otherwise you will reduce the clearance in the bearing and drastically shorten its life. If the shaft really is 17.05mm, then I am not surprised the bearings failed quickly.
 
The general rule for fitting bearings is the point load gets a slip fit and a circumferential load gets a transitional to interference fit. In your application, the inner ring is the point load since that shaft is stationary, so that should be a slip fit. The wheel rotates making the load rotational on the outer ring, so to keep the outer ring from spinning in the wheel as the load rotates with the wheel, it needs at least a transitional fit. The exception to this is cylindrical roller and tapered roller bearings since those are separable bearings and you typically use a transitional fit on each ring to keep them in place, and they are designed to have the proper clearance with those fits.

Typical fits for this would be for the shaft h6 (+0 -0.011mm) and the wheel would be something like a K6 or even an M6 fit, +0.002/-0.009 to -0.004/-0.015, these are all millimeters. K6 would be a transitional hole fit, and M6 would be more interference. Your 17.05mm measurement is WAY oversized even if the shaft was rotating, so it is no surprise you had trouble installing them.

You want to make sure you do not go too tight with the fits, or have tight fits on both the shaft and the hole, otherwise you will reduce the clearance in the bearing and drastically shorten its life. If the shaft really is 17.05mm, then I am not surprised the bearings failed quickly.
I called the manufacturer. They said the tolerance for this bearing was -.0008/0. good info in terms of tolerances. I am purchasing an outside micrometre today. a decent one.Also when I remeasured like 10 times with my crummy micrometreTaking care with my pressure. I got 7.02. So my initial 17.05 was wrong
 
They are a press fit on the shaft, because the shaft is rotating, always. In this case, the wheel is rotating, therefore the outer ring is the press fit, always.

From the bearing's point of view, how does it "know" which part is driving and which part is driven? It all looks like rotation to a set of balls and races. In those physics classes we'd draw little vectors to describe forces, and that diagram is the same no matter which part is driving or driven in this simple arrangement. As to which part fits loose, it's never acceptable to have slip, ever. The bearing must fit, always.
 
From the bearing's point of view, how does it "know" which part is driving and which part is driven? It all looks like rotation to a set of balls and races. In those physics classes we'd draw little vectors to describe forces, and that diagram is the same no matter which part is driving or driven in this simple arrangement. As to which part fits loose, it's never acceptable to have slip, ever. The bearing must fit, always.
Your last statement is completely wrong, most applications have one ring with a tighter fit, and one ring with a loose fit. It depends on which ring is rotating and which is stationary for which fit to choose for that seat. A typical electrical motor for example will have a shaft fit of k5 or k6 for the inner ring and H6 or H7 for the outer ring. If you don't believe that, maybe NTN will convince you.


It's very obvious to the rings when it is a point load vs a circumferential load. A point load will keep the ring from rotating with a slip fit since you have a force pushing at a single point the entire time, but if the load is rotating around the ring, like for an inner ring with a shaft that is rotating, it will very quickly start to spin that ring on the shaft unless it has a tight enough fit to keep it from rotating.
 
I called the manufacturer. They said the tolerance for this bearing was -.0008/0. good info in terms of tolerances. I am purchasing an outside micrometre today. a decent one.Also when I remeasured like 10 times with my crummy micrometreTaking care with my pressure. I got 7.02. So my initial 17.05 was wrong
I'm glad you are getting the micrometer, I would not trust a caliper to measure that accurately.
 
Update.
I got a used micrometer on marketplace. very happy with it. it's a Mitutoyo 102-121. I'm in love. I calibrated it. now getting decent measurements ( six or seven times) most are 0.6693" with some 0.6694" on the shaft . The only thing I don't have is a telescopic gauge so I went with the median value the bearing manufacturer provided

0.66897" low-end
0.66913 medium ( assumed)....... yes don't assume.
0.66930" high-end

which would be 0.66913" ( . that would put the almost exactly shaft 2/10 larger then the medianValue of the bore. Given that the fact is 17 mm, 300° temperature differential should have given me 0.00200" approximately. this should have slipped on. Maybe I just did something incorrect in the procedur. wonder if the shaft was a little too rough and needed a quick emery cloth sand. looked okay.

Or maybe I'm just an amateurAnd missed something blatantly obvious

I'm just north of Toronto. I have telescoping gauges and some micrometers and a press if you want to use them let me know.
give a shot on my own. if I'm running into trouble, I will let you know. this is very generous. Thornhill's around the corner from me. I'm in central North York.
 
If you heated the bearing up to 300F, throw it out and get a other one, it’s scrap. Heating it up that hot removes the hardness from the steel and it won’t last anywhere near as long as it should. You should not heat a bearing any higher than about 210-220F. You could have degraded the grease as well.

Your measurements are almost exactly 17mm, that is what the shaft should be for that application. You should only need a little bit of heat to get it on the shaft if the bearing is on the smaller end of the tolerance, if at maximum size, it might even slip on. I would take a look at the shaft and find out why the bearing is not going on, if someone banged on the end, it could be belled enough the bearing will never go on, the shaft could be oval, so check in three spots, ie 12 o’clock, 4 o’clock and 8 o’clock.
 
Back
Top