Shafts and bearings

Find a piece of allthread and two nuts. Put that in the "ears" on the left side (belt side), adjust the nuts so it's quite snug. This is to support the casting and keep it from flexing. The nuts go inside, not outside.

Make a plate with a hole in it to support the piece but clear the bottom end of the shaft. The hole should be slightly larger than the OD of the shaft.

Press on other end of shaft with hydraulic press. The shaft should move.

Don't hammer. Hammering can result in peening. Steady force is better.

If I understand you correctly the casing would be supported as well as the bottom bearing but the top bearing would be sacrificed? Also I am not quite sure what the "belt side" is: In this case the belt is looped in the middle over a pulley, between the two bearings. Or do you mean that these things are usually driven by a pulley on the left side?

I read it as the less well supported open side of the housing where the belt moves in and out.

Yes, unless the shaft wasn't all that tight in the bearing. That has usually been the case in my experience.

The cross section of the casting is sort of a U in one plane. I refer to the open end of the U.

I see. I went and found a Chinese catalogue with a search facility so I can find the numbers and dimensions in future. Now this may be a naive question but it puzzles me: A 5/8" shaft is 0.875 mm or 0.034" thicker than 15 mm. This seems quite a lot. Do all shafts have to be turned down to fit the bearings? The difference seems quite a lot to reconcile by using just heat and cold for a press fit.

You would find out that a caliper is very handy to have for checking bearing sizes.

Many (most) ball bearings are all metric, but some applications use a combination of metric and inch dimensions, with inch used as the I.D.

Generic (or used domestic brand) calipers are available for about $20, even for generic digital. Most digital calipers switch between metric and inch for instant conversions, and also can be reset to zero at any point along the beam.

Dial calipers are generally either metric, or inch, although there are some that indicate both on a single tool.

Shafts and housing openings are machined to sizes which are appropriate for the desired type of fit for ball bearings, otherwise bearing sizes are selected to fit existing parts.

WB ......... metalworking projects

I am not sure that I get your point. Are you saying that the tolerances on bearings are so loose that one has to get a handful and go through them with a caliper to find which matches your shaft?

Looking through McMaster-Carr web site there seems to be a wealth of both metric and imperial shafts as well as bearings. In my simple mind one would use a 5/8" ID bearing for a 5/8" OD shaft. The tolerances quoted are 0.0003" (converted from metric) for the bearing ID and 0.003" for the OD shaft (worst case).

I am puzzled why in the case of this particular piece of equipment there are

15 mm ID bearings supporting a 5/8" shaft (measured in the middle with calipers). To me it implies a degree of machining to achieve a fit, but why?

It's really pretty simple. Ball bearings were originally designed to metric dimensions because they were not developed in the USA or England and millions of them were designed into products. The vast majority of them are still made to metric dimensions although there are ones which are entirely or partly inch dimensioned. The USA manufacturers commonly machined their parts to fit the standard (metric) bearings. If you have a 5/8" shaft running in a X202XX bearing, the shaft has been machined down to fit. There has to be a shoulder of some kind to establish the axial location of the shaft and bearing.

To remove the shaft you have to remove one bearing with it. The bearing will slide out of the housing with a little force from the shaft, though there may be a snap-ring or other retainer that has to be removed first. It is not considered good practice to apply force thru the balls but often there is no other way to remove the bearing. A light tap on the shaft end from a soft faced hammer will usually do it. The bearing should be re-usable unless it is corroded in the housing and a lot of force is necessary to get it out.

Don Young

[ ... ]

If one went through them with a caliper, one would get very frustrated trying to find variations in a given part number. You will need something with a lot more resolution than a caliper can give you. :-)

And you expect to measure 0.0003" variations with a caliper?

There is a very good reason for this -- which the maker of an early double-sided 5.25" floppy drive did not understand. He mounted the spindle in a pair of flanged outer race bearings which slipped into a cylindrical bore. On one end of the shaft was the cup which drove the floppy. On the other end was simply a tapped hole. (1/4" shaft, FWIW). They slipped a pulley over the end of the shaft, and tightened a screw and washer to hold the drive pulley onto the shaft. They then painted over the end with Glyptol.

Some months later, I was finding double density floppies very difficult to read, while single density was still easy. Pulling off the drive belt, I discovered cogging in the spindle.

The bearings were radial thrust bearings, and the tightening of the screw had put axial thrust on them, causing them to wear out quite rapidly. When I got new bearings for the drive, I spent a little time on the lathe and turned up a spacer to go around the shaft between the bearings, and to hold them just a tiny bit farther apart than the cylinder in which the outer races mounted would do. This allowed me to tighten the screw firmly enough to keep the pulley from slipping without putting a serious axial load on the bearings.

In the case of your assembly, the 15mm shaft is turned down to

5/16" so the bearings will slide onto the shaft and then stop at a certain position. The reason for this is that there are some parts missing from your device which press against the outside end of the inner race, and support the wheel or buffer -- held in position by the nuts being tightened against a washer to hold the outer surface of the wheel or buffer.

Without that step in diameter from 5/8" to 15mm, this tightening would put a serious axial load on the bearings. Your bearings were made for radial thrust only. You *could* use angular thrust, but to take the various loads, they would need to be larger and more expensive. The full 15mm diameter of the inner portion of the shaft provides a spacer for the inner races so you can tighten the nuts as much as needed.

I suspect that you will find a step in the bore where the OD of the bearings mounts, to keep them from moving too far towards the center. Fairly light pressing should move the far bearing out of its bore, and the shaft inwards through the inner race on the near side. You will need to find a setscrew on the pulley to allow you to release it from the shaft.

Note that you will need to use a lathe to make the spacers between the outside end of the bearings and the wheels or buffers before you can use this. I doubt that you will be able to find the parts needed, as they are made for the task at hand. I have similar things on two grinders with built-in motors. They have similar steps on the shafts to keep the ball bearing assemblies from being pressed in too far and suffering similar fates.

Perhaps someone who has the same model can measure and draw up what you need to make. But you still need to make a new shaft to get around the bent end.

Good Luck, DoN.

Bearing tolerances are very tight measured in microns. Check out

to see.

I would imagine it was machined down to allow a shoulder to press the bearing inner race against. Most applications with a rotating shaft have a slip fit on the outer race and a press fit on the inner race. Check out these sites for info on shaft and housing design and shaft and housing fit.

I would not suggest reusing the bearings. They're pretty cheap and readily available and it's not worth the risk. I'm going from memory but I think you stated this was a 6202Z? If so, it most likely has two shields so you would need to order a 6202ZZ. The 6202Z is normally stamped on the shield because the factory doesn't know if it will be used in a single or double shield application. If you order with one Z you will only get a single shield. Good luck.

I didn't mean to suggest that ball bearings were individually hand selected from batches of bearings to achieve the correct fit.

A more-common all metric dimension bearing would likely be cheaper than a metric bearing with an inch I.D. in most cases. If the manufacturer had size and strength constraints a less common bearing may be more suitable.

When machine manufacturers are designing their parts that the bearings will mount/mate to, they will typically machine their shaft and housing dimensions to match a commonly available bearing.

When GLU (guys like us or maintenance/repair folks) are working with/repairing existing machines, the bearing mating parts are first measured, then an appropriate, commonly available (with any luck) bearing is ordered/selected to fit the application.

As you can see by the other responses, the buffing arbor shaft was machined for several reasons. Positioning the shaft's length relative to the arbor's housing primarily, and to fit the bearings' I.D.s, and to use common mounting hardware.. 1/2" nuts. The shaft remains stronger in the center (providing shoulders for the bearings), utilizing a commonly available pulley I.D., held in place laterally, and commonly available mounting hardware for the mounted accessories.

If the shaft were just 1/2" diameter, shaft collars or some similar hardware that wouldn't compromise the shaft strength would be required to lock or locate the shaft in place (more hardware generally means higher manufacturing cost), and the bearings would've been more expensive.

The arbor manufacturer had a responsibility to market a fairly safe product, one that wouldn't fly apart, safe enough for a DIY-type to take home and use.

There are ball bearings that have integral locking features for securing the inner races to shafts, but they are more expensive than an ordinary ball bearing assembly.

The machining that was performed on the arbor shaft wasn't an expensive operation, performed on automated machines it was probably completed in probably less than a minute from raw barstock to a finished, threaded part.

WB ......... metalworking projects

snippage

Who knew that such a simple piece of equipment would provide so many lessons in engineering and associated history and economics? Thanks everybody for such a detailed explanation. Some of the links provided are useful as reference. Again, my Google technique has proved deficient as I have been looking for them or something similar without success for the last two weeks.

In the final analysis the whole thing cost me $5 and most of the procedures suggested seem to involve greater costs than I can justify at present. I shall just keep looking at it 'cos it's pretty and an idea for good use will come to me in time :-)

Michael

Lots of education and a great experience and for only $5.00. Now get brave and take it apart and straighten the shaft and reassemble it. And put the arbor to use.

Bob AZ