Press fits for bearings

[Simon CollierParticipant @simoncollier74340]

I have a set of commercial, CNC machined wheels, axles and axle boxes with ball bearings. The bearing OD is 35 mm, ID 15 mm, 10 mm thick. These are a close sliding fit in the axle boxes. I measure the ID at 14.99 mm with internal mic. The axles are 15.03 mm. This seems excessive for a press fit. I have minimal experience with press fits, but I’m sure plenty of members here can advise.

[Simon CollierParticipant @simoncollier74340]

[David George 1Participant @davidgeorge1]

Hi Simon have a look on here.

https://www.skf.com/uk/products/rolling-bearings/principles-of-rolling-bearing-selection/bearing-selection-process/bearing-interfaces/tolerances-and-resultant-fits

It may be that you heat and shrink when cool rather than just press on.

David

[Martin ConnellyParticipant @martinconnelly55370]

The most recent bearing matching parts I have made have been designed (by someone else, not me) with a H8 (clearance) fit. For a Ø35mm hole this gives -0 +.039 and the Ø15mm shaft -.016 to -.034. However the chosen fit needs to match the materials and working conditions they will be subjected to in use. Since the thermal expansion of different metals varies different fits may be required for parts subjected to high temperatures in use. You don't want hard to access, or change, bearings to slip in their housings or on the shaft.

In order to measure the bores for the bearings I use a Bowers micrometer bore gauge, other methods are not as easy to use to get an accurate measurement.

Martin C

[Nicholas FarrParticipant @nicholasfarr14254]

Hi Simon. it's all to do with designs and applications and I don't know the exact answer you need. However, when I used to service gearboxes and worm drives the bearings were always a press fit on the shaft and a close sliding fit in the housings, some you could knock on the shaft with a suitable dolly, but most needed a hydraulic press although you could use controlled heating for the bearing to expand them so they would just slide on and shrink onto the shaft. But on vibrating screens and the like, the bearings were a press fit in the housings and a sliding fit on the shafts. I don't remember anything that had a press fit on both the shaft and housing or a sliding fit on both, but overall the bearings on most things were a press fit on the shaft and some were so tight they had to have extreme high pressure oil injected between the shaft and bearing while bearing was pressed on and oil had to also be injected to remove them, but these were few an far between and the company I worked for didn't have the equipment to do that, so those jobs were farmed out. I agree with Martin that it's all about keeping the clearances correct between the bearing elements for whatever their use is.

Regards Nick.

Edited By Nicholas Farr on 26/07/2022 08:45:06

[HopperParticipant @hopper]

You'll soon tell if the fit is too tight a fit. The bearing will tighten up on installation. If that happens, you can press it off again and reduce the shaft diameter.

[Ex contributorParticipant @mgnbuk]

You'll soon tell if the fit is too tight a fit. The bearing will tighten up on installation. If that happens, you can press it off again and reduce the shaft diameter.

Or use a bearing with more internal clearance.

On MZ motorcycle engines both inner & outer races of the bearings are an interference fit, requiring the crankcase to be heated to 120 C to fit the bearings, then the bearing inner race heating with a pre-heated plug to fit the crankshaft. The bearings need to be C4 clearance, as the effect of the interferences reduces the bearing internal clearance – using a "standard" clearance bearing gives noisy running & shortens bearing life..

From memory, the SKF bearing application manuals give indications of the degree of interference fit required for both housings & shafts for different applications – reciprocating loads or drivers required heavy interference on both IIRC to prevent the bearing from rotating in the housing or the shaft rotating in the bearing.

Nigel B.

[Simon CollierParticipant @simoncollier74340]

It is only for a 4 wheel driving truck so not a demanding application. The bearing pockets in the axle boxes are blind, so there could be no pressing anything off again with any ease. I had a tentative feel of the fit with my 3 ton arbor press (mostly used for macadamia nuts) and I think I would have to swing on it to get very far. I have not heard of heating bearings.
Are the press fits designed to also expand the outer race to lock in the housing?

[Nicholas FarrParticipant @nicholasfarr14254]

Hi Simon, no you wouldn't want the inner race to expand enough to squeeze the outer race into the housing, the sliding fits are to allow expansion during an increase in temperature when running, otherwise they are likely to seize up. It's all a bit technical and you have to know everything about their usage to design for the best performance etc. which I don't have, all I had to know is which bearing was used where in my old job, which was all documented in the maintenance logs.

Regards Nick.

[Clive FosterParticipant @clivefoster55965]

Simon

As you have commercial ball bearings it should be possible to unearth some official manufacturers specifications for that particular bearing.

Probably need your Google-Fu firing on all cylinders but I imagine some carefully worded searches along the lines of "shaft fit for bearing no ….." may be more rapidly productive than hunting through online catalogues. That said the data is usually relatively easy to find in the big "books" but understanding can be a different matter as the listings are for professionals with more understanding of what it all means than home shop guy.

Clive

[Martin ConnellyParticipant @martinconnelly55370]

Bearing heaters are a thing you can buy. SKF make them as well as others. We used to use them where I worked, we used a contact thermocouple on the bearing to check when it was up to the required temperature. It worked by acting as a transformer using the bearing as the secondary winding. This induced a high current/low voltage in the bearing. The energy put into it was based on the current squared times the resistance, watts= I²R

Martin C

[Ex contributorParticipant @mgnbuk]

I have not heard of heating bearings.

A widely used technique – SKF and others sell (ferociously expensive) induction heaters specifically for the purpose of heating inner races without using localised heat sources that could distort the race or interfere with the heat treatment. My previous employer used to hire one in from the bearing supplier to fit headstock bearings to large lathe spindles. The heaters we used had a temperature sensor that was attached to the race & the heater warmed it to a preset value. The other option was to use an oil bath, but that becomes awkward and messy (with attendant fire risk) with bearings of 2-3 foot bore.

The MZ manuals give dimensions for the inner race heating billets for the various bearings in their workshop manuals.

Another option (where heat isn't an option or appropriate) is deep cooling in liquid nitrogen. We used BOC Cryoservice (sp ? ) for that. Pop your shafts in liquid nitrogen for a minute or so & they would fall into the bearing !

I think more of "interference" fits rather than "press" fits – I don't particularly like pressing bearings if I can find another way (a bit of heat usually). The interference is to stop movment between the bearing and housing / shaft in service, so that the bearing carries to load & the housing / shaft don't get worn.

Nigel B.

[Simon CollierParticipant @simoncollier74340]

It is even more technical than I appreciated. Thanks for the education. Bearing heaters! As it is just a basic riding truck, I might turn the shaft down to an interference of 0.015-0.02 which should push in easily. By the way, the bearings have only numbers, no name.

[David JuppParticipant @davidjupp51506]

Bearing fit adhesive may be an alternative approach… also more forgiving if you machine away slightly too much material.

[ChrisLHParticipant @chrislh]

Consulting my admittedly ancient SKF catalogue your shaft needs to be between 15.005 and 14.997 mm. Quite a challenge for the measuring equipment even before you start removing metal ! Bearings are made to international standards so this applies whatever the make of bearing assuming a reputable manufacturer.

[John AbsonParticipant @johnabson65530]

Just a note of warning from an old SKF guy. Heating the bearing to ease assembly on the shaft is perfectly fine, but don't go over 120C. And needless to say (so why do I? ) press on to the shaft through the inner ring and not the outer ring.

[Martin ConnellyParticipant @martinconnelly55370]

This is the bearing heater we had at work. Not much to see as the cover is on it but you can see a couple of laminated bridging pieces on the top. Inside there was a U shaped armature with a (hidden) coil around it. You put the bearing (or hub in our case) over the bridging piece and placed it onto the U shaped armature to complete the magnetic circuit. We did not use it for bearings, we used it to fit the coupling hub to large motors, an example of which can be seen in the photo. If it did not go on properly it had to be ground off and a new one put on so best do it right first time.

Martin C

[Simon CollierParticipant @simoncollier74340]

Sounds like a high stress operation Martin. Reminds me of heat shrinking a driving wheel on and having it grab too quick before it located. I had to drill and bore the 4140 axle away and start again. Easier than grinding though.
Thanks ChrisLH, a simple number is what I wanted. And I happen to have some bearing seal but it wasn’t bought for bearings and is quite old.

[Neil LickfoldParticipant @neillickfold44316]

It does pay to measure the ID of the bearings. Some can be either 0.02mm larger than nominal diameter, while others can be 0.01smaller than nominal sizes for the no name brand bearings. Been caught by that one just recently, and fortunately the shaft was oversized for the oversized bearing and then needed to be skimmed for the new brand name bearings to fit.

[Simon CollierParticipant @simoncollier74340]

The job is done. The 5C collets got a rare outing. I turned the shafts to 15.005 mm. Bearing I.d. were 14.99 mm. Yes I checked with 5-25 internal micrometer. This gave a very satisfying feel with the arbor press, needing a firm but easily accomplished lean on the handle.
Thanks to all contributors.

[bernard towersParticipant @bernardtowers37738]

I think that fitting bearings by press fit at home is a task that requires a bit of luck as the greater percentage of us do not own certified inside and outside measuring instruments. the thing to remember is that the bearing is near frictionless so the fit does not need to be as tight as our mind would have us believe. The problem with overnight fits is that it shortens the life of the bearing by quite a dramatic amount. A tech man from SKF once told me that for small races (under 1 inch od) it is better to have a sliding fit and a DROP of loctite so there is no compression on the bearing case.

[Author]

Posts