Bearing Housing Tolerance

It is a very rare condition (I am yet to encounter one really) in case of SRB or TRB to have both cup and cone with interference. You may like to review.

The recommended fits (preferable) would be

p6 (very light interference) or n6 (moderate interference) in case of interference fits on both shaft or housing (capitalise in case of housings to P6 or N6)

The sliding clearance fit would be G6 (sliding fit with just a bit of clearance may be 0.01 or so for your diameter and maximum may be some where around 0.05mm) or H6 (very close sliding fit however minimum fit will be 0 ie both shaft and bearing ID may be same size and maximum will be around 0.04mm)

Unless we exactly know the application it will be difficult to recommend the fits. However an approximate recommendation will be found from any bearing hand-book

It depends also on type of load- rotating or not rotating and other factors. Also in case it is a machine tool or some similar application, where the precision is critical, the total requirements change (and I lack my expertise in that field)

The best way to solve your problem is to go on GOOGLE and try to find the technical books of SKF, FAG, Timken, INA or other bearing manufacturer they give ALL indications about tolerances as function of all possible parameters. Not wanting to be rude with participants but no one of us knows as much about this subject as bearing manufacturers know (the good ones and not the retailers who also look in the catalogue but some times forget to give all further).

You discuss fits for bevel bearings and the OP concerns spherical and double ball bearings, as far as I know working conditions are not at all the same. What he wants is to have one bearing fixed axially and the other able to slide in the housing if temperature changes. It is an usual design in order to avoid axial forces due to different elongations (temperature gradients and different coefficients). For such a case the bearing catalogue of SKF gives I am sure all required information. Bevel bearings that is an other story, in many catalogues one finds even the way to compute the pre-load at assembly. What is the way they are mounted ? Better make a sketch.

The problem is which one is defined as the axially fixed point. The other one must have on the outer diameter a sliding possibility.

In his example both have interference which will build up axial forces when a temperature gradient occurs between shaft and housing as usually is the case. Both bearings are sensitive to axial forces but the ball bearing could be the most sensitive. If the ball bearing is fixed then the roller bearing should have a sliding facility.

I have the feeling that the roller bearing is at the "outlet" (higher radial portance and tolerant to shaft angular deviations due to radial loads) so that it has to be fixed axially and the roller bearing is at the shaft end. In this configuration the ball bearing should be sliding and in no case with an interference.

Unfortunately SKF.com is in maintenance and the paper has neither nothing new nor the required details to estimate values.

may use this as the configuration ;

fixed 22212 N6-P6 outer

floating 3212 H6-H8 outer

i dont yet have data on the RIC for both bearings

output load on right hand end , direct coupled motor on left.

ideas people ?

Your 3212 is the locating bearing, retained in position by the lock nut since it seems that it is seating on the shaft collar (from your sketch.

22212 is tightened by lock nut till it abuts on the spacer (maintaining the spacing between the two bearings)

Is that right?

What is the type of load? Rotating Load? Vibrating ? Shocks ?

These will define what should be the interference fit - Outer race or inner race. (Note 3212 may still have inner race with interference fit. We have that in lot of our equipments - Lock nut is not used for locating, rather retaining it in position.

Even your sketch shows the thermal expansion gap for 3212

So The housing of 3212 should he usually H6 (if you can achieve) or H7 (however in case of the sliding action It could have been G6. The shaft could be easily n6 or p6 depending on the load and shock intensities.

The SRB could be more depending on load type.

What about the OEM recommended fits?

thanks

the load is 150 kg wheel @ 300 RPM which can have vibration due to imbalance and harmonics , radial force against the fixed outer is 4000 N maximum , and intermittent axial load of 3000 N.

the locknut is intended to stop the whole shaft being pulled out of the bearings axially and ties both bearings rigidly to the shaft to prevent spalling or spun bearings.

the locknut and thread has about 530,000 N axial load capacity , and pushes the 3212 against the inner sleeve , which pushes against the 22212 which pushes up against the shoulder on the shaft.

the 3212 in that drawing would be floating in the housing to accomodate temperature fluctuations , probably will give it around 0.3 mm room to float each direction in the housing.

From your picture still I have some doubts

a) When I looked closely the bearing 22212 is located on shoulder of shaft. The load being rotating and not too much of vibration, I expect a shaft fit of n6 should be enough.

b) The outer race seemed to be held in place by a retainer (cover) with seal on it. Since it seems to be the locating one, the cover step may need to be exactly fitted. I didn't check up the specification of 22212, it is a double row SRB with spacer between cups (ie with the bench endplay ensured by bearing manufacturer) ? In that case you may have to adjust the height of the cover/retainer to ensure that there is no axial play. To ensure assemble-ability of course you have to have H6 or H7 on the housing bore.

c) The bearing 3212 seems to be seating on the spacer as well as there seems to be step (shoulder) on the shaft? Of course it would not be possible for it to sit on both. So it must be sitting on the spacer which would retain 22212 axially and the 3212 would be held by the lock-nut.

Unless there is a taper sleeve, I would have preferred same n6 on shaft and may be G6/G7 on housing. If you warm the bearing while assembly the lock-nut may be tightened and the bearing gets located while still hot. Retightening of the lock-nut may be done after it cools down.

yes seal carrier is also a bearing locator , pushing against the bearing with 6 x M 12 bolts into the housing , bearing would be flush fit to it

yes the bearing inners touch the spacer sleeve and the locknut pulls them all together.

cheers

M.

From the drawing, which side is the working side of shaft? If I am correct it looks to be the left hand side! If this is so, then you may not want the expansion on this end. I would assume that your are mounting to drive motor via some sort of coupling, this would be the better end to take up the expansion of the shaft. But, thats how I see it. Not sure what your making or what it is to be used for.

no

left hand end has a direct coupled motor fitting to the shaft ( splined )

right hand end is the output end fitted with its rotating load , where the radial load is applied

hence i think the floating bearing is in the right place ?

Floating bearing is in right place since the as per my analysis (visual ) the SRB is the locating bearing and of course the coupling would be able to take the little thermal expansion and not stress the motor shaft.

1)-The problem has to be analyzed as follows: Main Load is mentioned as being the disk weight. Always vertical thus rotates with respect to inner ring, stationary with respect to outer ring. Secondary load consequences of disbalances, rotates with the shaft thus has a stationary position with respect to inner ring and rotates with respect to outer ring. Since in all rotary machines the disbalances are maintained to a low level its effect will be less important also because rpm are low. It is NOT known if other loads act due to the disk loading. If this is the case then same logics have to be applied. In the SKF cat you can find the indications and justifications of every decision step.

2)-Connection between motor and shaft: - is the rotor fixed on the shaft ? then the left bearing can be sliding within the gaps in the motor frame.

- the rotor is in the motor frame on bearings, then the left bearing can be sliding only if the coupling (which ever type) allows an axial relative displacement of the two shafts.

Depending on the work done with the disk the right bearing can or must be axially fixed.

If it is compulsory to have it as an axial fixed point then above conditions MUST be fulfilled.

3)-Follow the path in the SKF catalogue you will find , as I said at start, ALL indications and a fit calculator which shows how the choice of the tolerances influences the fit. It is a VERY helpful tool. Do not forget to verify the radial clearance in order to avoid overstressing of rolling elements and races.

Path in the cat: SKF.com / Products / Interactive Engineering Catalogue / Rolling bearings / Principles of bearing selection. / Application of bearings / Radial location of bearings / Application of bearings Radial location of bearings - Selection of fit - Recommended fits

4)-Remarks: - if you have a 55mm shaft the bearing MUST have at the end "11" and not "12" which corresponds to a 60 mm diameter. - as you designed the housing the assembly of the left bearing will NOT be easy. Better -if you still have a possibility - try to increase the outer diameter of the right bearing (22312 130x60x31 in place of 22121 110x60x28) so that you can assembly the shaft with the bearings and sleeve outside, introduce the assembled shaft in the housing and fix it in place with the cover.

Nick Name

yes you are correct 3211 is the 55 ID bearing , pardon my typo

the motor is female splined and can accomodate axial movement over the male splined shaft

undecided yet on the amount of axial movement to allow in the housing each side of the left side floating bearing

so thats next question , based on 4340 shaft and K1045 housing , in conditions where ambient temperature could be in the range of -4 to 47 Celcius , and heat input from the hydraulic oil passing through the hydraulic motor into the shaft could be up to 90 celcius.... what expansion do i allow for ?

the splined shafts will have copper based grease on them with a 3 cubic centimetres " reservoir" cavity at the end of the shaft to hold grease , the splined connection has a seal that stops the copper grease throwing out of the connection.

the output end is exposed to good air flow to dissipate heat through the large mating part bolted to it

thanks for the SKF links , both them and Timken are proving to have a wealth of information there

First of all I would consider the friction under load in the spline as axial load for the bearings as well for the motor as for the shaft. I would consider the shaft at the outside temperature as a safety measure and the motor at its highest value. In fact it is impossible to make an estimation without a complex computation considering all losses in the transmission and the nature of heat transfer at box surface so that for simplification one has to consider the worst case.

yes i think the friction in the spline could be a factor if the load was constant , but in this case the load is changing constantly , loading / unloading with variations in the torque load so i think the spline will be freed of axial pressures on the driven shaft.

the shaft expansion due to temperature fluctuations is going to be a slow phasing phenomenon in my opinion .

i am looking at using around 0.3 - 0.4 mm gap each side of the floating bearing