change C3 clrearance to CN (by change the bearing seat ) 8

[smaty]

We have a centrifugal water pump, 2900 rpm speed.
The water temperature is 20-30 centigrade degree.
In original design (KSB Germany) this pump has two 6413 C3 bearing with 0.023-0.043 mm radial clearance.
Now we can't find C3 bearing. Can we use CN (0.008-.0.028 mm radial clearance) bearing
by change the bearing seat tolerance on shaft/housing? The shaft tolerance on shaft is 65k6 (+0.002, +0.021m) and in
housing is 160K6(-0.021,+0.04mm).we use induction heater for mounting bearing on shaft. Thank you

 

[israelkk]

From your data it is seems that the CN bearing is more accurate therefore you can use it without any change to the shaft and housing dimension. When you heat the bearing you should be very carefull not to temper the bearing steel heat treatment.

 

[Hippo41]

Whoa!
C3 bearings (i.e. radial internal clearance greater than standard) are usually fitted to equipment where large temperature differences bewteen inner and outer ring can be expected on start-up. The inner ring usually expands a little quicker than the outer, therefore the internal clearance of the bearing can drop. Without sufficient clearance to start with (i.e. C3), the bearing may end up in a preloaded condition, which will create heat, therefore even more expansion etc etc until early failure.

If your machine has a C3, I would suggest that you get some.

You can relax the shaft and housing fits, but that seems to be a 'long' way of solving the problem.

Can't believe that 6413 C3 are not available to you. There are hundreds listed on the BearingNet.

Hot fitting will not damage the bearing steel in any way properly controlled. If you have to heat your bearing much more than 80C above ambient I would be surprised.

Hope this helps!

Lester Milton
Telford, Shropshire, UK

 

[rnd2]

smaty
You should read carefully what Hippo 41 says " Without sufficient clearance to start with (i.e. C3), the bearing may end up in a preloaded condition, which will create heat, therefore even more expansion etc etc until early failure."
There was good reason the original design incorporated C3 and Hippo 41 nailed it.
Good luck

 

[smaty]

Dear Hippo41, thanks for your help.
you Can't believe that 6413 C3 are not available, but believe because I live in a conutry that cant find C3.
Here in markets there are various kinds of bearings with different marks that may be many of those counterfeit and we can’t trust to those. I have problem to find C3 really, and we haven’t any possibility to buy from foreign countries. In your opinion, if a pump has 1450 RPM, Does it need C3 clearance for start-up?
Is pump among equipments where large temperature differences between inner and outer ring can be expected on start-up?
thank you
Smaty

 

[Hippo41]

Can you tell me where you are?

If you really can't get C3 then you have no choice other than to use what you can get. You may not get the problem developing - but if you do you'll know why!

Can you monitor the bearing temperatures on startup at all? Then you will some idea of what is going on. Remember the bearing temperature will climb rapidly on start-up, hopefully reach a peak (I would suggest you set a limit of less than 100 Celsius otherwise you will be cooking your lubricant!) and then settle back down to a normal operating temperature - possibly around the 40 Celsius mark. (or wherever it was before if you have that information).



Lester Milton
Telford, Shropshire, UK

 

[smaty]

I'm form Iran.
I try to monitor the bearing temperatures on startup.if i get an aanswer i'll tell you.

 

[Gerry45]

Is this pump a one-off unit undergoing maintenance ? If so, the chances are very high that fitting a CN will be quite satisfactory. Its unlikely that your shaft and housing will be in 'maximun metal condition' and the bearing bore and OD certainly wont be. Do the sums for the actual parts and you'll find at worst probably only a few microns of negative clearance in the brg after fitting. This is quite ok. Thermal runaway needs a lot more than this to give trouble. (If your housing is aluminium then its even better when your operating temperatures are taken into account.)

As regards fitting temperatures, there lots of unnecessary worries about this too. Most brgs are tempered to operate at around 125C. Above this, and you will start to create dimensional changes within the brg rings. However depending on the temperature, you will have to maintain it for hours or days to get any significant growth or shrinkage effect. Days and hours are easy to achieve in an application, but I doubt very much that you fitting process will take more than a few minutes !

Unfortunately, there are lots of old wives tales circulating in the bearing world, some of them even proliferated in bearing manufactures catalogues !

Gerry

 

[EnglishMuffin]

At the risk of being branded "an old wife", I disagree with Gerry45. I would not be so brazenly confident that "only a few microns of negative clearance is quite ok" and that "thermal runaway needs a lot more than this to give trouble". The inner race expands much more rapidly than the outer, especially during transient start up conditions. My experience, although somewhat anecdotal, is that once you run out of radial clearance at high speed, rolling element bearings are likely to become toast, although every application is different. But perhaps Gerry45 has some detailed theoretical and/or experimental data to back up his claim.

 

[Boomerang]

Gerry45 delivers an interesting discussion point. By definition, old wives last longer. I cannot resist speculating that it is so because they stay with what they know @

 

[rnd2]

Smaty says quote "We have a centrifugal water pump, 2900 rpm speed." end quote. Indicates a shaft speed above 9m/s. Even neutral clearance at start up is an accident waiting to happen.
Hippo41 & English Muffin are correct.

 

[electricpete]

I have watched this thread with interest. My approach has always been that internal clearance and fits should be strictly per OEM or other standard recommendation unless there is a thorough evaluation of the risks of changing it. With that said, I was also under the vague impression that C3 was intended for applications where higher temperatures may be expected, and the CN might[/] be acceptable alternative here.

On the subject of clearance or interference during operation, skf offers the following:

‘The radial internal clearance of a bearing is of considerable importance if satisfactory operation is to be obtained. As a general rule, ball bearings should always have an operational clearance which is virtually zero, or there may be a slight preload. Cylindrical and spherical roller bearings, on the other hand, should always have some residual clearance – however small – in operation. The same is true of taper roller bearings except in bearing arrangements where stiffness is desired, e.g. pinion bearing arrangements, where the bearings are mounted with a certain amount of preload, see under "Bearing preload".’

EM/rnd2 – what is the source of expansion during startup you mention? Bearing frciton heat?
Also, I have never heard speed mentioned in the context of requiring looser internal clearance. Is that also due to bearing friction heat?
just curious to learn a little more.

 

[electricpete]

please ignore the bolds.

 

[DrummerBoy]

Electricpete is correct. A slight preload is best. but slight is the main word. It quickly moves to a less life condition. C3 is also normally recommended when the speed exceeds around 75% of the speed rating of the bearing.

 

[EnglishMuffin]

Electricpete:
The thing is that if the speed is higher, the heat generation rate will be higher. But at high speeds, ball bearings generally have a greater heat generation rate on the inner race, if there is a contact angle involved, because the spin velocity is higher there. There is also usually less opportunity for heat dissipation on the inner race. So consequently the relative expansion of the inner race will be greater and the operating preload will increase. As has been mentioned above, this can then lead to a vicious circle of "thermal runaway". As has also been mentioned, the fit of the outer race can also come into it. Using FEA, it is possible to study this effect analytically in any given case, using FEA, but it gets rather involved becuase it's a coupled field situation. Poplawski associates has developed a program that runs in conjunction with Ansys, to solve this problem, and they have done a number of studies for people in special cases. There is a much older program called Shaberth, which used finite differences to do a similar thing. Most of my experience in this field has been with machine tool spindle bearings, where the speeds tend to be higher than run of the mill electric motors. I have come across cases where even spring preloaded angular contact bearings ran out of radial clearance at high speed - and they have failed instantly.
You can certainly run with preload in the case of angular contact bearings, of course. But your SKF excerpt is a bit of a cop-out where they say "As a general rule, ball bearings should always have an operational clearance which is virtually zero, or there may be a slight preload". The use of the word "operational" lets them off the hook, rendering the statement almost valueless from a practical standpoint.

 

[electricpete]

thx em. You get a star.

 

[rnd2]

star to Hippo 41 for making the call
star to English Muffin for the elegant explanation
star to Gerry45 for the phrase "thermal runaway"

 

[Gerry45]

Being and 'old wife' is really nothing to be ashamed of. I have been one myself on numerous occasions. But as with alcoholism I tell myself, admission is the first step to recovery! So anyway, back to CN versus C3 debate and what is the basis of my brazen statements .....

1)Deep groove / Conrad type ball brgs made by good manufacturers generally have bore and OD tolerances at the middle of the tolerance band (+/- a couple of microns). There is a little more 'spread' on the RIC but even this wont reach anywhere near 'catalogue' limits.

2)The processes used to make shafts and housings are by no means as well controlled as brgs, but the likelihood of a single machine taken in for repair having both shaft and housing features on a 'max metal' condition is quite remote.

3)normal surface quality on shafts and particularly housings means that, during the brg fitting process, significant amounts of the theoretical interference is lost.

4)Machines which have been in service will often exhibit additional wear/smoothing of the seating surfaces.

5)Even in a brg that is 'theoretically tight' after installation, normal elastic characteristics mean that when an external load is applied, a clearance zone is re-created within the brg.

6)In my experience, assumed shaft~hsg temperature differentials of often exagerated for safety due to the difficulty in attempting to predict or measure them.

Ok these are just general points which may or may not be applicable in every case. But certainly not taking into account the 'real' dimensions of the parts when doing fit calcs can be misleading.

The one issue I haven't addressed yet is the conditions required for 'thermal runaway' and my statement that 'a few microns' will not cause trouble. This is probably because I dont really understand the situation myself!!! But it is the case that 'moderate' speeds and 'moderate' negative clearances in ball brgs can and do work successfully. If they didn't, lots of other things wouldnt work either. Take machine tool spindles. These are often fitted up with pairs of angular contact brgs clamped together in an axially preloaded condition. From the rolling elements point of view, they cant tell the difference between an axial or radial preload, a zero or 15 deg contact angle. All they know is that they are trapped tightly between raceways. After startup, temperature differentials develop across the brg and things, theoretically at least, should go from bad to worse....except that it doesnt !

In the automotive world, there are lots of applications where conventional theory has been ignored without disasterous consequences. Some alternators for example, incorporate C2 clearance brgs(ie less than normal) heavily interference fitted into aluminium housings (to make sure that they dont become loose at 100 degs C). According to the theory, this arrangement shouldn't work either, but is does do, even when you freeze them down to -40C !

Im not suggesting that 'thermal runaways' never happen, just that they dont happen as frequently as the world would generally predict. I remain convinced that provided nothing dodgy happens on the lubrication side, Smaty will get away with fitting a CN. He'll probably also find that his pump becomes smoother and quieter than its ever been!

Gerry

 

[unclesyd]

Smaty will need to get the best lubricating oil that is available.

 

[EnglishMuffin]

"Take machine tool spindles.........After startup, temperature differentials develop across the brg and things, theoretically at least, should go from bad to worse....except that it doesn't !"
"From the rolling elements point of view, they can't tell the difference between an axial or radial preload, a zero or 15 deg contact angle."

I don't agree with either of these statements. In the case of the first, I can assure you that "it can and often does". That is why many high speed spindle bearings are spring or hydraulically preloaded, even though this unfortunately reduces stiffness. Even allowing the whole bearing to float in the housing can still lead to failure if the outer race expands enough, and spindle manufacturers have had to resort to mounting floating bearings in linear roller cages. However, in many of the less demanding cases you can get undoubtedly get away with it. There are also a number of cases that I have come across in gearboxes where any amount of excess preload does not seem to lead to thermal runaway, apparently because of the flexibility of the casing. Every situation is different, and I just don't think you can make blanket statements like that.
In the case of the second statement, there is often a big difference in practice between a zero contact angle and a non-zero contact angle, because bearing assemblies are usually much stiffer radially than axially. I stand by my comment that once you run out of radial clearance at high speed (ie achieve a zero contact angle) - watch it !