Bearing running temperature

[JeanMicheling]

Hello bearing lovers,

I'd like to discuss another phenomena with all of you. I think we all know that running precision bearing too high in temperature is not good (over 50 C) but what about them too cold. For angular contact bearing, if your bearings run too cold, it could mean that there is not enough preload. But what if the run-in final temp is ok let say 35 C (95F). I put that spindle on the field where the ambiant temp. is about 8 C. The spindle ran 4 hours loaded at 3600 RPM. The final temp. is 16 C (60 F) which is really too low to me. But since the temp. is the same on both end of the spindle and they have the same bearing and the temp is about the same for all the spindles on that machine, I can't say it's a preload issue or whatever could come from the spindle. I'd say it's because of the ambiant temp. (8 C) and the spindle's casing which is really thick. So this casing is cold down during the weekend and when they fire the spindle back up, it can't warm up enough because the casing dissipates too much heat. I'd like too know if anyone have ever had this kind of operation at low temp. and what are the consequences?

 

[EnglishMuffin]

You must measure the bearing temperatures close to the bearing races for the readings to mean anything. The usual method is to drill a hole into the bearing location bore so that a thermocouple/thermistor/RTD etc can be placed directly in contact with the outer race, usually in conjunction with some thermally conducting paste. What usually matters is the temperature differential between the rotating and non-rotating parts of the spindle, not the temperature itself - although possibly oil viscosity considerations might apply in extreme cases. If the design is such that higher temperatures do not result in more preload, then you generally don't have anything to worry about. But bear in mind you have to consider the radial and axial expansion effects - under transient conditions, radial expansion can be more significant than axial.

 

[JeanMicheling]

How can we set up a numerical analysis of the spindle behavior under different temperature conditions? Bearing books just say that you have to be careful but they don't tell how to do it. I know my housing temp. and my bearing running temp. so with those inf. I think I could have an idea of the impact on my spindle? I'll give a try and let you know what I found.

 

[Tmoose]

Are you having failures, or just concerned?

If you give the bearing manufacturer's tech dept your running temperatures and shaft and housing fits they can calculate the operating preload, and the spacer offsets necessary to make it right.

A rapid temperature rise in a given time can be a powerful indicator of impending runaway bearing preload. I'd limit that rather than a temperature limit. Gotta get close to that outer race.

I'd be making some kind of vibration measurement on the housing during warm up. Good old Spike Energy (IRD trade name) is pretty good at detecting the squeaky scratchy scream of bearings being squeezed to death and dragged through their oil film. Most other analyzers have functionally similar features.

 

[EnglishMuffin]

Yes, Tmoose is absolutely right - in fact, when you do get runaway, it is so fast that even an automated rate sensing shut down often cannot catch it. But from what you have explained about your new spherical bearing design with lots of clearance, it may be that thermal runaway is not going to be an issue. The ultimate software for investigating transient thermally induced bearing preload is probably that recently developed under a government grant by Poplawski Associates, which works in conjunction with ANSYS FEA software. It would probably produce more comprehensive data than you would get for free from any of the bearing manufacturers, but you may not want to spring for that much money. I think a PC version of the old finite difference mainframe program "SHABERTH" is available in the public domain, but it is very hard to use.

 

[JeanMicheling]

I took my measurements and here are the results :

On the bench test

Side Outer ring temp Housing temp

Fix 102F 95F
Free 100F 94F

The housing temp I recorded on site was :

After the week-end before starting the machine : 46.4 F
After 4 hours loaded : 60 F

So I can say I have around 40F difference between outer ring temp and onsite housing temp. But another way to see that is that because the housing dissipates a lot of heat, the outer ring on site temp. is maybe lower (around 65F). My grease has no problem running at that temp but my tolerances are not designed for that. Could it be problematic? I'll have a call to the manufacturer and let you know what I got.