slip in a cylindrical roller bearings

[alexsurf]

How could i determine a slip in a cylindrical roller bearings ?
in spectrum analysis or whaterver ?

zzlinks@hotmail.com

 

[electricpete]

Maybe you are talking about skidding? That can occur when there is not enough radial load (such as a motor designed for belt application run uncoupled without radial load).

It may or may not show up as a high-pitched squeal, or as a bearing fault in vib. (Or it may be intermittent and missed by vib).

It may be avoided by adding load (4% of static rating). Or also possibly by special lubrication preparation of the bearings.

 

[electricpete]

... also ramping the motor up to speed with variable speed start is better than d.o.l. start to prevent skidding.


All of the above comes from a Toshiba article on this subject.

 

[machinerys]

A Defect Frequency at No of balls X Rpm can be an indication of Slip in the race ways.

 

[electricpete]

Is the previous comment based upon experience or from physcial reasoning?

From physical reasoning I would expect that if the balls all skid at the same point in space, we would see BPFO=FTF*Number of balls. That's because the balls rotate around the bearing at FTF, not at 1x.

 

[alexsurf]

The machine is an electric motor 1000kW.

This bearing is a spherical roller bearing, so there is only rollers not balls.

DO you think the best way is to see the time signal to search sometimes high impacts at the roller's frequency ?

 

[MachineryWatch]

alexsurf,

The terms rollers and balls are typically used interchangeably for the rolling elements in rolling element bearings and the formulae used for analysis. I wouldn't expect to see impacts spaced at any particular calculated frequency if light shaft loading is the cause of the skidding. It would be too random for that in my opinion. You may have success looking in higher frequency ranges for the excitation of individual bearing component natural frequencies.

As for number of rollers times RPM, I have never seen that frequency in 25 years of data collection and analysis. That doesn't mean that someone else hasn't seen it. But think about it. It almost requires that the rollers all be welded solid to the inner race in order to generate that frequency.

Skip Hartman

http://www.machinerywatch.com

 

[alexsurf]

I am agree with you Skip Hartman.

I think if there is only skidding without rolling defect, it is impossible to see some frequency.

But if there is any defect in the bearing conponents it is possible to diagnostic a skidding phenomena.

the caracteristic frequencies of bearing conponents will be lightly higher or slower.

Are you OK ?

 

[Hatch]

Here is my suggestion. It depends on wether you want an automated solution or a manual visual solution.

For a manual solution and assuming that the machine operates at a constant speed, I would mount either reflective tape and a photoelectric sensor or a stud/hole and a proximity sensor on the shaft. This will give a once per revolution pulse. Then you look at the pulse train in the time domain. If the time between pulses is constant, either there is no slipage or it is constant. If the time between pulses changes, then there is slippage.

For an automated soulution, you can mount a shaft encoder or position sensor to do the same thing, but have the computer calculate the time per revolution.

Other variations may be required if the speed is always changing, but you can handle most of that if the data is reviewed manually. C. Hugh

 

[acmotorengineer]

for Alexsurf

If you have a vertical motor with spherical thrust bearings, you must have axial load, typically 2000 to 4000 pounds for successful operation.

no axial load, and the bearings skid - one skid and the bearing is damaged and will have short life.

 

[electricpete]

Call me picky, but I'd like to respond

Hatch - maybe your technique may detect belt slip or other problems, but I don't think it will detect bearing skidding.

acmotorengineer - What you say is true, but we are talking about cylindrical roller bearings.