Attached you will find a set of data that was taken on a FAG triple race bearing. The call that I made is also stated on the attachment. What was found was actually a pretty intersting conclusion, but just for learning puposes I would also like some additional possibilities. This Type of bearing has given us problems in the past so this is a baseline test taken on a roll with this bearing in a test stand. The cost of this one bearing is in excess of $80K aside from the downtime associated with it if it fails in place.
This bearing was replaced with another bearing, and "magically" the same thing appeared.

We have already inspected this bearing along with the bearing reps and I will give their opinion later.

Scott

CCR1600.doc (86 Kb, 77 downloads)

One other thing to add,
There was also a BPFI showing up on the inner race of the inner bearing. So, both defects were showing up on both stationary rings of the bearing.

The data that was presented actually had the bearing reps baffled as well. Initially, their vibration analyst agreed with my findings. One other oddity that I failed to mention was that I use headphones as well, and that oh so familiar buzzing sound was present that you normally hear on a stationary race defect. Once the same freqs. showed up on the second bearing, we had to start thinking a little differently due to the odds of a 2nd minor defect on another bearing lining up with the load zone again. (maybe I should let those guys at the roll shop buy my next lottery ticket)

What is believed to be happening is the rollers are skidding each time they enter the load zone probably due to decreased roller speed when it is out of the load zone. Since the load zone and the speeds remain constant, with the exception of the rollers, the BPFO is generated, or the BPFI on the internal ring of the inner bearing.
This will probably eventually create a spall in the area where the rollers are impacting, but it's anyones guess as to how long that will be.
This also probably explains why in the past we have always started out with "bearing defects" on the rolls that we have to watch until they eventually reach higher levels.
I don't know if this is just the nature of the beast for this type bearing or something else. I was just curious if anyone else has ever run into something similar.

Scott

Scott,

I'm not sure I understand how this type of bearing works. Do you have any pictures or drawings?

The velocity spectrum does have an abnormality in that the defect shows up with the first harmonics being the largest rather than in the usual pattern of higher harmonics being larger. To me that signals skidding or out of round races. If this bearing works like I think it does, it probably takes a really big load to prevent skidding.

Scott,

I did not get it from your description, but was this phenomenon actually confirmed upon bearing disassembly? In any case though this theory sounds plausable because it is hard to imagine that all bearings start exhibiting this behaviour from the beginning.

In general, all antifriction bearings experience to certain degree spin acceleration-decceleration as they enter-exit the load zone causing slippage and race wear. In your case slippage could be agrrevated IMO due to:
- excessive clearance
- poor lubrication or
- insufficient static preload on this bearing.

All the above could be verified.

Your data I believe should and looks like does show modulation of BPFO by run speed due to rotating unbalance force. This is also I think is visible from autocorellated PeakVue TWF and spectrum.

There is also another set of harmonics, is it BPFI?

David

Danny,
This bearing has 3 rings(4races). The center ring has the inner race of the outer bearing, and the outer race of the inner bearing. The center ring is bolted to the rotating shell of the roll. The outer ring is the "carrier bearing", the inner ring is a slip fit onto an inner bar that has hydraulic pressure to control the crown on the roll.

David,
The first bearing was not disassembled. The bearing reps just tried to inspect it by looking between the rollers and you know how easy it is to miss something that way. This roll weighs in excess of 80K lbs. so, I personnaly don't think insufficient loading to be the cause of what I was seeing. I do believe that there is something there in terms of excessive clearance in this type bearing. This is an oil fed bearing.

In most all of my test stand data, there is some looseness present due to not being able to secure the rolls as tight as they would be in the machine.

The other set of harmonics that you see is the BPFI of the inner race of the inner ring. These harmonics are actually more prevalent on the opposite end of the roll since I can actually place my accelerometer on the center piece of the roll.

If I get any more info I'll pass it on.

Scott