This post is on behalf of another person, so I don't have any data, just a procedure. Following below is his approach. He will appreciate yours.

" Set sufficiently high resolution in PeakVue.
Suppose sidebands are observed around the harmonics of bearing fault. If dryer roll is turning 200 rpm the sidebands should be normally spaced at 200 cpm because the bad spot on the inner race goes through the load zone
once per rev.

Let's say that the sidebands measure out to be in fact 190 cpm. That means that the crack is going through the load zone 190 times per minute. The only way that can happen is if the inner race is going at a different speed than the journal as the journal is turning 200 and the race is turning 190. In other words it means that the inner race is slipping 10 cpm. That's why you have to have high resolution."

Question:

How can we be sure that the inner race will turn at a constant speed of "190" rpm? Seems like if it were cracked and slippping on the journal it would be almost impossible to say it will turn a constant speed. Friction or the lack of it would regulate the speed of a slipping inner race wouldn't it?
Or is this a "1 time" phenomenon on this particular dryer?

Ralph raises a good point.

Even assuming the inner race rotates at a constant speed with respect to the shaft, then the BPFI will be altered (reduced), in addition to sideband spacing, wouldn't it?

This message has been edited. Last edited by: electricpete,

I agree with Ralph and pete that if the race is turning on the shaft, you should see harmonics of frequencies that don't match known defect frequencies. My theory is that you can calculate the turning rate by working that math backwards to calculate effective turning speed.

For example:

Given a known bpfi of 6.9 orders and running speed of 100 rpm, you find a peak at 670 cpm. 670/6.9=97.1. Therefore the race should be slipping at 2.9 rpm. In theory.

I have used this logic to call an outer race turning, but had no way to confirm the theory.

Good Luck,

Hey,
That is a direct copy/paste quote from an e-mail that I sent to a friend of mine this morning.

Trust me on this one, fellows; it works. If any of you have access to the papers presented at the Emerson Exchange in Orlando last year, you can look up my presentation on it and see the whole story.

David Eason

It makes sense. And a good tool to have in our toolbox if we see it.

Just curious - didn't the inner race frequency change as well? Maybe that is just not significant becase it's not as easy a change to pin down as the sideband spacing? (although if you see both that would increase the confidence in the diagnosis)

This message has been edited. Last edited by: electricpete,

Having a race turning on the shaft at a constant speed is a rare case, IMO.

It is more likely that slip speed is flactuating in a random fashion. But if resolution is high, then BPFI peaks should be smeared over several bins. Same is true in regards to 1x sidebands: they will be smeared and there are going to be multiple SBs as well.

The very fact of the above pattern may be a reliable indicator of a IR turning on the shaft.

There also have to be to some corraborating symptoms, such as looseness pattern in the spectrum and impacting present in the TWF.

This message has been edited. Last edited by: Dave_man,

quote:

But if resolution is high, then BPFI peaks should be smeared over several bins. Same is true in regards to 1x sidebands: they will be smeared and there are going to be multiple SBs as well.

Dave

I am assuming (which one should never do ) that you are saying here, that the inner race peaks should not align as harmonics if the slip is in a random fashion but should align if the slip is at a constant rate or if there is no slip at all. Same holds true for the sidebands?

Interesting information. We were lucky enough to have 2 bearings (130 mm bore and 400 mm bore) on one of our paper machines crack the inner race. Most unusual, to say the least. One of the indicators I look for is 2 sets of harmonics of running speed. When a race cracks, it becomes loose and generates multiple orders of running speed. By paying close attention to those spikes, you will see that there will be a couple sets of harmonics of running speed very close together. We also could see roller impacting in the time domain. App. 60% of the rollers would impact the crack each revolution which showed in the ring down. Was not a good week for that machine.

Another day in a papermill,

Gary B

quote:

Originally posted by Gary B:
When a race cracks, it becomes loose and generates multiple orders of running speed
Gary B

Gary, are you saying all bearings become loose when the inner race cracks? I have not noticed this always. I have not noticed this as some tapered bore spherical rollers bearings on our dryer cans show cracks but no looseness. When removed they dont slip easily off the journal either.
I think it is possible for a cracked inner race to actually tighten if the thrust is such that the bearing moves up the journal, further reducing clearance in the bearing.
I read some time back on this site about dominant 3x harmonics when running speed harmonics are also present could indicate looseness.
I began looking for this and I have noticed 2 dryer cans that I called for inspection because of increased running speed harmonics ( when I say increased I am saying "more" harmonics of running speed are present ) with the 3x being the highest.
Both bearings were loose on the journal, no cracks were present.They were loose enough that the inner race was beyond the end of the journal. Sleeving the tapered journals was completed and things back to normal with reduced running speed harmonics.
Has anyone ever noticed 2x BPFI showing up ? I have had 2 dryer bearings have 2x BPFI show up in my peakvue data. From past experience on these dryers if BPFI pattern is present in peakvue I immediate call for inspection for a cracked race. I wondered how 2xBPFI could show up and what both these bearings had in common was a solid cage with offset roller separators.
The rollers on each side were not in line. Is this coincidence or does this make since that this type of cage could produce 2x BPFI ?

One thing I have learned is there is sometimes never a "ALWAYS" will be this or that.

But this is what makes the job interesting!

Hope everyone has a good day, sorry for being long winded.

Mike

quote:

One thing I have learned is there is sometimes never a "ALWAYS" will be this or that.

Just like vibration analysis, that is perfectly clear

Danny

Mike,

It makes sense to me that 2X BPFI for a double row bearing with offset rows would be prominent. Each row is creating its own time trace at BPFI, and it wil look like 2X the impact rate, as if we doubled the number of balls. This is only for a defect that extends across both races, like a cracker IR on a dryer.

Then again, maybe I need another cup of coffee...

Bill,

Thanks for the response. The first time I saw the 2x BPFI show up in the dryer can peakvue data I was hesitant to call for inspection, but since it only takes removing the steam joint and pulling the bearing cap to inspect I wrote up to inspect on the next shutdown for cracked race.

The inner race was cracked and I noticed then the cage was solid one piece brass with offset roller design.

It hit me that with this design you could have twice the number of rollers passing the crack so it may make sense.

The next time I saw this I was more confident and the findings were exactly the same, cracked inner race with same cage design.

To reduce cracked inner races we have gone to a carburized inner race and C4 fit, while its fun finding the cracks its too easy to stop them from occuring with this type of bearing, and management would rather make product than cracked inner races



Have a good weekend.

Mike

This message has been edited. Last edited by: MikeH,