Hi Everyone,
I'm new to this forum. I have a question about CSI Peak View. When a defect is detected is the bearing change out called by what is seen in FFT or FFT and TWF. I use SKF Enveloping and have found tha when the defect is seen only in FFT it is not as reliable to make a call. However, when the defect is seen in both FFT and TWF it is definitive. Thank you for your repies

In my experience with CSI PeakVue, we watch the trend once the defect shows up in the FFT. CSI says not to pull the bearing on the PeakVue reading alone, but to wait until it shows up in the regular reading. We work in a paper mill, and we have seen bearings fail before they reach the regular reading. We take readings once every 2 weeks. When we see a Inner Race or Outer Race fault in PeakVue, we watch the trend, and pull the bearing when the readings start to increase. When we do see faults in the FFT and the TWF, we do the same thing, just maybe a little sooner. I had a broken inner race a couple weeks ago that really showed up on the FFT and the TWF in PeakVue, but hardly showed in the regular reading. We pulled the bearing and the inner race had cracked all the way across. There was a small spall where the crack started and it went across from there. Hope this helps some.

Thank you so much for your quick reply Jimb,

I also work in a paper mill and before using demodulated was not reliable at all, until I over heard the use of the enveloping 4th filter FFT & TWF scenerio. Previously I only used velocity points to detect I/R crack and was relatively successful. The unfortunate thing is that the demodulated frequencies are not always excited and I have evaluate the velocity points. Another thing is that I have not seen the "angel fish" patterns at all in any of the I/R crack calls. Has any one seen this pattern in their calls
Again thank you for your response.

quote:

Originally posted by JimB:
CSI says not to pull the bearing on the PeakVue reading alone, but to wait until it shows up in the regular reading. We work in a paper mill, and we have seen bearings fail before they reach the regular reading.

Hmmm. I too monitor equipment in a paper mill and I have called bearings with HFD data alone (gSE similar to PeakVue). I wonder if CSI means PeakVue "overall" alone. Do you collect spectral data as well? I remember a time when I was told by my supplier to use gSE overalls only because the spectral data was unreliable. I now use gSE spectral data with high resolution and gSE time waveforms to detect bearing faults. If I see it in the velocity spectrum (normal data?)it's toast.

JimB,

I've heard that rule applied to both PeakVue and Spike Energy. Experience has shown me that it is only applicable on standard motor and fan bearings where you have good proximity to the bearing and high enough speed and load that the defects will show clearly in the normal spectrum and twf. I cover many motors with say a 6313 bearing that I track outer race defects in PeakVue for years before making a change. But if it shows up in a slow running press roll or a felt roll, out it comes and there is the defect.

Like all rules of thumb, there are exceptions.

Dan..thanks..that makes sense.
Lukas, below is what we show on the press rolls with a broken inner race in PeakVue. This was verified when we pulled the roll out and examined the bearing. I'd like to make the post smaller, but not sure how.

hi friens

what is TWF

bearing defects can be detected early at ultrasonic frequencies , at fail at nine stages
u can easly detect bearibg failers using FFT and time waveform .

TWF=Time Waveform. The raw data from which the spectrum is derived.

JimB,

I noticed the "H" inner race faults do not line up on what looks like the inner race defect or are the "mounds" something else?

Ralph,
Good observation. For some reason our speed in RBM wear doesn't remain constant. The speed should be 3015 fpm and 352 RPM's. Plus after we pulled the bearing, we noticed that it had been spinning on the journal, so even with the correct speed, the faults probably wouldnt have lined up. the first mound is the correct inner race frequency and sidebands.

quote:

Good observation. For some reason our speed in RBM wear doesn't remain constant. The speed should be 3015 fpm and 352 RPM's.

Do you have it (RBMware) set to something in globals that causes this?

Here is my setting:

Ralph, I did havbe one checked that you don't. I unchecked it and will see if that works.

Thanks for the advice

Bet it will work better.

quote:

Originally posted by JimB:
I had a broken inner race a couple weeks ago that really showed up on the FFT and the TWF in PeakVue, but hardly showed in the regular reading.

Jim

From the PeakVue the fault is obvious. Out of curiousity, could you post the regular data. Do you think that failure of the defect to appear in there has to do with the applied AP setting?

Thanks,
David

I recommend you rely heavily on the g-level in the PeakVue waveform, e.g., for the machine running at 352 RPM, the generic default alert level would be 1.5 g's and fault at 3 g's for an inner race fault. You are seeing greater than 7 g's. Also, if you would look at the circular plot (with correct RPM value) you will find a very interesting pattern for cracked inner race faults. Namely the impacting level will be low every where except in the proximity of the load zone.

I have seen many severity calls made in PeakVue based only on the fault only showing up in PeakVue and not ever showing up in the normal data. The key is in relying on the observed g-level and its trend.

David G'
Here is the regular Spectrum and Waveform:

As far as the AP set, this is the way I have it set: FFT 0-1000HZ,1600LOR/4Avs, Hanning Window. the bearing had been spinning on the journal, so I think that is part of the reason it didn't show up as much on the regular reading.

2ND_PRESS_BOTTOM_ROLL_REGULAR.doc (68 Kb, 33 downloads)

Robinson,
Do you have or know of a chart with the generic fault levels on them? Also, we have never done much with circular plots here. Our instructor touched on them in our advanced vibration course, but never got in depth with them.

Circular plotting or the waveform is a very good tool. I use it to show roll cover problems as well as gear and bearing problems.

There is not much to it. Just select a single waveform, turn off the trend and spectrum and select circular plot. You can display a single rotation or whatever you like. Just make sure your speed is right or it will not make any sense.

Great Thread,

Thanks

Danny,
You said, "There is not much to it. Just select a single waveform, turn off the trend and spectrum and select circular plot."

Click in the waveform or hit the down arrow key to move the cursor in the waveform, then hit the "F" key. Hit the "S" key to go back to your regular view. The default will stay on the circular plot as long as you use the S key to back out, otherwise it will change the default to "do not use."

David Eason