On rolls like this I would normally collect data with about the same settings as Ralph and Jim and monitor the spectra as I collect the data (if i am working with only one channel i will monitor the spectrum and waveform). If I suspect anything I wll collect data with fmax at 3000 cpm, 1600 or 3200 lor, 3 averages, and a 500 hz hp filter. Defects show up very clearly in both the spectrum and waveform and if you are precise enough in inputting your running speed match the predicted fault frequencies. Even in cases when they don't exactly match the frequencies, if you see a clear set of harmonics in the range of any of the defect frequencies, you have probably been given the wrong bearing data.

I cannot really tell from the 400 hz fmax pv spectrum you posted if there is anything there or not. Maybe if you expanded it and looked around inner or ouoter race defect frequencies you might find sometihing. Can you expand it and repost and include the timewaveform?

Good Luck,

Danny Harvey

Yes, I did set the RPM and check to see if my bearing defect frequencies lined up. They didn't line up with what I calculated the RPM to be. I know the diameter of the roll and I check to see what ft/min it is running prior to taking the data. Unless I am getting bad info to calculate I should be very close. It works out to be .91 Hz or 55 RPM. Also when I change my regular spectrum to velocity, I was able to find a spike at .91 Hz. I had wondered about the spikes in my PeakVue but when they didn't line up I wasn't sure of the significants of them.

Ronnie

Whew!! I answered one reply and before I could post it, two more came in. I wasn't expecting such a response but really glad to see such interest.
David G., I didn't see the bearing myself but was told there were very visible spalls in the outter race. If possible, I will see if I can get access to the bearing and take picures. But there's a good chance the bearing is scrap metal by now. I was curious about a lower filter setting? Maybe I should aquire some PeakVues with different filters to see what it shows.
Ronnie

Ronnie - you've been getting many good recommendations but, in my opinion as a professinal vibration test systems designer, the comments and recommendations by Jim Crowe and Ralph Stewart are especially good. I very much recommend following them. Ralph made a comment about calling inner race flaws in PeakVue. On large, low speed bearings, there's so much attenuation of the inner race signal compared to the outer race signal, that if you find anything from the inner race at all, it's probably a significant problem. Outer race problems are much easier to detect and the signals from outer race flaws may be much higher without being so much of a problem.

Duncan,
I looked up the Low Frequency accelerometer I am using, it's a Wilcoxon model 799 LF, and it does have a low pass filter. Although it didn't list the setting of the filter.
Thanks,
Ronnie

TA has a very good book on Peak Vue
Also never take more than one average when using Peak Vue

Brad

This is a follow up to show the results of following the advise of all those who pointed me in the right direction with PeakVue on slow speed rolls. I am attaching a spectral comparison showing before and after changing my setup. I am now using a wilcoxon 793 accelerometer, 100 Hz fmax, 800 LOR, 500 HP filter. This is one of several points that I see a marked difference, more than enough to believe that this is not a bearing that has gone bad since last survey date. On this particular bearing, I don't know the speed or bearing ID so I can't identify the defect frequency but there is a rough feel to this bearing so I know it is bad. Thanks to all of you for the help.

Thanks,
Ronnie

Doc2.doc (24 Kb, 47 downloads)