Topics

Mounding around BPFO

I have a motor bearing (Deep groove ball bearing 6324) that has been showing increasing BPFO in the demodulated reading for about 8 months and now it is showing mounding around the BPFO.  I assume this means that the condition has worsened, but do you think it is skidding?  What do you guys think?

 

What does a normal spectrum and waveform show?

Could be a little smoothing of the surfaces from wear if they are surface defects versus subsurface flaws.

The normal spectrum shows low level BPFO and a waveform with modulation at approximately 2190 CPM.  Running speed of this motor is 1196 CPM.  I attached them to this post

    Attached Files

  • docx Doc1.docx 106896 17 downloads

indeed it has very low amplitudes, any sideband around 1xBPF0?

There could be low frequency sidebands in the Enveloped reading.  Possibly around 82 CPM.  They are not very well defined, I should probably take a higher resolution enveloped reading on the motor next time I run the route.  The motor used to be open to the air, but it has recently been tented.  They are sandblasting paint off of the structure around it.  My fear is that the lubricant has sand in it.  Would that cause mounding like that?

Normally in gE if we see any lubrication issue its a raised random noise floor with little to no defined peaks, there are the impacts in the regular twf but as others have said its low amplitude and not what I would be immediately concerned about, was the twf similar in the earlier readings you mentioned? if so then looks like those irregularities have smoothed off. have you looked at this velocity in acceleration (hit button 'A' on keyboard) 
Sand is deceptive as the smaller dust from it can work its way easily into seals, especially the Drive end/Inboard.

It's entirely possible it's getting small particles of sand ingress however you could also get the same result just being in the open air. Fine particle dust is almost impossible to prevent, especially in an industrial setting.

The levels seem low and in my situation would only warrant watching more closely as I got data and increasing the frequency as levels increased. The question would be what is the history of that machine? Are vibration levels increasing, increasing at an increasing rate or just slowly drifting higher? 

The mounding could be from several possibilities. The bearing is wearing and the edges of any defects smoothing out so the impacting it is imparting more broadband noise into the waveform. Slight changes in speed. Remember, the enveloped signal is a much higher filtered signal that is heterodyned into the 0-xx CPM/Hz measured range. Any modulation of the speed, from a VFD or just variations in the load, while the time sample is taken can affect the signal as it is processed through the bandpass filters and the mixer to produce the output. Or you could have multiple small defects developing. I've never positively seen a slipping/sliding of balls in the race but I imagine the waveform would certainly be more distorted that you are showing. I'd like to see some examples myself. Perhaps I/we should Google and see if there are any case studies with graphs on the web.

Trapper: I had never thought about the changing speed thing. That's a great point. I will see if I can find a case study for ball skidding or roller skidding and if I find something I'll post it on the thread.

MarkL: it went from a very well defined peak at BPFO in the enveloped spectrum to a lower mound like thing around BPFO. I just put the normal velocity spectrum into acceleration and BPFO is there, but it's tiny. This is likely nothing at all to worry about. I just thought it might be a bad sign that it mounded out.

I think trapper may be on to something with the slight change of speed affecting it. I remember that this machine is difficult to tell if it's loaded, so a load change is likely

Thanks everyone for your responses and help.

Have you tried using different filter? try filter 4 maybe. It may give you a different signature.

Varying speed mid data collection cylce can have a huge effect on the readings as previously mentioned, so try and make sure the machines can be locked to a speed for the duration of the data collection cycle.

2Per makes a valid point trying envelope 4 you will get a different pattern though as its focusing on a different range@ Sheridan did I send you the skf paper on envelope filter selection/setup previously? something tells me I did, I'm just not sure, I can re-send it again if it's of any benefit?

Is it this article I linked after this paragraph?  I don't think I have seen this before, but I found it after searching "enveloped" on this forum.  I think you posted it on a different thread.  I also have one about Condition Monitoring of Slow Speed Assets.  This one I put in the link below is a great explanation of what's going on in enveloped readings.  I think it suggests using logarithmic view of the spectrum to diagnose severity on page 12.  Is that something that you do?  I have just been diagnosing the severity by seeing what defect type it is in the spectrum and then keeping that in mind when I look at the waveform's amplitude and patterning and then comparing it to previous readings (If I have them)

http://www.maintenance-engineering.eu/downloads/public/envelope%20bearing.pdf

MarkL.  Thanks for your help again, and if this link isn't what you're talking about I would be grateful to get a link to what you are talking about.

2Per, I will try using filter 4 and will check out the difference.  Unfortunately, this motor is at the top of a giant silo, so I probably wont be up there again for another two months or so.  I don't mind trying whatever you guys suggest once I get to the top, it's just the many flights of stairs and 3 hours of driving that it keeping me from doing it right now.  Thank you for your suggestion and I will let you know how it works out

Hi Sheridan I think I sent it to your Dad or your Brother previously, 
I will send you a Pm with it. If anyone else wants it drop me a mail on here.

Reply Cancel