Sign up Calendar Latest Topics Donate
 
 
 


Reply
  Author   Comment  
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #1 
I look at a large number of pumps at one of my sites, and my sense is the motor shop they are using is not doing a very good job on rebuilds with bearing installs.  This is (I think) an example of this, but I wanted to run it by you guys before making an "issue" of this.

They did a complete reinstallation of this pump back in 2015 including an all new base (concrete), new pump, and rebuilt motor.  The overall vibration went down, but vibration at 4.12X was immediately apparent, and while not increasing, has remained high (pump doesn't run often, so only 3 data points since the rebuild).

Took a hi-res spectrum last week, and the peak is at 4.119X compared to 4.093 for an outer race defect on a 6220 bearing (per the motor nameplate).  What would account for the difference between 'actual' and 'calculated' BPFO?  What mistake might the motor shop be making to cause the defect to show up on a new bearing, and cause the difference in frequencies?

My goal here is to contact the motor shop directly to help them get better (they know me, so I think they'll listen).  I'm not looking to hang anyone out to dry, just to fix the problem.


Pump 5 Trend.png 
6220 BPFO.png 


__________________
"The trend is your friend"
John from PA

Sr. Member
Registered:
Posts: 963
Reply with quote  #2 
Rusty,

Various literature documents that the failure rate of rolling bearings due to installation deficiencies is quite high.  See http://www.vibrotek.com/articles/sv95/part2/index.htm, specifically the paragraph titled "Cause of Bearing Failures During Initial Operation".  I have also seen "high" numbers, like over 50% reported by bearing manufacturers. However, I don't think its the machine OEM's and their repair facilities that are lacking.  I know at my former employer, bearing suppliers would volunteer to come in and instruct personnel, sometimes even one on one, on how to properly mount and set bearings.  But the environment of the OEM also includes the need to purchase thousands of bearings annually so the services of the bearing manufacturer are readily available.

At the repair shop level, where the mechanics aren't trained as well, it is amazing to see the practices.  Not too distant from a bench grinder, I have seen bearings heated in oil for mounting, but the oil container was uncovered.  I borrowed a magnetic pickup tool from a mechanic and dragged it across the bottom of the container and readily collected ferrous debris.

On another instance, a customer was puzzled by the failure of a simple gearbox about every six months.  When I watched the rebuild process, a mechanic was driving the bearing onto the shaft (a light shrink fit) with a block of wood that impacted the outer race instead of the inner race.  He was surprised that was an issue!  The gearbox when rebuilt properly ran at least five more years, at least I personally was aware of that duration.
Danny Harvey

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,403
Reply with quote  #3 
Rusty,

Do you have data to a higher fmax?

In my experience, high amplitude at 1 x bpfo means an impingement in the outer race like a metal chip either between the rollers and the race or between the outer race and housing causing a high point in the outer race. It quite often seen in the axial data too because a chip behind the race throws the ball path out of line.

Changes in the defect frequencies can come from excessive clearance, from a ball path that varies from the root of the groove, races slipping.  Races slipping would make it operate at effectively lower speed since the defects are based on relative race speeds. I haven't really thought about what the effect of the other two would be but I'm inclined to think it would be the same since both would make the ball path longer.

RHP makes a 6220 with a 4.130 bpfo.  Could it have been bored and sleeved so that it uses some oddball bearing?

Either way, my inclination is to make a call even when the numbers don't match up exactly. You've seen thousands of bad bearings and know that there just isn't much else outside of 2 x lf in that range. Plus, I know you've stuck your yellow handled analyzer on it and heard all you need to hear.[cool]
Sinski

Avatar / Picture

Sr. Member
Registered:
Posts: 394
Reply with quote  #4 
Just noticed your trend data is motor inboard bearing and the spectrum is motor outboard bearing. Are both points showing similar data? Also you are only showing the fundamental frequency so are there many harmonics as well? In saying this I think Danny might be on to it as thats what I would have thought as well. Also I assume that is far enough away from 2 x lf as well as that can be just over 4x. 
RustyCas

Avatar / Picture

Admin
Registered:
Posts: 1,810
Reply with quote  #5 
1X = 29.83 Hz so 4.12X = 122.9 Hz The amplitude is actually higher at MIH. I just always start my collection at MOH so did the zoom spectrum there.
__________________
"The trend is your friend"
Beatnik

Sr. Member / Supporter
Registered:
Posts: 183
Reply with quote  #6 
I've been taught in my Mobius CAT 2 course that a ball bearing with an axial load (end play too small) will have different bearing frequencies. I cannot remember why exactly but I think it's because of the contact angle.
Danny Harvey

Sr. Member / Moderator / Supporter
Registered:
Posts: 1,403
Reply with quote  #7 
Beatnik,

I believe that is correct.  I'd have to take a look at the math to see if the change in pitch line diameter is accounted for by the change in contact angle.
Walt Strong

Sr. Member
Registered:
Posts: 889
Reply with quote  #8 

I developed a test system for a manufacturer of precision ball bearings. The system measured shaft speed and Ball Pass frequency and calculated Contact Angle. The Contact Angle was used to indicate whether the new bearing fell within acceptable dimension tolerance or not. Based on this testing and experience over the years, I have found the following general guides when bearing fault frequencies do not match the calculated values:

1) inaccurate shaft speed measurement
2) vibration frequency resolution or interpolation is not accurate
3) installed bearing is not what you think it is
4) bearing inner or outer race/ring is slipping on shaft or housing
5) bearing dimensions are out of tolerance from manufacturer (cheap bearings?)
6) thrust or preload force is excessive

I'm always willing to learn something new if anyone has any suggestions.

Walt

electricpete

Sr. Member
Registered:
Posts: 651
Reply with quote  #9 

Yes, the bearing frequencies do change with contact angle.  Here's how I remember direction of change qualitatively.

  • Let's say we have zero-contact angle fault frequencies, the inner and outer must add to number of rolling elements.
  • Example BPFO = 4.093 and BPFI = 5.906  for 10-ball bearing.
  • What happens when we apply a load to the bearing to move the ball off-center within the rings?
  • The effective radius of the outer race decreases and effective radius of the inner ring decreases.  The two radii become more the same and their fault frequencies become closer to the same. So BPFO increases and BPFI decreases (they still add to integer like 10).

Good list by Walt.

It's worth double checking 2*LF as Sinski said although I bet you checked that.

If in fact it is BPFO fundamental without much in the way of BPFO harmonics (you don't mention harmonics so I assumed there are none), that's one scenario we've talked about before and lots of theories… I have some thoughts and not sure I have nailed it down.  I currently have one machine on my radar that showed the particular symptom high motor 1*BPFO (without bpfo harmonics) on motor bearing during coupled run after motor rebuild, but interestingly did not show during the uncoupled run a few hours before.  We asked ourselves why it would show during the coupled run but not the uncoupled run:  the first theory we came up with was that the motor must not have room to grow thermally between bearings.  So we uncoupled, loosened the inboard bearing inner cap and did an "endplay" (growth allowance) check… found plenty of room for growth… that theory is dead.  Coupled up and once again showing the same pattern.  The new theory is that thrust is being put on the motor through the coupling by virtue of coupling hubs that have a tendency to wander on the shaft (we have found that condition on these machines before).  We plan to go in and inspect the couplings, will post the results when we do. This particular motor has been through a LOT, and it could be a very  interesting case study when it's all over, but we're not there yet.

Walt Strong

Sr. Member
Registered:
Posts: 889
Reply with quote  #10 
The question about bearing thrust load came up during the commissioning of the bearing quality control test system. The "bearing guys" said that some people (experts?) thought there was no relationship between thrust and Contact Angle. We conducted a simple test by changing the axial force on the test system and found there was a direct and instantaneous change in Contact Angle. They were impressed with the sensitivity of the test system, and the fact that it agreed with what they thought should happen! The Contact Angle is generally unknown when calculating bearing vibration fault frequencies. A nominal value of perhaps 20 to 30° for Contact Angle should be adequate for fault frequency calculations. Frankly, I do not worry about minor differences between the measured vibration frequencies and calculated bearing fault frequencies. It is not an exact science unless you test tens or hundreds of brand-new bearings on a test stand under controlled conditions.

Walt
RRS_Dave

Sr. Member / Supporter
Registered:
Posts: 216
Reply with quote  #11 
I think your outer race is slipping and fretting around a slightly too big or out of round bore.

D
Jim Crowe

Member
Registered:
Posts: 53
Reply with quote  #12 
The formula to calculate BPFO includes speed, ball diameter, pitch diameter, and contact angle. Any of these variables that are different will change the defect frequency, and yes axial load changing the contact angle will change the frequency.
dmcmmc

Avatar / Picture

Member
Registered:
Posts: 69
Reply with quote  #13 
The difference between actual and calculated BPFO frequencies could be that the speed slightly off.

I agree with Danny that bearing outer race could have an impingement. I have a felt roll that has a bearing housing that is out of round. The BPFO very well defined in PeakVue  but nothing in velocity. It's been running that way for 6 years.  
Ron Brook

Member
Registered:
Posts: 35
Reply with quote  #14 
Rusty,
Had a problem like this that was caused by the new shaft diameter being oversized under the bearing and causing preload on the inner race that produced a clean BPFI peak.  If the bore in the bracket was undersized for the bearing, the same thing would happen with BPFO.
Ron
Previous Topic | Next Topic
Print
Reply

Quick Navigation:

Easily create a Forum Website with Website Toolbox.