bearing spinning in housing 5

[electricpete]

In your experience, is there typically vibration evidence of rolling bearing spinning in the housing? (raised noise floor)

I realize it can be intermittent spinning, but we cases of two 4000hp vertical motors (sister units) with lower bearing extremely loose in the housing... one 0.003" and the other 0.007" (supposed to be an interference fit ~ 0.0005&quot. Evidence of severe spinning that I think would have been continuous.

 

[electricpete]

Implied but not stated above: no evidence of spinning was seen in our vibration spectra.

The interesting thing was that the PUMP manufacturer's manual included instructions for a "swing check" test. You push in both directions horizontally at the bottom of the shaft. Measure total movement by dial indicator... is supposed to be less than 0.002".

The limit kind of makes sense. The lower bearing is an angle contact ball bearing (preloaded with a spring). Should have relatively little movement. The upper bearing above the thrust bearing is segmented pad radial bearings, with 0.005&quot; nominal radial clearance. Using a little geometry IF there is not much slop in the lower bearing I expect <0.002&quot;.

We found 0.011&quot; and 0.015&quot; using this swing check test on these two vertical motors. And this is part of the reason we pulled the motors out and found the loose bearing housings. But I am still a little skeptical whether this is a valid test. Curious that it comes from the pump manufacturer manual... I think this is what the pump manufacturer wants the motor to do for his pump, not what he thinks is indication of problem for the motor. Motor manufacturer is long out of business.

Has anyone ever done this &quot;swing check&quot; test? Any comments on its validity in diagnosing bearing problems?

 

[electricpete]

Part of my uncertainty in the swing check is predicting how a normal lower bearing would act. Given that it is an angle contact bearing, the ability to provide tight radial clearance depends on the axial preload. Not sure exactly how much force should be applied during the test. Unfortunately we didn't measure these two motors after repaired. I will measure them the next time we get a chance. Will attempt to plot force vs distance.

 

[don01]

electricpete,
I recall that on abearing training course by SKF the mentioned measuring the play on the shaft both in line with the axis and perpendicular to it. Now that was to check the bearings in the context of the course we were doing.

But wouldn't the bearing manufacturers specs hold true and this would indicate the need for bearing check ~ change out.
I have not heard of the NDT boys reading loose bearings but I will ask them next time I se them. If any of the other members now some NDT guys they might ask the question too.
It might be worth talking to one of the more knowledgable bearing reps in your area, they may have some experience here

regards
Don

 

[induction]

I have seen this swing test done by machanics that I have worked with. From what I was told this test tells how much play is in the endbell they had no wieght or force requirements. They would push on it and take the measurment. Depending on the bearing design some of the outer races acctually move when they are cold. Like don said I would ask a bearing rep what those fits are to be. thats a lot of wieght to push around and the distance between the two bearing may have a lot to do with the manuals numbers for play.

 

[GusD]

Hello electricpete

Normal bearings, ball or rollers, that are loose in the housings (even brand new bearings)show 1xrpm vibration with
a lot of harmonics of the fundamental.The greater the looseness the higher the vibration and # of Harmonics.
A lift check with a dial will work quite well for most Horizontal motors or motors that are equipped with standard Ball/Roller bearings.Their internal clearances are around .003&quot; to .005&quot; for average sizes.
The angular contact ball bearings require that they be subject to some thrust,to be able to check lift.
Your 4000 hp motor could have a large bearing with a 150 to 170 mm bore.These large angular contact ball bearings in the sizes 7326 to 7334 (fag/SKF)could have large internal clearances.
As for the typical vibration spectra showing loose fits
in the housings ,I have never seen much other than 1xRpm and multiples.IN the HI-Frequency Acceleration spectra(gSE orOther) the Trending Raized Floor, with Discrete
peaks of 1x rpm and Multiples could be a direct indication.
Electropete.
When everything else fails.,the only way to be sure, is to analyse the last failure or removal from service.Measure the housings and the bearings OR. Outer race Slip fits should drop in the housings but shouldn't allow much radial movement if any.

Take care,.

GusD

 

[GusD]

Hello electricpete

Normal bearings, ball or rollers, that are loose in the housings (even brand new bearings)show 1xrpm vibration with
a lot of harmonics of the fundamental.The greater the looseness the higher the vibration and # of Harmonics.
A lift check with a dial will work quite well for most Horizontal motors or motors that are equipped with standard Ball/Roller bearings.Their internal clearances are around .003&quot; to .005&quot; for most average sizes.
The angular contact ball bearings require that they be subject to some thrust,to be able to check lift.
Your 4000 hp motor could have a large bearing with a 150 to 170 mm bore.These large angular contact ball bearings in the sizes 7326 to 7334 (fag/SKF)could have large internal clearances as well.It may appear loose,but not nessecarily so.
As for the typical vibration spectra showing loose fits
in the housings ,I have never seen much other than 1xRpm and multiples.IN the HI-Frequency Acceleration spectra(gSE orOther) the Trending Raized Floor, with Discrete
peaks of 1x rpm and Multiples could be a direct indication of this condition.
Electropete.
When everything else fails.,the only way to be sure, is to analyse the last failure or removal from service.Measure the housings and the bearings OR. Outer race Slip fits should drop in the housings but shouldn't allow much radial movement if any.

Good luck

GusD

 

[EdDanzer]

The motor you are working on seems to have some poor design. A single angular contact bearing for thrust and radial load should not have any movement on the shaft, there should be some preload, other wise the shaft could float causing the race to rotate some each time the balls contact after lifting off.
A motor this large should have the bottom bearing fixed to both the shaft and housing and the top floating. Depending on bearing size there should be enough press to offset housing deflection under load and temperature rise. There should be enough float on the top bearing to allow for expansion and some clearance.

 

[electricpete]

good comments, Ed. Although I think the lower bearing should be floating to allow for thermal expansion.

The motor is designed with requirement to withstand momentary upthrust. 27000 pounds momentary upthrust. (also 54,100 continuous downthrust and 83500 momentary downthrust.)

Upper bearing has 17&quot; Kingsbury to handle the downthrust, and a thrust collar to handle momentary upthrust, with 1/32&quot; of shaft end-ply between upthrust and downthrust positions.

The lower bearing is an angle-contact bearing AFBMA 160-BH-02 = 232R style placed in direction to take upthrust.
On bottom of this lower bearing, the inner race is fixed to the shaft by lockwasher. On top of this lower bearing,
there is a flimsy wavey style washer (looks homemade) identified as an &quot;endplay&quot; washer which is placed between top of outer race of bottom bearing and the top of the bearing housing cavity. Purpose of this washer I believe is to provide preload. Unfortunately I don't have the info to identify how much force it provides. I believe the upper bearing thrust collar is intended to take the bulk of the upthrust and the lower bearing acts like a shock absorber to slow the upward shaft movement before it slams into the upper thrust collar. Still I think this function could be performed with a conrad ball bearing and our lower bearing would not be require any axial preload (we would leave the wavey washer in there for shock-absorber action).

We will be rewinding a sister motor within a month. I plan to ask them to take careful measurements during disassembly to establish how much clearance is available for that wavey washer. Based on calculations of wavey washer compression at the point where upper thrust collar contact, we can determine max load which may be placed on lower bearing. If this is within the capacity of a conrad, I'd like to replace it with a conrad so we are not so dependent upon this wavey washer. As I said it looks flimsy, appears homemade, bearing supplier could not identify any replacement wavey washer that big.

Motor OEM is out of business so I am trusting our repair shop to keep us from making a big mistake.

Do you guys think that lack of axial preload on the angle contact lower bearing could have caused the spinning? Note that the races and balls were in very good condition. I know that lack of preload is bad and can lead to skidding of balls on races, but I don't see offhand how it might cause spinning within the housing.

 

[EdDanzer]

Since the majority of the load is down and the load it at the bottom, the bottom bearing should be fixed. With the bottom bearing fixed any expansion of the motor components will not effect the output shaft position. There should be back to back angular contact bearings or tapered roller bearings on the bottom that are designed to be clamped into a housing for preload. The top bearing will need to have axial float with the outer race looser than the inner, but enough press to stop rotation of the outer race.
When placing bearings to control thrust, keep them close to the item that shouldn’t move, if shaft is long, or large in comparison to the housing, keep all components that need alignment on the same end as the fixed bearing.
The other option is to float the bottom bearing, but use a single row deep groove bearing.
It might be wise to have someone reengineer the bearing system on these motors besides the motor shop.

 

[electricpete]

I may not have explained our setup very well but we have a Kingsbury thrust bearing on top. That will control the position during normal downthrust operation. Therefore I think the lower bearing needs to allow some float.

 

[induction]

Yes pete that is correct the upper bearing actuallyholds the load from what I have seen during the first run after a new bear we would lift the shaft to take off the stress from the first run. I do not know the amount of float it needs but it must have some. That is a common bearing setup for large pump most pump house motors use this design. The wave washer acts as a preload for that bear so it does not slam form what I have seen of this setup from the guys at work is that it is very important to ajust the top bearing correctly. Does this one have a large lock nut on the top bearing, it has been a while since I have seen one.

 

[EdDanzer]

I’m not familiar with the term Kingsbury.
If the Kingsbury will hold thrust both directions you can just float the bottom as long as any expansion differences don’t effect what is below. If the Kingsbury doesn’t hold thrust in both directions, a redesign is the only way to have a long term fix.
The outer housing and armature will expand differently so calculating the expansion will be difficult. Knowing the materials and length of each should give an approximation.

 

[GusD]

Hi electricpete

This type of assembly as you describe it is fairly normal for this size of equipment.The Kingsbury takes the large downword thrusts and the angular contact with the preload &quot;mickeymouse&quot; washer takes the smaller upword trusts and whatever radial loads are in the system.
The spring washer allows for some thermal growth and preloads the OuterRace of the Angular contact ball bearing.If the washer is weakened, you could have an outer race that could float and without the spring preload it could spin in the housing.I think the spinning would be more like a creeping unless the housing is out of spec or worn.On some of our drives we have a ball and a roller on the NonDR or Dend ,whichever the case.The outer race of the Ball bearing is preloaded by springs contacting it.
In these case the Outer race doesn't even contact the housing.The roller bearing keeps the shaft in place radially.
Electricpete.as you mentioned,the sister motor will be worked on soon and that will be the best time to do the precise measurements for housings and bearings.


GusD

 

[Farkel]

I'll bet it is a U.S. vertical motor!

 

[electricpete]

Yes, US. Why... do they do it different overseas?

 

[jbartos]

Suggestion: Visit

http://www.joliet-equipment.com/electric_motor_repair.htm

http://www.usmotors.com/PB500/2001/Mods_A-D.pdf

etc. for more info

 

[volvibeguy]

electropete,
Iv'e just read the discussions on this motor and a question comes to mind, what rpm is this motor turning? coming from a electric motor repair shop background there has always been a tolerance chart for performing endbell and bearing fit tolerances, however I occassionaly encountered fits that were to tight or to loose; some from the motor mfg., some from another motor repair dealer, and of course, some from fatigue and/or wear, now when we considered a fit bore and bush I always took into consideration rpm and load, the reason is thermodynamic/thermal growth, the higher the rpm the looser the fit allowance; and the lower the rpm the tighter the fit.
And in relation to the vibration data, How much vibration is there? 1xts harmonics are not always bad or good it all depends on the amplitude envolved whether it's mill/ips/g's
once again i go back to load and rpm, and about the multiples of 1xts it could be looseness however, i would also check the vibration at the instant power is removed, i have personally seen signs of looseness especially in a vert. mounted motor turn out to be broken rotor bars which immediatly deminish with the removal of electrical power,also look for sub synchronous harmonics such as 1/2, 1 1/2 etc. of turn speed they usually always indicate a problem also watch fo high 2xts for misalignment possibly being a bearing cocked in the fit and use the waveform data to help give way to severity.
Hope this response is appropriate,
only trying to help.
Sincerely,
JB

 

[jbartos]

Suggestion: The subsynchronous harmonics and interharmonics are relatively new areas slowly coming into the industry standards.

 

[jbartos]

Suggestion: Visit

http://www.corningfrequency.com/hirel/mil_specs/prf55310.pdf

for Figure 10 to observe harmonics and subharmonics
and Section 6.4.19 presenting harmonics and subharmonics.