.***********History ******************

1/20/05 - motor running normally COUPLED to the pump (sorry, I had incorrectly written uncoupled here before).

March 05 - We removed the motor to work on the pump (no work done to motor).

Reinstalled the motor without alignment shims.

3/30/05 - uncoupled run with no alignment shims. Motor may not have been tight.

Installed alignment shims and torqued holddown bolts.

4/3/05 - uncoupled run with alignment shims and bolts known to be torqued.

*******Spectra are in the attached powerpoint and summarized as follows:********

1/20/04 (coupled) -
1H1: overall - 0.18 ips; 0.5x - 0; 1x - 0.17 ips; 1.5x - 0; 2x - 0.05 ips.
1H2: overall - 0.21ips; 0.5x - 0 ; 1x - 0.18 ips; 1.5x - 0; 2x - 0.09 ips.

3/30/05 (uncoupled; no shims; possibly not torqued)
1H1: Overall - 0.22 ips; 0.5x - 0.04ips; 1x - 0.21ips; 1.5x 0.02 ips; 2x - 0 ips.
1H2: Overall - 0.26 ips; 0.5x - 0.05ips; 1x - 0.25ips; 1.5x - 0.024 ips; 2x - 0 ips.

4/3/05 (uncoupled; with shims, hold-down bolts torqued)
1H1: overall-0.22ips; 0.43x-0.13ips; 1x-0.17 ips; 1.57x-0.04 ips; 2x - 0ips.
1H2: overall-0.28ips; 0.43x-0.2ips; 1x-0.17ips; 1.57x-0.06 ips; 2x 0.01ips

*********** Discussion ********************
The interesting thing is that 0.5x and 1.5x show up on 3/30, even though we have not worked on the motor. Then on 4/3/05 these frequencies appear to shift to 0.43x and 1.57x. The symmetry around 1x makes us suspect sidebands around 1x.

This is vertical 1800 rpm motor 800hp. Lower bearing is conrad oil bath. Upper bearing is 97320UPD. I don't know what that is but I think maybe angle contact or spherical roller. It is also in an oil bath.
Sister motors to this one have been run successfully uncoupled before without abnormal vibration and abnormal frequencies.

We plan to take some higher-resolution data, also higher frequency range, also TWF. In the meantime, are there any thoughts what might cause this?

This message has been edited. Last edited by: electricpete,

LPHD11.ppt (296 Kb, 45 downloads)

3 bearings identified on the nameplate.

100BC03 - I assume this is lower conrad bearing.

100BT03DXXD, 100BZ03DXXD - I assume this is upper bearing. Looking up BT and BZ....

My secret AFBMA decoder tells me 100BT03 is angle-contact ball thrust bearing.

I would have thought if there were 2 bearings on top they would both be angle contact in some duplex arrangement. But What is 100BZ03?

I have a note recorded a long time ago that the bearings are:

MRC 9320UP AND 7320P TANDEM MOUNT

7320P I assume is angle contact.
What is a 9320UP?

electricpete
Are you sure that the top trust bearings end play has been taken out. This ajustment is under the pump ajustment lock nut on the top of the motor.
I have seen times when the motor with no load the rotor will try to lift up when running. And this will let the trust bearing move around side to side and up. Not the outer race but the rollers and inside race.
Greg

Pete,

the 9320UP is likely an MRC. This is an angular contact (40 degree) with a split inner race.

dj

Greg - No - we didn't do any adjustements to remove the endplay prior to the uncoupled run. I understand this is good practice. I have a suspicion that not doing this may be a source of our problem but I still have a question:

We have done many uncoupled runs on this motor and it's sister... never made any adjustments to the endplay, and never had any abnormal frequencies or magnitudes. I went back and looked at uncoupled runs on this exact motor in 1995 and 1997 to verify this (no work on this motor since then). Also I looked at 3 other sister motor uncoupled runs and no abnormal frequencies or magnitudes.

If lack of adjustment to endplay in preparation for uncoupled run is the source of our problem, why didn't we see abnormal vibration in previous uncoupled runs?

Electricpete. How long is your pump? My experience has shown a .43X when the spider bearings are wearing out on the pump shaft. Phil

We have seen some frequencies 0.3 - 0.4x for a different set of large vertical pumps... different story.

In this case, the motor is not coupled to the pump.

quote:

Originally posted by D James:
Pete,

the 9320UP is likely an MRC. This is an angular contact (40 degree) with a split inner race.

dj

Thanks djames. I also heard that this 9320 has a "gothic arch" which allows it take take thrust in both directions. Maybe the split ring allows them to assemble the bearing with high shoulders on both sides.

This message has been edited. Last edited by: electricpete,

Another piece of info - we find among this pump and sister pumps we have had low-level peaks spaced symmetrically on either side of 1x at non-sync frequencies wandering around in the general neighborhood of 0.5x and 1.5x.

The highest of these occured on a sister pump when the 0.57x jumped up to 0.13 ips for 1 reading and then disappeared.

I looked at the plant computer for this period of time and the pump was under abnormally low flow. That leads me to beleive it may have had abnormally low thrust load from the pump... This seems very consistent with the idea that we now have similar symptom from operating without downthrust while uncoupled. Still I'm wondering why we never encountered this during previous uncoupled run.

On advice from our motor repair shop, when we plan to collect the above data to double-check for possible bearing problem (don't think so) and also do a coast-down test. The shop says if the vib goes away within a few 100 rpm, then it is very likely the rotor whirling floating due to insufficient downthrust. The magnetic force apparently plays a role in this whirling (along with lack of thrust which makes the anlge bearings less stiff radially). The repair shop checks for this condition by doing reduced-voltage test - not an option for us in the field - so coast-down is next-best thing.

Some further info attached to complete the case study.

Slide 1 shows high-res spectrum which consists of multiple sidebands 0.57x around running speed harmonics. Most frequencies are N +/- K*0.57 as detailed in slide 2.

Velocity twf slide 4. It is hard to recognize this as modulation but I think that's what it is (modulation of 1x unbalance vibration due to whirl at 0.57x). It sounds good anyway.

Slide 5 shows high frequency range. Some normal rbpf +/- 2*LF pattern. Slide 6 is acceleration twf - nothing more than a few tenths of a g. Slides 5 and 6 helped us rule out bearing fault as a source of the non-sync frequencies.

During run on 4/4/05 we noticed vibration slowly increasing. We were at 0.3 ips on 1H2 at the end of the 20 minute run.

We did coast-down test and noted the vib decreased from 0.3ips to 0.18 ips within 2 or 3 seconds of cutting the power. We believe this helps confirm the diagnosis of rotor floating/whirling due to lack of thrust load.

We have decided not to take any corrective action based on the uncoupled test run. We will await further run after the pump is coupled and fluid system available for run in a few weeks. Will let you guys know the results.

In the meantime (while we wait for loaded coupled run).... any other comments or predictions? A pool on vibration levels after coupled?

LPHD11post.ppt (232 Kb, 27 downloads)

Pete the 100BX03-- is a Ball bearing single-
row, Anglar contact with a two piece inner
ring. And your are correct it will take thrust
in both directions.

Pete,

9000series are combined with a 7000 series to supply dual thrust capability...it should be both to maximize this.

Copied from MRC webpage...
MRC...supplies a split inner ring 9000 Series angular contact bearing. These bearings are usually matched with a 7000 Series bearing of the same size and contact angle. This arrangement offers two bearing thrust capacity in one direction and one bearing thrust capacity in the opposite direction.

I have the frequency data on an MRC 9320UP & 7320P if you need it.

And while Phil is correct about cutlass bearing wear, your spectra don't quite look like that, could be, but all the harmonics don't look quite right, especially on the uncoupled spectra. That's where it drifts from the .5x to the .42x . All the extra "stuff" disappears when it is coupled, the pump load may be acting to stabilize the motor somewhat.

Would you take a 24k Fmax with 6400 line Spectrum and post it along with the gSE spectra.
Like on most bearings, 42% of 1x is awful close to FTF. FTF on both a 7320P and a 9320UP is ~.41 orders.


Dana

Thanks Dana. I looked for the frequency data and couldn't find it. If you have it handy that would be helpful.

We won't have an opportunity to collect anymore uncoupled data... we already told the plant we believe the problem will disappear when coupled so no more runs until coupled sometime next week.

The 2nd file I posted above (LPHD11post.ppt) has both high-resolution (slide1) and high-Fmax (slide5 ) spectra... I'm not sure what else we gain with 24k Fmax with 6400 line Spectrum.

I don't think we got gse spectrum... wish I had. Will double check on Monday. Still with high-Fmax spectrum (slide 5) and the acceleration time waveform (slide 6) I see nothing remotely resembling bearing fault. I see in the low frequency range evidence of multiple sidebands spaced at multiples of 0.593x from the running speed and its harmonics. In higher Fmax I see only normal RBPF +/- 2*LF. In velocity TWF I see evidence of the 1x modulated by 0.57x consistent wiht spectrum. In acceleration TWF I think we see added on high-frequency stuff looks like noise on that plot presumably associated with normal rbpf pattern. Maybe I need to zoom in a little more to confirm that but no peaks exceeding a few tenths of a g in acceleration twf. To me it all adds up to rotor whirling as if in enlarged clearance presumably due to lack of load on upper bearing. But I want to keep an open mind and rule out bearing positively so will be interested to check those fault frequencies against the data I already collected. Thx.

I am happy to report that the non-sync frequencies tremendously decreased when the motor was coupled. Vibration while coupled is now even lower than coupled runs prior to maintenance (presumably due to pump rebuild which also occurred during this period). But it seems clear that the non-sync vibration in this motor decreases when coupled.

Summary data as follows
1/20/04 (coupled) -
1H1: overall - 0.18 ips; 0.5x - 0; 1x - 0.17 ips; 1.5x - 0; 2x - 0.05 ips.
1H2: overall - 0.21ips; 0.5x - 0 ; 1x - 0.18 ips; 1.5x - 0; 2x - 0.09 ips.

3/30/05 (uncoupled; no shims)
1H1: overall - 0.22 ips; 0.5x - 0.04ips; 1x - 0.21ips; 1.5x 0.02 ips; 2x - 0 ips.
1H2: overall - 0.26 ips; 0.5x - 0.05ips; 1x - 0.25ips; 1.5x - 0.024 ips; 2x - 0 ips.
New vib at 0.5x and 1.5x. Increase in 1x

4/3/05 (uncoupled; with shims)
1H1: overall - 0.22 ips; 0.43x-0.13 ips; 1x - 0.17 ips; 1.57x-0.04 ips; 2x - 0ips.
1H2: overall - 0.28 ips; 0.43x-0.2ips; 1x - 0.17ips, 1.57x-0.06 ips; 2x 0.01ips
New vib shifts to 0.43x and 1.57x and increases. 1x decreases towards at-power level

4/18/05 (coupled, after pump rebuild) [THIS IS THE NEW DATA]
1H1: overall - 0.091ips; 0.54x-0.03ips; 1x-0.07ips; 2x- 0.01 ips
1H2: overall - 0.10ips; 0.54x-0.02ips; 1x-0.08ips; 2x-0.03ips

Powerpoint is linked below. Waterfall plot on slides 2 and 3

LPHD11post4.ppt (534 Kb, 19 downloads)

Pete,

Sorry about taking so long....past two weeks have been totally insane.

MRC______________9320UP - 7320P
________________________________________
ftf/cage_________0.412 - 0.408 orders
roller/ballspin___2.089 - 2.004 "
roller defect____4.178 - 4.008 "
outer race_______4.934 - 4.892 "
inner race_______7.065 - 7.108 "

I agree... the loaded/unloaded state has a lot of bearing on how the bearing responds. We had a newly rebuilt hollo-shaft vertical with 220' of pump shaft hanging from it showing classic race damage with ball-spin/defect freq's after three days of use. It went back to the motor shop where they said there was absolutely no bearing defect on either of the angular contact top bearings. They were running it on a test stand unloaded! We had them send the bearings back with the rebearing'd motor and guess what we found...a major cage-shaped spall on one ball and spall shaped dents in both races. Reckon someone missed with a hammer during installation ?!? Our guess was that running it unloaded allowed the mag field to evidently pull the rotor upward enough to where we almost literally had an "air bearing" for the top bearing. We also found brown rust stains on the edges of these "new" bearings. Kind of think someone had a storage and handling issue.

The one-half & one-and-a-half multiples still suggest a little wear in the pump bearings[IMvHO].

Good Luck,
Dana

This message has been edited. Last edited by: Dana,