Hi Guys,

Could use your help troubleshooting a problematic condensate pump.

This is a 16x25 - 7 stage condensate pump in a coal fired power plant. The unit runs in parallel with 1 of 3 sister pumps.... never runs in solo condition. The pump is constant speed and runs at 20 Hz speed.

This pump was placed into service about 2 years ago, and it reported to continually increase in vibration amplitude (1x dominant) at the MOH position (upper motor bearing). When I was able to get data on the installation, I found that there was a dominant 1x peak at the top of the motor that was around .45 IPS peak consistently but that when the power plant decreased MW down to minimum load the vibration increased significantly to up around .7 IPS peak. 

I did check for imbalance but there was approximately a 180 degree phase shift between the horizontal and vertical planes of both the upper and lower motor bearings.

When I did an ODS, I found that there was a torsional like motion on the motor. Based on my impact testing, I did not believe there was any resonant excitation taking place. So my thought was that as the power plant reduced MW's, the pump would reduce flow moving away from BEP, and the differential pressure across the impellers would considerably increase. If the rotor was in contact with any stationary components below, the rotor may try to take a dog leg shape which would then transfer the motion up to the motor resulting in the torsional like motion I saw in the ODS.

We pulled the pump and found that there were two impellers in contact with the bowls (lift was set too high) so I felt like my hypothesis was confirmed. The damage was cleaned up and the pump returned to service many months later.

When we put the pump back into service, I recorded vibration most of the day and through the night.... luckily the power plant cycled load so I was able to see both maximum and minimum MW performance.

I also repeated the ODS.

I found that the 1x amplitude at the MOH position was up around .55 IPS Peak and that the swing in power plant load had no effect on the vibration. The ODS data shows that the torsional like motion is gone, but there is a side to side rocking motion present.

Once again, there does not appear to be any contribution from imbalance as the phase shift from H to V at the top of the motor was approximately 150 degrees.

I tried to repeat the impact testing during this second visit, however my impact hammer wouldn't work... go figure.

After looking at the impact test data again, there may be a resonance in the horizontal plane of the motor.... I am hesitant however to blame this on resonance... I have attached a power point with a bunch of data... if you look at slide 10-12, you can see the large impact peak at 20 Hz.... I kinda wrote this off because there are 2 sister pumps running at 20 Hz but if you look at the phase spectrum on page 12 there does look to be a phase shift there of about 135 degrees (MOH Z).... I wanted other people to look at this because I don't want to start looking for resonance just to have something to blame this on.

Could you please review the attached data and let me know what you think? I truly appreciate the assistance.

Please let me know any questions you may have.

Thanks again,






Original Post

Why do you doubt yourself?

For my money, Sheet 11 is about as strong case of resonance as I have seen. The fact that there was some "rubbing" contact found during your recent inspection may explain why the 1X peak shifted in amplitude over time and, after correcting for proper lift, it is fairly steady....yet still high.

If the only change your maintenance effort did was correct for lift, then it would be safe to assume that resonance is still a primary suspect.  Not totally sure whether the ODS shows a bending mode or a rocking mode (if resonance is found to be the culprit, the "fix" will depend on its mode shape)....  you might look at the base & flooring support for amp/phase.  This might also be a good candidate for the new visual amplification process.

Question:  Can you take impact testing on the other two sister pumps to see any difference?

Good luck

Jim P

 post edit: BTW - Don't worry at this time about requiring impact hammer data, just use the impact hammer or large rubber mallet or block of 2x4 wood during a "peak hold" and compare all 3 pumps response amplitude

Last edited by ivibr8

Hi Jim!

Thanks so much for looking at this.... always appreciate your insight and expertise.

I guess I am a little worried to call this a resonance even though in both the linear and log scale impact spectrums it sure looks like a 20 Hz resonance is present. My concern is that I expected there to be more of a "clear' phase shift at that frequency? The phase is what worries me the most.... there are 3 other pumps running at 20 Hz there, and I guess I am just worried that this may be back ground noise given the way the phase looks? What do you think?

I have added another power point that includes the three impacts that I did in August of last year. On the first impact (pretending I hadn't seen the spectrums), I would lean towards resonance at 20 Hz on the first impact.... the second impact would not make me say resonance.... the third impact concerns me as well.

In the power point you can also see a spectrum while the pump was offline showing the back ground noise.... there is a very noticeable 60 Hz and 20 Hz peak.

I am just a little worried about chasing "back ground noise" as resonance if that makes sense.

I do agree bump testing the other two pumps would be a good idea.

Please let me know what you think on the above.

Thanks again,




For an impact test, I wouldn't be too concerned about phase unless you start comparing to other parts of the structure - which I don't think you did.  Are you perhaps concerned about a 180 deg phase shift?  This is something you would look for in a coast-down plot...not an impact test.

IMHO (please others  comment), pg 14 of your last PPT is clear.  The 60 Hz noise I could understand - but the 20 Hz peak?  Ask why does it show up?   Is there common suction/discharge piping that "couples" well with this pump?  still not sure why this particular pump (vs. other two).   Could it be the other two have a similar resonance but at a different frequency not coinciding with running speed?

It still may be a case where this particular pump has higher unbalance as well as a sharply defined response at running speed; but I would work on the high response at 20 Hz first before balancing.   If possible, I would try to take impact "bump" test data on all 3 pumps when none of the pumps are running to see if there are similarities in response and frequency location.

Interesting case....keep us posted


BTW - Backround noise is 0.003 ips rms. Your running data is >0.5 ips      Big difference    

Last edited by ivibr8

Do you have an image of the actual installation? Recently had what I believe to be a similar problem. Found lower end bell on the motor to be soft. The bolt diameter on the lower motor end bell was so small it could not hold motor still. Adding wood wedges we were able to shift the natural frequency. We had to use the motor end bell bolts to hold the motor still, as in lift the motor, drill holes in pump head at end bell bolt pattern/diameter. 

Also found improper grout is contributing to the natural frequency issue. Spec for this project calls for 120% of run speed lowest natural frequency. With proper grout we were able to meet this, pump with improper grout did not meet spec. Like from 900 cpm natural frequency down to 780 cpm with bad grout!



Photos (2)


Really like the wood wedges/stiffness that you provided the pictures of.... man that was super creative.

I am still looking into this one.... I am really starting to subscribe to my issue being resonant.... Jim P had some really convincing points.

Thanks for your feed back.... stay safe and healthy.




It's interesting to know that torsional vibration can be measured by casing single axis sensors. It's difficult for me to imagine this type of ods if the sensors are mounted lateral to the axis of the machine (their axis is pointing to the shaft centerline). My basic understanding is that a twist is a motion that needs a different degree of freedom than lateral vibration and each has a different coordinate if they are considered separate from each other for the same point.

But I could be totally wrong as in many cases. 


Can you please clarify your measurements that were used to generate the ods? Did you yse use triaxials, any special brackets, did you have enough measurements to avoid aliasing?

Regards- Ali M. Al-Shurafa 

Add Reply

Link copied to your clipboard.