Gents
What do you think the reason for unsteady vibration level on a vertical mounded motor? During the monitor overall the amplitude goes to .25 IPS then it drops to .14 IPS on repeated manner

Is there another motor(s) near the motor (common base or piping) operating at the same speed? It could be a beat frequency, where two signals close in frequency add and subtract each other as they move in/out of phase. Take a long time waveform and look for the beating. Also, take the highest resolution you can with 100 hz Fmax to separate the 2 signals. You could also monitor 1x amplitude and phase with a narrow filter.
Does the pump move fluid from a reservoir/tank or is it fed another way? Can you monitor flow or pressure via guage or transmitter? That may point to system design or operation problems.

Good luck, let us know your results or if you need assistance.

I agree with Bill. Most likely either beating of vibration from two different sources or modulation of vibration (for example pole pass modulation) and the key to recgonize either case will be a high resolution of the spectrum looking to find those peaks close together.

Since you didn't tell us the frequency of your vibration, I should mention there is also a possibility that the frequencies causing the bearing might be 2*LF and the closest running speed harmonic (for example beating between 2*LF and 2x in 2-pole motor, bearing between 2*LF and 4x in a 4-pole motor.) So if you have peaks at/near 2LF, you should get high resolution there as well.

Unfotunately I have the same problem with horizontal multi stage KSB pump with frequecy control when it works close to 40 HZ, ampliyude increases with a recognized sound to 0.40 ips nearly each second period,

I guess something like beat vibration, I know it is a concide between 2 vibrations leads to amplification but what is the 2nd source beside a bearing fault???

In order to solve such a problem, one has to recognize the source of the two frequencies which implicitly necessitates the recognition of the exact value of the frequencies.

Say the amplitude fluctuation (at 40 Hz)is cyclic every 10 second (by collecting the overall amplitude trend).

The frequency difference between the two sources is:

del f= 1/10sec =0.1 Hz

Hence the two frequencies are 39.9 and 40 or 40 and 40.1 Hz.

So if you are a capturing 100 Hz spectrum, min no of lines is

N= 100/0.1= 1000 lines but this no is not available in the usual analyzers; so use the option 1600 lines.

Hint:

To isolate the external sources, take the background vibration as a reference.

It would appear that the pump in question in Saudia Arabia is too distant from the Center of the Reliability Universe in Knoxville, TN USA!!

Walt

quote:

It would appear that the pump in question in Saudia Arabia is too distant from the Center of the Reliability Universe in Knoxville, TN USA!!

Huh? What is the max distance from Knoxville is one allowed to be?

Ralph,

Ask Bill Kilby about the Reliability Universe!

Seriously, the original posting is rather vague by indicating "unsteady" and "repeated manner". The amplitude variation may be repeated from min to max, but the time period may not be periodic. An example would be a natural frequency close to service speed that causes amplitude amplification.

Walt

This message has been edited. Last edited by: Walt Strong,

quote:

Ralph,

Ask Bill Kilby about the Reliability Universe!

Just picking at you Walt.

I haven't heard of the "Reliability Universe". Must be a new nickname for Knoxville.

SA is quite a hop from Knoxville.

k.kahtib - Gaza

Yours sounds more like a simple resonance than a beat, although it could be either or both. When it operates at 40 hz, does the amplitude remain at .4 ips or does it regularly rise and fall while at 40hz?