I did testing Friday on a motor with suspected rotor bar issues. The vibration data shows 2Xslip sidebands around the first three orders of run speed and the motor sounds like it has broken bars on start up. My gut says the rotor is bad. My problem is - when I pulled current data to confirm what I saw in the vib data, there was no indication of a rotor bar issue. In fact, the dB difference was greater than 65. I've always put a lot of faith in the current test and this one has me worried. Also, I normally see elevated 1X when I find bad bars, this time the largest increase is at 2X turning speed (not 2XLF) I am also seeing excessive vibration at 6X turning speed in the axial direction and I'm not sure where it's comin from. This is a 500 HP 2 pole motor driving a centrifugal air compressor. Vibration at all of the compressor points is at very good levels. As I said, my gut says it's a rotor bar problem but my data doesn't back me up as well as I'd like. I would appreciate any opinions.

Thanks.

You said it is on a compressor - was the compressor loaded? Did you check for a phase imbalance? How about rpm? I learned from a motor repair guy to check the rpm. If rotor bars are going bad, they slow the rotor down.

If it was not loaded, see if it loses rpm when going from unloaded to loaded.

You said you had a 6xRPM in the axial. What is motor speed? Is it 1790 or so? If so, is it 6x, or 3x line freq.?

Stan Riddle

Stan,

The compressor was fully loaded. Motor was turning 3575 rpm (no appreciable loss of rpm)and was pulling full load (60) amps. Phases were within 2 amps of each other. this is a 4160V motor. The axial 6X peak is definitely 6X turning speed.

Thanks for your interest.

Is this a cast rotor? A while back, we had a 4160V 350HP motor with rotor bar problems showing up in the vibration spectrum very similiar to what you describe. The current signature had rotor bar problem peaks but not up to the 65 db point we use as a early warning limit. When we sent the rotor off to a motor shop, the rotor loop test thermography indicated several broken rotor bars, which was confirmed by visual inspection. We have since decided to lower our warning threshold for rotor bar peaks in the current signature for cast type rotors, especially if vibration is indicating problems.

Yes Ed, this is a cast rotor. Thanks for the information. This is the kind of confirmation I was hoping for. Has anyone else experienced confirmed rotor bar problems that showed up in the vib data but not in the current data?

Joe,

I sucessfully diagnosed a rotor fault on a cast aluminum motor rotor on an Induced Draft Fan that had the classic 2xSlip frequency in both vibration and current spectrum data. Porosity in aluminum rotor should have same characteristics (causes stray current) as cracked/broken rotor bar or end ring joint.

I don't have a reason to explain why your data doesn't agree. I would recheck/remeasure both data sets and be sure the analysis technique is correct.

Walt

I have experienced a case when vibration data showed classic pole pass sidebands which was not confirmed by the PdMA off-line rotor bar check. Unfortunately, I did not have a chance to follow up. In general, I'd trust the vibration pattern but trend it for a while before calling it.

Regarding 6x on the motor... Is it related to the number of lobes in the screw and possible resonance?

Whose MCSA/ESA device did you use?

What type of resolution do you have set in the low frequency spectra around the line frequency range?

You have a very tight slip frequency. If the resolution is not set right, then you will see the 2Sf frequencies in the line frequency peak. If it is load related, you will also see either a large number of peaks around the line frequency or a 'raised noise floor.'

Are you able to perform an inductive rotor test? Normally performed with either the PdMA (RIC Test) or ALL-TEST Pro (RCA test). This can be used to confirm.

Finally, can you make the signatures available on this forum or by email? (howard@motordoc.net).

Sincerely,
Howard

Article on MCSA (from when I was with ALL-TEST Pro prior to 2005).

Note: One of the case studies relates directly to what you are running into, although ESA did detect the rotor.

Howard

Motor_Current_Signature_Analysis_and_Interpretation.pdf (451 Kb, 71 downloads)

Doc,

All of my data is collected with a CSI 2120-2. Both the vib and current data were taken with 6400 lines of resolution. I'll try to attach the data but if it doesn't work I'll email you copies directly.

Walt,

I retested both vib and current three times with the same results. I've been doing this type of testing for 21 years and have found a few rotor problems over the years with the vib and current data confirming each other. I have a great deal of confidence in my testing procedures but this time I'm getting very mixed signals. This ones got me baffled.

Dave,

This is a centrifugal compressor so lobe pass isn't an issue. I can't find any force that I can readily relate to 6X.

I appreciate everyones input.

This message has been edited. Last edited by: Joe Gurney,

Rotor_Bar.doc (117 Kb, 65 downloads)

Joe,

It would seam that two data sets (observation and vibration) agree and that MCA does not. I don't put a clamp-on transformer on 4160-volt lines, so how did you get the current spectrum? Was the current waveform peaks flat or show amplitude modulation? Does the "current sensor" have appropriate wideband response?

If the current spectrum is accurate, then there is no significant mechanical or electrical torsional modulation of motor rotor. So the vibration modulation would have to be caused by a radial or axial mechanical force. How can the vibrations be caused by motor rotor stray current with cast aluminum rotor or not? I think more detective work and measurements are needed. It is getting close to Halloween, so ghosts and goblins are possible!

Walt

Actually, you do have rotor related frequencies in this spectra.

However, Walt brings up a good point, that was also identified in the article. Did you take your data from the control CT? This is important, as a control CT may not have the same frequency response as the CT used by the instrument (assuming you are using CSI's CT). This results in a 'dampening' effect (much like trying to take vibration spectra on a sleeve bearing). The result is that the 2xSF may not be exactly correct and the peak may be lower than if you took the reading directly.

In either case, it looks like you have sidebands with harmonics. These will relate to either a casting void or some other rotor (including load) related issue.

Can you provide a current spectra from 0 to 120Hz? And one that shows from 0 to about 1400 Hz?

Howard

Doc - you are suggesting that the CT has a lower response to 59hz and 61hz than to 60hz? Enough to wipe out the sidebands? That explanation doesn't seem reasonable to me.

Another log on the fire:

Is it possible that the motor current data was taken with compressor at low load, while the vibration data was taken at high/full load?

Walt

How about a cyclic load??? Our compressors cycle full-load / half-load every ~ 3 - 5 minutes. Was there a soft foot check performed on this motor?

Walt,

The CT was on a line lead. We shut the unit down, opened the breaker, installed the CT, closed the breaker and restarted and loaded the compressor to full load. The compressor tech monitored load for me to insure it didn't change. The customer took another compressor off line to help insure demand. I'm on my home computer now so I can't get to the waveform and I don't want to guess at the answer about the peaks. I'll check it in the morning and get back to you. As far as the sensor, I've been using the same one for years without problem so I have to say it's not a CT sensor issue.

Doc,

I didn't collect any current data with a higher fmax and the compressor is 60 miles from me but if I can get back there I'll get more data. I always use an fmax of 75 Hz to insure adequate resolution. What are you looking for at the higher frequencies?

Jamie,

I installed and aligned this motor in 2001 and I can tell you it was right at that time. As far as I know it's not been moved but to answer your question, I didn't do any soft foot checks during this trip.

Joe:

A number of mechanical and electrical-related fault frequencies can be detected using MCSA and ESA. Please find a presentation at:

http://www.motordiagnostics.com/Video/ESA%20Part%201.html

If you are unable to view the presentation, you can access the information by going to http://www.motordiagnostics.com/presentations.htm

There is a significant amount of information that you can download for free.

Sincerely,
Howard

Joe,

Broken rotor bar problem is not the only one that causes slip frequency times # of poles sidebands around running speed harmonics.

Think of dynamic air gap variation... Synchronous magnetic field runs faster then rotor by amount of slip frequency times # of poles, thus, modulating the unbalance (centrifugal) force at this rate due to unbalanced magnetic pull.

Therefore, if current analysis did not show sidebands while vibration did, then it is not a rotor bar problem.

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