Ronnie - That is surprising. The single-phase test has always been the one recommended by the shop repair old-timers.

Did you guys come up with any theory why it didn't work on this particular motor? Maybe a strange winding configuration like a delta vs typical wye? Maybe there were three breaks spaced at the exact distance where exactly one of the three would always be over a given phase. Mabye several adjacent broken bars cover an arc equal to 360/p where p is number of poles, so that it spans 3 adjacent pole-phase groups and when it leaves one pole phase group it is arriving at another pole phase group of the same phase. Maybe this motor had a very large number of parallel paths through the winding which minimized the effect of variation in one path ? Just wondering what the explanation might be.

Sorry Pete. That has been a few years back and we did not investigate any futher after seeing the end ring separation. Being from an electric motor background, I was taught single phase test would show an open rotor. It was the only test I knew of for a long time. And I have seen it work. But since then I have placed a lot more reliability on current analysis when sufficient load is present. As for the RIC test, I just don't know. I personally have not seen any open rotors caught with this test. But I do not do this kind of test and have only been involved a couple of times when it was used. My bet goes to current signature analysis under load of at least 60%. And I have been told 60% is a minimum. The higher the load, the more likely the rotor bar problem will show up. Sorry I cannot provide more insight.

Thanks Guys
This is very infomative to the electrically uninformed.The biggest motors we have here are 700 hp. I would love to see some of the bigger units described like the 23000 hp unit that Ronnie spoke of. E.Pete a boroscope would be a nice addition to our pdm tool list and could be utilized in many ways. Again Art has exposed us to his vast experience through his programs. This man should either be cloned or something because we should never lose his knowledge base.
I've never been more influenced by one man's knowledge than by Mr. Crawford. Sorry but I have to add my Dad Before Art.Thanks to all for the info shared.

Thanks Ronnie. At least now I will be careful or skeptical about the single-phase test. As always, it's better when you have the luxury of multiple tests telling you the same thing.

Reply to all, especially MotorDoc,
I don't know what software you use to identify sidebands in a current signature. I have always just set up a point in my route and used a current probe with an AC output into my dataPAC. There are most definitely multiple sidebands even in the current signature. I have attached an example case (cast rotor no less!) that not only identifies the multiple sidebands in BOTH the vibration and current sigature, but also illustrates the problem with taking data below 80 % load.
Enjoy!
Ron Brook

rotorbarcase1.doc (1,275 Kb, 38 downloads) Broken rotor bar

Sorry,
Correction to the last part of my attachment. I accidently reversed what you should be looking for in vibration (sidebands surrounding 1x,2x, 3x) with current (sidebands around the fundamental current frequency). I also think I answered my own question to MotorDoc. Just symantics. There are multiple sidebands in the current signature, but the only one you need to consider is the first sideband to the left of the fundamental current frequency, which was in this case 60 Hertz.
Hope everyone had plenty of Turkey or Tofurkey if that is your lifestyle

Ron,
Good posts. I have disassembled a few motors that looked like that. Anyway, thanks for posting the data. You are taking the current readings the same way we do only we are using CSI. I just purchase a Fluke I2000flex with the flelible clamp with a 7.5 inch diameter. Comes with all the attachments to plug right into the analyzer and multimeter. Can't hardly wait for the opportunity to use it.

Thanks,
Ronnie

I agree, that's a great case study covering current, vibration, multiple load levels, and best of all.... inspection results. Thanks for sharing that Ron.

Does anyone know what causes a cast rotor to fail like that? Same factors as cause fabricated rotor failures (frequent sever starts causing mechanical degradation over time)? Or maybe individual starts were so severe that they lasted long enough to allow the rotor aluminum to reach its melting point?

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

Dear Ron:

Yes, the PPF sidebands around line frequency will show you the rotor bar issue. However, additional sidebands around line frequency will usually relate to situations from the rotor out, such as unbalance, bent shaft, and load related problems.

Howard

I would like to revisit this important thread. There were a numerous good points discussed, but the Ron Brook's point of multiple sidebands did not get the attention it deserved.
I am attaching a current signature from 400 hp, 1800 rpm motor. The motor was sent to the shop for reasons unrelated to the rotor problem. During the routine tests (for this size of the motor) it was discovered that one rotor bar is broken. The rotor is a cast aluminum, closed slots, 48 bars. There is absolutely no doubt that it is one and only one bar. Even the growler test showed 1 only bar broken beyond any reasonable doubt.
Upon the request from the customer the motor was coupled to the test generator, a 500 kW calibrated unit, and it was tested basically in laboratory conditions.
The load was a full 100%. Loading the motor with the DC generator can be considered the "smoothest" load possible. The variations of the load during one revolution can be considered nil.

And yet there are numerous sidebands, and numerous they are!

I can supply more current signatures with lots of sidebands. I would have trouble supplying any current signature with 2 sidebands only. It is my opinion, and it is my experience, that the numerous sidebands due to broken rotor bars is a rule, not the exception. In the following thread I post the theoretical reason for the multiple sidebands.
Based on the above (and the theory) I consider the statements such as: You only get 1xPPF sidebands around the line frequency for broken bars..." just another motor myth.

jank

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

400HP,1775RPM(10348).pdf (212 Kb, 16 downloads) 400 hp, full load

For those mathematically inclined here is the proof that the multiple sidebans in current signature is the rule, not the exception.

jank

MULTIPLE_SIDEBANDS.pdf (100 Kb, 19 downloads) why are the sidebands so numerous in current signature

Wow, so what you are saying is that this:



Cannot happen, according to your formulae. Which also means that all of the work from Oak Ridge National Labs, the motor diagnostics industry and the universities involved in such studies are incorrect.

This is not the first time that you and your friend have made statements such as the one that you are making. And when others, such as myself, attempt to answer, we are attacked. I am very disappointed in this type of stuff and the confusion that people have expressed to me at the recent PdM and IMC conferences.

What you have identified in your latest 'theories' is that there would be multiple sidebands in any machine that had broken rotor bars. Then you show some formulae for frequency modulation (FM). If you understood how ESA/MCSA operated, you would understand that you are looking at AM (Amplitude Modulation),which is how sidebands occur (see the attached slide). Basic physics. This also means that in the demodulated spectrum you would receive one peak at the Pole Pass Frequency.

The type of Current that develops this type of sideband looks like this (from the same machine as above), taken from the CT of this medium voltage, aluminum rotor cast machine. There are two broken rotor bars:



Additionally, if things were to operate the way that you state, you would not be able to observe the rocking of the needle on an analog amprobe as the broken bars pass each pole.

In your data, you have a fairly broad base on your line frequency. This normally means that there is some type of torsional issue in the load (ie: reciprocating compressor, punch press, rock crusher, etc. etc.). I would also wonder where the speed of the motor was determined because the type of signature showing on this machine appears to be more load related, because, with the amplitude of the claimed PPF, you would have far more than two broken rotor bars.

a20_ESA_Analysis.pdf (35 Kb, 12 downloads)

quote:

In your data, you have a fairly broad base on your line frequency. This normally means that there is some type of torsional issue in the load (ie: reciprocating compressor, punch press, rock crusher, etc. etc.).

He already told us what the load was (a generator, not a reciprocating machine punch press or rock crusher). What you are seeing is called spectral leakage related to the digital processing performed on the waveform. That spectral leakage is present in your spectrum as well. It is more noticeable in Jan’s spectrum because he has a lower noise floor due to higher frequency resolution.

quote:

I would also wonder where the speed of the motor was determined because the type of signature showing on this machine appears to be more load related, because, with the amplitude of the claimed PPF, you would have far more than two broken rotor bars.

I see a ratio of maybe a factor of 200 (46 Db) between the fundamental and the highest sideband. If you are saying that a 46 dB difference guarantees “far more than two broken rotor bars” then I disagree.

quote:

What you have identified in your latest 'theories' is that there would be multiple sidebands in any machine that had broken rotor bars. Then you show some formulae for frequency modulation (FM). If you understood how ESA/MCSA operated, you would understand that you are looking at AM (Amplitude Modulation),which is how sidebands occur (see the attached slide). Basic physics.

Jan’s explanation was based on AM, not FM. For an explanation of the difference between AM and FM, see www.dspguide.com or ask the knowledgeable people on this forum.

quote:

The type of Current that develops this type of sideband [57hz and 63hz] looks like this [shows a time waveform with a total duration of 0.05 seconds]:

The most distinguishing feature of the TWF for a broken bar would be the amplitude variation which occurs over the long period of the 3hz envelope frequency (approx 0.3 seconds). But you didn’t give us anywhere near 0.3 seconds. You only gave us (0.05 seconds = 3 cycles of 60hz but only one sixth of a cycle of pole pass modulation). From the brief snapshot of TWF you provide, we can’t distinguish any modulation. The distortion shown in your TWF appears periodic at 60hz (syncronous to the stator field rather than the rotor angle) and is therefore not associated with any rotor defect. (I’ll bet you have harmonics at 5LF and, 7LF in your spectrum reflecting saturation or supply voltage distortion.).

quote:

Originally posted by MotorDoc:
This is not the first time that you and your friend have made statements such as the one that you are making. And when others, such as myself, attempt to answer, we are attacked. I am very disappointed in this type of stuff and the confusion that people have expressed to me at the recent PdM and IMC conferences.

I haven’t expressed any opinion on this subject up until now, but it looks to me like Jan has come up with a case study to contradict your statement made 14 November 2006 08:11 AM “You only get 1x PPF sidebands around line frequency for broken rotor bars – that's it, no more, no less.”

If someone came up with a case study that contradicted a rule that I had stated absolutely (“that's it, no more, no less.”), then I would be curious to understand more about it. Your response is predictably different.

As for creating confusion through misinformation, the most blatant example I have ever seen is the thread entitled “quiz - torque producing force acts on bars or slots” and it is not Jan or I that is providing the misinformation:
http://maintenanceforums.com/eve/forums/a/tpc/f/7161085912/m/3641027862/p/4

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

Yes, actually, a lot of misinformation. Throwing together a few formulae from books - and I have read those books - as proofs is not engineering and very misleading.

I read all of your insults in the other string and was waiting for a warning from those owners of the Forum. Seeing as they are not prepared to stop the lunacy, I think it is time to put this to a stop.

Unfortunately, I normally do not like coming into a battle of wits with unarmed opponents, but I think I will make the exception. I have been far too nice after a threatening phone call from someone when dealing with certain individuals last year, from this forum. It concerned my family, so I have backed off when receiving personal attacks, but no longer.

I had also reviewed your material on the other string and found it lacking. So, we will have to correct you on that one, as well. (Although, I found it a little interesting how obsessed you have become with me, which is a little disturbing, it is ONLY a forum, after all).

Have a great evening. I will do my best to find the time to provide a little training for you.

So, hurl the insults, etc. all you wish. The truth always prevails.

Howard

quote:

Originally posted by MotorDoc:
I had also reviewed your material on the other string and found it lacking.

Well it has been about two months since you first said that, and the amount of scientific rebuttal you have come up with is still 0.

quote:

I will do my best to find the time to provide a little training for you.

I will look forward to it, since I am sure the outcome will be the same as every other time you have tried to provide a response. When can I expect you to provide a response?

Oh, not to worry. You begged for my attention. Now you have it. But don't think what is coming is for your benefit.

I reviewed your responses in an airport last night, I am responding to you from an airport right now. I am sure I can scratch up a little free time.

motordoc makes the following statement.

quote:

The type of Current that develops this type of sideband looks like this (from the same machine as above), taken from the CT of this medium voltage, aluminum rotor cast machine. There are two broken rotor bars:

Then he shows a picture of a waveform.

Could this be the wrong wave form, motordoc? Not to say anything about the conclusions, but this waveform is periodic, and would not have (what I know of as) sidebands. For we all know that a periodic signal consists of a fundamental and harmonics, only.

Depending upon the window used, this could be spectral leakage as suggested by another.

http://www4.ncsu.edu/~chow/Publication_folder/Conferenc...on03_IFDIM_Ayhan.pdf contains a paper pertaining to some of the spectral ideas and compares 3 non-parametric spectral methods for analyzing motor signatures.

Originally posted by MotorDoc:
quote:
__________________________________________
Wow, so what you are saying is that this:
_________________________________________

No, I was saying nothing. Everything what was said after the above sentebce was said by you.

quote:
__________________________________________
Cannot happen, according to your formulae.
___________________________________________


Of course it can happen, but I consider it highly unlikely. If the C=D=E=.....=0, it will happen and I said it.

quote:
____________________________________________
Which also means that all of the work from Oak Ridge National Labs...
________________________________________________

I merely said that I am finding lots of sidebands and the 2 sidebands only looks like a myth. And I have a bunch of spectra to support this finding. Why don't you go to your database and find similar bunch of spectra with one sideband only? The Oak Ridge - those are the guys that developed the bomb, aren't they? I don't need any one of them on my tail!

quote:
___________________________________________________________
Then you show some formulae for frequency modulation (FM).
____________________________________________________________

Bill and Electricpete have already refuted this error. It is indeed the amplitude modulation what I was talking about. I considered the impedance variable, but not because of variation of the frequency.

quote:
______________________________________________________________________
The type of Current that develops this type of sideband looks like this...
_______________________________________________________________________

No it does not. Again it was refuted before. It is the waveform taken with upper frequency something like 5000 Hz. I can present you tons of them from my database.

quote:
__________________________________________________________________________
Additionally, if things were to operate the way that you state, you would not be able to observe the rocking of the needle on an analog amprobe as the broken bars pass each pole.
__________________________________________________________________________

In your mind if the needle is to move it has to be the exact sine variation. No, any vartiation will move the needle.

quote:
____________________________________________________________________
In your data, you have a fairly broad base on your line frequency. This normally means that there is some type of torsional issue in the load (ie: reciprocating compressor, punch press, rock crusher, etc. etc.).
________________________________________________________________________

No again! The load was a DC generator. The base of the line frequency widens due to slow events. If you unload the motor, the speed increases, hence the slip frequency decreases and the sideband moves closer to the line frequency peak. The impacting from the reciprocating pump or similar equipment is way too fast, and shows far away from the line frequency peak, something like 20 to 40 or so Hz. I have an excellent case study to prove it. But it is not in presentable form yet.


quote:
________________________________________________________________
I would also wonder where the speed of the motor was determined...
________________________________________________________________

The speed was measured with the strobe light. From the measured speed I calculated Fc (calculated position of the 1.sideband, see under the sectrum) and compared it with the real position of the sidaband. The correlation was excellent.

quote:
__________________________________________________________________________
with the amplitude of the claimed PPF, you would have far more than two broken rotor bars.
_________________________________________________________________________

There is only one broken bar. Period. The decibel ratio is 50.5 dB, see on the spectrum.

I am attaching spectrum (with multiple sidebands)of a 600 hp motor. She is in bad shape. Every time when they press the button they have their fingers crossed. The details are on the spectrum.

jank

600_hp,3600rpm,_pump.pdf (185 Kb, 21 downloads) 600 hp multistage centrifugal pump

quote:

Originally posted by MotorDoc 10 December 2006 04:37 PM :
What you have identified in your latest 'theories' is that there would be multiple sidebands in any machine that had broken rotor bars. Then you show some formulae for frequency modulation (FM). If you understood how ESA/MCSA operated, you would understand that you are looking at AM (Amplitude Modulation),which is how sidebands occur (see the attached slide). Basic physics.

Well Howard, you have suggested that someone has a misunderstanding, and now it turns out that it is you who had the misunderstanding. The proof was in fact based on AM.

You have had time to make posts in other places, but apparently no time to come back to rectify this situation.

Is it a case that you still don't understand the difference between AM and FM? Or is it that you don't think that you owe other people the common courtesy of apologizing when you insult them as a result of your own misunderstanding?