Hello guys (and girls)

This is actually my first posting here, as im a freshman in this vibration thing.
So bear over with me

We are having a problem at our plant with a small motor running at appr. 2946 rpm.
Measuring in the axial direction we get a high vibration at appr. 4,2 mm/s rms close to 3X.

This is a brand new motor, and it is driving a flare blower.

As we can se from the spectrum, we the first peak at 1x, with some harmonics, and the highest peak lies just above 3X.
The vibration also modulates/beats with a freq of appr. 4-5 hz. As all of us at this plant have no experience regards to vibration, we are at this point still trying to study our way towards a solution on this one.
I therefore hope some of you can give me some clues regarding this issue...
I have attached 2 figures, showing the spectrum from 0-500 hz and also the waveform with a timebase of 150 ms.

Doc1.doc (109 Kb, 175 downloads)

Are you using a 3 jaw coupling? If so, could be some misalignment or improper coupling gap. I've seen this before when the coupling halves are too close together (no gap) causing the flex member to be "squeezed" by the two 3 jaw hubs.

Can you give us more detail on the setup of this blower (coupling, motor shaft mounted fan?, # of fan baldes, etc.)?

My eyeball tells me it might not be exactly 3x. It looks like maybe 3.1x with another peak out there at twice this frequency or 6.2x. Can you check the exact frequencies?

Also maybe some additional data would help fill in the picture. I like TWF in acceleration. Also spectrum with high Fmax, maybe 300kcpm and 3200 lines.

There is no coupling on this, the motor is driving the blower directly.

And pete, you are correct. It is just above 3X, and also as you say, a harmonic of this one at appr.6,2x.
At this point we have no other readings of this one.


The blower has 10 blades. The air is suct in axially ,and then pushed out centrifugally.

TWF-Time waveform?
It is actually measured with accelerometer integrated to velocity.
But i'll take a note of that, and do a reading without integration next time.

This message has been edited. Last edited by: Svein Hansen,

It could be a bearing fault frequency. Maybe BFPO on an 8-element rolling bearing.

Do you have the bearing part numbers? In that case we could see if the frequencies match up.

It would be unusual to see that high magnitude fundamental bearing fault frequency, but you never know. There are not many sources of non-syncronous vibration.

Bearing type is 6309. BSFO is 3,0367 x 49,375 which gives a freq of 149,9 hz.
It is in the area of the peak.
So one theory is that it is the BSFO or maybe the BSFO driving a resonance in the same area.
Anyhow, it seems like theres is a harmonic of the peak as earlier said around 6,2x and im not certain about if either the resonance or the BSFO could give a harmonic?

It is very common to see many harmonics of outer race frequency in the case of an outer race fault. Typically the highest harmonics are in the range 30,000 cpm - 150,000 cpm of the velocity spectrum, which will correspond to a resonant frequency of the bearing. The harmonics come from the impacting (vs sinusoidal) nature of the waveform in the case of bearing defects.

As I mentioned, a high fundamental defect frequency is not as common and a little more difficult to interpret. When there is a high fundamental BPFO, sometimes we suspect a housing deformation. Also I have found it on one occasion to be associated with bearing overheating (which is easy to check by checking the housing external temperature while observing proper safety to avoid contacting rotating parts).

From what little we have I would say it is definitely a condition that requires further investigation. Maybe others have some more ideas.

Well as said pete, im all new to this, so my information may be limited. However, we are planning on doing new measurements on it, for both checking for resonance and other things that may arise.
At this point im still trying to gather enought information, to know what i shall look for during the next reading.
However, i find it strange, and not likely in the first instance, that there could be a bearing damage on this brand new equipment.
It has only been tested and driven through the commisioning phase at out plant, and has not been set into operation yet.

For information,we have also another motor with the same type of pattern, running a little high on vibrations. That one is coupled up with a hot oil pump.
It shows almost the exact same values at the same spot. And also the same peak a little above 3x.
So this is as i see at problem that may arise with more motors. I think the motor is identical but just a bit smaller than the one im talking about now.

But when you say that you suspect a high BSFO when there is a housing deformation, it might be there has been some wrong installation of the bearing?
This could probably also be routed to the other motor we have...

Anyways, if there are any more advices out there, im more than happy to hear them.

I tried to set up the readings with sidebands to see if there was any match. As you can see the only sideband match i get, a little above 300 hz and down at 0,6 hz. That means there is a spacing at 151 hz?!
Makes any sense??

Doc2.doc (80 Kb, 68 downloads)

I think you are not seeing a sideband but the second order of the frequency. You say you get the same frequency on another motor, is it near this one? Can you take a reading with the fan off to see if it is coming from somewhere else? And I have seen a lot of new equipment with bearing defects.

Hello Jim, thank you for paying interest in this.
No, the other motor is actually half a mile away, so no interference there i guess.
And no, at this point we probably cant take any readings with the fan off, because it is directly driven by the motor.
If there will be a need for dismantling, sure i would guess that this can be done.
Probably, if we cant find any solution nearby in time, we would have to take it off in some way, to rule out the possibility of any problems in the fan.

But at the other hand, i guess that because the vibration is in the axial direction, there could be some force within the fan creating it.
Unbalance in blades maybe?
Or wrong fit on the shaft?

I'm not sure why you captured data on this machine, but for typical monitoring purposes, given an anti-friction bearing, the frequency range seems limited. I would think you would want an fmax of 8x BPFO or about30-40 x running speed of the machine in general

Hello Joe!
Well yeah, it might be.
Actually did them for baseline readings, but you are right about the freq, could be higher.
Still this was one of the first readings i did, and i was actually glad because i thought the readings looked good. Not to much noise and all that, except for the high peak
Yup, still in the learning phase!

I cant recall why we did them at that freq, but i think we got the advice from are vibration vendor, that we should set up the reading for generally 1 khz span.
Guess you can't live by that rule in all cases...

Svein,
What is that second peak? Is it at 305 Hz, or is it at 302.5 Hz? Are you sure that the tall peak is the primary peak? Could it be a sideband of the little peak to the left? Is that little peak dead on 150 Hz? If so, you could have a primary peak at 150 Hz with a harmonic peak at 300 Hz and they both could have one sideband to the right that is 2.5 Hz away. If this is the case, could this be electrical instead of mechanical?

David Eason

Hi David.

All those are interesting thoughts.
I im currently "off duty", staying at home with my family. But i will certainly check up on those things as soon as im back on work, and then you will get a reply.
Think i checked all the frequences for sidebands, but i might have missed that on because...well i didnt see it that way
But im very glad for all the responses i have received, so if there are any other, please post. I all they way back at the learning curve, and everything will be of help.

Hej Svein,

I saw something similar a few weeks ago while balancing a brand new cooling fan assy on a thruster drive. We had problems balancing that finally was a issue of too much clearance in fanmount on shaft but during the process the free end of the motor indicated hi bearing values and hi axial vibration specially after a few minutes operation. Finally it was with some persuation, since it was brand new from factory, inspected. It was found as endbell is aluminium cast etc., bearing that would have axial movement by sliding in the endbell for temp. expansion was stuck. So a new was fitted and all was fine. Just a idea, maybe similar to your thing? I think if i remember correctly that 3xRPM was one of the peaks that was large so I was looking for some loose things, not things that got stuck. God jakt efter felet, är motorn från det blå fabrikatet?! Olov

Tjenare Olov!

Blå er helt korrekt

Your problem, was it with a blue motor?!
As all of our mainly are blue, i guess this could be an issue with more of them.
As alsready mentioned above, we have one more motor with these symptoms in our plant.
And also today, i measures a smaller motor;Running fine with only 1x and some harmonics.Then after some minutes it started to modulate/beat, and i saw what i suspect is a rising 3x(hard to see exactly on the instrument).
So... what i see in common now for all of them, is this tendency that the vibration starts to modulate/beat.But this modulation has no certain time pattern to it. Seems more like something is coming into a load zone of some sort, with unfrequent patterns.

But this information is good olov, i will discuss them with my collegues, and try to find out if this could be a possibility.

Hej Svein,

Yep, it was a blue one.
There is a problem with aluminum caste endbells, not only blue ones that the bearing "easily" have a "tendency" to rotate in the aluminium and grind away, if it´s not found in these modern designs the rotor touch the stator as the clearance is small and that´s end of story. So I guess as always specs in cast details is a problem. And while whining, look at the cast feet on modern toys that never had a final grinding so they point anywhere from the casting, not talking about motors having so weak feet they need a concrete foundation to keep up the motorframe, otherwise it all sags, giving uneven rotor-stator clearance. Enough. Keep filling the foundations with concrete! Of the expanding type don´t forget. Olov

Yeah. I guess we would have to keep our heads up for this.
Also, as already mentioned, im pretty fresh in this.
Started out 4-5 months ago.
Im still trying to gather ends from the classes i've done, to make it all make sense in some way. There is SO much to learn, and you simply have to put your mind and soul into this to get going.
Anyhow, we have had some "problems" with several motors. From ISO norm, we would expect readings of 2-3 mm/s rms.
However, some of them run above this, and there is also discussion everytime about whether we should use peak or rms. Im still trying to figure that one out.
We did a solo run on a 2400 KW motor. Rms; it went around 0.6-1.5 mm/s.
Peak; 6-14 mm/s.
Of course, no one of them is wrong to use, you just have to know the difference.
However, vendors and many more people solely depends on rms readings, and when you are in a situation where the peak is 14 mm/s... you have some explaining to do.

Well, formally for acceptance test the ISO standards say in the standard case measure velocity, 10-1000Hz, True RMS. Idea is that it does correspond to the forces and there is a long time of experience that support the connection btw. the accepted levels around 3 mm/s and a long life of the machine. Start checking for a reason for the increased level at 7 mm/s and fix it back to 3 for that long machine life. Don´t run longer than necessery btw. 7 and 17 mm/s and don´t run the machine above 17 without fixing it and if you find 200 mm/s that is our top reading for rotating machinery, don´t stay around longer than needed..... Peak readings should not be so much off as you describe unless something strange is happening, there must be some transients that give that extra high peak levels, a look at the time signal may give some clue. Comparing RMS to Peak value have a fancy name and do have a kind of digital behaviour as I have seen it, first there is nothing and then suddenly it is a large difference and that indicate clearly some problems, that can be bearing faults, metal-metal contact of some kind, looseness etc. So it should be taken seriously (if the frequency range is similar for the different readings). Looking at the FFT should normally give some clue to what the problem is. (Unless it´s bad transducer cable, bad contact transducer to machine, not so good selection of measurement point etc.) Try to compare what you find in the data to what is found if you lift the top of the bearing house for inspection or other inspections and or repairing required when the machine is shut down so you can build a feeling for the level of trouble and the readings you find in the data. Happy hunting, start with the machines that breakdown often and find the reason(s) for that and fix them, then make sure the most important machines for the plant operation are cared for and then pick them in some importance and or trouble making sorting procedure and if you can look at trending, check those that increase drastically from reading to reading before those that have had higher readings for fairly long time. Just some 2c or what I would try to start to do. Olov