Hi All,
One DC motor was noticed with high intermittent noise/vibration. It showed up and disappeared randomly. Can you help to explain the symptom? High energy around 25 and 75Hz while noise appeared? Are they natural frequency? What cause the symptom and how to resolve it? Why randomly?

Thanks, Kevin Ku.

XE04-PRS2-belt-6-23-22009.pdf (493 KB, 66 downloads) pictures and spectrum

Can you zoom in on the area shown?
I guess this is the are showing when there is noise, huh?

Has this motor been recently lubricated? I've seen this before when the bearing rolling elements skid instead of roll. Wierd to see.
Just my first thought.

Hope this helps,

Regards,

Hooch

No lubrication problem I think, it can be confirmed from Peakvue spectrum.
Auto-correlation showed time waveform vibration around 25Hz.(While Fmax set @ 30Hz)

Thanks, Kevin Ku.

Thanks Kevin.

I see the date shows about 15 minutes difference in the one I posted above and the one you correlated. Does the 10:30 correlated one have the same pattern as the 10:15 one?

Made an error here, too much hurry. SORRY.

Looks like the difference in frequency is 25 Hz. What is located near 25 Hz in the spectrum? Also what is located near 275 Hz?

Can you post the Peakvue spectrum with a .1 vertical scale instead of autoscaled?

What is the cycle time of the noise, steady or random?

This message has been edited. Last edited by: Ralph Stewart,

I saw cases like this but without a belt. They were small size motors (around 50 hp) with rolling element bearings. In the cases I saw the problems were bearing related. We did regreasing but the problem persisted. Yet, when the bearing was removed, the defect was not clear (maybe because, the problems were captured at an early stage).

In your case, please notice that all values of the crest factor are relatively high.

However, the pictures you attached are interesting and encouraging to check belt related faults. One of the common belt frequencies is subsynchronous and its harmonics. Have done the calculations? Have you inspected the belt?

Let us know your findings.

Kevin,

Likely, something is exciting belt resonance, possibly due to variable load when tension is changing. The belt could also be worn out.

If you have belt tensioning device ( from Gates or somebody else ) check current tension and resonant frequency, then tension properly.

David

Cool pictures and a weird case.

If I am reading it right:
1 - In the first spectrum, it is 20 mm/sec at 25hz and 100/sec at 75 hz. (is it exactly 25 hz or could it be 26hz = twice running speed).
2 - running speed 13hz
3 - line frequency (input to dc supply) is 50hz?

Two guesses come to mind.
First - could be a bearing. I think we have all heard mysterious rolling bearing noises that randomly come and go. Can be skidding, or an outer race rolling within housing or inner ring sliding on shaft, or perhaps ball defect changing orientation so it hits and misses. Perhaps the violent broadband vibration from a ring sliding in the housing excites a belt resonance.

Second, if line frequency on input to your dc supply is really 50hz and those frequencies are exactly 25 and 75hz, than I think those related frequencies are too coincidental to be a bearing - must be electrical. Take a look at the dc voltage waveform to see if it shows a 25hz ripple for some reason.

Hi

Last week i have came across a similar problem, a VFD drive motor with toothed belt, I have checked and inform the workshop guys its belt misalignment but after 2 day of running the noise and vibration reduced.

When the plan stop we remove the cover to confirm and its was the belt which have cause the motor shaft worn out(at the inner ring of the front brg, and thats explains the reduction of noise cause it has pull the shaft to near align

So in your case make sure its is not the belt misalignment because it can damage the shaft and since its expose you can try to spray some silicon while its vibrating see if its reducing.

good luck

Kevin,
First, excellent data and excellent pictures! It is not often we get this much MEAT to chew on.
First, this problem has no middle ground. It is either smooth or bad, correct?
Second, the problem is not really related to bearing frequency calculations, correct/
I just sent an email to all of the field techs with my firm and I believe you are seeing the same thing.
Motor/VFD, is a system. The system is controlled by two very important signals at the drive, the reference signal that tells the drive what the process wants out of the motor and the tach signal from the motor that tells the drive what the motor is currently doing. Once these signals are compared, depending on the settings in the drive, corrects in speed are sent to the motor windings. Depending on the amount of error in the signals and the attack rate the drive has been sent to, these changes in speed can be violent.
Remember to always look at the entire system. Check the tach output and see if you are getting intermittent errors. Then check the encoder board to see if it is doing it's job properly in comparing the reference and the tach signals.
Good Luck!
Ron Brook
Again, thanks for the great data and pics!

I have a couple of plants with older power houses that still generate some 25hz supply. In at least one instance, they are required by the power utility to continue producing a small degree at 25 hz.

Just to fill out some details:

What is being driven by this belt?

The speed ratio 1:1, correct?

Are both halves of the belt flapping at the same time or just one half flapping at a time? If just one half is it the slack side?

EP,
Go back and look at the time waveform and ask yourself what could cause an almost instantaneous change in amplitude, over and over again. That's the key. Mechanical issues, just because of damping, don't usually do that.
Ron

I think your comments are on target as usual Ron and your theory is probably the most likely.

I just realized we have missed some significant data we usually have (speed of both machines, nature of driven load). If the driven machine is running near 25hz, that would obviously point in a different direction (eyeball of the sheave size tells me it's a 1:1 speed ratio, but I just wanted to double-check). If the driven machine is a type that is expected to undergo large rapid change in torque load (something like a woodchipper whose load changes based on product input), then that could be relevant also.

Also I was trying to understand the significance of the belt flap. Belt flap on slack side could correspond to increase in torque load. Belt flap on both sides would be something else

EP,
Good points.
Kevin,
Let us know what you find.
Ron

Hi All,

One more data taken this morning while motor speed increase to 1086 rpm. The similar vibration pattern persisted, and belt flapped more often than Yesterday. Amazing thing happen this afternoon after we sprayed "belt dressing". The major noise and vibration gone immediately(motor speed 1121 rpm). Widespread spectrum around 25Hz and 75 Hz gone and overall vibration dropped dramatically. See attached spectrum. Any explanation you have?

Thanks, Kevin Ku.

XE04-PRS2-belt-6-24-22009.pdf (108 KB, 23 downloads) spectrum

Likely 25HZ, and 75 HZ ( probably 50 HZ as well ) are belt resonant frequencies with associated mode shapes excited by poor meshing and which later was corrected by "belt dressing". These natural freqiuencies are still there.

I am curious as to what is # of pulley teeth * RPM value?

Why wouldn't you bump the belt?

vellangana was the one who suggested the belt spray. I don't really understand why but I'm interested to hear more about why he suggested it.

At the link below is a brief discussion of the idea of using a sprayed on solution to troubleshoot a vee belt. If the noise goes away it's supposed to indicate belt problem like excess slip. When you spray it its' still slipping but the friction changes to be quieter (perhaps stick/slip goes away?). The explanation kind of makes sense for vee belts. But for sync belts which are supposed to have no slip it's a mystery to me it would ever make a difference in the noise/vibration
http://www.eng-tips.com/viewthread.cfm?qid=210528

Hi,

electricpete, You are right in the sense the sync belt should not have slip.

But in this case as it was mention intermittent noise/vibration of the belt, which i suspect that the belt are moving or sliding and rubbing againts the pulley,s coller and since the motor pulleys coller are stiff enough eventually it will push the belt back to the centre of the pulley.
(as you could see it from the photo the coller is welded and in original its only press fit, which will pushes off the collers in case of excessive belt misalign.

So what really happen when the belt dressings are sprayed it lubricated the surface between the belt edge and pulleys coller and as time goes the problem will back again. A good belt alignment are needed to solve the problem.

OK, now I am at a loss. Why, if this is belt misalignment, isn't there more vibration at 1x and 2x operating speed? And if the belt were rubbing the 'retainer' on the motor sheave, that should also set up a vibration at 1x running speed on the motor ib bearing, as well as excite the belt resonance.
As far as why the belt dressing eliminated the vibration at the assumed belt resonance frequencies, don't know. That would eliminate any torsional issues that could be created in the motor by the belt rub on the 'retainer'. I would still like to see the encoder output tested.