Motor Beating 1

[Jamino87]

Hi All

I have a number of small 1 KW electric motors running at 240V 50Hz. these are usually attached to pumps, but I am fairly sure this noise isn't emanating from the pump, but the motor itself.
The noise we're getting is a beating (adding/subtracting of multiple frequencies), not a single frequency. it's not all that loud, but because it's fluctuating, it's very annoying.
The noise of one particular motor can be changed from good to bad by loosening one of the screws that hold the stator shell to the pump wet end (axial screw, 1 of many). It's interesting too that this causes the noise to go from beating to a high frequency whine at Rotor Bar Pass Frequency. I'm unsure what frequencies are causimg the beating.
(see attachment, black is beating, red is the high frequency RBPF)

I understand that both these noise conditions can be caused by eccentricity of some sort, but can anyone explain why there'd be such a difference in tightening one mounting screw (it moves the mounting foot by only 0.2mm axially).

Thanks in advance for any assistance.

 

[GregLocock]

That ridiculously large file is completely unhelpful for figuring out audible heterodyne. You are looking for two peaks of roughly equal amplitude within about 10 Hz. Band passed time series analysis is a much easier way to look at beating (heterodyning).

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[Tmoose]

Please post pictures of this arrangement.

==================================

I'd be interested in the nominal frequency(s) that are audibly beating. It may not be the largest amplitude on a bearing housing vibration measurement.

One of these may help narrow it down a bit.

https://www.youtube.com/watch?v=ukQ6OSs3dWo

https://www.youtube.com/watch?v=M8gjsefNYpE

Of course processing your actual sound gor frequency content would be better. A decent microphone and your FFT analyzer probably can be made to work.
Applying A weighting would be better for comparison to understand what you are hearing.

Loosening just one of many screws allows the housing to move 0.2 mm / 0.008" ? Yikes.

=======================

Please post pictures of this arrangement.

 

[Strong]

Apparently your vibration spectra show two conditions loosen and tight? It would be far more useful to present two or more plots with different frequency ranges and label the known frequencies. If someone dumped a load of hay at your doorstep and asked you for the to look for a needle, what would you do? Do you know how many rotor bars and stator slots are present? If yes then label the plot with those frequencies. The sound you hear and the vibrations that are measured may not be easily matched without some effort. So far you've only presented casual observations about your machine rather than a scientific or engineering experiment or test.
Walt

 

[Jamino87]

Hi All
Thanks everyone for your replies. I have taken your advice and focused on finding the beating frequency, using Band passed time series analysis (thanks GregLocock). Please see the attached images (smaller than last time); 1 of the original time series graph, band-passed to the specific 2 frequencies; and the FFT spectrum showing the 0-1000 Hz and the 4 specific peaks that I believe are causing the beating (heterodyning).
So I believe it to be 739.305 Hz and 736.387 Hz that are beating, and these happen to be 2 x slip apart. Can someone tell me why this is occurring at around 737 Hz? It doesn’t seem to correspond to a multiple of rotor bars, stator slots or rotating frequency.
I am a bit green at this, so please excuse any shortcoming in my knowledge, that’s why I’m here.

"A" weighting was difficult to determine the spikes, but i know that is how we hear, not like the plots i've presented.

Ask away

 

[Jamino87]

the band-passed image didn't attach, try this