wahts the reason of a peak at 50Hz in the spectrum 1

[serbek]

hi,

can anybody help me

I measure vibration of an roller bearing with defects in different shaft speeds with a vibration sensor, i convert waveform to frequency spectrum by using FFT , and analyse the result in frequency spectrum,
but i found a peak at 50Hz, I think it can be caused by electric motor but I use a DC motor,

can a dc motor cause peak at 50 hz?

 

[MikeyP]

I presume your local ac supply is 50 Hz?

You can usually tell if the 50 Hz peak is due to a real effect or ac mains interference by looking at the width of the peak. If it is very narrow compared to other peaks on the spectrum, then it is probably mains interference.

Common causes are:
1) Damaged shielding on the cable linking the accelerometer to the signal amplifier or from the signal amplifier to the data logger/analyser.
Solution: replace cables, use differential amplifiers.

2) Poor earthing of mains powered equipment
Solution: Get your electrical safety officer to test the earthing of your various pieces of equipment.

3) Ground loops
Solution: Try to earth all your equipment to the same nominal earth point.

M

--
Dr Michael F Platten

 

[pjhype]

I once knew a guy who traveled all over the world "measuring" vibration levels in semiconductor facilities. When I analyzed the data, I found only peaks at 50 and 60 Hz. It turned out that he was using ICP accelerometers and forgot to turn on the ICP power supply in his analyzer!

pj

 

[GregLocock]

The quick way to check for mains is to switch the system under test off. If the 50 Hz stays on the screen you are picking up stray mains.

However that may not help in your case since it is possible that the DC motor is driven via a mains power supply, and it is the thyristors (or whatever) in this that are radiating mains frequency.

Can you describe your instrumentation setup?





Cheers

Greg Locock

 

[serbek]

thank you for help

I use test an equipment which has bearing-shaft system, driven by a dc motor. shaft is connected to motor using a belt. and bearing is 6205 ball bearing

as I investigate the spectrum in details, the magnitude of peak values at 50 hz is greater in healthy bearing and this magnitude increases with increasing shaft speeds.(from 1000rpm-4000rpm)

another point is, lower peaks at 50hz occur in defected bearings and in some cases it can not be recognized and hide in background noise.

yes our local ac supply is 50 hz but

I think if these peaks caused by the transduser or amplifier or earthing, peaks must have approximately same magnitude in defected and healthy beaaring, but its no.


and another question is why magnitude of peaks increase with
sahft speeds

regards...

 

[MikeyP]

I was helping out a colleague with a modal analysis of a non-linear plate last week. Not only did we have additional spikes at 150 and 250 Hz but running them through the curve-fit actually yielded sensible looking modeshapes!

M

--
Dr Michael F Platten

 

[mcdm]

just curios your machine isn't running at 5oHz/3000cpm is it?

 

[mcdm]

I’m just curios is your machine running at 50Hz/3000cpm?

 

[Franko]

If you have an FFT, zoom in to be sure it is EXACTLY 50 Hz - if not it's a resonance - at least do a rap test. More vibration with defective bearing could be giving more broadband excitation to increase amplitude of 50 Hz. If it is exactly 50.0.... Hz, the coupling with mains frequency could increase as power is increased during speed increase and still be instrumentation error. Try using a battery-powered accelerometer -

http://www.pcb.com/

 

[steve383]

If it was a bearing defect(ball spin I presume)it will travel up the spectrum as speed increases. If it is electrical it will stay at 50hz. In DC motors I have seen increased line frequencies with power isues. Like loose connections and failing SCR cards. I have also seen line frequency issues with brush\com. faults. If brushes are hanging up line frequency will increase. Check your brush wear pattern, brush tension and brush contact.