We have discussed the significance of 2 x line frequency in PeakVue and I see the common pattern all the time. It is usually stronger (ac induction motors btw) on the back end and is usually accompanied by rotor bar passing frequency with sidebands of 2 x lf in the velocity spectrum.

The data I'm posting caught my eye because of the sudden onset and the location.

I do have a trainee who collected this reading, and I was distracted by the really bad fan bearings, so that might be a consideration here. I really don't know what he could have done to have caused this though.

Any ideas?

2xlf.doc (52 Kb, 48 downloads)

Danny,

You need to have a discussion with Trainee and verify where he/she put the accelerometer. I would take no action on machine or even tell no one until you personnally verify the measurements. It looks like there might be sidebands around 7200 cpm (1xSS??) to make it more interesting.

Walt

What is your HP Filter, 1K, 2K? Do all the spectrums shown here have the same HP Filter?

Why were there 3 spectrums taken on November 18? Suspected problem?

What is your Fmax on the velocity spectrum? Does the slotpass show in it? Does the Peakvue overlap the velocity spectrum?

Does anyone have an opinion that when we see electrical vibrations in spectra, they may not be physical vibs at all, but induced in the accelerometer amplifier or piezo crytal by a disturbed flux field?

As a further poser, I was taking some measurements on a small motor the other day with a magnetic mount, and I could feel a vibration a good 5mm before I got clamped to the motor. Again, I suspect that the field in the motor was interacting, with the magnet this time, rather than the amplifier.

Walt,

I think you are right about reporting this. It is possible that this reading doesn't even belong to this motor. I will probably get their on-site guy to check it out (if he can figure out how to turn his data collector on).
Yes, very small ts sidebands. Good eye.

Ralph,

All except one of the Nov. 18 readings were with 120000 cpm fmax and 2k hp filter. Yes I suspected a problem back then, but determined that it was background from the very bad fan bearings.

Velocity spectrum is to 120000 w/ peak at 55 x and sidebands spaced at 2 x lf.

I'm not sure what you mean by your last question.

Danny,

In theory any force in an AC motor related to magnetic pull between rotor and stator, such as RBPF modulated by 2xLF should be detectable on stator body only ( not bearing) since the body is flexing as result.

So agree with previous statement that, assuming same data acquisition setup, the trainee just did exactly this way as oppose to previous measurement location.

Danny,

quote:

I'm not sure what you mean by your last question.

quote:

Does the Peakvue overlap the velocity spectrum?

I have seen some people take a velocity fmax of say, 2K Hz and a Peakvue with a HP Filter of 1K, which starts the Peakvue 1K Hz before the end of the velocity Fmax. This is good for seperating what is between the 1K and 2K area, if a problem might be seen there, but I would normally setup a Peakvue point starting at the least where my velocity Fmax ends and not before.

On your motor, it might be possible that if you have a 55X peak in the velocity data which is before the 2K Hz Fmax, there might be a harmonic of this beyond the 2K Fmax and this is what Peakvue is seeing.

Only my opinion and I could be wrong.

David,

This motor does have extended endbell supported bearings and we did have quite a crowd there and he may not have been able to get to the bearing. The next logical point is on the stator so that is a good possibility. I'll ask him next time we speak.

Your explanation also gives me the answer to a long-unanswered question-why do I see this pattern on the back end and not on the front? I take most back end readings on the stator because of the fan shroud.

Ralph,

I was taught in PeakVue School to use the hp filter >= the fmax, ie 1000hz fmax, 1000hz hp filter, 1001 hz fmax, 2000 hz hp filter. I will have to give the overlap thing some thought.

You are probably right about something higher and I would guess that it is 110 x running speed modulated by 2 x lf.

Thanks for your help. You guys really drilled down to the problem here. I'll let you know what the rookie (my son) has to say for himself.

Danny

Bust him Danny !
And I'll buy dinner next week.

Dave

Sweeeeeet! I'm having pie!

Danny

quote:

Originally posted by Danny Harvey:

Your explanation also gives me the answer to a long-unanswered question-why do I see this pattern on the back end and not on the front? I take most back end readings on the stator because of the fan shroud.

Danny,

Most people do this way because of the shroud. The most common conclusion is: 2xLF is better detectable on OB bearing. Wrong! 2xLF is better detectable on the motor body.

To have this effect more pronounced take readings in the middle of motor length and on the end bell. Please post data.

It's normal for demodulated signals on cases and end bells that include ac ( especially 3 phase) motor slot pass frequencies typically show a great amount of normal variability. The default frequency ranges that PeakView and some other manufacturers use are really the worst way to make motor bearing measurements if you use the cases or the end bells.

GIGO

Duncan