dear all
Hi!
I encounter a lot of cases that the side bands are strange. below are some examples:
1- unbalalnced side bands,e.g. there are peaks at the right side of center, but there are not noticeable peaks at the left.
2- 4Hz and 5Hz side bands around 1xRPM and other harmonics in motors that directly connect to pumps (there is no gears or belts ...)
3-side bands that are greater than center frequency.

Have anyone experience with such side bands and may give me some insights about them?

Regards.
Farahani

First and foremost, I would not draw too much importance from unequal sidebands. But since you ask the question I'll give some thoughts (others are welcome to add)

Sidebands as you know can result from amplitude modulation (of a sinusoid or of ampcts).

Sidebands can also result from speed oscillations of at least part of the machine if there are torque oscillations present and low rotating inertia or torsional flexibility between parts of a machine.

One possibility that jumps to mind is simultaneous AM and FM. If the upper sidebands add and the lower sidebands subtract (out of phase), then you get unequal sidebands.

Another possibility is multiple sources of AM vibration within the same machine, again with the same type of phase difference.

Another less interesting explanation is that the signal starts with equal sidebands and as it is transmitted through the system, different components are attenuated/amplified differently.

Or maybe there is no good explanation other than that is the fourier transform of a general the waveform which is too complicated to be expressed as a simple AM combination of sinusoids (there are plenty of other non-linear transformations of signals that can occur besides simple AM).

If I had a specific problem in front of me, I would be inclined to check the time waveform to see if there were any clues on what the spectrum represents. Also strobe can confirm or disprove F.M. if you can get to the relevant rotating parts.

Final thought - don't read too much into it.

dear electricpete
Hi!
very thanks for your explanation. but, may you please explain the freq. modulation and its difference from AM? and my second question is reminded. side bands of 5Hz for a simple motor-pump. see below spectrum.

regards.

Frequency modulation means that the frequency of the vibration is oscillating due to speed of machine oscillation. Amplitude modulation means the amplitude of the vibration is oscillating.

For sidebands at 5hz, I would think there might be a valve fluttering at 5hz or other fluid system oscillation at 5hz. In that case, you might also see flow or pressure oscillating at the exact same frequency.

If VFD-driven machine, the vfd control might create torque oscillations or speed oscillations

In the spectrum you show where 2x, 3x, 4x each larger when normally we often expect them to be getting smaller, it might be the case that vibrations are transmitted (from the point or origin to point of measurement) better at higher frequencies and attenuated more at lower frequencies which might explain why the upper sidebands are higher (or not).

An interesting side trip on this last aspect - if your sidebands were equal in displacement, the upper one would be higher in velocity. Another frequency-dependent transformation of the signal which can increase the USB

I have one last item to mention. I have seen spring-mounted fan which had sidebands of 200cpm around running speed and harmonics. 200cpm was also found to be the bump test frequency of the base. I'm not sure of the explanation for that one.

This message has been edited. Last edited by: electricpete,

I may be looking at your example plot in a different aspect than what you are, so:

In the example you have:
where in the spectrum are you referring to?

quote:

3-side bands that are greater than center frequency.

But anyway, This is what I see in your plot, even though higher resolution might make the plot a little clearer and also some "marked" peaks with cursor sidebands marked might help.

Notice in the attached plot than I am only asking, not stating as a fact, yet.

Of course I could be completely wrong.

I hope I'm not out of place in commenting on a question that was not directed to me. It looks to me like the original question was general (mentioned "a lot of cases"). The spectrum was a specific example, which happens to demonstrate 1 and 2, but not 3. I interpret the spectrum the same way you did as far as location of the running speed harmonics.

This message has been edited. Last edited by: electricpete,

An example of #3 here - the upper pole pass sidebands at 5.04, 6.04, 7.04, 9.04, 10.04 orders are higher than the running speed harmonics at 5,6,7,9,10:
http://maintenanceforums.com/eve/forums/a/tpc/f/3751089...031079702#3031079702

quote:

I hope I'm not out of place in responding to Ralph here. It looks like the original question was general (mentioned "a lot of cases"). The spectrum was a specific example, which happens to demonstrate 1 and 2, but not 3. I interpret the spectrum the same way you did Ralph

You are not in any way responding out of place here Pete.

I realized from the start that he was referring to "many cases" in the original post, but it is much easier for me as an analyst to analyze each on its own merits rather than to gereralize all of his or anyone's "cases" as equal or not equal to a generalized basic rule. There are too many variables that can enter into the type question he asked using the term "many cases" and trying to compare all to one. The example shown here may in no way be related to any of the other "cases" he or anyone else might be seeing and to generalize and have a fixed or mind-set attitude that "this is what I see in all the plots", might be misleading.

BTW, I did refer, in my reply to his post, an "example" plot.

Sorry if I am not one to generalize data into categories or classes without first doing a complete analysis of the data available and maybe not even after that.

My apoligies to Farahani and you, Pete, if I have replied in the wrong frame of mind.
But I am what I am and that's all I am. I think we each have a right to respond as we see fit so long as it does not in any way low-rate or degrade another's reply or character.

Again, sorry if my reply offended you or Farahani . Let me know and I will more than gladly delete it.

Once again I'm sorry Ralph for jumping in. Glad to see you have a lot of smiley face which means you're not ticked off. Next time I'll let the OP respond.

quote:

Next time I'll let the OP respond.

What is "OP"? Old People? Original Poster.

Aw! MAN! respond when you want! I enjoy your posts. Wish more of the members would respond. Have you noticed how many members have never posted anything? I forget right off how many "pages" of members have "0" "karma" points (if this relates to the number of posts). Seems like it is about 38+ pages (3800+ members). What a world of knowledge there must be, just waiting to be shared!!!!!

So you're saying I don't post enough? I'll have to work on that ;-)

I enjoy your posts Ralph and I agree that more comments are always better and those lurkers should feel free to chime in with thoughts or questions of any kind.

quote:

So you're saying I don't post enough? I'll have to work on that ;-)

YEAH!!!! That's what I am saying!!!!!

Seriously, You do a good job. You put a lot of thought in what you say and post very interesting replies.

Dear Ralph and Pete
I appreciate your kindly responses, the questions I introduce are in my mind from past and I don't face them newly. So they are general questions. But the given spectrum relates to a new problem that I face it some days before. It have unbalanced side bands of 5.2Hz at several harmonics of RPM (2xRPM into 7xRPM). The pump was checked and it has severe misalignment.Below are a close and a general look before solving alignment problem. after that the vibration decreased and side bands disappeared. See the third spectrum.
Regards.

the spectrum in general look.

The spectrum after correcting misalignment. There are some peaks (1xFline, harmonics of RPM,...) but they are little.

Dear Farahani,

Post mortem: Your machine is a nominal 50 Hz machine but 1X seems to be at something like 2990RPM. The peak close to 50 Hz must be your 1X and the peak at 100Hz is a true magnetic 100hz. Then we see 3X, 4X with the sidebands (but oddly no 2X as other authors have indicated). Sidebands must have been pole pass frequencies or multiples thereof caused by asymmetry of motor airgap due to excessive misalignment. Probably misalignment was so high that it was causing the body to warp. Very simple to diagnose now that you have shown us what the problem was.

Would I have called it a misalignment? Without the 2X and 1X, with lot of reservation. Anyway thanks for this excellent example that teaches us that there are no absolute rules in the world of machine vibration diagnostics. (This is why my blood freezes when I see those single page diagnostic charts handed out to newcomers.)

Thanks again and hoda hafez!
Ibrahim

Dear Farahani,

The symptoms you are seeing in this motor is due to rotor defect. Kindly check up for a coast down run of this motor at decoupled condition to confirm this.

Regards

Dear Anbu
What do mean from rotor defect? (e.g. bars,laminations,shafts,...).Anyhow I will do MCSA and also run down test as soon as possible.

Dear guys!
See below spectrums that relate to a 400V AC motor at no load test (uncoupled). What do you think?
Regards.

continue, with harminic cursers. It is a 2 pole motor.

continue, with side band cursers about 3xRPM