One more question:
Have any of you ever noticed looseness showing itself first in the 6-7-8th harmonics first.. and then degrading from there
i.e. "connecting up" with the lower harmonics?
Particularly Axially
See Attached
The reason I ask is because I had a case last year like this, and the final descent into failure was very short. (They left it running when I told them to switch to back up)

The sidebands are 96 cpm Pete. :-D I knew You'd ask.....

Forgot the attachment

Martin,

What type of machine is this and what type of bearings and seals does it have? Given the limited information and the spectral data provided, have you considered rubbing as a possible cause of the vibration? I say this because of the subsynchronous vibration and the modulation.

Zack

This is an Air Handler. I don't have the bearing numbers. They are likely Sealmaster or Linkbelt rolling element, roughly 2 to 2 11/16th bore size, but that is the best info I can offer.
The Subsync is 2x belt freq. 739cpm

the 256 peak (.211 orders), I don't know, but I haven't concerned myself with it.
Doesn't a rub include .5x????
I don't have a lot of journal bearings to deal with....(never seen a rub)

This looks simliar to a fan I had go bad a few years back. The locking nut backed off on the drive side bearing. Another possibility is that the shaft is starting to wear down.

Isn't a rub normally associated with sleeved bearings?

What exactly failed? Was their any autopsy performed?

btw, thanks for those sideband numbers. That wasn't pole pass was it?

Did you ever get take Infrared level 2? I took it here in Houston and guess who was my instructor?... Nobo. We had a good time remembering the first class in San Francisco and all the people we met there. Of course Houston was a much better place for a course than SF!

Martin,

I think the pattern presented is indicative of a different type of looseness, likely more severe.

Traditional looseness spectral pattern features high 1x and 6-10 much smaller harmonics. In your case 1x is small and the rest of harmonics (except 2x) is comparable with it. This will be typical for a short duration impacting pattern in the TWF. Short duration impacting vs non-impacting by itself describes looseness type.

It will be interesting to see the TWF in both G's and ips units.

Also, was any evidence obtained after failure analysis?

David

In the previous case, we did not do a post failure Anaysis. There wasn't much left to analyze.. a couple of balls on the floor and an outer race that was flat as a pancake. The only thing intact was the bearing housing. .. But get this.. it was still running!

Anyway, I am trying to decide whether to move this to weekly monitoring, or call it for service. The signature is the same.. but the previous case had a .71 Overall amplitude, this case is only .113 overall amplitude. So although the pattern is troubling, I think I will monitor weekly for a few weeks.

Agree/Disagree?

Pete: Although I have been doing some limited Thermography, I have not yet went for the level II cert, Last year i went to get Level II vib cert instead, because vibration here is expanding rapidly, I can only squeeze in Therm when it is really opportune....

Hey I have my own grievances living in the city of SF, but Houston better??? I've lived in both cities, and you see where I am....

Pete: no on the pole pass

quote:

I think the pattern presented is indicative of a different type of looseness, likely more severe.

quote:

Short duration impacting vs non-impacting by itself describes looseness type.

Why more severe?
Explain the term "Loosness Type"

The wf is truncated Axially, but not Vertically.
The IPS waveform is similar to the acceleration waveform... truncated turning speed low amplitude impacting.

I treat looseness pattern which is causing impacting and low 1x as a special case in comparison to that where 1x is dominant and no impacting..

There may be be different opinions on that...

What I am asking David, is how do you account for this difference mechanically, and why.

I understand an impacting looseness is more significant than a non-impacting one, but why do you assign a low 1x a higher severity than a high 1x with descending harmonics?

I am certainly open to it, since at the beginning of this thread I suggested that I had seen this pattern before and it was not good.

Is this a belt driven fan? What with the speed being 1212 rpms (12 rpms plus the slip) more than a 1200 rpm motor?

How many blades does it have? 7 ?

Since this data is from the axial direction, and assuming (which one should never do ), that this is the direction of the highest amplitude, both spectral and waveformish, and the sudden stoppage of the amplitude in the waveform in one direction (negative), there may be a "giving" in where the bearing sits. But not actually knowing how the machine is made, I am just guessing.

Only my opinion and I could be totally wrong again.

Belt Driven: Yes
Motor 1x=1765
Blades = 9
Axial is highest amplitude, the amplitude was not the concern, more the charcter of the spectra.

quote:

Originally posted by Martin Moore:


I understand an impacting looseness is more significant than a non-impacting one, but why do you assign a low 1x a higher severity than a high 1x with descending harmonics?

Martin,

In my opinion if you look at a spectrum having high 1x with descending harmonics the TWF won't show impacting. If the above is true and impacting is an indication of looseness severity the equipment in the example you have demonstrated is in trouble.

TIOMOAICBTW.

David

Just when you think you are starting to get a bit of a handle on this sort of thing....

The attached is from a felt roll where the withdrawal sleeve has come loose.

Over that last 6 years or so, we've had 20 or more instances of this, and in all cases, we've seen the "classic" looseness indicators; a bunch of running speed harmonics, and impacting in the waveform at once per rev (makes sense, really...)

Except this one.

I guess this is one of the reasons I like this job - keeps you on your toes.

Sleeve_looseness.doc (102 Kb, 29 downloads) Sleeve looseness?

quote:

In my opinion if you look at a spectrum having high 1x with descending harmonics the TWF won't show impacting

I would think that the impacting in the wf would depend on the particular relative amplitudes of the harmonics.

ie. the more harmonics, the more the energy is dissapated in other-than-turning-speed events.
Does this make sense to you guys....??

Often times impacting gives a series of harmonics which peaks near the resonant frequency of the thing being impacted. Typically the resonant frequency is a lot higher than the impact rate and the higher frequency harmonics are higher magnitude than the fundamental. So I agree the pattern that David describes is what I usually associate with impacting. I also agree with Martin that the more harmonics we see, the more likely it's impacting, even if the fundamental happens to be the highest. But those are just generalizations and there are always exceptions.

The units displacement, velocity, or acceleration may be worth mentioning as well. Displaying a vibration signal in displacment makes it "look" much less like impacting in both spectrum and twf than displaying the same signal in acceleration. Let's say I have an acceleration signal of impacting at 1,200 cpm with acceleration harmonics peaking at 60,000 cpm (50th harmonic). Let's say the 60,000 cpm peak were 10 times as big as the fundamental 1,200cpm in acceleration. If I switch to velocity, the fundamental jumps up by factor of 50 compared to the 60,000 cpm component. Now the 1,200 cpm fundamental is 5 times bigger than the 60,000 cpm component.

As a purely academic excercize, attached is a waveform where each harmonic is 3/4 of the magnitude of the one before it. It looks something like impacting.

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

DecreasingHarmonics.pdf (25 Kb, 29 downloads)