Hi everyone,

We manufacture large gearboxes. I am a vibration analyst but strictly theoretical - I model systems in simulation software at the design stage. I have not received formal training in analysis of measured vibration data, however I have been charged with the task of trying to troubleshoot a noise problem with one of our gearboxes. We had a consultant take some vibration data and we have done a quick preview of them to see if any problems popped out at us.

The input shaft is running at 1246 rpm. The output shaft is running at 3473 rpm. In terms of measured data, we notice that the axial velocity spectrum on the INPUT SHAFT shows a spike at 413 Hz. This frequency, however, corresponds to a bearing fault frequency (fixed ring) for a bearing locaed on the OUTPUT SHAFT. As well, there are two sidebands to the left of the fault frequency that are difference frequencies between the Bearing fault frequency and the output shaft speed.

The measurements taken on the output shaft show very little vibration at these frequencies.

Does this make any sense? Why would a bearing on the high speed shaft transmit a clearer measurement of the fault frequency on the input shaft? As well, do the sidebands with difference frequencies indicate a looseness problem? Anything else anybody can think of?

Note: I DO see some spikes at 1 x input speed and 1 x output speed at all measurement locations and directions, however the levels are consistent with those from our good "not noisy" boxes.

Regards,

D. Denny

D. Denny--

I am not an engineer, but I have seen similar phenomena through the years (i.e. problems w/one shaft showing up more clearly on another). For lack of any better understanding or explanation, I've just assumed it's related to transmission path / structural components in the gearbox itself and let it go at that. Be sure to eliminate any other possible source for what you're seeing; if still the only thing left it definitely deserves attention.

Not sure about the sidebands--the fault frequencies would be enough alarm for me! I confess I'm guilty of trying to avoid "over-analyzing" the situation--sometimes at the expense of "clues" that might've been helpful in identifying either other or root cause problems! Tony

In my world there is a strong correlation btw. the sideband frequency and the speed of the shaft where the problem exist. Even once got half the running speed of one of the shafts in a vertical aluminium alloy gbx. indicating possibly a not round gear or warped gbx. frame, it was swapped and scrapped so I can´t say either. So I vote for that you are having a problem at the output shaft if that matches the side band freq. Why it seems to be popping up at the other end, there may be many reasons but data is data. I have been trying to track a bearing fault in a 14 spindle drilling assembly, all the same speed and a noise that could awaken dead humans all ower the place. I found a marginally higher signature close to one spindle and the mechanic keeping me out of trouble said "Yes I think so too it´s a darn much warmer than the others" so use common sense and all available information. Good luck. Olov

Oli, once again your no nonsense down to earth response comes through. Thanks for the reminder and keep it up!!

Thanks for all your comments. This is a NEW gearbox, and the techs have disassembled the box quite a few times and really see nothing wrong with the bearings. The sidebands make me wonder more about a balance problem or maybe the bearing is loose on the shaft? As I said, I have NO experience in this. I have simply read material on this in books!

We are investigating into buying a data collector/analyzer. If I could give some guidance to the techs to help them figure this out, based on the limited amount of data and my limited analysis capabilities...then I KNOW I could sell the managers/owners on the cost benefit of obtaining such equipment.

D.Denny

De nada, you are welcome. I learn a lot too, every day I am happy to say. If I can share the things I found IRL, I hope to get support on the funny things I find, that make no sense to me, but possibly to some one else. Thank´s Ron. Olov

D. Denny--

Note that you can get/see bearing "fault frequencies" when no "real" bearing defects exist due to inordinate loading / preloads. In a motor shop where I used to work, such frequencies especially in a rebuilt vertical hollow shaft motor signalled us to back off the locking nut on the top of the hollow shaft to reduce thrust load on the inner race of the lower (if I remember right) bearing.

Another thought, a bearing can/will produce higher frequency fault frequencies far before there's any visible damage to the bearings!

Tony

In my experience this side banding only can be seen when you have to bad balancing conditions. Are the couplings balanced w. half wedge and is the shafts also? A common problem despite standards also on parts from the same company. Next is that the sidebands exists, at least in dryer cylinders in papermachines where I seen this most is a more significant indicator for bearing problems than if you find all fault freq. you can have in the bearing. Can you see multiples of the bearing freq? I always confirm that I find at least 3 matches out of 10 multiples of the bearing fault, then I am happy.
Can you see any other faults like BSF, BPFI?
There are always exceptions, latest issue a testrig where the number of balls in the drive shaft bearing that was bad was the same as the number of teeth in the gear, generating noise that rejected 100% production. So use common sense and all tools you have, did you listen to the various brg. positions w. the yellow handle screwdriver analyzer? Did you compare brg. temp. readings? Have you checked the outgoing coupling? Does it behave the same uncoupled? Have currently a feed pump where a worn coupling excite resonances in the gbx. not so hi readings except for general bearing methods, but it got real hi noise. Happy Hunting for the cause. Olov

D. Denny, you mentioned the 413 hz (24,780 cpm). What are your tooth counts? Have you ruled out GMF? You also mentioned something about a bearing might be loose on the shaft. Just take a look at the shaft OD and bearing ID measurements to double check you have the right interference fits. What do the tape lifts of tooth contact look like? What type of gears are you running.........and bearings? These vibration readings are on the shop test stand correct? Sorry for all the questions, just need more info. Don

Lots of good comments so far. I would second what OLI brought up about resonance. This is probably the major cause of excessive noise on gearboxes. Do large flat surfaces exist? A cover plate that does not have well matched surfaces and is preloaded when bolted down may tune its natural frequency to a source in the box. The poor fit may account for its natural frequency response being different than other similar boxes.

I would also agree with Tony in reminding that bearing fault frequencies appear with improper fit in bearings that create excess loads, even though no bearing defect exists.

Skip Hartman

Just to throw one more wrench in the works...if you did not take the data how do you know where it came from? For all you know the vibration was taken on the consultant's Ford pickup...

Make really sure you know how the data was acquired and from where.