Gearmesh Frequency/Amplitude, why in Gs?

[rfong]

Ok maybe it has been a while since I've opened my vibrations books. I understand the basic concepts and the relationships (integrate/derivative) between displacement, velocity, and acceleration. I have taken some vibration routes but fail to understand why the amplitude (y axis) of the gearmesh frequencies come out in G-s. I know this is an acceleration (ft/s/s) and I understand what acceleration is (rate at which velocity changes). I can even say I can visualize what displacement and velocity looks like in a physical sense. But I fail to see what acceleration looks like in a gear mesh? What could be changing (rate)? I'm very confused. Any help would be greatly appreciated.

Thanks,
rfong

 

[EnglishMuffin]

Well, if you have torsional oscillation in a gear train, for example, then you will obviously have acceleration. This ultimately produces transverse vibrations, both in the form of airborne noise and lateral movement at bearings etc, which can be picked up with microphones and accelerometers. One source of such torsional oscillations is inherent gear tooth inaccuracy, and another is gear tooth deflection - both of which cause velocity variations because of the deviation from a true involute profile. It is very hard to detect torsional oscillations of gears visually, even with a stroboscope (unmagnified), because they are so small.

 

[MachineryWatch]

Rfong,

You do not mention what data acquisition system you are working with, or who set up the database that defines the frequency ranges. Were the data acquisition parameters used when you are collecting data defined by someone other than yourself?

It is often the case that data acquisition parameters for gearboxes include spectrum maximum frequencies that are approximately 3.5X Gear Mesh Frequency (GMF). GMF is #teeth times gear RPM, so this can result in a very high frequency range being defined for data acquisition. At these high frequencies velocity begins to lose its sensitivity. This means that acceleration becomes a preferable vibration parameter to trend for detection of problems that generate high frequency vibration.

The vibration data parameters that you collect with your data collector was either defined by someone who understood this relationship, or you have a data collector that defaults to acceleration when high frequency ranges are defined.

http://www.machinerywatch.com

 

[electricpete]

I agree with Skip (MachineryWatch) on the subject that we typically always use displacement for low frequency vib (perhaps below 1000cpm) velocity from 1,000-24,000cpm and acceleration above 24,000 cpm.

The reason is that the conversion between displacement, velocity, acceleration is frequency dependent.

Typical machinery vib spec would have a limit on all three. For example NEMA MG-1 (going from memory) displacment 0.002" (pk-pk), velocity 0.12 ips (pk-0), acceleration 1 g's (pk-0).

If your vibration is in the range above 24,000cpm, the acceleration limit will be hit first. If vibration below 1,000cpm the displacement limit would be hit first. If vibration in the middle the velocity limit will be hit first. So you look at the parameter which will hit it's limit first, based on the known frequency range.

Another aspect is the time waveform. I agree that displacement would appear to have the most physical meaning in understanding how the part is moving. But if you compare displacement and acceleration time waveforms, you will see that the acceleration waveform accentuates sharp impacts, while the displacement emphasizes slowly changing components (like 1x unbalance)