Greetings,

Last week I had the opportunity to admire a medium size centrifugal belt driven fan (200 HP) with particular attention given to the vibration isolators underneath. Both the motor and fan operate very close to 1785 CPM although there is about 5 CPM difference between the two. I'm not going go into great detail at this time but based on some early calculations, find the frequency ratio between the forcing frequency of turning speed and the natural frequency (impacted and found fn to be around 3.75 Hz) to be quite high. Static deflection was calculated to be just under 0.700"! These isolators have been in place since 1987 and there is apparent visual signs that the isolators (8 total) are sagging a bit. The complaint is actually turning speed vibration (around 3.5 mils) that is very sensitive to balance. There was no apparent natural frequency around turning speed but a very strong peak component at 3.75 Hz. Any ideas? Thanks in advance for any input!

Was the static deflection calculated from the resonant frequency? There is good agreement between the measured natural frequency (3.74 Hz with a measured of 3.75 Hz) and the static deflection. Are there other natural freuencies? Have you checked all directions, including twisting? Dirt build up?

Sensitivity to balance goes along with resonant behavior. A natural frequency at 3.75 Hz should give isolation, unless there is another natural frequecy closer to 1X.

Mr. Foiles,

Thank you for your reply. Yes, the static deflection utilized the natural frequency. The unit was impacted horizontal, vertical and axial directions; although not in the "twisting" direction. This 3.75 Hz or there abouts is present in all directions and upon impacting, will oscillate for at least 20 seconds. I also shared your thoughts regarding balance sensitivity and resonance. That was actually my focus until data was collected. I can pass along some data if interested... Thanks again for your input!

It sounds like you were hammering on the frame and/or structure. Maybe you are running close to a rotor critical? Were you able to impact the shaft/wheel? If the fan is on rolling element bearings, I would think this would give you a decent estimate for the rotor criticals. If you knew mass and dimensions you could also do a back of the envelope calculation to see if this is even a remote possibility. Does the lag angle indicate you are running near a balance resonance?

Michael Titone

Hey Mike,

Yep, I was hammering away on the sheave side bearing of the fan but did impact the shaft where it entered the housing (fan housing); nothing earth-shattering was noted. I'll see if I can obtain the balance data and I'll pass along...

Just for further information, I've added a portion of the impact test data as well have retrieved the balance data. Hmmmmm. Balance speed 1787 RPM with 1.51 Mils @ 81 degrees. Placed a one ounce trial weight at 270 degrees with the result of 11.2 Mils @ 64 degrees. Any further input?

Thought I'd add the Spectral data as well...

It looks like a resonance may be providing some amplification here.

What is the angular locations of the transducer compared to the trigger device?

Along with the one ounce, one needs the radius. This sounds like alot of amplification depending upon the radius. Look at the balance weight as a change in mass eccentricity for the rotor (balance weight)*radius/Mass of rotor, and compare this to the 11 mils to see what type of amplification you have.

Really, do the vector math of with and without weight with the mass eccentricity to get amplification.

Mr. Foiles,

I placed the transducer on the Sheave side bearing in the horizontal direction and impacted the lower portion of the structure roughly three feet below in the same direction... I'll attempt to locate additional information for the mass eccentricity comparison... Thanks again!

Amplification looks something like 1/(f^2-1) where f is the ratio of the speed divided by the natural frequency as you probably know. To get isolation one needs f^2-1>1 or the natural frequency less than speed/sqrt(2).

The impact looks like you may have a resonance much closer to speed than the 41% separation needed for isolation. The size of the peaks in an impact spectrum depends upon the mode shape.

Hey Allen--it's been a long time since the VI MVAIII course in Syria, VA!

Have you identified/calculated your belt frequencies to see if it might be a contributor here? Your 3.75Hz sounds like it could be close to a 1X Belt Freq--not what I'd normally expect (compared to 2X or higher), but I have seen it.

Also (and you've probably thought about this already), is the beat that I'd expect (between motor and fan speeds) possibly causing some amplitude and phase fluctuation? Just some thoughts--for what they're worth!

Tony