I just balanced an overhung fan and achieved good results on the fan bearings but the vibration on the motor points went up about 100%. How can this be? The waveforms look very sinusoidal and the 2x trend did not rise. The coupling is a Falk grid type. I am going to check phase readings and compare them to last time. I suspected misalignment but the data I have taken so far doesn't seem to indicate it.

Basically I've made it a rule to check alignment prior to balancing such a unit. The absence of 2X doesn't mean alignment is OK. But you mentioned the vibration doubled but not at 2X - does this mean OA levels or 1X & 4X or?

Suspecting misalignment of some form would be a likely start.

Thanks Sam,
I am attaching a spectra for the worst motor point showing the difference from the last data I took. I took some cross channel phase data and saw no more than 45 degrees difference across the coupling horizontal or vertical saw only very small differences from end to end on each component. I did find a slight rub on the fan where the cone comes in, but can't seem to get it out of it without moving the bearings. We are on a short shutdown and I don't have time to do an alignment.

P-16_FAN_DATA.doc (24 Kb, 118 downloads)

Randal,
Have you looked at that motor point using a log scale for amplitude? Also, did the horizontal readings do the same thing? You may have pedestal resonance issue, although if the fan bearings were the main issue and you balanced them down successfully, it doesn't make sense that the MOV would go up.
I have to agree with Sam misalignment has several flavors. If you have a bad angular alignment, that would appear as a bent shaft would, give high 1x, high axial and 180 degree phase shift axially across the coupling. I take it you didn't see any of this.
What was the change in the three directions at the four locations?
You did mention a rub. What if the rub is exciting the MOV resonance?
Ron Brook

Ten minutes for an alignment check should be performed. Less than an hour for alignment unless problems out of the norm are encountered. Typical with dial indicators would be 30 min on a clean, easy-to-get-to base.

Thanks Ron,
Log scale spectrum is attached limited to 500Hz. I have not used log scale, what is it you are expecting to see there?
The amplitude changes are as follows for 1X:
Motor inboard hor.: decreased .213-.139
Motor inboard vert.: decreased .384-.304
Motor inboard axial: same .406-.400
Motor outboard vert.: increased .269-.643
Motor outboard hor.: increased .162-.295
I can't get an axial fan bearing reading because of the guard so I did not do cross channel phase readings in the axial plane. I will check the coupling alignment on the next shutdown if possible because misalignment still seems like the most likely culprit. Base resonance is also possible as this base is old and poorly mounted.
Sam, I wish I had your expertise in alignment, it always seems to take me longer. I am sort of obsessive when it comes to making sure everything is right before I start (soft foot checks, base and foot cleaning, etc.) and it slows me down. The environment in our plant usually means extra time spent cleaning as well.

p-16-2.doc (24 Kb, 63 downloads)

Randle; where are you in TN or rather who are you with - I like your area (I'm from TN). Check two adjacent feet, do run-outs on shaft and coupling. Small anit-friction bearing units may use Rim & Face method when Indicator Reverse isn't as practicle or size constraints rule. Dirt is another issue depending on what it is.

Randle,

When you balanced the fan, considering that both the motor and fan are sitting on the same pedestal, you have effectively added/subtracted the "unbalance" vector to the motor vibration. If you had phase data before balancing you would've seen a change.

David

Randle,
If you look at the base of the 1x peak, there is a 'bulge' to the left side. There is a 'breadth' to most of your peaks so there may have been some speed fluctuation. The data might show it a little better with more averaging (16 averages), but it does appear to indicate the 1x speed is just above a support resonance. Also, your vertical and axial readings, being higher than your horizontals may also point in this direction. Let us know what your alignment readings find.

Thanks for all responses.
Sam,Our plant is located about halfway between Knoxville and Chattanooga in the foothills of the mountains.
David,it would be great to have phase data of each machine but I haven't been able to find the time. I am a one man operation in a big plant, but I am trying my best to streamline my data collection and analysis so I will have more time for diagnostic work.
Ron, I am intrigued by what you were able to read from the log scale spectrum. If you could explain further how you came to these conclusions or point me to some reference material I would be grateful. Resonance issues are probably a factor that I will be investigating further.

Randle,
Here is a plot from a 1000 hp four pole motor. The first plot has a linear y axis. Looks like just unbalance. The second plot is with a log y axis. Now the resonance response is clearly identified. It does show a high degree of damping (width of the frequency spread), but I now know that part of the response I get at 1x is due to some amplification from the resonance.
Ron

lin_vs_log.doc (86 Kb, 73 downloads)

Randle,
Here is a sample of a natural frequency resonse test performed with an instrumented modal hammer and three accelerometers. I only include it so you can see the similarities in the response due to machine vibration vs. the instrumented modal hammer.

FRF.doc (96 Kb, 65 downloads)

Sorry,
Here is the impact from the same crane structrure as the FRF data.

Impact_for_comparison.doc (41 Kb, 49 downloads)

Thank you Ron for your information, I have been very busy and not had much time to study it but I think I am starting to understand. I have been involved with vibration for a little over 5 years, collecting data, analyzing, training, and am just getting to the point where I understand enough to do some effective diagnostic work. I am grateful for this forum as a resource and people like you who take the time to help. If I understand correctly the broad "noisy" areas in the log scale spectra indicate amplifiction due to resonance, and the width of the base shows the influence of damping on the system. Is this how you figure out if a structure needs to be stiffened or damped to fix a resonance?
I have done some more investigation on my fan and made some minor changes with mixed results. I took more cross channel phase readings before and after I tightened the base bolts and I did a running soft foot check. The running soft foot check showed no changes as I loosened each foot in succession. The phase readings are attached. The tightening of the base bolts on one side caused vibration to go up so I stopped and went to the other side, tightening on the other side brought vertical vibration down (I was monitoring vertical on the motor). Afterwards I did my regular route check and noticed that vibration had gone up on some points and down on others. I am now seeing indications of misalignment and will be doing an alignment check on the next shutdown. I don't know if the changes I am seeing are due to changes in the resonance of the structure from tightening the bolts or if I actually changed the alignment while doing so. By the way, the base of this machine is a mess, it was replaced years ago and it looks like somethng that was meant to be temporary and never done right. It is mounted on hard rubber pads.

P-_16_FAN_PHASE_DATA.pdf (13 Kb, 38 downloads)

Just curious but did anything change when you did the balance job on the fan, did you pull apart the coupling. Then reason I ask is I have had a few issues with mechanics putting together falk grid coupling with the keyways lined up, In addition to the fact that they don't cut the keys down to size gives me an unbalance issue usually on the motor.

Perhaps you had a dynamic unbalance across the machine and now that the fan is balanced properly the motor unbalance is free to run where it wants.

I agree that resonance could be an issue. But wouldn't resonance have made it very tough to do a balance job? Was your phase steady while taking readings?

Thanks for your response Torptorp,
I didn't break the coupling on this job but I did do the alignment last time. I usually make sure to put the keys at 180 apart but may have overlooked it. The motor vibration has always been higher than the fan bearings. I used the 4 run method without phase to balance the fan. I put my accel on the inboard fan bearing as that was where the highest vibration was. I have used this method with great success on several tough cases where phase readings are erratic. As a matter of fact I have begun using this method as my primary balancing method baecause it is usually faster and easier and works great even when there is resonance present. No balancing method will fix resonance but it will take most of the forcing function away This fan may be one of the ones that the four run method is not so well suited to, but it seemed like a good candidate because it has a large thin impeller.

Randle,
Your last email turned on some lights! Your base is 'floating' on hard rubber feet. There is a problem with performing alignment when the motor housing is stiffer than the base it's bolted to. What you end up doing is bringing the rails up to meet the motor feet instead of the other way around. This is the same problem as aligning diesel generator sets in container cranes. I always warn my field service personnel to NEVER fully tighten the foot bolts and check for soft foot. You will not have a clue where the rails and the motor feet really ended up. Leave all of the motor feet loose, shim to get close (straight edge) and then see if you can move or pivot the shims under any one foot. The motor weight should be enough to hold the shims, unless you do have a soft foot. Correct it by shimming appropriately, but again, with all the foot bolts loose. Once you have any soft foot corrected, shim the motor to be approximately 0.005" high (assuming you haven't used more than 3 or 4 shims under any one foot.
Now, correct side to side and then watch your side to side as you slowly tighten the motor foot bolts and move immediately to another foot if the reading moves more than 0.004". This is just like torquing down an engine head. Once you have successfully tighted all of the foot bolts, take a new set of readings. You should be in the ball park!

Randle,
Just looked at your phase data. Along with the information concerning the hard rubber feet, I believe you are dealing with a nightmare of a base.
I noticed that after you tightened the motor bolts, a high reading on the fan 'squirted' out on the motor. Begs that the base is not a stable working platform.

Ron, thanks for your reply. I am attaching some pictures of my fan so you can see what the base looks like. When this fan was replaced years ago there was a plate bolted to the old foundation and then the fan was bolted down to the plate with the rubber pads in between. one problem is, this fan is bigger than the old one and hangs off the foundation. The grout under the plate (if there ever was any) seems to have disappeared. The bolt heads that hold the plate down are under the fan rail making them very hard to check for tightness. If the pads are worn out and allowing the fan to rest on the plate bolts this may also cause a problem.
Having seen the structure I am dealing do you have any suggestions how should I proceed to make this machine as good as possible?. I don't there is any chance of getting enough downtime to rebuild the base and mount it correctly. Our shutdowns are limited to 48 hrs max, much less lately. Do you think I should replace all the rubber pads and realign?

p-16_fan_pics.doc (500 Kb, 42 downloads)

Randle,

After looking at your pictures I have a question. Since your motor's amplitude of .6 in/sec is in the vertical direction, what is the amplitude on the base where it is bolted to the bottom plate (on both sides)? What I am wondering is the motor moving independent from the base at the floor or is the whole thing moving. What about the amplitude and phase difference between the vertical on the motor and the plate it is sitting on.

Have you tried wedging something between the open space where the "rubber" bushing are to see what happens?

The overhanging frame under the fan housing, where you said the new fan is bigger than the old one, have you checked to see if wedging there changed anything?

Why is it mounted on rubber wedges? Is it still necessary since the pedestal has been changed from the original installation?