Shaft unbalance caused by oversize key 1

[nisqually]

Hello,

I am in need of assistance. I am very new at this. I am told a larger than normal key on a shaft must be balanced with some mass added on the other side of the shaft. That's fine. But we are supposed to use a vibration test to verify the shaft is balanced properly. I assume the location of the extra mass would be opposite the key (180 degrees). Someone suggested the 4 run method can be used to verify balance. Can we attach some mass and just look at vibration in/sec in the frequency domain and assume the equipment is balanced if the overall vibration is below, say .11 in/sec- 0 to peak? I don't understand the units of permissible residual unbalance as in Appendix C of ANSI/AMCA 204. I do have a Commtest vb1000-T analyzer available to use.

This test is Monday. What do you suggest? I can't become an expert in a day so I'm looking for practical suggestions if possible.

Nisqually

 

[diamondjim]

Nisqually,
If I am following the logic, you are saying
that the key pulls whatever is being keyed off
center causing a slight shift or eccentricity.
if it possible to just file a flat on the high
side untill it balances? No extra stresses on
the shaft this way. I once read that if you
are using two keys, the second should be about
120 degrees from the other rather than at 90
degrees. I assume you are using a setscrew to
hold the key in place or are you using a tapered
key? If you are using a setscrew, it would tend
to pull the mass off center, but the setscrew hole
above the end of the setscrew to the od would reduce
the mass also, so maybe you could c'drill the set
screw hole. Sounds like an interesting problem.
Keep us posted. There is nothing better than
experience to help us on our way.

 

[jmack]

nisqually,
A couple of things here. The first is an easy one...profile the key to match the shaft and the coupling symmetry. This is much easier than trying to compensate for a key that is not profiled. The second thing is that these type of balances are largely dependent upon the individual circumstance, ie. it would be much more important to balance a 3600 rpm centrifugal pump assembly say 100 hp because of a large key as opposed to a 450 rpm vertical pump because of the obvious speed differences and the overall mass ratios of the keys compared to the rotating assembly. One other thing to remember here is that misalignment can also be a major contributor to your overall vibration and can influence your 1X reading, so always deal with that first before balancing to justify the balance operation in the first place. Just a few thoughts.

J MacK

 

[nisqually]

With little more info. maybe I can get a little more help? the equipment is a horizontal fan (shaft) driven (via belts/sheaves) by a 5HP motor (that's five, pretty small I guess). I believe the motor speed is about 3600 RPM. I'm not sure of the fan shaft speed until the test but it is less than 3600. (1)Because it is belt driven (not direct coupled) doesn't that mean alignment isn't a big issue? I have not had the opportunity to see the oversized key yet myself. (2)I'm not sure if it applies in this case but is anyone familiar with how to test for residual unbalance? jmack, I think you are saying make the key right in the first place so as not to make unbalance. Of course that makes sense but I am told we must use this oversized key. I will not see the equipment until tomorrow (Monday March 10th). IF we are stuck with excess key material on the shaft, does it make sense to add equal material to the shaft on the opposite side?

 

[jmack]

Nisqually,
First and foremost with 3600 rpm equipment, alignment is always a big issue even with belt drive.You MUST align the sheaves properly first before you undertake any kind of bqalancing operation. Misalignment on belt drives can produce peaks predominately at 1X, so it is especially hard to discern alignment issues from balance issues from the FFT. You could make an alignment check with a relative phase measurement across the IB drive and IB fan bearings, you are looking for a phase difference of approx. 180 degrees to verify misalignment(I would assume that you would have reflective tape or a strobe if you are going to continue with a balance operation). If your fan is supported on both sides by bearings (not overhung) and the key is not symmetric at about the center position of the fan, and this is your only source of unbalance (which is most unlikely) then yes add the unbalance weight in one plane on the opposite side. I would always suggest a two plane balance, because No.1 ...the static unbalance case is rare, and No.2 ...you are dealing with high speed equipment (unbalance force is proportional to the SQUARE of speed). If the fan is overhung, you must balance in two planes (dynamic balancing).

J MacK

 

[micjk]

Nisqually,
I have a technical article concerning keyed shafts and balance. If I can find it I will send it to you via e-mail if you provide me with an address. Roy Gariepy
Maintenance and Reliability Dept.
Bayer Corporation Dorlastan Fibers Div.
Goose Creek, South Carolina USA

 

[Ralph2]

The size of the key has very little to do with the balance of your assembly..... Providing.... that the key completely fills all the voids in both the shaft and the coupling, impeller etc. So, if your key seat is longer than the coupling the void must be filled, if the key protrudes past the coupling but fills the key seat then it is a simple matter to remove the excess flush with the shaft.
When the coupling is initially balanced the void is calculated and an appropriate mass is added with the intent of arriving at a condition where it would be the same as if there was NO key-way at all. This can get complicated if the coupling is of a different material and some worry about the mass missing because of the gap (clearance) over the key. In the big picture this is insignifigant ;-)
Good luck
Ralph

 

[nisqually]

Hi. I just came back from the field. The situation is the keyway extends all the way to the end of the shaft so there is missing material (not extra material like I thought). What is usually done in this case? I did take vibration readings in the horizontal and vertical directions on both shaft bearings. This was with no extra weight and with some keystock hose clamped on. We move the keystock from inside the keyway, to 120 degrees away from the keyway and then 270 away from the keyway. Am I on the right track to decide where to add some keystock? Doesn't it seem like the keystock should be put in the keyway where the missing mass is? Thank you for your help.

 

[GregLocock]

Yes you are on the right track. You need to read up on static and dynamic balancing, or more properly single plane and two plane balancing.

Here's a quick summary of the minimum set of measurements one would make. Vibration is recorded at each pedestal for all runs.

1) Initial condition.

2) Trial imbalance at zero degrees on plane 1

3) Trial imbalance at zero degrees on plane 2

The trial imbalance should be big enough that the vibration on the nearset bearing should change by at least 45 degrees and/or 30% in magnitude (roughly) for each test. You don't have to use the same trial mass for each plane, but it is simpler.

This will give you a two plane solution, if all the emasurements are accurate. More practically you might like to test with two levels of imbalance, and mount the trial masses at 0, 90, 180 and 270 degrees. This makes plotting the vector diagram much more accurate, and may reveal linearity problems or the presence of other sources of first order vibration.

Before asking more questions, please read up on balancing, we are happy to help but you need to be confident of the maths and concepts.

Cheers

Greg Locock

 

[Ralph2]

Why are you concerned with this key problem? If you have had an increase in vibration and now want to "field" balance the unit make your corrections where they will do the most good, on the fan.
The force of the missing mass in the key seat can be calculated, but chances are that the forces created by the runout of your coupling will aproach that of the missing key. where its influence may add or subtract the net force.
If your missing mass was say 1/2 x 1" then you would have some 16 gram inches missing. This, at 3600 RPM would exert roughly a 13 pound radial pull. If your coupling was say 20 pounds and had a tir of .003" (.0015" runout) it would create some 13 gram inches of unbalance with a resulting radial pull of 10 pounds.
This unbalance force will show up as a dynamic unbalance on your fan. If it is possible to balance your fan by adding (removing) weight to the coupling then by all means do it, you may be lucky ;-)
Good luck
Ralph

 

[nisqually]

-
- pulley
fan - -------------|||---shaft (with unfilled keyway
- /\ /\ | on this side of pulley)
- brg4 brg3 | belt
|
motor|||pulley

OK. Tell me. Is a picture really worth a thousand words?

Not sure why the major concern with the empty keyway Ralph2. This particular work place is hard to describe in a few words. Does this picture shed any light at all?--Nisqually--

 

[jmack]

nisqually,
Yes a picture is worth a thousand words, and you will have to dynamically balance this type of overhung setup if you want to be effective in reducing vibration due to imbalance, it is basically the same for overhung pumps as well. Is it possible to cut the shaft off to eliminate your problem?

J MacK.

 

[GregLocock]

For a two plane balance I'd put balance weights between the pulley and bearing 3, and between the fan and bearing 4

Cheers

Greg Locock

 

[ChrisOlsen]

Speaking as a guy that does alot of balancing, perhaps I can offer some advice. Your choice of instruments is good, I have a VB-2000T myself and they work great.

I am guessing you are trying to balance at the sheave because the vibration amplitudes are highest here. It may be unbalance, but it might also be sheave eccentricity. Perhaps you should remove the belts and use a dial indicator to check for sheave run-out.

Another cause can be a bent shaft, some folks like to tighten belts TOO tight. Yes, belts CAN bend shafts.

Something I would suggest is to look closely at the vibration amplitudes of the fan versus the motor. If the motor is shaking at fan speed AND at amplitudes higher than on the fan, you have an eccentric problem. (Picture a cam instead of a pulley.)

If you determine that it really is unbalance due to the keyway (0 or 180 deg from keyway), I would advise your customer to use the proper key. The rule of thumb for key lengths is 1/2 the length of keyway.

And a few words from experience ... please don't use those lead weights with double stick tape, they are a serious hazard when used on sheaves.

 

[jmack]

Ralph2,
It would seem that some of the more recent issues to you are insignificant, or should be "ignored". I can see why we have petrochemical plants having major disasters with this type of attitude towards precision. My advice would be to read some of the advice that Locock is writing...you may learn something. Sorry if this offends, but it should be brought to your attention, as people rely on this advice and might make decisions with serious consequences based upon sub-optimal information.

Just an observation,

J MacK

 

[Ralph2]

Sorry jmack.... What is important is the big picture, Precision is a good thing but it needs to be kept in perspective. Something engineers have trouble with ;-)Nisqually's original question was a concern about a excessive key (turned to be a void in the shaft where the key did not protrude or fill the key seat). There was never a question as to why this fan needed re balancing, was it good at one time? If it was good then "something" changed. Look for it! Was it a new installation that was never run? Was the fan balanced before it was installed?
The likely scenario here is some technology student noticed a bump on the shaft next to the coupling and is getting excited for nothing. Insignificant, without a better idea of the big picture who knows?
The accepted process for balancing a fan in a balance machine is to dynamically balance without the coupling using a fully formed half key to completely eliminate the void of the key seat. Then (if one has it) install the coupling with a key that both fills the void in the coupling and fills the excess key seat in the shaft. Recheck the balance, if it is not in tolerance then make the correction on the coupling. Most times one does not have the coupling.... the balancer now must relly on the trades person to assemble the coupling with a correct key. Here he has a choice. He can make a formed key (full but with the excess removed or he can add a full size key and have it fill 1/2 of the length of the exposed (unused) key seat. Both methods will suffice, but the latter will leave a "lump" where someone could get concerned.
The reason for having a longer keyseat than required (typical of a lot of small fan assemblies)is to allow axial movement of the coupling to align the belts to the motor.
And.. To dynamically balance that fan one needs two correction planes. If you want to use the coupling as one then the other should be on the center of gravity of the assembly (typically just inside the back of the fan. Or use both sides of the fan, the outside (suction) and the back face, these are the planes that would have been used when originally balanced.
Cheers
Ralph

 

[nisqually]

Hi. No this wasn't a student noticing a bump. I was called into this at the last minute. I was told the fan needed balancing because of a keyway problem. This fan is used to ventilate large tank of nuclear waste. I'm not sure what maintenance activity had been done previously but probably no balancing. We can't mess with the impeller because it is hot. Radioactive hot. This was my first time involved with balancing. I am just trying to do a job. I needed help and I asked questions. There you have it.

 

[Ralph2]

To balance any rotor requires access to the correction planes. In a situation like this you could define your own correction planes on the exposed shaft, make them as far apart as possible without impinging or interfering with any existing shrouding. You may need to modify the guard to include these disks.
Have 2 balance rings made from a piece of plate that can be split and bolted back together. A disk some 8 inches in diameter by .875 wide would likely suffice. Counter bore for (4) socket head screws (3/8) grade 8 or better two in each side with the depth of both sets equal (and opposite) to equalize the mass removed (two bolts will go in rightside up and two will go in upside down) The disks must be bored to a interface fit (on the shaft) while clamped together. Have as many 5/8 holes drilled and tapped radially into the periphery as can be evenly spaced. Have the disks clamped on to a mandrel the same size as the shaft and statically balanced on a balance machine.
Now install the disks on your fan, torque the bolts to the designed specifications and proceed to field balance the assembly making corrections by adding 5/8 set screws in the balance rings. If your balance requires adding a substantial amount of weight (i.e. several holes are completely full) you might be advised to rotate the disk so that the split is in line with the heavy / light side. This will minimize the stress on the 4 bolts holding the ring together. When complete and satisfied stake the setscrews to prevent them turning out.
BUT....... I reiterate...... if the vibration has increased over what was the "norm" SOMETHING has changed. And hot or not, if a blade has come off the fan; if a buildup on the fan has come off un evenly; if your shaft is bent; your balance attempts will be an exercise in frustration and the results will not last long.
Good Luck ;-)
Ralph

 

[Ralph2]

Perhaps I should have also stated that allthough this is a method I have used the proceedure described for your case is NOT engineered. And considering the location perhaps it should be.
Ralph

 

[GregLocock]

nisqually - given the somewhat exotic environmental conditions of your problem, and your lack of experience with balancing, might it not be worth getting some outside help?

Cheers

Greg Locock