Thought I would do a"poll" and see what we get on this balance job as far as placement of the first trial weight is concerned, based on the data listed here.
It is no trick question, was an actual balance job. just like to see who says what. Don't let the majority influence you if you disagree.

See the coastdown also.

Question:
Where would you put the first trial weight having only 8 discrete positions available, based on:

Reference = 3.5 mils @ 97 degrees

Transducer and Tachlight in the vertical direction facing down

Rotation is CCW when looking from the rotor end toward the motor

Direct drive

Overhung rotor

Single plane balance

Choices:
0 Degrees?
45 Degrees?
90 Degrees?
135 Degrees?
180 Degrees?
225 Degrees?
270 Degrees?
315 Degrees?
360 Degrees?
None of the above?

 

Questions before guessing:
1. Ball/sleeve bearings?
2. Mounting (springs, concrete, steel weldment)?
3. Why vertical direction?
4. Which bearing is this?
5. Was alignment checked?
6. What is the lower end speed on bode plot?
7. Phase V-V, H-H between bearings?
8. Shaft diameter?
9. Picture?

Guessing will get you in trouble without enough data...

quote:

Questions before guessing:
1. Ball/sleeve bearings?
2. Mounting (springs, concrete, steel weldment)?
3. Why vertical direction?
4. Which bearing is this?
5. Was alignment checked?
6. What is the lower end speed on bode plot?
7. Phase V-V, H-H between bearings?
8. Shaft diameter?
9. Picture?

Guessing will get you in trouble without enough data...

Bill Kilbey, Director of Training Mobius Institute- Modern, Visually Interactive Reliability Training

Bill,

Are you saying if one doesn't answer all your questions, one is wrong in "guessing" where to put the first trial weight? I'll try to answer them then, but I disagree and had to balance this long distance with only the info I gave. Too many questions has a tendency to get one totally confused, sort of like answering all the stuff some analyzers' balance programs ask.

1- Ball ( where would the angle have been if sleeve? Never heard of this having an affect on the first trial weight, HMMM.)
2- No Springs, not sure of all the others asked, but some was welds and some were bolts.
3- Why not the vertical? Is there a "rule" against using a direction other than the horizontal? OH Yeah! Seems like someone did say balance can only be done in the horizontal, but that was so many years ago, I can't remember who said it.
4- Next to the wheel
5- Yes, I am sure it was, but I wasn't there in person.
6- 500 rpm
7- V_V 0 H_H 0
8- Unknown
9- Sorry no picture

I'll assume the bottom of the phase plot is 0 degrees.

My best guess on the coast-down plot is that the machine is operating far below resonance. Therefore I'd think to put the trial weight 180 degrees opposite the initial vibration. That would be 90 deg CCW from the point on the rotor where you're reading triggering (keyway?).

To provide a little margin for the unknown, one might guess that you are anywhere at or below resonance, from 90 degrees lag down to 0 degrees lag. Split that in the middle and assume a 45 degree lag. Therefore to get an effect 180 degrees opposite the initial vib, you'd have to put your weight 135 CCW from the point of the trigger.

Ralph:

OK, I’ll bite…why does it matter where you put ‘the first trial weight’? I would think a more important issue would be determining the ‘weight’ of the trial weight based upon fan configuration considerations.

One other thing…if it is a one-plane balance, there shouldn’t be a ‘second’ trial weight unless things have changed since the last time I balanced a fan!!

Of course, I could be totally wrong….

Gary Forsythe

What transducer may matter: 3.5 mils - were you using an eddy current probe, accelerometer or velocity transducer?

Radius and speed? Your zero reference was in-line with a blade and key @zero and your keyphasor read 97 degrees plus from zero with negative going CCW at _______RPM and TW raduis _________ inches.

I'm kinda in tune with Gary F. But on the phone would ask these questions like rotor weight to guess the trail wgt.

So, like Gary, I'll bite 50 grams @270 degrees. It's gotta be somethin'

Hi Guys,

I try to answer your questions when I get a little time to do it.

Sam, I noticed you did not enter your "guess" in the "poll" when you replied. At least there was no "270" entered at 10:00 CST.

The trandsducer was an accelerometer. Fastbal II program.

This message has been edited. Last edited by: Ralph Stewart,

I'm with Gary, there are 359 wrong places and one right place. If I had my choice it would be the right place! The wrong weight could be troublesome, but the location, if wrong probably won't hurt anyone. Thanks, Jim

In what would be a typical scenario for me(working in displacement)...
Using Lasertach and accelerometer, place photo tape adjacent to a blade for easy reference. With your data from initial run, stop the rotor, move the rotor so the tape is even with accel position. Now move the wheel in the direction it turns in operation. Count off your phase reading. (you would actually be counting against rotation)For eight blades, each blade would represent 45 degrees. Your phase reading is 97 degree. Place the trial weight on the second blade from your reference point. This will give you 90 degrees. For our balance program which references location counting in direction of rotation, this would actually be 270 degrees. If we enter 8 positions instead of 360 into our program this would actually be blade number 7 counting in direction of rotation.

If my first solution is not successfull, then I might consider resonance or looseness as a possible issue.

I agree with ElecticPete re: coastdown review which seems to indicate machine is well below resonance. The amount of lag is a bit of an unknown...my initial guess is to assume closer to 0 deg rather than 45 deg.
Since measurement is displacment (dont have to account for integration phase shift) and the phase indicates rotation to high spot (with assumption of 0 deg lag), I would put TW on the 270 deg (~97 + 180)location.

But, I could be off plus or minus 180 deg :-)
Hows that for a safe answer

Hello all,

I voted for 0° - In my books and from years of experience with in situ balancing, I am positive that where you put the trial weight on does not matter at all.

Look at it this way...

Supposing that you put it accidently on where the heavy spot of the unbalance is. Then it will give you a larger displacement reading and the angle should remain the same or very very close to it. The reverse is also true if the weight is place opposite the heavy spot!

If, per chance, it is placed at any other point in reference to the actual unbalance lies, the results will either be larger or smaller in amplitude and the phase angle will have changed.

And that is regardless of phase lag so long as the same speed is reached everytime.

But, we all are aware that the unbalance is not concentrated on a single spot - It is a component of a multitude of unbalance points.

When these new numbers are plotted or inputed into your balancing instruments, the eventual suggested corrective weight and angle will be right... perhaps a second trim run may be necessary.

Balancing is not a science by itself.
The instruments are there to give you the results of your interactions but it is the logic that lies beneath all this that matters, else there would be no need for a technician, nor a graduate engineer to be there in the first place.

I believe that too much emphasis is put on the fancy approaches to what is, in essence, a very simple operation. Too many numbers will clutter the brain !
We should all use the K.I.S.S. principle.

My 2 cents worth,

Markoleo

Ralph,

Do you ever take shaft stick readings? The shaft and housing often show a big discrepancy in phase/amplitude, so yes, it may be important.

Horizontal is typically the highest direction, so that is the one I usually balance from. Vertical is very subject to base stiffness/bending, but usually works OK as well.
The phase plot looks like it may shift a little over the range you plotted. What is that shift? 30-45 degrees lag is my guess so about 45 is heavy spot, 225 would be a good trial weight location.

For those who think this is not important, I had a colleague guess wrong on the TW amount/angle. When the machine (ID fan) flew apart and broke both his legs and hips, the TW really mattered. lots of surgeries and 6 months in a body cast and he still walks with a limp 9 years later.

From my experience I've found that at times it's better to balance in the direction of least vibration, as many times the direction of the most vibration may be so due to resonance in that direction. I can't say that I have a preference for either horizontal or vertical by direction alone.

Bill Quote:

quote:

Do you ever take shaft stick readings? ...............................................For those who think this is not important, I had a colleague guess wrong on the TW amount/angle. When the machine (ID fan) flew apart and broke both his legs and hips, the TW really mattered. lots of surgeries and 6 months in a body cast and he still walks with a limp 9 years later.

Bill,

Some times I take shaft stick readings, especially on sleeve bearings.

I once saw a guy (outside contractor) just "throw a weight on anywhere" on a large dryer fan (1200 rpm), which already had about 20 mils amplitude, and when it was started up, the fan shook dust from the entire machine room (about 30,000 square foot area), had people running for cover (not from the dust) but from the noise and vibration. The guy said he saw the amplitude reading reach over 100 mils before the power could be cut off. Turned out, the fan was running nearly dead on top of its critical speed (or resonant whichever one wants to call it) and he had "accidently" placed the weight on the wrong spot (needless to say ). The weight did not appear too heavy for the size fan it was, but it was too heavy for the machine running at its critical. This was before I got into balancing myself full time.

This is why I take time to do my best to figure out where I want to place my first trial weight and make sure it is not too heavy. Of course this is only my opinion on this particular phase of balancing, and I could be wrong, but do not intend to stop.

All,

So "throwing" a weight on anywhere, to me, is not a very good practice. Again, I might be wrong, but after seeing a couple of bad incidents, well.......

I don't know how the "amount of weight" got into the "poll".???? The original question was, where, based on the coastdown and Reference phase would one perform a S.W.A.G. as to place the first weight. This "position" question is irrelevant to the amount of weight (at least, it is to me). That is a completely different part of the equation. It takes less time to calculate the first position than it does to calculate the first weight from the weight of the rotor.

Good points entered so far by all, but each is based solely on the "path" that each individual wants to take. Some I do not agree with, some I don't quite understand and some seem to be along my train of thought.

Like one of our members has said before, "this is not rocket science".

Thanks for participating. Keep on with the replies and "votes", great to see so many different opinions.

Job done?!. So, how much wgt was put where?

This is a very simple case.... coast down data indicates a well-behaved machine, running below resonance, with what looks like about a 50 degree lag. High spot at 97 deg, heavy spot at 47 deg, and light spot at 227 deg... that position will be within +/- 20 degs of final weight placement. Took about 30 seconds to figure this.

Never one to by shy, the idea of "it doesn't matter where you put your first trial weight" is just dumb. Of course it matters! Your first trial weight should always result in a reduction in vibration for numerous reasons... if you can't figure that out on your own, you shouldn't be trying to balance anything. One of the most important reasons which never gets mentioned is that high amplitude vibration will cause a machine to respond in a non-linear fashion, which can invalidate your balance data. Balancing is based on a linear response of the machine to the trial weight.

If you begin with moderate vibration, put the trial weight in the wrong place (or use too large a weight) and then have very high trial run vibration, your trial run data is probably invalid. You will end up chasing your tail.

If you start with very high vibration and are successful reducing the vibration on the trial run, your balance run will probably give less than optimum results because your reference run data is probably not good becasue the machine was non-linear at that vibration level. If you start with very high vibration, go ahead and weld a weight in place if your trial run was considerably better. Then start over with a new reference run.

This message has been edited. Last edited by: rustythevibeguy,

Rusty,

Am I missing something? You're stating to estimate close, weld a wgt in place and start over? Weld a final wgt? Get close again, start over again? Never heard of such of an approach! Did I read it wrong?

I think a good trick is: alway record what you do and save the hard copy. "I put the transducer at __________ in-line with the key phasor mounted _______________ on an over-hung fan rotor with two pillow-block bearings belt driven (the rotor was 1P, 2' radius w/8 blades). Keep one of these for every job and pull it out for the next similar job. If your last was a New York Blower overhung as mentioned then by using keyways as indexes and repeatability you can even sometimes trim in one-shot. By being consistent in your setup you can often one-shot a machine or get lucky or like I mentioned to this lady last weekend, some like to rub me for luck - wanta try? Not all Genie's look alike.

Why no scales on the plots? One cannot determine phase lag without a scale. Good to know the bottom speed was 500 rpm.

I think Rusty is in the correct location assuming we are talking lag angles for the measurements (lacking a scale to judge the phase lag). Polar plots really make this easier.

225 to 240 degrees looks good (without scales). Machine design also enters - could the weight placement be the other side of a node from the measurement location? In that case subtract 180 degrees.

Of course it matters where one places the weight. Why would I hire a consultant who 'guesses' for the first weight? Too much risk for me to take. Ever look at the cost for missing on a big turbine-generator? What if you couldn't get to speed or sustain speed/load on one of ElPete's TG's - I would love to have the money that would cost.

I don't fully understand the rest of what Rusty was trying to say. I think he meant that he would get the vibration to a reasonable level and start again with a balance? For me I would probably try to bring it in with one correction after the trial, anyway. Most people don't like to do a lot of welding of weights.

Sam, Bill, what I'm referring to is when you have a machine that is -- for whatever reason -- "shaking the world". Your reference run data is possibly non-linear because of the extreme vibration. Often when you try and balance such a machine, after your trial run, you will get a balance calculation that seems reasonable, but when you put that weight on (whether welded or clamped), you don't get the predicted result. If you did everything correctly (didn't put the weight "180 out" for example), the shot "misses" because the high-amplitude reference run was non-linear.

For example, I have a large (1500 hp, 10,000 lbs) fan running 15 mils at 900 rpm.... it is "shaking the world." Chances are, it is also non-linear. So I put on a trial weight using a 30 deg. lag and knock the vibration down to 7 mils. I will then go ahead and weld the trial weight on (or something equivalent)and make a new reference run because at only 7 mils I am fairly certain it is linear now and I will get good reference run data that is valid going forward. Certainly I don't do this in every instance...

I will repeat this, as a fact, if you try to do a "textbook" balance where you have very high, "shaking the world" vibration, you will usually (80% of the time) not get predictable results, because the machine will likely behave in a non-linear fashion in extreme imbalance situations.

quote:

Originally posted by rustythevibeguy:
...coast down data indicates a well-behaved machine, running below resonance, with what looks like about a 50 degree lag. High spot at 97 deg, heavy spot at 47 deg, and light spot at 227 deg...

I wonder if one can use coastdown data for determination of the lag angle.

Let's assume the coastdown showed a 50 deg phase change, but this is just a CHANGE. Beside mechanical, the measured phase angle also includes electronic phase lag, so to claim that heavy spot is at 97-50=47 deg may carry a big error.

My guess that was the problem when TW was installed at 225 or 270 deg.

(Some people say that there is no electronic lag now days... . I am not sure how accurate this is.)