Balance two stages overhung rotor

Thanks Oli,

I know that the position should be marked to be sure the balance is the same with the balancing time. Actually, I do not have experience on balancing machine. So, could I:

- Balance each impeller and then do it for assembly. (Difficulties: have to manufacture bushing to fix each impeller when do balancing, I do not know how to do the rotor setting, static or dynamic?)

- Do it for whole assembly at one time.

Pls see file attached.





Photos (1)

I would do it all in 1 go with both discs, it will reduce mounting errors I think, are they not axially movable? Is it a defibrator of some kind for pulp processing so in operation the discs are close about 0.1mm or less? So I would balance it with discs close and that give you basically a 1 plane balancing and you could split the weights 50/50 on the discs if it is much weight. Are they worn? Or am I assuming to much and totally wrong? What would the operating speed be?

Thanks Oli,

In the last balance for this rotor, I did it as assembly. I set static first and try to correct static unbalance at stage 2 and then set dynamic for final balance.

I read in API 687 mentioned that for overhung type, it should balance for each impeller first and then do it as assembly. That why I ask for advice because I did not  use balancing machine before.

As you mentioned above, our procedure is not wrong, is it?

As I know, on the balancing machine have compensation mode like BIAS and Key, INDEXING. But in our machine, this function is locked with password. I do not know the passwords. So...


Oh so it's a petrochemical thing, it looked very similar to a defibrator for pulp. So it may be hi speed? Anyway, API apart I would do as you tried but I would part the static to be 50/50 on the rotors 1 and 2 if possible that is the only thing I do different. Unless you have a reason to put it on only one that is. You enter a lot of more unknowns regarding mounting etc. by doing it the other way but it is only my view. I don't think you really need to use the password protected features. They can sometimes more safe be handled manually. Again only my view as I like to see what happens and do try to avoid black boxes to interfere with my work and if something obviously is wrong I like it to be something I have done so I can understand it, sometimes. What balancing grade would you try to achieve?

Thanks Oli for your sharing.

This pump is small vertical pump with speed 2960rpm. I am try to get the residual unbalancing to G2.5 (almost the pump is applied G.5, some pump G.1  with speed around 8000rpm, turbine & compressor are applied API grade).

you are right that It is not necessary to keep the compensation function password. But some guys do it and now i can not assess in. Normally, I am not a person in charge for  balancing job on machine (the other guy do it) but I am in charge to find the root cause of high vibration in this pump (the pump vibration from 10 to 12mm/s) and we do some maintenance time before but the problem still remain. I take the spectrum and see that the main frequency is 83.7Hz. 1X is not much. It still question for me about this frequency.

When we open the pump and take the rotor to do the balancing check on the machine and see that the unbalance value is 180 times of G2.5 which before maintenance, some guys balanced it to 2 times of G2.5. By this, some people are doubt on the reliability of balancing machine (due to same rotor). And i am in role to check the procedure of balance and as well as the reason of unbalance is 180 times. At the time, i want to use the bias and indexing function to check the performance of system (because we do not have test rotor). But i can not do, so, i try to one rotor balanced in the past but still stored in house, just try to run this rotor 3 time to see phase shifting and turn rotor 180o and check for changing phase and amplitude (phase shifting 1-2o). I think it is fine, i suggest the maintenance to check run out of the pump rotor and see that shaft is bent at coupling hub position with run out value of 0.10mm (could be happened when technician install the coupling hub without heat it). So, it have a reason to explain the unbalance issue but is it 83.7 Hz in spectrum or not still question me.

And now, i try to balance the rotor again.

Yes my suggestion was not to use the builtin function but as you did, do it manually. I think this password is standard in the machine and always apply and maybe found in the manual it may be some random number but I don't remember. If I find a strange freq. that do not compute I look for a rub and that may also contribute to change in balance. A rub would excite a resonance at a random freq. and that is what you find in the data so balancing procedure may be ok. Would you have a water lubed bearing? Is it ok? I am always suspicious of those as they often sit in a difficult place. Is the shaft long? So yes it may be difficult to assemble those. 

This pump is vertical pump, small one using SKF 7311 BECBM, 6212 C3 with oil mist lubrication . The length of shaft is about 808mm (you can see file attached).

You are right that rubbing could excite the natural frequency and sometimes it is very strange one. I also think about the resonance issue and intend to double check when the machine taken into service. Now, this pump is still repaired for some corrosion issues. 

Anyways, when you do balancing for rotor, have you put the cooling fan to balance with rotor as assembly as sometime the fan could make effect when turning?


Normally the fan is a part of the motor rotor or in some cases not normally vertical it is also a part of the gbx. To cut it short, I would have the fan treated as a separate case and a thing to be balanced separately, normally just cleaning is enough unless any other damage occurs.

Yesterday, I do balancing for this rotor but it is very unusual when the appeared value of unbalance on machine is not the same with correction amount. It means that the balance is not linear. The machine showed that rotor got 34g at plane 1(near bearing) and 25.4g for the plane 2 but when I put 5g for plane 2 the unbalance reduce to around 20g plane 1 and 15g plane 2 but when I increase 2g of clays more, it make increase unbalance again.

i also try static/dynamics to balance with static first and then dynamic but it still not work. 

So, I think it have some thing wrong here. I remove all clay and recheck rotor unbalance again and it show the same unbalanced value and same phase to original. 

I balance this with fan installed on rotor. So, it could be effect to but it still question for that:

1. Is the rotor flexible in speed 500 rpm? Why it is unlinear. 

2. Is this the fan actually effect to unbalance value of rotor when it turn.

3. The distance from two disks to bearing is quite long (a=480, c=370, b=160, rotor weight: 30kg). So, it swing like beam.

Just some stupid ideal. I hope you can share your ideal in the case.

Normal in such case is that something is loose alternative is that the speed selected for unbalance is "unlucky" test another speed. You can take away the fan to see if it makes a difference or package in plastic foil to avoid turbulence.

Hi Oli,

I have thought about loose but after I remove the clay weight and recheck, the result show the same value as the original. I also require to turn rotor 180o to check phase shifting and the result is 180o shifting.

I will take the fan out and see.

Dear William,

The radian for correction weight is correct.

1. Check that the distances entered in measuring unit of balancing machine (CAB 700) are correct physically. i.e. a=distance of correction plane 1 on the rotor from bearing#1, b=distance between both correction planes, and c=distance of correction plane 2 from bearing # 2. Also check that there must be no loose component(s) on rotor. 

2. Check again the radii of correction points on correction planes.

3. After balancing the rotor, mark the match mark, disassemble it, again assemble it according to match mark and then test it on balancing machine. It must show balancing readings on both planes within 2 to 3 times of the previous readings achieved.

4. Are hard bearings being used for this rotor suitable for diameters of rotor?

5. Check that end support on axial direction (also a bearing) working properly or adding its own friction during balance run?

6. Is this belt driven type arrangement or universal drive?


Dear All,

Now, this issue is solved.

After doing two case, Case 1: balancing rotor with two planes at two impeller and Case 2: balancing rotor with one plane at fan and one plane at both impeller. I discovered that the balancing with two planes at two impeller is correct and it can eliminate the axial force due to couple unbalance of two impeller.

And the reason for unsuccessful balancing is due to tolerance of connector which used to connect shaft end and universal coupling.

Thanks for your contribution!


Add Reply