I am trying to do the calculation and I am not sure that I am doing this correctly.

Here are what I assume to be the important numbers.

Weight or rotor = 200lb
Speed 3000 rpm
radius of balance weight= 8"
Correction/trial weight 1.5 ounces
Lets assume original was 3.8 mils at 0
and final was .3 mils at 0.
I calculate G=2.5 which meets the 6.3 criteria....is this correct ???
i am using equations from Eshlemans paper in the recent "Vibrations" Magazine.

fwiw, I calculate the same as you (attached, using free program "smath")

jbaolkdoUB.pdf (39 KB, 72 downloads)

I'm a little bored. I recalculated another way as attached. This way ASSUMES M*e = m*r (UNREALISTIC ASSUMPTION). Then calculates velocity in mm/sec under that assumption. This idealized non-realistic calculated velocity in mm/sec has the same numerical value as the G grade (2.6).

Anticipating getting beat up, flamed and bruised as has happened in the past when I presented this concept with similar copious caveats, I'll say again that the velocity so calculated is NOT the velocity we expect to measure anywhere on the machine. But this method imo provides the most straightforward way to analyse the definition of ISO balance grade.

jbalkoUB2.pdf (45 KB, 49 downloads)

Thanks for the quick response.

Hi Jbalko,

Are you balancing some parts ?

Dan..Fans. Dont understand the question.

ep,
Geschwindigkeit! And that isn't a blessing because you sneezed! It took 36 years in the business for someone to finally explain to me that the 'G' in that balance spec doesn't stand for 'Grade', is stands for Geschwindigkeit, or German for velocity.
The spec assumes that if you balance to the level on the chart, then when that rotor is placed in it's given machine and the machine is placed on a sufficient foundation, it will operate at that velocity level in mm/sec!
Ron

Ron

I have to admit, I never knew that.

Chris

Have you looked at what the ISO documents say?

quote:

Balance quality grades G are designated according to the magnitude of the product eper Ω expressed in
millimetres per second (mm/s). If the magnitude is equal to 6,3 mm/s, the balance quality grade is designated
G 6,3.

I tend to believe what the standard says and how the people in TC 108 WG 31 talk about it in the meetings.

Bill,
Thank you for providing all of the tinsel to the topic.
mm/sec....hmmmmmmm, YES! That is units for measuring velocity in metric terms. Then, ergo, a G 6.3 mm/s vibration is expected from balancing to the graph limits for G 6.3!

If there are 2 impellers on 1 shaft, are there 2 numbers? Or 1 cumulative ?

I have found this IRD Balancing paper very helpful, especially the simplified calculations.

quote:

Then, ergo, a G 6.3 mm/s vibration is expected from balancing to the graph limits for G 6.3!

I would respectfully disagree.

As was mentioned above, some not so insignificant caveats. The calculation is an artificial mathematical one. It relies on M*e = m*r, which is an equation that would be valid only for an SDOF system with total mass M = rotor mass, at a speed far above rotor support resonance (such that the effect of any rotor support stiffness disappears and rotor is moving in mass-controlled fashion). Many machines do not operate far above resonance. Further, other masses beside rotor mass may affect the dynamics (different than SDOF). Additionally, when we talk about velocity we usually mean housing vibration which surely will introduce additional error in difference between shaft and housing.

I think it is useful to understand that the ISO balancing grase (G number) has a connection to velocity in a abstract or simplistic sense, but we should not expect to measure that level of vibration anywhere on our machine.

By the way, you might measure a peak-to-zero rotor displacement somewhat close to the theoretical value of e=m*r/M on a soft balance machine (with some machine calibration factors associated with extra moving mass neglected) as was discussed in another thread. But that is not typically done at normal operating speed which is the speed where the velocity is mathematically (artificially) associated with ISO Grade..I mean Geschwindigkeit (v = e*Omega)

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

Balance quality is calculated as a function of w (omega) which is a function of 1x RPM. But isn't vibration usually a function of 'force' which is a function of RPM^2?

It's always seemed odd to me that Balance Quality is a function of Speed, and not Speed^2. Or am I missing something?

quote:

Then, ergo, a G 6.3 mm/s vibration is expected from balancing to the graph limits for G 6.3!

ElPete has answered this well.

I think it would be good if everyone taling about the ISO document get a copy and read it. I must be part of your business.

Regarding vibration response from ISO 1940-1.

quote:

A simplified method seems to be applicable in easy cases, but a proven basis is not yet available.

quote:

Balance quality is calculated as a function of w (omega) which is a function of 1x RPM. But isn't vibration usually a function of 'force' which is a function of RPM^2?

ISO talks about this. Again, it is worth geting and reading. This is not the first document with this approach.

One might reason that the permissible mass eccentricity times the rotor angular velocity is constant. API uses this approach also in the 4W/N approach. US standards have used this for some time. Mass eccentricity times angular velocity being constant is the same as allowable mass eccentricity varying with the inverse of speed.

You do know that the correction plane imbalance is not necessarily the balance tolerance from the tolerance planes?

"You do know that the correction plane imbalance is not necessarily the balance tolerance from the tolerance planes?"

I think I know what you mean. Could you elaborate on this important point?

quote:

Originally posted by jbalko:
If there are 2 impellers on 1 shaft, are there numbers? Or 1 cumulative ?

.
Beats me. Maybe others will chime in.

quote:

You do know that the correction plane imbalance is not necessarily the balance tolerance from the tolerance planes?

I feel that I might be giving the plot away and kill your interest in getting the 'book' and reading it. Suffice it to say that section 4.4 discusses 'Reference planes for balance tolerances,' and section 4.5 discusses 'Correction planes.' Plus, there is more in the standard.

One note, for those who have taken ISO certification courses (probably stated poorly) knowledge of the standards is required for higher levels. This might be one of the advantages that training and certification has.

If one owns a balance machine and sells balancing services, this seems important to me as a potential customer for such services.

Not every consultant does this. So, broad general knowlege might be expected, but for those who sell these services more detailed knowlege would be expected.

quote:

Balance quality is calculated as a function of w (omega) which is a function of 1x RPM. But isn't vibration usually a function of 'force' which is a function of RPM^2?

It's always seemed odd to me that Balance Quality is a function of Speed, and not Speed^2. Or am I missing something?

I think it’s a good question.

Bill mentioned "proven basis is not yet available."

But still worth trying to think through it.

The existing approach leads to a limit something like:
m*r ~ M*G / w
where m*r is residual unbalance, M rotor mass, w is speed, and G is ISO Gesundheit level. (an easier to remember German word, means healthiness... not too far off).

Since unbalance force is F=m*r*w^2, this leads to unbalance force F of:
F ~ M*G / w * w^2 = M*G*w

In other words, the unbalance force resulting from the specified allowable residual unbalance for a given mass rotor and given Gesundheit level increases linearly with speed.

So comparing two machines with assumed same rotor mass but different operating speeds such as 3600rpm and 1200rpm, we end up with three times as high allowable force on a 3600rpm machine as a 1200rpm machine.

Assuming similar bearings for both machines (doesn’t seem unreasonable for same-mass rotors) and both machines operating far below resonance where the vast majority of unbalance force transfers through the bearings, we could do a bearing life calculation. The 3600rpm machine gets penalized once by having a higher unbalance force by factor of 3. Then it gets penalized yet again because it accumulates three times as many revolutions per time (L10 life is based on revolutions). OK, maybe I glossed over combining the unbalance force with other forces that may be larger, but just making a point: which is that if there is a machine that should have lower allowable force, it is the higher speed machine, not the lower speed.

I have a hard time coming up with any logic that could defend this particular approach based purely on the machinery.

The only thing I can come up with is this is a compromise approach between balance effort and machine health.... to achieve maximum combined reliability bang for a given reliability buck.

Realizing there are not unlimited time and resources for any activity, let’s say we have 100 minutes total to spend on balancing both machines 1200 and 3600 rpm machine (same rotor mass).

To analyse, we need to assume a relationship between balancing time and residual m*r. I’m going to assume they are inversely proportional. (need to spend 3 times as much time balancing in order to cut the residual imbalance to 1/3)). I'm not saying that's realistic (I'm sure it's not), but it's simple and illustrates a concept.

How am I going to spend my limited 100 minutes?....

Does it make more sense to spend:
Option A - 10 minutes on the 1200rpm machine and 90 minutes on the 3600 rpm machine (ratio 9:1 required to achieve the same unbalance force on both rotors).
OR
Option B - 25 minutes on the 1200rpm machine and 75 minutes on the 3600rpm machine (ratio 3:1 in accordance with existing ISO approach)

In option B, we have dramatic improvement in time available to balance the 1200rpm rotor, since we have 250% of the time spent balancing it compared to option A (25 minutes vs 10 minutes) . What is the price of this improvement on the 1200rpm rotor? We spend 75minutes vs 90 minutes on the 3600rpm rotor (13% decrease).

Maybe the dramatic improvement in the 1200rpm rotor is worth the rather small sacrifice on the 3600rpm rotor? (bigger bang for the buck, even though the 3600rpm suffers slightly).

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

But Pete, it's a "standard" developed by experts.... it has to be correct! Who are we to question the logic of it?

Bill is gently suggesting that I study the standard, since I own a balancing machine and offer balancing services, and he's correct, and I will (I have found the IRD paper very helpful in the interim).... OK, just purchased 1940-1 & -2 ($232)

So far, all of the balancing I've ever done, the balance quality I achieve is so much better than required by the standard, that the standard is rendered useless for practical purposes. A serious balancer should use the standard as a starting point, not an ending point. I suppose it's "good enough" for equipment manufacturers, apparently. But I wouldn't accept the ISO 'recommendations' for most classes of equipment if I were a purchaser. The 'standard' applies only to calculation of balance quality level.... 'application' of a particular balance quality level to a class of machines is a 'recommendation' only.

This message has been edited. Last edited by: Rusty Castleman,