This is rather kind of discussion then a specific question.

Phase readings rarely make any sense in my experience when it comes to balancing. At high amplitudes I just go for balancing regardless of phase and most of the time it works.

Here is another case of phase. Hopefully someone can make sense out of it.
It is a small fan. The impeller (D/W=4)is mounted cantelivered on motor's extended shaft. All phases in H and V, on DE and NDE are practically in phase (+/-6 deg). All amplitudes are about same and equal to 0.4 ips.

These phase readings mean that motor is moving as a whole body along a straight line.

I'm trying to picture mentally as to what would cause such a motion? No shaft misalignment to blame. Worst case scenario - dynamic unbalance. But it should not result in all synchronous phases. Could it be an indication of internal bearing misalignment?

I'll likely do balancing and it will likely reduce vibration. But I am just trying to understand.

David

If there was a mode shape involving pivoting at diagonally opposite feet with resonant frequency near running speed, then it would be excited by unbalance or other running speed excitation.

David,

I'm a little confused. You state that the phase of the fan is in-phase, but then you say the motor is moving as a whole body. Did you mean to write fan or is this a solid coupling?

I'm assuming you are expecting an overhung fan to be out of phase due to the overhung weight, and this has been my normal experience too.

You state it's a small fan. If you look at the phase during the startup or shutdown, does the phase change significantly? On some overhung fans, with relatively small impellers, I have seen the phase change as centrifugal forces over come gravity and both ends go in phase. Of course, like most fans, they run in their first mode.

Hope this helps,

Zack

Pete,

The fan is running 35% above resonance.

Zack,

The impeller is mounted on the motor extended shaft end. There is no coupling. As I mentioned, all measurement points (V, H on DE and NDE) are almost in synch, meaning that the whole body is moving in space along a straight line. Although it is not relative shaft movement, but still I expected some kind of absolute circular movement of the motor's body. Am I wrong?

I would like to return to the idea of resonance.

We have a force shape associated with the excitation and a deflection shape associated with the modes. If you're exciting force occurs at a frequency of one of the modal frequencies, the resulting response shape will tend to look like the modal response shape, even if the exciting force shape did not look anything like that.

One example is traditional pure H or V resonance. Let's choose running near H resonance. The excitation may be unbalance equal in H or V direction,, but the response looks like the H resonant modeshape (all vibration in the H direction).

We describe this particular vibration pattern as "directional", but really it is just motion in a straight line that happens to be almost pure horizontal in this case. But some other types of resonant mode shapes may still approximate a straight line motion, but not necessarily be restricted to horizontal or vertical plane . One example I mentioned was rocking pivoting on diagonally opposite feet which would include both horizontal and vertical components. If the rocking motion creates straight line motion at the measurement points, the vibration will be at 0/180 phase.

Like you guys, I am picturing an overhung imbalance can excite this mode.

You mentioned a single resonant frequency. As you know if we considered it a rigid body we have those 6 rigid body modes with 6 resonant frequencies and flexible ones can have more.

Just a thought. There may be other explanations as well and I'm interested to hear yours or others.

Hi David,
Could it just be that the unbalance forces deformed the motor's base? That way the motor would be vibrating in a rigid mode.

Just a thought,
Best regard, Marcel

What bothers me, as I mentioned, is the fact that in a case of solo machine I should expect close to 90 deg phase difference between H and V at 1x in a given plane, since unbalance is the only possible fault at 1x in this case. That is regardless of whether or not there is a resonance (which is not ).

But what happens in my example is that in both planes any point on machine appears to be moving along a straight line instead of an oval.

It sounds like you disagree with my explanation. That's ok, maybe I'm wrong. But I'd like to explain/argue my point further.

For illustration, let's take an extreme resonant case of vibration in horizontal is 10x vibration in vertical. The shape is an long narrow oval, but it looks very close to a line. Since the axis of that oval is along the H/V axis, the H/V phase difference is 90 degrees.

But now take that same long narrow oval and rotate it 45 degrees. It's still the same shape, slightly oval, but now the angle between H and V is very close to 0/180.

There was no change in the shape of the orbit, only a change in the orientation. So my point is the phenomenon we call "directionality" is the same as the phenomenon of H/V ~ 0/180, but just that the axis of the orbit is tilted.

If the ratio is more realistic maybe 3:1, then the same logic holds, but instead of 0/180 it might come out something like 10/190.

Rather than talking in terms of resonances, maybe we can just say that the motor base has low dynamic stiffness to bending in the direction pivoting about diagonal feet. Even though the force pattern (unbalance) doesn't look like that, the defomration/movement takes that shape because it has has low dynamic stiffness.

We can start with our H/V symmetric unbalance force, and by manipulating the stiffness of the support of the machine, we could make that same unbalance force result in any of the following shapes: primarily vertical motion, primarily horizontal motion, or primarily diagonal motion which has the 0/180 phase shift.

When we look at vibration like this we are always looking at forced response. However if we look at mode shapes, which are close to forced response near a resonance - but not exactly, then directionality is built into the modeshape.

When one takes an introductory vibration course (formal course like at a University) one learns about normal modes, which is what systems have under certain assumptions that approximately hold in general situations, but not quite true.

The orthogonal degrees of freedom (measurement points if you like) have relative phases of 0 or 180 degrees (because they are 'real' modes). Later one does learn about 'complex' modes, but that is a deeper well.

All this means that near natural frequencies the response tends to be directional.

David, I've read thru all the responses and my eye caught your answer to Zack as shown below.

It is the first time the word 'IMPELLER' was used by you other than in your first of the messages... FAN was always used.

It is very common for impellers to develop axial thrust as well as radial.
That may be what is causing this longitudinal movement of the entire rotor, assuming that this force is substantial and irregular.

I would look along this avenue if it were me.

Just my 2 cents.

MarkoLeo

quote:

Zack,

The impeller is mounted on the motor extended shaft end. There is no coupling. As I mentioned, all measurement points (V, H on DE and NDE) are almost in synch, meaning that the whole body is moving in space along a straight line. Although it is not relative shaft movement, but still I expected some kind of absolute circular movement of the motor's body. Am I wrong?

quote:

Originally posted by William_C._Foiles:
All this means that near natural frequencies the response tends to be directional.

Just to clarify - would your definition of "directional" in this context include comparable H and V magnitudes having 0/180 phase such as described in the original post? (that is the way I view it).

A question as strange as the phase:::
What is the amplitude at a point half way between the vertical and the horizontal?

Yes to comparable H and V as a possibility. Suppose you had completely directional vibration, straight line vibration. Rotate the coordinate system so that the line is between the senors; now the X and Y would be comparable (equal). Rotate so that the line is in the direction of one of the sensors; now you have a high multiple of X to Y, infinite.