Rotary Balancing

[LudoLelo]

Hi,

I am using the Sound and Vibration Suites to do some rotor balancing. I am trying to use the Order Tracking to get the amplitude and phase of the first order for each of my accelerometer.

However, during my reading, the phase is decreasing at a constant rate slowly in time (around 10 deg /minutes for example). I tried to modify my setup, to position the tachometer at different location, but I still get this phase shift.

The amplitudes remain constant in time but not the phase.

Nothing change during the runs so I am confident is not coming from my equipment. I tried different setup and still experience the same thing.

I would like your help to understand what could cause this to happen.

Here are the equipment I am currently using:

- (2) 3-axis Accelerometers PCB Piezotronics (356A15)

- (2) NI 9234 ((1) for the accelerometers, (1) for the tachometer)

- (1) Remote Optical Sensor (POS-P from Monarch Instrument) with a Self Powered Sensor Interface Module.

(I also try to acquire the tachometer using a NI9205... got the same behavior)

Thanks in advance for your help.

PS: I looked at the "Two Plane Balancing Example with DAQmx." and it seems to get constant reading 

I would like to understand the difference between this example and the Order Tracking 

[RL]

I ended up solving my balancing with a non NI solution.  Using a simultaneous reading will eliminate skew which is good.  I have not used PCB piezo accelerometers  but I have used dozens of their pressure transducers which can handle very short rise times but have a rather rapid signal decay unless there are pressure changes to recharge them.  I would think that those fluctuations are present when balancing.  Are you using only the axis that is perpendicular to the center of your object being balanced?  I now use two CMCP1100 General Purpose Industrial Accelerometer which are fast enough when you consider that my maximum balancing speed is around 2,700 rpm.

My angular position is done with reading the reflection of a Banner QS30 laser beam.  I have also used a infra-red transmitter receiver.

I guess your priority question re: Order Tracking has not been answered. 

[MickeyD]

Hi,

I believe the example shows the speed results. I am assuming the speed remained constant during this run? If it did not then that could be the culprit. If for some reason the measured speed and thus the speed profile then the phase relationship of the first order would change as well.

You mention you have 2 346A15. Each of those is 3 signals. Are you seeing this phase decrease on all the axis? Or only some of them?

I am assuming as well the 9234 devices are in a cdaq chassis together? If they are both in a single-slot USB sleeve then that could cause some issue.

The 2 plane balancing example is different in many ways. I didn't believe it continuously showed a phase result? But could be incorrect. Can you provide a little clarity when you say it gets a constant reading? I will try to take a look at the code just to familiarize myself with how it does its calculations.

[LudoLelo]

MickeyD & RL

- the speed is indeed constant

- I read only the radial axis of the 2 accelerometers. Phase shift is seen on all 4 measurements

- I use a chassis

I have just tried to use only (1) NI9234 for both the tachometer and one accelerometer and it seems to solve my issue.

Would it be possible that the fact I am using (2) different module NI9234 for the accelerometers and the tachometer cause some phase shift ?

[MickeyD]

That is a good catch. It should not the mudles should synchronize for you but what version of sound and vibration are you running? Thank you

[LudoLelo]

I am using 2012 Version.

I also try something else...

So far I was using 2 different steps (I am using the Sound and Vibration assistant) to acquire the accelerometers on one hand and the tachometer on the other hand.

But I do only one step for the 5 signals ((1) tach + 2 x (2) accelerometers) than it solves the problem as well.

So i guess, having different steps was causing some delay/time shift.

What do you think ?

[peper2001]

Hi,

I'm not sure I understood your setup corretly.

Are you making the acquisiton of the accelerometers, and the tachometer using NI 9234 modules? With all modules in the same chassis?

If not, I think it would be better to do so, in order to make sure the clocks of all acquisition devices are synchronised. Also be aware that 9234 modules have an anti alisiasing filter that will introduce a small delay compared to other modules like 9205. So if you need to have absolute phase measurement, I would recommand acquiring the tachometer with a 9234 as well.

In software, make sure that the acquisition of all channels (including tachometer) belong to the same DaqMx task. Again it is to make sure everyone has the same reference clock.

Best regards,

peper

[MickeyD]

Hi,

That makes perfect sense. I had not realized you were using the Sound and Vibration Assistant. 

Each 9234 has an oscillator in it that can create an oversample clock that drives the ADCs inside it. While each one is very close every oscillator is unique and some drift between them is unavoidable.

To allows for sychnronized measurements that oversample clock can be shared across the chassis between multiple modules. (This is a simplification, feel free to request more info in the hardware forums about delta sigma synchronization if you like.)

When you create 2 steps in the Sound and Vibration Assistant the DAQmx driver is creating 2 tasks. Those tasks do not have to be synchronized and thus the routing to make them synchronized is not happening in the hardware.

When you create a single DAQ step with all of the channels the driver attempts to make them synchronized, which is why you see a difference your hardware is now being configured properly for the multi module acquisition.

The example of 2 plane balancing did not provide this choice. There is also a LabVIEW examle that comes with the Sound and Vibration Measurement Suite for order tracking that provides identical analysis as the Sound and Vibration Assistant, but the hardware configuration is built like the 2 plane balancing example you mentioned above.

It sounds like everything is working now. The phase you were seeing was the drift between the 2 oscillators on the 9234, which won't occur when you use a single DAQ step in the Sound and Vibration Assistant

[LudoLelo]

MickeyD

Thanks a lot for the explanation!

Make complete sense now. I hope it will help other people as well.

It is working fine for me now.

Thanks everyone for your help.

Have a great day