Vibration V3 sensor question

Hi,

I have configured a V3 vibration sensor following the instructions in this video :

The data exchange between the sensor and the receiver seems to work well, except for the following 2 points:

  1. Only the x-axis data is displayed although all 3 axes have been activated (see pictures below)

  1. The 4 blade fan on which the sensor is installed is running at 665 RPM. Under these circumstances, the recorded frequency peaks (562, 1112, 6412) on the x-axis seem high to me. This may be due to the fact that it is the Z axis and not the X axis that is perpendicular to the fan shaft. What do you think about this?

I’d like to get your opinion on this.

Thanks

Hi Bert,
if the RPM is 665 then the ODR value should be 200.

100sps (Range 1.56 – 25 Hz) 0x07

200sps (Range 3.125 – 50 Hz) 0x08
400sps (Range 6.25 – 100 Hz) 0x09
800sps (Range 12.5 – 200 Hz) 0x0A
1600sps (Range 25 – 400 Hz) 0x0B
3200sps (Range 50 – 800 Hz) 0x0C
6400sps (Range 100 – 1600 Hz) 0x0D
12800sps (Range 200 – 3200 Hz) 0x0E

it ignores the values below 35mg during FFT ( dead band). that will be the reason why it wont read anything on Y and Z axis

Thanks

Hi Bhaskar,

I’m not sure if I fully understand. Is ODR synonymous with sampling rate when talking about vibration analysis? If so I don’t understand how you arrive at this value. In your table, what is the difference between sps and Hz?

In the field of vibration analysis, this equation is commonly used:
Ns = Fs X Ts
where Fs is the maximum frequency of the vibration to be measured in Hz X 2.56 (Nyquist Theorem),
Ts is the sampling time, usually the time required for the shaft to perform 10 rotations,
Ns is the number of samples required

If, for example, we want to measure the vibration caused by the bearings of a fan rotating at 1800 RPM, how would you set up the sensor considering that the vibration frequency of a defective bearing can be of the order of 20 kHz, by determining Fs, Ts and Ns as in the previous paragraph.

I would really appreciate if you could answer with the settings of the sensor taking into account the previous example as well as with the equation Ns = Fs X Ts so that I understand how to configure these sensors according to the real use cases that we have.

I know that my question is very picky. It is important to understand that we are working hard to get our customers to adopt this technology instead of traditional analysis equipment.And to do so, we must offer them sensors whose settings allow them to achieve the same result as their current vibration analysis equipment.

Thanks

Bert,
Here is how the algo works. once raw samples are obtained they go through hardware low and high pass filter to remove noise and unwanted signals . in this filter we define the high and low cut of frequencies.
If the user defined ODR is set to 100sps. it means the low cut off frequency will be 1.56hz and the high cut off freq will be 25hz. so any machine which works within this range should choose this ODR. the filters cut of freq are designed to satisfy the Nyquist theorem.

According to Nyquist theorem the vibration sampling freq should be always greater than atleast 2.56 times of machine frequency. as you can see the max frequency is still complying the Nyquist equation in all cases.
for an example a fan running at 1800 RPM ( 30hz) should be samples at min 75hz. in this case the users can choose 200sps,400sps,800sps,1600sps.
the higher the sps the lower the lower freq will be.

thanks

Hi Bhaskar,

Thank you. I understand better now.

If we continue with the example of the fan running at 1800 RPM :

You say that a frequency of 75 Hz is enough (30 Hz X Nyquist factor). Since 30 Hz is 1X the rotation speed, I guess you are only measuring the symptoms of unbalance in this case, right?

If we assume for the example that the fan has 8 blades and we want to detect a blade problem, we would have to choose 6400 sps instead, right? Indeed, the vibration frequencies of a blade problem correspond to 1 X Blades Numbers and 2 X Blade Numbers. In our case, the highest frequency corresponds to 2 X 8 blades X 30Hz X 2.56 = 1229. So ODR should be set to 6400 sps, is this correct?

If my reasoning is correct, is it possible that these sensors are better suited to detect problems occurring at low frequencies than the vibration problems associated with bearings, which cause frequencies of 20 kHz and more ?

Bert,
200SPS is the minimum recommended sample rate in case of 30hz machine. 1600SPS will cover 1x,2x,3x,4x,5x.
yes, the fan blade assumption is correct.
this sensor sample rate is limited to 12.8K.
The mems sensors mechanically can not reader higher frequency than this.
Bearing issues start to show at 6X,9X,12X
Rapid change in computed acceleration is one of the key indicator of bearing problems as well.

Thanks

Hi Bhaskar,

Thank you very much for the clarification. Everything is much clearer for me now.

Last question for today Bhaskar:

Is the calculation of velocity values from the acceleration values given by the MEMS module independent of the ODR settings. This question refers to the use of the sensor on a variable speed machine. For example, I have in mind the example of a warehouse elevator where the speed can vary by a factor of 10. This is an important point since we use velocity according to ISO 20816 to predict the degradation pattern of the equiment.

Thank you :slight_smile:

Bert,
Yes, velocity will depend on the ODR as well. Most of the motors are controlled by VFD and they have pre defined range. The ODR should be selected based on this range. if the machine frequency is outside the sensor settings, it will still read the signal but some of the vibration values will get attenuated.
we added support for 25600hz ODR as well. this will give bit more flexibility in such cases.

Thanks

Thank you Bhaskar :slight_smile: