Reading accelerometers using NI9230

[qusaimn]

Hello community. I am new to accelerometers and  am trying to measure vibration on rotating equipment. I have a B&K Vibro accelerometer (datasheet attached) . I connected it to an NI9230 and was able to get a signal out of it. Although I am not able to understand how to interpret what I am looking at. I would expect g values to show on y axis and show a constant ~1g when my accelerometer is vertical. Although what I see is values close to 0. It does respond to vibrations but I dont understand what I am looking at on the y axis. Do I need to scale the sensor readings? It does take 'g' as unit by default though.

Any help is appreciated! I have attached relevant snapshot.

 

 

 

[mcduff · Accepted solution]

---

@qusaimn wrote:

I would expect g values to show on y axis and show a constant ~1g when my accelerometer is vertical. Although what I see is values close to 0.

---

The acceleration due to gravity is "constant". If your accelerometer reads it, it measures a constant DC voltage. However, your accelerometer is AC coupled, which means it blocks DC voltages. That is why you are not seeing a constant of 1 g.

[qusaimn]

Thank you for the response! Your answer prompted me to search AC coupled accelerometers and I found this which makes perfect sense to me now. 

Although I am still wondering what are the values I am seeing on the Y axis then? Is it just the amplitude of displacement of the vibrations then? How do I interpret that?

[mcduff]

---

@qusaimn wrote:

Although I am still wondering what are the values I am seeing on the Y axis then? Is it just the amplitude of displacement of the vibrations then? How do I interpret that?

---

You are reading 1 sample at a 1 Hz rate. You are seeing noise and low frequency drift. Sample at 1000 Hz, read at least 100 samples. Tap the table; do you see a spike in the data? You may have to take at least 1000 samples to see the spike easily, you may miss it on a shorter  timescale. Bonus - take the FFT of the data, do you see any vibrations in the background?

[jason-cce · Accepted solution]

The accelerometer measures the vibration (accelerated motion) of an object. When the object is still, the output of the accelerometer is 0.
Generally, the vibration of a rotating part should below 0.1 g. If it exceeds 0.3 g, the vibration may be too large and the structure may be easily damaged.

[qusaimn]

That makes sense. I understand that. I was able to increase sampling rate and was able to see the spike.

However, what I am trying to understand is what is the units of the measurement on the y axis? Am I still measuring g forces(m/s2) or am I measuring the amplitude of whatever vibration I am detecting? Say I detect vibration spike in my data at 1Hz, what would that mean for the yaxis?

[mcduff · Accepted solution]

---

@qusaimn wrote:

That makes sense. I understand that. I was able to increase sampling rate and was able to see the spike.

However, what I am trying to understand is what is the units of the measurement on the y axis? Am I still measuring g forces(m/s2) or am I measuring the amplitude of whatever vibration I am detecting? Say I detect vibration spike in my data at 1Hz, what would that mean for the yaxis?

---

You set the scale and conversion in DAQ Assistant, so the Y-axis should be units of g. This assumes the calibration is correct and the accelerometer was not damaged when dropped. The units of the Y-axis are always going to be g if you use the same settings in DAQ assistant. Mostly likely your accelerometer has a frequency dependent response curve, in the calibrated frequency span, the response is ~100mV/g. If 1 Hz is in this span, then it has the same response, if 1 Hz is outside of this span the response may be lower or higher, that is, 100 mV may not be 1 g. You need to look at the literature that comes with the accelerometer.

[qusaimn]

That makes perfect sense. Thank you for the explanation!

[Henrik_Volkers · Accepted solution]

All sensors I saw from Bruel&Kjaer (now HBK) came AT LEAST with a (production) calibration certificate (at least one frequency point).

The spec of your sensor say 100mV/g +-10%  and another +-10% in the frequency range 1.5Hz to 10kHz.

So without the individual cal-sheet value you can expect  10%  increasing to ~15% (20% worst case) to the lower end (High pass filter 0.5Hz) and higher end (resonance effects) of the frequency range.

With a calibration in the frequency range (usually 1/3 octave) you migth come down to 1-3% (depending on the cal-lab)  ... if needed.

Kodos for the well documented question!

If you use LabVIEW, have a look at the Sound&Vibration Tools.

[qusaimn]

Hello, 

I have posted another question on accelerometers - Accelerometer + LabView + Physics question - NI Community.

If you had any input for me I would greatly appreciate it. Thank you!