Impact test in concrete beams with accelerometers

[Henrik_Volkers]

Another test without the risk of overloading your sensor by dropping it:

Mount the sensor on an end of a metal rouler (or any other (spring) steel bar )

Clamp it in a wise or on the table and adjust the frequency to 16Hz (vary the length and / or vary the weigth) 

Now rougthly every mm peak amplitude should give you 1g

Bend the spring a known distance ...   release 🙂

5Hz and 10mm peak (or 20mm peak to peak) should work for 1g too

To avoid transverse motion you can fix the sensor in the center of the spring/rouler and clamp/fix it on both sides.... 

[billko]

If you haven't done so, confirm your test results on hardware first, so you can get an idea of what values to expect in your software.  In this case, 300 Gs sounds pretty high.  But like I said, check it with hardware first, so when you do it in software, you can have reasonable assurance that you are doing it right.  🙂

[StathPol]

Many thanks for your kindly replies Bill and Henrik. I will try this technique tomorrow in the lab. 

I have the setup as follows:

LabVIEW (On Laptop) - PXI - NI PXI6361 - Coupler (power supply) - Accelerometer...I have set a known voltage on the card of 5Volts to check whether the hardware is ok and labview reads 5Volts. 

Next, an oscilloscope was connected with the coupler and the accelerometer. Once I hit the accelerometer, it picks signal as well. 

I have a weird feeling about the settings. The coupler provides the DC power in the IEPE accelerometer. I have input the sensitivity and the range in the power supply. So if the full scale output is 10 VOlts , and the range is 1g, then it will result in a + 10 Volt output for 1 g input. On the other hand the IEPE accelerometer is powered by a constant current from the coupler where a min. current of 2 mA is recommended for general sensor operation. 

However,  since the power supply gives an output voltage, thats the reason I set an analog input vi for voltage in LabVIEW as can be seen in my attached vi, to set labview to expect signals from +/- 10Volts with a single ended accelerometer configuration. I am not sure what else shall I do with this accelerometer. 

[StathPol]

Hi users, 

I have tried the test with the ruller as suggested, but it did not give 1g. 

I set the full scale output to +/- 10 volts and the range to +/-10,000 g again. this IEPE sensor powers up by a constant current of 4 mA which comes from the power supply. The graph again was starting at - 5 g..very very weird...I set a tare function to zero the measurement and it worked. What I noticed is that by setting a high pass filter, the measurement approaches to zero in rest position. However, I dont understand why 1) the sensor does not starts itself at 0, and 2) why 1g is not diplayed by dropping the sensor in free fall. 

Regards. 

[billko]

Sorry to keep on this subject, but these measurements are taken with an O-scope?  Until you can validate the hardware setup, there's no chance to get the software correct.

[StathPol]

Thanks Bill.

I tested the hardware with 5 volts external source and LabVIEW picks the signal. Do you mean to validate the accelerometer readings with the O-scope while the accelerometer is connected with the card? Can you please provide me with any specific steps, just in case I am doing it wrong.  The way they are coonected by order as as follows: LabVIEW, PXI, NI 6361, Power supply, accelerometer.

Thank you.

StathPol

[billko]

---

@StathPol wrote:

Thanks Bill.

I tested the hardware with 5 volts external source and LabVIEW picks the signal. Do you mean to validate the accelerometer readings with the O-scope while the accelerometer is connected with the card? Can you please provide me with any specific steps, just in case I am doing it wrong.  The way they are coonected by order as as follows: LabVIEW, PXI, NI 6361, Power supply, accelerometer.

Thank you.

StathPol

---

Before we go any further, have you taken a look at this link?  http://www.ni.com/getting-started/set-up-hardware/data-acquisition/accelerometers

[Henrik_Volkers]

---

@StathPol wrote:

Hi users, 

I have tried the test with the ruller as suggested, but it did not give 1g. 

I set the full scale output to +/- 10 volts and the range to +/-10,000 g again. this IEPE sensor powers up by a constant current of 4 mA which comes from the power supply. The graph again was starting at - 5 g..very very weird...I set a tare function to zero the measurement and it worked. What I noticed is that by setting a high pass filter, the measurement approaches to zero in rest position. However, I dont understand why 1) the sensor does not starts itself at 0, and 2) why 1g is not diplayed by dropping the sensor in free fall. 

Regards. 

---

At what frequency did you run the rouler test?

With you settings 1g is 1mV ,  and some couplers (and you DAQ) have a sligth zero offset , keep that in mind 😉

Allow the setup to warm up 30min and to settle for 1min after a shock ... the high pass filter can influence the peak acceleration on 'long' shocks, but this is mostly relevant if you want to do calibrations... 

If you want to post some data, do it together with your vi , if you have the data in a graph hit edit- make current values default -   save and post the vi.

Another test: To avaid overloading of your sensor: Set the coupler for the first tests to an output range of +/- 20000g , in your posted graph the 12000g peak migth have been clipped by the DAQ (the +-10V range is actually slightly more) . You should run the sensor only up to 10000g 😉  and on concrete these values can quickly be reached.  

[StathPol]

Hey Henrik and Bill, 

Bill: I have used the voltage example; not the accelerometer one because the card is not DSA and not compatible with this type of setup.

I reconnected everything from the scratch. The accelerometers are in rest position fixed in mounting blocks as can be seen in the picture. In addition, the scaling is shown in MAX. Thanks for the suggestion Henrik of leaving the setup to warm up for 30 min. I have attached the vi, Acceleration resutls and data for your convenience. 

However, the results indicate that the acceleration value is around1g in rest position..However, this is not very clear as the results contain a lot of noise. We can do differently the scaling or even is there any way to proceed the results to avoid this noise ?

Many thanks for your replies. 

[billko]

I would imagine the accelerometer at rest to be 1g.  0g is free-fall.  🙂