Bullet with mounting accelerometer hits the load cell and does not provide the same readings

[StathPol]

Please find attached the graph. I connected a scope with the output from the signal conditioner of the accelerometer and it provides the same result with labview. 

The impact force is the peak acceleration multiplied by the mass of the bullet. The bullet is hemispherical with round nose yes. The load cell has a domed surface. 

Y axis top graph  - Newton

Y axis bottom graph  - g (Acceleration)

[Alexander_Sobolev]

I would say maximum impact force should not be constant. Impact energy (= bullet mass x drop height x g) should be. Hemispherical bullet can rotate and peak force will change depending whether impact force is along the bullet center of mass or not. With spherical bullet it always is (if hitting the center of cell dome) and trial results will be more consistent. 

To calculate energy you need to integrate the signal - I would say first peak of the force graph.

Also

Force calculations look like a running average of acceleration. It changes the shape of the signal and distorts maximum value - if you are interested in force maximum and shape, integration should be  shorter.

[Alexander_Sobolev]

>> Force calculations look like a running average of acceleration. It changes the shape of the signal and distorts maximum value - if you are interested in force maximum and shape, integration should be  shorter.

To correct myself, it looks like frequency window filtered - no offset, not high frequencies. 

>> Take a look at the power spectrum of your accelerometer signal .... compare it with the spec of your load cell 

a little earlier in the thread, Henrik_Volkers

Processes happening should be covered by accelerometer (looks ok) and not overlap with cell resonants, otherwise it is hard to get good results.

[StathPol]

The impulse from the drop test is calculated by integrating the area under the force versus time curve.

basically, I am interested for the accelerometer readings. The gain automatically changes on the signal conditioner once the user change the range (up tp +/- 10,000 g) and the sensitivity (0.000516 V/g) of the accelerometer. Does this <reading> of the gain which in this case is 1.97 needs to be taken into account?

[Henrik_Volkers]

So, please give some more information on your sensors. 

accelerometer is ICP .... type?    Bandwith of signal conditioner? type?

loadcell? type? bandwidth? ... signal conditioning?

Your mass seems to have a resonace at ~4kHz and your loadcell at ~2kHz ?

If you put both on a shaker your simple peak approach migth work up to maybe 800Hz ,   personal guess ...

here is a paper on model based force transducer shock calibration from a colleague

[StathPol]

Accelerometer certificate

1 axis (vertical)

Single ended

Range = +/- 10,000 g

Sensitivity =0.532 mV/g = 0.000532 V/g

Resonant frequency = 100 kHz

Bias Voltage = 10,8 V

Signal Conditioner certificate

Load cell is just a full bridge load cell with 788 ohm input resistance

[StathPol]

Anyone?