Accelerometer signal (+/- 10,000 g) help

[Henrik_Volkers]

Well some sort of verification... I don't want to call it calibretion 😉

For 10kg aehh...  10 000g a Hopkinsonbar would be the way to go, however I was thinking of a classical 1 g calibration:

place the sensot together with a small heavy non magentic part on a vibrating surface.  increase the amplitude, at the point where the acceleration is greater 1g the item will lift off and that can be seen on the (1mV) output (THD if sine exitation) ....

[StathPol]

Thanks for the suggestion Henrik. I will have a go tomorrow (Despite the fact that my background is  civil engineering :P) and get back to you. 

Thanks again, 

Pepis21

[StathPol]

https://answers.yahoo.com/question/index?qid=20081213171236AAo9v5I

It makes sense to verify the calibration and settings. A simple test is to record the value as you have been doing, and drop the operating transducer in free fall, maybe about a meter, catching it gently at the end of the fall. I expect to see a step when it is dropped that is 1G, close enough. Ignore the polarity. There will be a reverse (bigger) step when it stops as you catch it. If it is a transducer to measure very high G acceleration you may have to get more inventive. You can get higher accelerations by dropping onto a ballon etc., so long as it is all kept square (acceleration is a vector).

You can drop it a known distance, calculate the speed using time which is in your file:

V = V0 + (a x t)

Where:
V is final velocity
V0 is initial velocity = 0m/s
a = 1G = 9.8m/s^2
t = time in seconds
So if it takes 0.1s speed is 0.98m/s (and it falls 0.98m)

Now it is moving at 0.98m/s, stop it in a shorter distance (balloon, inner tube, etc).

acceleration = Vinitial / t
= 0.98m/s/.01s
= 98m/s^s

[Henrik_Volkers]

Well, I use laser doppler vibrometers to calibrate accelerometers 😄 and have a hopkinson bar for accelerations up to 100 000 m/s² ...

Drop test:

The tricky part of the drop test is to measure the residual velocity (bouncing).. the stop time could be estinated by the sensorsignal itself and the integral of your acceleration signal should match the velocity difference. that formular assumes a constant acceleration not only for the free fall acceleration but also for the stop acceleration...  and if the 'break' periode is to long the high-pass of your signal chain will give you an additional error.......

and as a homework: What is the free fall distance after 0.1s?  

Lookup in a physiks book or at wikipedia free fall

[StathPol]

Hi Henrik_Volkers, 

The problem is not with the calibration as the accelerometers are already calirbrated. The problem is with the correct scale of the units. Each 1 volt corresponds to 1000 g. I have posted the complete question in the labview forum for an impact test. I have tried one test, but the accelerometer gives very unusual results. Y-axis is g force and x-axis is time in seconds. This was the outcome with only a single touch. 

 

[Henrik_Volkers]

How I interpret your signal:

@~5.805 release of  the sensor: signal before was 1g steady, but due to the high pass it settled to zero. the signal at free fall is now -1g

@-5.815 sensor bounces to surface, you see some ringing, either from the mechanics of the mounting or of the seimic mass in the sensor, after that again free fall

@~ 5.825 second touch down

...

looks usual, just like what I would expect .... 😄

What is the question?