LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Labview Integration Results Not Matching

Hello, 

    I am trying to create a program in LV 8.2 (full version only) that displays the acceleration, velocity, and displacement of an analog accelerometer. In order for this to be accomplished cascaded integrators working off of one signal must be used because displacement is a function of velocity and velocity is a function of acceleration.  I am using an analog accelerometer on an SCXI 1500 for my input.  I used a mathcad application to double check calculated results  (with no DC offset or filtering) and thought I would achieve the same results using Labview under perfect simulated conditions; however, I cannot.  I have tried many different avenues but nothing seems to give me the right results.  Can anyone shed some light on this situation, or am I just missing something simple??  I just started programming with Labview recently so anything else that seems funny feel free to point out  See attached

Regards,
Mike

0 Kudos
Message 1 of 4
(3,149 Views)
- not matching by how much?

- which one is right? (for a set of data corresponding to a known function)


0 Kudos
Message 2 of 4
(3,127 Views)
gwd,

    Usually my results are off by about 200 to 300 times my calculated result.  I have tried using a constant input for dt at the integration blocks. That seems to help get closer to the desired result, but I should be using the dt off of the waveform input.  If I understand your second question, the only function that has the correct results is the acceleration function.  This function doesn't have any math calculations in it just measurements.   The velocity function is an integration from acceleration and the displacement function is an integration from velocity.

Regards,
Mike
0 Kudos
Message 3 of 4
(3,117 Views)
200 X the result might not be bad...for example, if you were integrating an entire period of a sine wave.

I suspect either you are integrating different data OR you are using different algorithms. My (old) Labview has four choices of algorithm, all of which will give either slightly or greatly different results depending on the data. Another possibility is that the data is different. If Mathcad is getting the data from a text file, it may not be getting all digits of the data.

So if they disagree, by an amount that seems excessive, the next step is to integrate with both methods a set of test data where you know the answer.
0 Kudos
Message 4 of 4
(3,092 Views)