From acceleration to FFT and to displacement - how to?

[RPJ]

Hey..., good question.

[johnsold]

There are several ways to remove DC components.

In the time domain: 1. Use a high pass filter.  The cutoff frequency must be lower than the frequency of any signal component you want to measure. Filters may add some delay or phase shift.

2. Take the mean of the dataset and subtract the mean.  This avoids most of the phase shift issues but can only be applied to blocks of data. It is not suitable for point by point calculations.

In the frequency domain: Set the DC component of the FFT to zero.  If some spectral leakage has occurred, several bins may need to be set to zero at the risk of losing some low frequency information.

Lynn

[rwbrook@integratedps.com]

I have tried filtering to the max and when I have my accels input into the SC2345 accelerometer modules, the output is bone ugly.  I am now considering just using the formula subVI and dividing by j omega.

[RPJ]

Good luck!

[rwbrook@integratedps.com]

Sir,

Again and again everyone on this thread refers to 'zero the DC component in the fft' without any direction on how to do it.  Does it involve changing the data from dynamic to the array, performing the DC removal on that bin and then changing the data back to dynamic?

Regards,

Ron