LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

Highpass filtering

Hi,

 

I'm having an issue in Labview (I'm using 8.2) with being able to apply a filter to some data, so I have recorded a 2D array of double (64-bit digit) where one column represents a time in ms, and the second a standard deviation of position, I wish to apply a highpass filter to this so that motions with a frequency of >10Hz are removed from the data, unfortunately I have not had a lot of luck doing this, the filters either seem to require a dynamic data  input or alternatively a 1D array with a set value for dt, due to the slightly variable sample time this isn't sufficient for what I want to do... would be very greatful if anyone has any ideas.

 

Cheers

 

RP

0 Kudos
Message 1 of 8
(4,763 Views)

RP,

 

Your terminology is unclear.  A highpass filter pass high frequency components and removes low frequency components.  You indicate that you want to remove frequencies greater than 10 Hz.  That sounds like a low pass filter.

 

What do you mean by "motions with a frequency of >10Hz?"  Is the motion periodic like a point on the circumference of a wheel or like a pendulum?  Are you referring to velocity?

 

If your data contains time and  standard deviation of position, how do you expect to get any information about the motion or its "frequency?" I think you will need the actual position information.

 

What is the source of the data?  What is the physical meaning of what you are trying to measure or calculate?

 

Lynn

0 Kudos
Message 2 of 8
(4,749 Views)

Hi sorry I meant motions with frequency <10 Hz.

 

the motion is of a particle trapped in a roughly harmonic oscillator, using video microscopy the centre of mass position of the particle is being tracked, it has been difficult to isolate the system from mechanical noise which occurs in the 1-10 Hz range typically. so the particle position will be plotted against time at a rate determined by the frame rate of the camera (around 1800 fps). I hope that helps 🙂

0 Kudos
Message 3 of 8
(4,737 Views)

Hey RP,

 

Can we get at look at the code so far, or alternatively perhaps a copy of data results so we are able to show you how to implement filtering in labview.

Regards

Andrew George @ NI UK
0 Kudos
Message 4 of 8
(4,735 Views)

Hi Andy,

 

Here's the relevant part of the code and one of my data files - any suggestions would be greatly appreciated, if you open the file the x/y graph will show you the signal i want to filter.

 

Cheers

 

RP

Download All
0 Kudos
Message 5 of 8
(4,732 Views)

Hi RP,

 

Looking at your data it is all in the same frequency range, i.e. all the data you have collected is at 2K Hz (difference between each one of your time signals). Therefore there should be no frequencies below the 10Hz range and thus a high pass filter would be useless.

Regards

Andrew George @ NI UK
0 Kudos
Message 6 of 8
(4,728 Views)

The data is collected at around 2 kHz, but in terms of the data there are correlated motions of the particle over all frequencies from 1-500 Hz (the knee frequency of the lorentzian) as can be seen from a typical power spectrum (See attached) the motions at low Hz will still be recorded by a 2 kHz sample time (although anything over 1000 won't be), what I need to do is filter out correlated motions that occur at 1-10 Hz. Sorry If I'm not making sense

0 Kudos
Message 7 of 8
(4,719 Views)

Hi RP,

 

In order for you to implement the filter in LabVIEW you will need to have the data you have collected  in the frequency domain. Unfortunately the data you have provided is all separated by the same time interval meaning that when I perform a Fourier transform on the data all the results are within the same frequency, as a result all of the data points reside in the same region. If you have the raw data in the frequency domain then we will be able to filter it.

Regards

Andrew George @ NI UK
0 Kudos
Message 8 of 8
(4,697 Views)