Noise data Filtering

[Munna232]

Hello all,

Please help me if anyone know about this.

I want to filter the Noise data based on Cutoff freq. after Filter I want to validate if data is properly filtered/not.

In my sample VI, I’m using Low pass RC Filter and LabVIEW filter Logic. But both results are not matching.

Anyone have idea to judge the Filter data is correct/not.

BR,

Munna

[1984]

What you mean on "correct"? The express VI will always do the same operation on your samples, so if you set the filter properly, your output will always be filtered accordingly.

[Munna232]

Yes .. Express VIs will gives correct values.

I want to validate same filter data with someother way (with formaula).

as per my VI i'm appalying Filtering in both the ways but Results are not same.

[1984]

Well, if the express VI gives you the correct values then the math you applied is incorrect. I have checked the equation, you realized what you wrote to the textbox so most likely your equation is wrong. I am not an expert in filters at all, but by instinct I would say the equation for a butterworth filter should look more complex.

Why dont you just dig down in the express VI source code to see where are the differences? You can right click on the express VI and select "open front panel" and then open it as a normal VI. Deep down you will find the math I guess.

(also I have found this page which looks easy to follow. http://www.electronics-tutorials.ws/filter/filter_8.html)

[Munna232]

Just I used 1st Order RC Filtering. Equation is correct.

http://www.pdx.edu/nanogroup/sites/www.pdx.edu.nanogroup/files/2014_Labview%20example_0.pdf

But I have to Check Express VI "Butterworth" logic.

[1984]

That is applied to a "clean" sine wave with one given frequency. Your signal has multiple frequency components so you can not apply one as you do currently. Dig down in the express VI, you will find the solution.

[Munna232]

Yes correct, But I want to validate those Results Because I dont have confidence to say my filtered data is correct.

[1984]

I am not sure about this... the filtering express VIs are correct. The math they apply can be complex, but they do what they need to do. So you can duplicate the math, but it makes no sense. If you dont trust it then you can start questioning all signal processing VIs.

Its up to you though, I dont think I can add anything else. 

Good luck!

[johnsold]

Munna,

The value you use in the filter formula only applies at one frequency, 1 MHz, because that is the value you entered into the formula. That appears to be the sampling frequency, not the frequency of any signal component.  For the formula to work you need to apply the value of the transfer function at EACH frequency to the magnitude of the signal component AT THAT frequency.

Your signal has a DC component which is ~30 dB larger than any AC component. That will create problems for the filter. The transient response of the filter probably lasts 5-20 ms, which is  far longer than your data set. When I subtract the mean before any other processing, things look more like what you might expect. It appears that your data has very little enegy below 30 kHz so I would not expect a lowpass filter at 100 Hz or 169 Hz to do much.

The image below shows the results after removing the mean. I also plotted the spectrum of the signal.

I do not see very much noise and I have no idea what you consider to be the signal in that data. Please describe in more detail what information you are trying to extract from the data.

Lynn

[Munna232]

Thank you so mcuh for your help.

In my project, I’m Injecting External Noise to DC Signal. Here External noise Generating by Two-way Switch.

This Noise Data with DC Signal displaying on DL850Scope. From scope I'm Reading this data (1000 samples).

So, Now I want to filter noise (based on Cutoff freq). After that I want to validate Filtered Data with Formula.

I hope this is clear for you.

Can you send VI of graphs which you posted.