LabVIEW
Dynamic Data Time Axes for Bandpass Filter
Solved!
Hi,
Does anybody know why my bandpass filter VI is not incorporating time to the waveform chart? Probing the filtered signal produces a voltage but all I'm seeing is are the values moving vertically up down. Also what's with the configuration of the time? I want to make it look like the waveform chart above. Thanks!
Hi od,
@od805 wrote:
Does anybody know why my bandpass filter VI is not incorporating time to the waveform chart?
How do you know the filter is working correctly after all?
You create an array of exactly one (1) element. Then you build a waveform from this one-element-array. And then you apply a filter on a waveform containing just one (1) sample: which result do you expect from your filter?
Suggestion 1: filters work much better when they are applied to an array of more than one sample…
Suggestion 2: Your VI will be much more maintainable once you get rid of ExpressVIs. There is a palette full of (plain) filter functions…
Hi Gerd,
Thanks for the feedback. Attached is my updated VI. I'm able to get some data, although idk how significant. Any tips on EEG signal processing? Probably need to implement 60 Hz notch filter and maybe a high pass to get rid of ECG waveform. I'm using this an EEG click coupled with an Arduino. Link below. Thanks!
https://www.mikroe.com/eeg-click
Hi od,
you still create an array consisting of exactly one sample: this does not look right to me…
Additionally it seems you use the output of one Bandpass filter function as input for the next bandpass filter: this also doesn't look right to me!
Usually I apply several bandpass filters on the original signal to get several filtered bands of the original signal…
Hi od,
@od805 wrote:
I’m not sure what else I can do besides an array with one sample. I am reading an analog pin of arduino to capture my data so naturally yes it is only one array. Do you have any suggestions?
Well, have you ever tried to create an array containing more than one sample?
Like using a loop to read 1000 samples before applying your filters on the resulting array?
@od805 wrote:
I purposely made that input design for the band pass filters because I think EEG signal extraction/decomposition builds off filtering from the previous signal in the broader frequency range.
I'm not an EEG expert but in my experience you don't apply bandpass filters on already bandpassed signals…
Attached is a photo of what I've done since to reproduce what you have recommended. But my this loop is not executable by my computer and there is something evidently wrong, I just don't know what.
Hi od,
@od805 wrote:
Attached is a photo of what I've done since to reproduce what you have recommended. But my this loop is not executable by my computer and there is something evidently wrong, I just don't know what.
Yes, there is something wrong…
In the inner loop you initialize an array of 10000 elements with the value of that single point measurement. Why?
Then you use a concatenating tunnel to create a larger array of 10000×10000 elements. Why?
(Is your computer capable of managing that array of 100M elements? Does it make sense to display those 100M points on a chart???)
Did you do any of those basic Training courses to learn simple array handling?
