Unexpected noise with Daisy

Hello!

Today I started trying the Daisy module for 16 channels.
The soldering went fine thanks to the tips here in the forum.

After assembly, I could check that the channels on board and daisy were working fine, with the expected drop in "actual sampling frequency" to half.
So the Nyquist Frequency (NF) would be 125/2=62.5.
However, I could verify that besides the expected 50 Hz and 75 Hz peaks [62.5 + (62.5-50) = 75], I also have the 25Hz peak you can see in the figure below. What would be the reason for this? I wasn't expecting to have "Subharmonic noise" being the orignal noise below the NF...

Tests with ECG signal

Thanks

Comments

  • wjcroftwjcroft Mount Shasta, CA
    How did you get TWO notch filters. Did you add 2nd filter code to the GUI? What does the 16 channel signal look like with the stock GUI and only one notch filter at 50 hz?

    It's possible that this is some side effect of the DSP filtering. Does the 25 hz peak change when you try various combinations of turning off/on the various bandpass / notch filter combinations?

    Have you tried looking at your EEG or ECG signals with other interface programs such as neuromore, OpenViBE, etc.?

    https://www.google.com/search?q=subharmonic+eeg

    Your Dropbox images above have TONS more mains noise than those shown on the docs.openbci.com tutorial image (below). Why would that be? Where are your leads positioned, Ground / Bias, etc. Do you have any power supplies or AC cables running nearby? Try changing the location of your experiment. Sometimes AC conduits run through certain floor or wall areas.


    image


    http://docs.openbci.com/tutorials/01-GettingStarted#getting-started-w-openbci-v-connect-yourself-to-openbci-4-launch-the-gui-and-adjust-your-channel-settings

    http://docs.openbci.com/tutorials/01-GettingStarted#getting-started-w-openbci-v-connect-yourself-to-openbci-5-minimizing-noise

    William


  • edited December 2016
    Hi @wjcroft

    Thanks for the reply.
    Yes I added the second notch in the GUI.

    The signals I showed have much more noise than the ones in the tutorial image but I wasn't very concerned about that because in the first test I didn't even use the condutive gel. I assumed that much of the noise is also a consequence of that.

    Today I did the tests again (now with conductive gel), with one lead on my right ankle and the ground and ref on my left ankle. The peak at 25 Hz is not appearing today, as you can see in the image below. I don't know if it was some noise source that is not active today or something like that...
    But the 75Hz component is clearly removed with a notch.


    From this image, I conclude that it should be recommended to add a 75 Hz notch filter together with the 50 Hz notch when using daisy. Probably also a 65 Hz notch for the 60 Hz case.

    [EDIT]

    As I disconnected the leads and they were "on air", the 25 Hz peak appeared again, so I believe that this noise is avoided by correctly connecting the leads...

Sign In or Register to comment.