very high DC offset, maxes out ADC

wjcroftwjcroft Mount Shasta, CA
This discussion was created from comments split from: large millivolt data values / FbEEG Full Band EEG.

Comments

  • Hi! during long sessions my DC slowly drifting out of acceptable signal range and sticks to it (187500 microVolts). Can i do something with that?
    i use cython board, data looks like that:
    image
  • I mean theres no signal, when DC drift out of range 187500. All datapoints there become 187500,187500,187500,187500....etc till it drifts down, below 187500
    see pic, there are complete flat periods of signal
    image
  • Thanks for patience William! I read all thread before posting and i use DSP to build bandpass filters. But i dont understand how filtering will help me, as i get this flat line in the raw data from GUI, so i get same flat line after any filter, as there is no signal at all - is it correct? And as it's hardware limitation the only way to deal with it is to reduce channel gain factor - can you please point out how to do that? 
  • wjcroftwjcroft Mount Shasta, CA
    Poddubny, my misunderstanding.  If you are getting very high DC offset potentials in the raw data (maxing out the ADC range), at the default gain of x24, that means that your electrodes are causing this issue. Or something to do with your skin preparation, paste / gel, electrode metals being used, etc. Normally at the default x24 gain, the DC offset never goes off the top of the ADC range like this. Try using plain gold cups with Ten20 paste. Also check that the subject is not using some kind of unusual hair conditioner, lotion, or shampoo. Since this is happening with all your channels, I would suspect the reference electrode.

    You can as I said before, drop the gain to see if the offset will stay in range. But the offset should never get this big in the first place.

    I'm going to try to split this thread off into another topic, since it is really an electrode issue and not about software or FbEEG. Wait until I setup that new thread before you add any further comments. You should receive a notification after I've created the new thread.

    Regards,

    William

  • wjcroftwjcroft Mount Shasta, CA
    @poddubny, hi.

    OK, the new thread is here. Let me know if you can get rid of the huge DC offset by experimenting with your electrode or scalp setup.

    Regards,

    William

  • Thanks for link, it helped!
  • wjcroftwjcroft Mount Shasta, CA
    edited August 2018
    You should be able to fix the high offset (eventually), so that you can just use the default x24 gain factor. After you figure out what is unusual about your electrode situation, that is causing the high offsets. Since they are on all channels, suspect your reference lead.
  • wjcroftwjcroft Mount Shasta, CA
    Also make sure your Bias/Ground connection is solid. This centers the differential amplifiers.
  • wjcroftwjcroft Mount Shasta, CA
    Yet one other possible cause could be MIXING electrodes using different types of metals, this will cause a battery / galvanic effect along with the electrolyte. OR, if you are using plated electrodes (such as gold, which is plated on tin or silver: and the plating has worn or chipped off, exposing the other metal type and generating galvanic effects.
  • actually i use standard electrodes from MARK IV https://shop.openbci.com/products/ultracortex-mark-iv
  • wjcroftwjcroft Mount Shasta, CA
    Did you try the test with the plain gold cups and paste? Did you examine your reference electrode / lead / connectors, perhaps swapping with the other ear clip? Did you try some ohmmeter tests? Etc.
  • No, as i dont have these at the moment :(

    Can you please help me with command to decrease gain from 24->12 via python interface? I send command 'x1120000X' to switch gain on 1st electrode, but it shuts off. 

    And mb there are some docs about details on this command, moreover than mentioned on https://github.com/OpenBCI/OpenBCI_Python

    For example, writing

    -->x3020000X will do the following:

    ‘x’ enters Channel Settings mode. Channel 3 is set up to be powered up, with gain of 2, normal input, removed from BIAS generation, removed from SRB2, removed from SRB1. The final ‘X’ latches the settings to the ADS1299 channel settings register.

  • wjcroftwjcroft Mount Shasta, CA
    Your command, 'x1120000X', is wrong, you are turning the channel off. You want instead, 'x1020000X'. Note the inverted sense of the on/off channel digit.

    Also be sure that you send all characters quickly before the timeout.

  • x1020000X sets gain of 2, but i need gain of 12..
  • wjcroftwjcroft Mount Shasta, CA
    Gain 12 for channel 1 would be '5' in the gain digit position: 'x1050000X'


    The gain control is just a single digit, in the 3rd position.

    Regards,

    William
Sign In or Register to comment.