Ganglion high pass filter problem


According to some notions ( please see that for reference: ) there is a design flaw, that leads to 2.4Hz high pass filter and strong signal attenuation below 10 Hz. Actually it is not acceptable for correct EEG processing, since it has physiologically important spectrum from 1Hz and up.
So, what are perspectives to resolve this issue in the nearest future?




  • wjcroftwjcroft Mount Shasta, CA
    Mentioning @openbci @biomurph.

  • biomurphbiomurph Brooklyn, NY
    The issue that Alex brought up in his blog has been resolved.
    We adjusted the RC filter to bring the 3dB cutoff down to 0.3Hz.
  • Dear @wjcroft, thank you for bringing attention to the question and 2 @biomurph for prompt reply and great news!

    So, all current Ganglion stock in openbci store has only units with updated schematics, right?

    Also, Please correct according section of schematics description. It still contains old RC filter parameters.

    Thanks again,

  • Does this mean that Ganglion boards already purchased need to be modified if the flatter filter response is required?

  • @newdendrite depending on when the ganglion was purchased, yes, they need to be modified. you can use a digital multi meter to verify your hardware filters are correct if need be!
  • Will openBCI be performing upgrades or offering parts kits for the modifications?

    Best regards,
  • Hi,

    I'm wondering what came of this, seems some stuff didn't get answered.
    1.  What was the date when production units began to be built with the revision.  If there's no clear date, datecode, etc., then the best you have.
    2.  What was the change -- probably only component values? which components and to what values.
    3.  I looked at a few schematics including github-ganglion-hardware kicad and pdf, but I don't see a change.  Could be no schematic was ever generated. Anyway could that happen now, with the schematic including date & some revision designation that makes some sense, and then upload?
    4.  Would I be correct in answering newdendrite that openBCI won't be performing upgrades or offering parts kits.

    I developed my own change, but my Ganglion is shy and wants to be like everybody else.   :)

    Alex:  Thanks very much for finding this, good catch!

          -- Bruce P.

  • I'd like to offer some feedback to openBCI on this topic.  I have
    several Ganglion boards, as well as other products from openBCI, so I'm
    obviously an investor in the project with a stake in its success.  The
    projects's handling of this development is truly discouraging..

    Customers should be made aware that this is a serious problem with the Ganglion board.  The
    incorrect filter parameters mean that Ganglion output is not comparable
    to commercial or research EEG hardware, or to published research
    findings that use conventional EEG hardware.  It does
    not meet the "research-quality biosensing" claim on the openBCI home
    page.  The problems described fall squarely in a frequency band that is
    often of interest to users.

    My operational Ganglion board has
    been gathering dust since I read about the issue early in October. 
    After receiving no response to my question, I stopped using it and
    ceased making any recommendations or references to the Ganglion for the
    last two and half months.

    The least that openBCI should do is to
    publish a full disclosure of the issue with an updated schematic and a
    tutorial on how to make corrective modifications.  In other DIY
    communities, the vendor would have immediately shipped the necessary
    parts to owners free of charge.  I have seen this done often in both the
    amateur radio and electronic music DIY communities.  Since the Ganglion
    comes assembled, something also needs to be done for users not capable
    of performing surface mount modifications themselves.

    The answer lies in thinking about user needs and applying common sense about good customer service.

    Best regards,
  • @biomurph @pushtheworld Hi, I purchased a ganglion in Dec 2017. Has RC filter issue been resolved on this stock?
  • The terse "yes" seems like a sufficient answer to ccake's specific question.
    Plus, I think/hope there's an implication that all shipments purchased on or after Dec 1 2017 do/will have the fix.
    (By saying that, I don't mean to imply that no shipments purchased prior to that date had the fix).

    Is there anyone who is supposed to, and can and will, give knowledgeable answers to the other questions asked in this discussion thread.
    For example the 4 questions asked by bpwork -- me -- on December 15 2017.

    I may be misunderstanding.  This is called a Forum, not a Customer Service submittal-thing.  Perhaps that means it's a place where users can communicate with each other; and official openbci people ..... e.g. employees? .... may also pop in occasionally but without any responsibility to answer, or to take other action, soon-or-at-all?  (In case that sounds sarcastic, it's not).  
    If that can be clarified for me here, it would would help me .... but if I should instead ask that particular question in a more general Forum place like "General Discussion", and if that other place is monitored such that it would be seen and answered within let's say a week, I'm all ears.

    Thanks in advance.

       -- Bruce P.

  • biomurphbiomurph Brooklyn, NY
    The change made to the Ganglion BOM affects 4 capacitors in the Voltage Reference circuits for each input channel.
    The change was made in our production run on 3/15/17. The updated units entered our fulfillment warehouse in mid April of 2017. 

    The relevant schematics change is here
    On page 3. The capacitors that were changed are C9,C37,C39,C44. They were changed FROM 3.3uF TO 22uF. 
    The BOM also reflects this change. Our recommended part is Murata capacitor GRM188R60J226MEA0D. It has the same package size as the initial capacitor.

    I am going to put up a detailed discussion of the change required, and post a link to it on this thread.

Sign In or Register to comment.