TLC272 DIY active electrodes, bias connection saturate the output value

Abhay7417Abhay7417
edited April 2018 in Hardware
I have designed concentration measuring BCI system using openbci and 2 dry active electrodes (Fp1 and Fp2). Current working system uses dry active electrode made using TLC272 (Picture atttached) referred from


Openbci is used in dual supply mode to power dry active electrodes. Along with 2 dry active electrodes One bias dry electrode (FRI dry AgCl electrodes) on right earlobe and second dry electrode on left earlobe as reference. 

In order to decrease the form factor of the PCB, we used SMD components in the new PCB. It is a 2 layer board with copper pour on top layer which is connected to OUT Signal. The bottom layer is just a single pad which is connected to IN signal. We haven't added a 3rd layer in between the top and bottom to serve as a shield.  Raw values are getting saturated once I connect bias electrode on right earlobe. Without connecting to right earlobe it gives values with less amplitude than old setup. Request help to resolve the issue. Results are attached in  folder link below

1) OldSetup.txt - values are good and system working perfectly with all connections attached
2) NewPcbtest.txt - values with SMD version, initially without bias connection then with bias connection which saturates both channel values to 187500.00
3) PCB schematic design image

Comments

  • wjcroftwjcroft Mount Shasta, CA
    Abhay, hi.

    The question is, does this active electrode amplify the signal or not? Some do and some don't. If there is ANY amplification, then your reference lead as well as the channel lead will need an active electrode. Because EEG is a differential amplifier.

    Not sure you need Bias in the case of Actives.

    William

  • Abhay7417Abhay7417
    Thanks for reply william,

    Active dry electrode which we are using is unity gain buffer which does not amplify signal. Do you mean to say if I am using dry active electrodes I don't need bias electrode connected to right earlobe? My reference electrode is not an active electrode, will it create problem?
  • wjcroftwjcroft Mount Shasta, CA
    My understanding of the links you gave, referenced the design for the OpenEEG modular system. Which (I'm pretty sure) used actives for both channel and reference.

    Because of the way Ground / Bias interacts with the ADS1299 differential amps (centering them), the ground lead may not function as expected in your situation. Try a test with your actives on the channel and reference leads. If you only have two actives, that would be only one channel.

  • Abhay7417Abhay7417
    Thanks William,
    After using 3 active electrodes (2 Channels and 1 reference), saturation problem is solved. 
    1.  Even after exploring lot of links. I am not able to understand the role of Bias electrode. For now I am using dry electrode for Bias connection of OpenBCI board. Request you to tell exact working of Bias electrode and its application in EEG.

    2. For now dry active electrode PCB is having a 10mm exposed pad(HAL finish) coating which touch the skin and send input to TLC272. Could you please suggest an idea to add component with existing active PCB so that it works for haired skin too. (Kind of spike contacts)
  • wjcroftwjcroft Mount Shasta, CA
    Abhay, sounds great.

    Here is an old thread with a link to the TI tech note.


    In addition to the DRL function, the Bias / Ground also "centers" the differential inputs in some amplifiers. Not sure if that is the case with ADS1299.

    There is a Dritz snap fitting that can accept the back side of the FRI dry electrodes. But it's also rather large.


    Regards,

    William
  • Abhay7417Abhay7417
    Thanks William, In process of using all 8 channels on openbci board. I have connected 9 active electrodes(6 with spikes, 2 with non spikes, 1 for refrence at Fp2 location). Now while measuring data sometimes values are saturating for few or all channels. Even sometimes all channels doesn't saturate. I am not able to figure out the issue. Request your help to resolve the issue.
  • wjcroftwjcroft Mount Shasta, CA
    Didn't you have the saturation problem resolved earlier in your April 19 post with 2 channels? What has changed? Is it possible some kind of high-gain elements are acting in a feedback fashion, stimulating other active amps to saturate? You can adjust the channel gains, so as not to produce the 'Railed' label, and instead show you what patterns are happening, microvolts / millivolts wise.

    Maybe there is also some need for shielding, if the output leads of your actives are somehow also coupling back to the inputs.

    I think one possible advantage of actives that use greater than unity gain, is that they reduce the noise picked up by capacitive coupling effects / cable motion, that traditional non-active electrode leads suffer from.

    Regards,

  • Abhay7417Abhay7417
    Hi William,

    We have figured out that if the reference is Fp1 or Fp2, channels overhead are getting saturated. So we changed the electrode location like this:
    Ref - AFz
    Ch1 - C3
    Ch2 - C4
    Ch3 - Fc3
    Ch4 -  Fc4

    Now values start from some values in range 5000-15000 and decrease till it becomes constant. I am not able to figure out why it becomes constant after a few seconds. If any idea please suggest.



Sign In or Register to comment.

© Vanilla Keystone Theme 2026