Ch 1-8 OK, Ch 9-16 Railing?

Dreaming_GeekDreaming_Geek Atlanta
edited December 2015 in Cyton
I was hoping that when I switched from a velcro headset to an Ultracortex3 that my railing issues would end.
Channels 9 through 16 are railed or near railed.
How might I fix this?

Comments

  • Dreaming_GeekDreaming_Geek Atlanta
    Are there other details I should give?
  • wjcroftwjcroft Mount Shasta, CA
    Just checking, do you have both SRB2 pins on the two boards connected with a Y-cable that then goes to your ear reference electrode?

    In many cases there will also be a second Y-cable allowing for "linked ear" reference. But that is optional. It is however used for many BCI / neurofeedback applications. Linked ear references give a more balanced reference that is not lateralized.

    Has the daisy board at any time ever worked correctly, or has it always railed? I see you had some comments on a previous thread,

    http://openbci.com/forum/index.php?p=/discussion/430/connecting-electrodes-with-the-fri-19-channel-headband

  • Dreaming_GeekDreaming_Geek Atlanta
    Oh, I thought I was supposed to connect one ear to SRB1 and the other to BIAS on the mainboard.
    This is wrong?

    Yes, the daisy board has always railed.
    I thought it was the crappy headset I had before.
    But, now I wonder.
  • wjcroftwjcroft Mount Shasta, CA
    SRB2 (the one closest to the main board) is the default reference pin. If you do not connect this manually with a Y cable to SRB2 on the daisy, the daisy board will be without a reference line. And hence all the daisy channels will go crazy.

    http://openbci.com/forum/index.php?p=/discussion/254/soldering-16-channel-openbci-kit

  • Dreaming_GeekDreaming_Geek Atlanta
    Thanks.
    Connected the other ear to the daisy, and now have all channels.
  • ntlanhntlanh
    edited December 2015

    I have nearly the same problem. At the beginning of running the board (of 8 channels only), all channels are railing.

    Could you explain what railing value means? I feel fatigue about its meaning.
  • Dreaming_GeekDreaming_Geek Atlanta
    Also, I am getting some wonkiness when I connect BIAS to an electrode on the top of my head.
  • wjcroftwjcroft Mount Shasta, CA
    The GUI shows "Railed" when microvolt magnitudes for the channels are off the top end of the scale. In other words, generally it means something wrong with the differential voltage measured between the channel and the reference. Either one could have a poor connection leading to "Railed".

    One way to check consistency is to connect all three leads together, SRB2, Bias, and the channel(s) you are trying to measure. You can do this with a glob of Ten20 paste or alligator clips. Once you have all leads connected, that should produce 0 microvolts, or close to it.

    To respond to @Dreaming_Geek, if the electrode chosen for Bias or it's cable / connectors are flaky, it will be an open circuit (no connection), hence no Bias really connected. Try swapping out with another lead.

    Also, @ntlanh note, when the board is first turned on, all channels Rail until the differential (Delta-SIgma) amplifier circuits settle down, this only takes a second or two. After that your microvolt values should drop into normal range. Provided you have good connections.

    William

  • henrykhenryk
    Hi All,

    Hopefully I did my homework before I posted my note here. I searched thru OpenBCI forum looking for answers. I found this post (or its topic) which is most related to my problem hence I use it. Unfortunately it does not help with finding solution to it.

    Setup:
    • 16 channel setup – main board + daisy module.
    • Connect Main_BIAS to Daisy_BIAS and then connect it to ear 1 (Y cable).
    • Connect Main_SRB2 to Daisy_SRB2 and then connect it to the war 2 (Y cable).
    • Fix latency to 1mS as per description here: http://docs.openbci.com/tutorials/10-OpenBCI_on_Windows
    • Try two different machines: one with Windows 10 and one with Windows XP. Both behave the same way.
    • Try two different software’s – Processing and OpenVIBE. Both behave the same way.
    Normal operation (8 channels):
    • When I use my hardware in 8 channel mode (with daisy module attached) everything seems to work fine. I used snapshots of OpenVIBE software to show what is going on. Below is a snapshot with 8-channel system. (Noise is coming from the fact the headset was laying on my desk). Three bottom channels represent accelerometer.

      image
    The display used in OpenVIBE was configured as follow:

    image


    Problem:
    • It starts when I configure OpenVIBE (or Processing) to work with 16 channels. In such scenario every few hundred mS I can see railing of all daisy module channels. Below are examples of such railing
    image

    or

    image


    Observations:

    • As one can see there railing appears to happen in some sort of periodic fashion.
    • It can be observed that accelerometer data or data from 8 channels on main board are not affected. That might indicate that Bluetooth link is OK.
    • I tried (using Processing) disable channels - on main board only and on both main board and daisy module. With both configurations I could still observe the same periodic railing.
    • Do You think that the problem here might come from daisy board hardware faults? Is there a simple way to prove that hardware is fine?
    • Or maybe PIC firmware messes up data from daisy module? I'm using code which I downloaded around December 2015. Were there any changes to t since then? Looking at github repository (https://github.com/OpenBCI/OpenBCI_32bit) it looks like code is stable.

    Any ideas here are welcome.


    Thanks,

    Henryk



  • wjcroftwjcroft Mount Shasta, CA
    edited May 2016
    Henry, hi.

    So you are saying the screens look similar when using OpenBCI_GUI vs. OpenViBE? OpenBCI_GUI is generally preferred for examples since the lab uses that to checkout hardware.

    Your connection of the two Bias leads with the Y-cable is not needed, and may be causing problems. Just use one Bias, for example the lower Bias pin on the mainboard.

    http://openbci.com/forum/index.php?p=/discussion/254/soldering-16-channel-openbci-kit

    William

  • henrykhenryk
    Hi William,

    Thanks for Your comment.
    To answer Your question, yes in general the OpenViBE looks and works
    the same as OpenBCI_GUI. The advantage of OpenViBE GUI is control over
    time line (can display more data).

    Here is what I have done.
    • I removed Y connection between BIAS.
    • I did few more tests using OpenBCI_GUI for Your reference. Results are below.

    Headset on my desk.

    All channels are ON.

    image


    Headset on my desk.
    All channels are OFF.
    I noticed that amplitude of the railing is smaller.
    I also noticed that periodicity of the railing is more random in such scenario.

    image


    Headset on my desk.
    Single channel on daisy module is ON.
    Railing amplitude slightly increased.

    image


    Headset on my desk.
    All channel on daisy are ON.
    Railing amplitude increased even more.

    image


    This time I put Headset on my head.
    I configured OpenBCI GUI to work with 8 channels only.
    I could notice that from time to time there is an amplitude jump in the main 8 channels.
    There is some issue with 3rd channel so for now I turned it off.

    image

    Again I put Headset on my head.
    This time I configured OpenBCI GUI to work with all 16 channels (except channel 3 which is OFF).
    I could notice that from time to time there is an amplitude jump in the main 8 channels.
    Railing on daisy module is definitely there. It is rather periodic.
    I can also notice small disturbance happening on all lines. This one is not a periodic one.

    image

    image


    Would this issue might have something to do with power supply or Xtal on main board?
    Any thoughts? Any experiemnts to try?

    Thanks,
    Henryk
  • wjcroftwjcroft Mount Shasta, CA
    Henry, thanks. I'm going to mention Joel @biomurph on your question. This is starting to look like some kind of glitch with your daisy module. You do say that with the daisy unplugged, the 1st 8 channels work fine.

    Has the daisy module ever worked successfully for you? That is when you first received the kit? It's possible there may be some kind of loose or shorting solder joint in the pins connecting the boards. It's important that the header pins be clean and gold plated where they plug into the sockets. If the pins have solder or flux on them, that can interfere with the reliability of the connector connection.

    William

  • wjcroftwjcroft Mount Shasta, CA
    Henry, Can you clarify that everything works fine (8 channels) when daisy is unplugged? I re-read your first post and I think all the screen shots you have shown so far had the daisy plugged in. But your OpenViBE or OpenBCI_GUI was either configured for 8 or 16 channels.

    What power supply are you using? Assuming you've tried a baseline test with the battery pack alone.

    When you "put the headset on my head", you always have all 16 electrodes connected? Have you ever tested with the electrodes shorted together as mentioned in the above post,


    In such a configuration you should get close to 0 uV on each channel.
  • biomurphbiomurph Brooklyn, NY
    Henry,

    This is not something that we have encountered before. 
    Can you try to run the setup in different locations? Could this be an environmental noise issue?
    wjcroft is right about the Bias. The Daisy is designed so that it's inputs are included in the main board bias, so you only need to connect to the main board bias. the Y cable on SRB2 is necessary. 

    One test you could do is to record the data to an SD card, and then compare that data to the GUI or VIBE recorded data to see if there is some difference.

    Please send a photo of the hardware setup, so we can see if there is anything glaring in that.
Sign In or Register to comment.

© Vanilla Keystone Theme 2024