Really bad noise only when I am connected.

AdamMiltonBarkerAdamMiltonBarker spain
edited April 2017 in Cyton
I have been testing some suggestions since not being able to update to V2 which is described in my last thread.

http://openbci.com/forum/index.php?p=/discussion/954/uploading-v1-v2-firmware-read-page-failed#latest

WJCroft suggested to try by  attaching 1 channel and pasting the bias and srb2 together with it. When I do this, I get no noise, it shows correctly when I touch my finger on the electrode and activity shows on the data stream but not the head, everything looks like the screenshots provided in the getting started guide. As soon as I connect my self with the same channel, the data goes absolutely crazy. Can anyone suggest anything to try please, screenshots as bellow, I have also tried the same with channel 2 and channel 7 with the same results:

Connecting channel 2 with bias and SRB2

image

With me connected:

image

This is a bit better than my original thought the device was broken, looks it is just me that is broken ;) Thanks in advance for any advice.

Comments

  • wjcroftwjcroft Mount Shasta, CA
    Your FFT plot shows a massive amount of 50 hertz mains noise. Is it possible that your environment with the computer etc., has substantial EMF (electromagnetic field) noise? 

    When doing EEG, it is always wise to position yourself away from any power cords, wall warts, power packs, extension cords. Try another physical location. Try operating your laptop on battery power. Get away from wall outlets or even power conduits that might be running under the floor or through the wall. Be sure that your skin is not touching or near any metal objects, like a metal chair, metal table, etc.

    Possibly even try a test on battery power outside your current building.
  • wjcroftwjcroft Mount Shasta, CA
    Dang, I just noticed your notch filter is set at 60 hz. Should be 50 hz for your mains freq.
  • AdamMiltonBarkerAdamMiltonBarker spain
    Hi thanks for the reply, it is possible yes, I have 3 pcs, a Mac, 1 laptop, 4 monitors, 3 3d printers, 2 WiFi routers, 4 mobile devices, an Ipad and quite a lot of development boards in the area of my desk. The reason I thought it was something else was because when it was not connected to me in exactly but in exactly the same location it is fine, as in the first image. It is running of my PC not laptop, the laptop is knackered at the moment, how far can the device be away from the dongle to still work?, I could try sitting on the opposite side of the room if it will reach that far. Based on this though, it appears that the hardware is not broken ? 
  • wjcroftwjcroft Mount Shasta, CA
    Let me REPEAT. It's your dang notch filter, it's set wrong. Press the button on the GUI to change it to 50 Hertz.
  • AdamMiltonBarkerAdamMiltonBarker spain
    Hi thanks that message had not come through when I replied. However I gave tried both 50 and 60 with no difference.
  • AdamMiltonBarkerAdamMiltonBarker spain
    I have moved to middle of the room in the sofa away from electrical stuff apart from the light above my head. Changed notch to 50 but still no difference. It does not detect teeth grind, blink or eyes shut either.
  • wjcroftwjcroft Mount Shasta, CA
    Leave it set on 50 hz. That should help somewhat. As far as EMF. You need to try a test outside your current environment. The RFduinos have quite a bit of range.

    With all that gear it seems like you should be able to borrow a friends laptop for a test outside your lab.

  • AdamMiltonBarkerAdamMiltonBarker spain
    I live in the middle of no where I have no friends lol
  • AdamMiltonBarkerAdamMiltonBarker spain
    edited February 2017
    If that is the case though then it's a pretty useless device for me.
  • AdamMiltonBarkerAdamMiltonBarker spain
    One last question why does it not do this when it is not connected to me.
  • wjcroftwjcroft Mount Shasta, CA
    With channel, reference, and bias all connected together -- the differential amplifier sees no voltage. With electrode cables connected up to your head, each wire can act like an antenna, picking up any mains interference in the area.

    It's a matter of degree. Most locations are fine. Some have excessive noise. What location were you in for your previous tests that worked last year? If it was the same as this one, then I'm stumped. What has changed since then? Sometimes fluorescent light ballasts or other power supplies can go bad, spewing out EMF in all directions. If you had an EMF meter there you could sniff it out see what is spiking the 50 hz noise.

    https://www.google.com/search?q=emf+meter

    https://www.google.com/search?q=emf+meter+app
  • AdamMiltonBarkerAdamMiltonBarker spain
    edited February 2017
    To be honest I was that excited with stuff last year it may of always been like this. But since then the equipment I own has more than doubled. Moving to another location is no good for me it needs to be near my desk is there no way to shield it like tinfoil or something ? Is the interference the reason why it's not picking up activity ?
  • AdamMiltonBarkerAdamMiltonBarker spain
    I am downloading a emf reader.
  • wjcroftwjcroft Mount Shasta, CA
    You can also try shutting down everything in your lab, except the desktop computer, one monitor. Really pay attention to the lights, compact fluorescent's are notorious for spewing out EMF. Or your overhead lights.

    Your noise could be coming from a conduit under your floor or wall, power bricks, etc.

    Cheap Windows laptops can be had for $200 these days.
  • AdamMiltonBarkerAdamMiltonBarker spain
    So where I am now is about 25 microteslas. The screens and stuff and my file server goes up to about 200 they are on the other side of the room. The issue I am saying is I need to use this at my desk it is no good for the purpose we bought it for if I have to sit outside to use it. What level is OK for emf ?
  • AdamMiltonBarkerAdamMiltonBarker spain
    It is mainly about 30 - 40 except for right up close against the computer screens and right down by the plug sockets
  • wjcroftwjcroft Mount Shasta, CA
    Calibration of EMF apps is not so reliable, because each phone uses a different type of magnetometer. I don't know what to suggest at this point. Other than to try outside your lab in an electrically quiet area. Use battery power.



  • AdamMiltonBarkerAdamMiltonBarker spain
    I understand that I can get a laptop but that does not solve the issue. If it cannot be used in the office it useless to me. Any way of shielding it at all ?
  • wjcroftwjcroft Mount Shasta, CA
    Mains EMF cannot be easily shielded for EEG purposes. You need a quiet environment for EEG. Though it doesnt have to be 'perfect'. If your OpenBCI worked in the tutorial tests last year, then something has changed. It's remotely possible the mainboard has developed some failure mode, but I havent seen something like your situation previously. When the OpenBCI lab comes back Monday AM, I hope they will respond to your emails to contact at openbci.com . With more suggestions.

    Another test to reduce noise level is to put the mainboard / battery in a shirt pocket, then use a clothespin or similar clamp to route the cables over your shoulder and onto your head. This reduces cable motion, which can also cause artifacts.

    Cable length makes a difference as well, the shorter the better. If you look at the Ultracortex photos, the cables there are each less than one foot and go directly back to the mainboard. This cuts noise.

  • AdamMiltonBarkerAdamMiltonBarker spain
    edited February 2017
    Hi thanks like I said it is highly possible I did not do things correctly last year and did not really spend much time on it, this issue could of been there last year, it is more that likely that case, I did not have a lot of time to mess with it. I will try your suggestions. After all the hassle and costs we went through to get the kit here, it is pretty annoying to be in this situation now, sitting outside to develop our prototype is not an option. 
  • AdamMiltonBarkerAdamMiltonBarker spain
    edited February 2017
    I have shut down everything in the room and it has made no difference what so ever. Will wait for openbci to reply these types of things should made clear on the product page as ultimately close to 2000 has been spent on equipment that seems useless for our use case
  • AdamMiltonBarkerAdamMiltonBarker spain
    Thank you for the help by the way was not meaning to sound rude I appreciate your help this weekend.
  • AdamMiltonBarkerAdamMiltonBarker spain
    Came back to this last night, I have tried it a room with absolutely no electronics in my new home and it is exactly the same, guess the board is broken.
  • AdamMiltonBarkerAdamMiltonBarker spain
    I am thinking could humidity / water damage do this, my old place seemed to get a lot corrosion on devices being close to the sea.
  • wjcroftwjcroft Mount Shasta, CA
    Your first post had two Image links, that referenced files on your website,


    But these are now gone. Can you reload those files?

    Is the situation in the new location now the same: you get tons of 50 hz noise when electrodes on the head. But close to zero microvolts when you short together channel, SRB2 and Bias?

    Did the contact at OpenBCI.com ever get back to you with further suggestions? 

    William

Sign In or Register to comment.

© Vanilla Keystone Theme 2026