UltraCortex MK4 - Unreliable readings

knightv8knightv8
in Headware
Hi,
I was using a combination of electro-cap and Cyton 32-bit board and everything worked fine; I have also published some papers using this combo.

I decided to upgrade to a UC MK4 headset, however the readings I am getting are unreliable to say the least. Let me explain my setup:
Currently using only 8 channels (even though I have the daisy module for 16 channels if needed). So to keep it simple let's eliminate the daisy module and keep the basic setup of 8 channel.
I have C3, CZ, C4, P3, Pz, P4, O1, O2 as channels from 1-8. I have the BIAS on my left ear (or FPz, tried both) and the ground on my right ear.

I switch on the GUI, and wear the cap and start to tighten the electrodes, until I feel them gently pressing, however most readings where either RAILED, NEAR RAILED or with a high UV (14+) or fluctuating(9-23uV, which is definitely unreliable). With the electro cap I usually had readings ranging from 7uV-14uV, depending on the electrode position and activity at that time, but when idle they where consistent.

Another thing is that when I try to measure electrode impedance from the GUI, on some channels (not on all channels and not always on the same channel, expect channel 1 which is always consistent) it shows a consistent spike in the 31-33Hz range, sort of the 50/60Hz hum you have for electrical interference. Is this normal? When I switch the check impedance off on that particular channel, the 31-33Hz spike goes off.

Honestly I am at a standstill. I have tried tightening the electrodes, but it makes the signal worst. Sometimes I may get the O1 OR O2 to a decent signal, by leaving it running for a few minutes without doing anything.

Any ideas?

Thanks
Patrick

Comments

  • wjcroftwjcroft Mount Shasta, CA
    Patrick, hi.

    If you are getting noise on ALL channels, it usually means you have problems with your Reference (SRB2) electrode. Because this is the lead that is in common with all channels. Can you try swapping it with the Bias lead, checking the ohms of the ear clip leads with a digital ohmmeter? They should be very close to zero. 

    If the lead connections are intermittent (bad wire, connector, or screw down connection), this will also cause noise. Again use your meter and wiggle wires while measuring.

    William

  • knightv8knightv8
    Hey William,
    nice to speak to you again. Thanks for your prompt replies as always.

    I have just tried swapping SRB2 with the BIAS, same thing. I have tried changing the wires, but same result.

    Let me show you with some screenshots:

    Screenshot 1 - Shows only O1 and O2 active. Fluctuating Ohms and Uv.

    image

    Screenshot 2 - When I enable Channel 1 (C3) to detect the electrode impedance, you can see a spike in the 31-33Hz range on the FFT. When I disable the electrode impedance from the GUI, this spike goes away.

    image

    Screenshot 3 - All channels enabled... some railed and others don't make sense.
    image

    Might be something stupid that I'm not thinking of. But sometimes when you spend a lot of time on a problem, you need another set of eyes / ideas.

    Could you kindly guide me how to test the Ohms with digital ohmeter ? I have a digital multimeter that can detect Ohms, is that good? Where to I place to positive and negative leads?

    Thanks for all your help
    Patrick
  • wjcroftwjcroft Mount Shasta, CA
    https://www.google.com/search?q=how+to+use+an+ohmmeter

    You basically want to check for a (close to) zero ohms connection between the electrode and the connector (female) the plugs onto the OpenBCI mainboard header. As you are checking the ohms, wiggle various parts of the wire, connector, electrode, to be sure that the ohms remain zero. If the readings fluctuation with movement, then you need to resolve what is breaking the connection. This is done by 'fixing' the broken connection with tightening screws, resoldering, getting a new lead, etc.

  • knightv8knightv8
    Hi William, 
    I have tested 3 of the electrodes and showing a stable 01.7 on each one with (200 Ohm setting). .With a 2K Ohm setting it shows a stable .001.

    Is that within acceptable limits? Electronics aren't my stronghold.

    If these are within range, any more ideas ?

    Thanks for your time.

    Patrick
  • wjcroftwjcroft Mount Shasta, CA
    Did you check the ear clip leads? DId you wiggle the junction between the wire ends and the screw connector inside the electrode holders? Etc. You must have loose connections someplace. Could even be on your mainboard header I guess, but the header is very solidly soldered on there. And your mainboard worked fine with the touchproof converter and the Electro-cap.

    How did you get the UC headset? Was it prebuilt, did you build it, etc.

  • knightv8knightv8
    Ear clip same as the other electrodes, stable at 01.7 (with 200 Ohm setting). Just for my future reference are these values within acceptable range ?

    Yes I did. However it would be very unlucky to have eight electrodes that need wiggling/screw tightening etc? since i cannot manage to get a clear reading on all 8 electrodes. Or would just 1 electrode mess up the others ?

    Yes it worked perfectly with the electro-cap, in-fact I published 2 papers using that combo. The only difference is that I soldered the headers for the daisy module lately. Can that affect it? Tomorrow I have a very busy day lecturing, however I will make it a point to re-retry the electro-cap on Friday, to re-check its still working fine.

    I bought it un-assembled and built it myself. Attached 2 images of my setup. Don't mind the cable management, been unplugging and re-plugging them for a while now, trying different things.

    image

    image

    Thanks
    Patrick
  • wjcroftwjcroft Mount Shasta, CA
    Yes, you are getting roughly zero ohms with the .001 reading. Not sure why the other reading is 1.7 ohms, but that's also essentially zero. What you are looking for are readings that vary with jiggling the elements of the wire, connectors, screws, etc.

    Is it possible that your wire connections at your screw terminals are inconsistent? Your O1 O2 look fine. Why are the others so noisy? Could there be issues with the various male-female connectors that are in series with the electrode wire?

    I suggest emailing contact at openbci.com, and asking some of these questions. Post back here if they have other suggestions that helped.

    Regards, William

  • knightv8knightv8
    Thanks re:ohms. Understood will triple check on Friday.

    The electrodes come pre-assembled, you just screw them on the headset. I will tighten all the electrodes screws just in case. The O1 and O2 are inconsistent (check 1st screenshot). Sometimes they are around 8-12uV as required and sometimes they spike to 30ishUV. At first they always start 30ish uV and it goes down after a few minutes idle. Also O1 captures the Alpha waves and O2 doesn't and sometimes its vice versa; as I said inconsistent. Moreover why does the 31-33Hz spike when I check the electrode impedance of Channel 1? Tried changing electrode and cable, same result.

    Re: noise, yeah it doesn't make sense. Even more why are others RAILED as well ? All wires are from OpenBCI, nothing is bought separately. I also used wires from the 3 bunches (short/medium and long), so I doubt all are faulty. I also connected an electrode directly with the board and placed it on my forehead for blinking and same inconsistent result, so I doubt its the male-female connectors.

    OK will email them and keep you posted on any updates.

    Thanks for your time.

    Patrick
  • knightv8knightv8
    Hi William,
    I attached the electro-cap today and this is what I get, as expected, a clear signal.on all channels.

    Contact email haven't replied to me yet. If you have any more ideas in the meantime, please do let me know. I will update if they reply with a positive solution.

    image

    Thanks
    Patrick
  • wjcroftwjcroft Mount Shasta, CA
    I'm going to include @Conor and Joel @biomurph here, because we need some more suggestions. 

    Guys, Patrick is getting solid Cyton signals with Electro-cap, but wildly intermittent channels with the headset (8 channels). We've already tried ohmmeter checks of his wires / connectors. Tightening down the bolts, etc. Could he have marginal FRI combs? 

    Patrick are you sure you are tightening the electrode holders (spiky & flat units) down sufficiently onto the scalp? There will be some degree of "discomfort", they won't just be lightly grazing the skin.

  • knightv8knightv8
    Hey William,
    thanks for including them :)

    Yes I tried light, medium and hard tightening on the scalp. In fact when I tight them (when they start hurting quite a bit), the signal goes railed. I loose the tightening a bit and sometimes the inconsistent signal comes back up.

    After a few minutes of being idle O1 (ch 7) and O2 (ch 8) become a decent signal (without doing anything). At one point I manage to get signals on all 8 channels (inconsistent but not railed). Removed the cap and re-put it that same instant, and half the channels went railed again. I haven't managed to get any particular pattern of the errors. The only stable error that occurs all the time, is that if I enable the electrode impedance on channel 1, it always shows a spike in the 31-33Hz on the FFT (image shown in one of my attached images).

    I have been at it for 2 days, I tried all sorts of things and am at a complete loss. I have the R&D kit, so swapped electrodes, wiring... everything.

    Thanks
    Patrick
  • ConorConor
    edited July 2017
    Hi Patrick,

    Sorry for the issue, and sorry for the delay in my response. Can you take a look at your Ultracortex electrodes. If you see any discoloration at the tips of the spikes, it is likely an issue with the electrodes themselves. We do our best to screen for this manufacturing defect but it does happen from time to time. The discoloration will appear as a white-ish color. If you see this, it means that the Ag-AgCl coating has worn off or was not applied thoroughly enough to the ABS substrate. 

    If this is the case, we would be happy to send you replacement electrodes. If you want, take a photo of the spikey tips and add it to this thread. That could be helpful for others experiencing a similar issue.

    Best
    -Conor
  • ThreeFormThreeForm NYC
    Hello Patrick,
    I am also curious about the results you're seeing. I know you tried changing various things, but can you confirm that you switched electrodes/wires from a working channel into a non-working channel? Since I don't see that you stated that specifically, that result might help to narrow things down a bit.

    Also, I since designed the geometry of the headset itself, so I am mostly familiar with that aspect, and would like as much information as possible about the context. I have tried the headset on many people in public demonstrations and the biggest variable I'm noticing in results using our dry electrode system related to the subjects hair. There are different electrode comb designs, and picking one that works for everyone is challenging. I hope multiple designs can be offered in the future for optimum performance. If you don't mind, having at least a description of your hair (length/texture/density) might be helpful to me. If you're enthusiastic about that idea picture would be even more appreciated (from the back with headset on would be ok).
    Thanks,
      -Aaron
Sign In or Register to comment.

© Vanilla Keystone Theme 2026