Cyton picking up noise only on one channel?
Hi all,
I have an issue with one of the channels of the my Cyton board.
Here is the situation, the tests I have made and some documentations on the tests:
I have in my possession two Cyton boards (let's call them Cyton A and Cyton B ). As reported in another post here https://openbci.com/forum/index.php?p=/discussion/3726/electrode-impedance-test-using-the-gui-widget#latest , I had issues with the impedance measurement when trying different resistors instead of scalp. Those issues were reported and reproduced by Philipp Pitts. I made all those tests only the channel 1 of Cyton A, Cyton B being left completely intact.
Now that I am working on something else, I tried to use the Cyton A to record alpha waves (as I was able to do previous to the impedance measurement tests). I was unsuccessful.
This led me to conduct the following debugging experiments:
I used a waveform generator with a sine at 10 Hz along with a voltage divider, to produce a voltage below 10uV. I also plugged the SRB2 and BIAS pins to the GND of my circuit (which is the GND of my waveform generator).
I plugged the output of my circuit to directly to channels 1 and 2 of both my Cyton A and B.
I obtained the exact same results with both boards.
Then I attached an electrode and connected it "mechanically" using wet EEG gel and tape to my circuit output again. I did it for channel 1 and 2 using the exact same cables and electrode. I got the following results for Cyton A:
Cyton B had the expected results with symmetrical behavior whether I was using channel 1 or 2 (with same cable/electrode as Cyton A).
What do you think could cause the discrepancy between channels?
I suspect that maybe the RC filter on the analog front-end could be maybe damaged, I have no clue. But the noise that I am seeing is high enough and wide enough in terms of spectrum that it basically partially covers the alpha wave signals I am trying to monitor.
Comments
I'm not quite sure I'm following the tests you did after that point (bold face above). The first test worked as expected. Thus I would highly suspect your subsequent wiring. In particular you mention:
It seems that this may be the problem area. Why did you find it necessary to simulate a wet gel connection, if the original test worked just fine?? Have you considered that human subjects differ widely in the amount of alpha generated. And that testing with other subjects might give different results? Are your subjects being recording using parietal or occipital locations, where the alpha is highest?
Regards, William
I wanted to limit the amount of pictures in one message but for the second test what I wanted to see was to further investigate some potential issues regarding electrodes and the way the board handles my electrodes.
What I did is the following:
I plugged a male-male jumper wire at the output of my voltage divider circuit while still having the SRB2 and BIAS pins connected to the GND of my circuit.
In order to input the 10 uV / 10 Hz signal into my electrode I used a cup filled with conductive EEG gel and placed my electrode inside it while also placing the male part of my jumper wire inside. See following picture
I doubt that my wiring is the issue as the following screenshots show
These are the results I have with Cyton A (the defectuous board.)
Using the exact same setup but with the board that has never been used in any of the impedance measurement test:
These are the results I have with Cyton B (the control board).
This led me to do a way simpler test: What signal do I have when absolutely nothing is plugged? All the channels of Cyton A are immediately railed and the signal goes to 0 uV except for the channel 1 which sees a signal around 15 uV RMS regardless of what is plugged there. Of course, every channels of Cyton B are also railed and do not pick up any random noise.
Do you see a reason why a channel would see such a signal while all the other ones are immediately railed and see nothing?
Regards
Additionnally, I have considered that human differ widely in terms of alpha waves generation but I have no problem seeing alpha waves when I use every channel from 2 to 8, the only one who causes troubles is the channel 1. I locate the electrode on the O1 point and look at the spectrogram after a few eyes open / eyes closed.
If you believe your board is defective, you can email (contact at openbci.com) to log a ticket with tech support.
In your cup filled with gel test, using an electrode and a male-male jumper -- are you aware that certain (usually differing) metals in the presence of a conductive solution (or gel), produce what is called a galvanic battery effect?
https://www.google.com/search?q=galvanic+metals+voltage
re: "What signal do I have when absolutely nothing is plugged?"
You cannot leave the Cyton inputs free floating and connected to nothing. You must have at least the channel in question, reference and Bias (ground) -- connected to a subject. Otherwise the microvolt sensitive inputs tend to collect noise from the EMF electromagnetic field environment. Especially if non-connected wires are attached to the input pins, those then act as antennas picking up even more EMF.
All biosignal amps use differential amplifiers which subtract the two voltages to obtain the microvolt signal level.
Thank you for your quick reply.
I had no clue about the galvanic battery effect. Thank you for the reference.
This is not so much the picked up noise that is my issue but the channel-to-channel discrepancy. Noise does not discriminate, it should be noisy everywhere or railed to 0 uV everywhere. Plus, my "control board" measurement do not match the other one, despite being in the exact same testing conditions.
A fix for the impedance issue is forthcoming shortly. Sorry for the delay in resolving this.
https://openbci.com/forum/index.php?p=/discussion/3790/high-impedance#latest
William