Ganglion data collection "resetting" every 1 second (200 sample points)
I've been playing around with the Ganglion + EEG headband kit (using 2 flat and 2 spiky dry electrodes) for a couple weeks now, and have come across a problem with the data collected "resetting" every 200 sample points or 1 second. I noticed that this effect is only noticeable when I attach an ear clip electrode to the REF pin.
Below are screenshots of the OpenBCI GUI showing the 1Hz noise, as well as the text file of the collected data.
As can be seen with the blue highlight in the text file, the EEG channel data (columns 2, 3, 4 and 5) jump suddenly between sampling point 200 and 0 (column 1).
I have tried the following:
1. Changed batteries.
Initially using Energizer's rechargeable NiMH batteries, but changed to Panasonic's disposable batteries. No change.
2. Checked if electrodes/wires were faulty.
Measured resistance values of the used electrodes were 30 to 80 ohms for the two ear electrodes, and 2 ohms or less for the flat and spiky dry electrodes.
Also tried pretending that the flat electrodes were the REF electrodes (instead of using the ear clips) by changing the pin connections, but the same "resetting" noise was observed.
3. Re-uploaded the DefaultGanglion program to the Ganglion chip.
Followed the Ganglion programming tutorial to make a zip file and upload the default program to the chip. No change.
I am starting to run out of ideas on what to try next.
Any help would be greatly appreciated.
Comments
Hello,
Re-opening this thread with a new comment on behalf of a customer who is seeing the sawtooth 1Hz noise, and who had previously consulted this thread. She was sent a new Ganglion but is still seeing the 1 Hz noise.
The sawtooth in the data and the 1Hz noise seem to popping up here and there in the forum, as evidenced by an older post in April: https://openbci.com/forum/index.php?p=/discussion/2097/ganglion-to-brainbay
@Shirley and @Pipi, hi.
As I mentioned in the previous comment on this thread, the raw values and screen shot you posted, shows extremely large sample values coming from your (supposedly EEG) electrodes. Normally EEG values for Ganglion are less than 1 or 2 hundred microvolts, and more normally in the below 50 microvolt range.
The fact that your voltages are constantly rising into the 20,000 and 30,000 microvolts (20 to 30 millivolts) range -- generally are indicative of problems with your electrodes. Do you have gold cup electrodes and paste there? You mention these tests are with the dry electrodes. It's possible either the ear clip or flat / spikey electrodes have some damage such as chipped surface. These dry type electrodes are a plastic base, first plated with silver, then chlorided on top of the silver. If the chloride layer is chipped or missing, it will expose silver underneath. And mixing electrode metal types can cause a battery / galvanic effect, with large voltages generated.
Since your Ganglion has already been replaced, it would seem that the culprit is not the Ganglion, but something in the electrodes.
The reason the the 1 Hz "sawtooth" type wave is seen in these cases with Ganglion, is the ADC analog to digital converter has a limited range. Every one second (200 samples) the Ganglion resets it's differential (delta compression) voltage transmission packets, with an 'absolute' set of 4 samples. So the 'reset' effect is explained by this. See,
https://docs.openbci.com/docs/03Ganglion/GanglionDataFormat
So my suggestion to you is to follow closely all the steps in the Ganglion tutorial, using possibly new or different electrodes.
https://docs.openbci.com/docs/01GettingStarted/01-Boards/GanglionGS
The gold cups and paste are a reliable standard that is known to work. Perhaps the dry electrodes you have could have been damaged. I would certainly pickup some cup electrodes and a small tube of paste and give that a try.
Regards, William
Hello
First of all, thank you William for your suggestions. I have not been able to get around to trying all of them out yet, but will try a bit at a time.
I've finally found a bit of time to test the Ganglion board once again, and have tried running the Ganglion boards with no electrodes at all. This was a suggestion given through one of my e-mails with the staff at OpenBCI. If I am to jump to the results, the two Ganglion boards, without electrodes, behaved much the same, and both signals did not quite "look right".
The signals started out quite noisy with odd jumps in the first 30 seconds or so. Then I felt that it settled down, exhibiting what looked I would have expected the signals to look like without electrodes for 15 to 20 seconds at a time. However, I let both boards run for about 5 minutes or so while staring at the OpenBCI GUI, and found that it would have a jump in signals every now and then (see pictures below: First picture is from the old board, the second picture is from the new board).
So now, the jump is not regular, but still exists. However, I'm not sure whether this is the same noise as before, just visible under a different perspective, or a different type of noise altogether.
Running any EEG system with no electrodes connected to the scalp or skin, gives spurious and random results. This is because the open pins act like antennas for noise. Depending on your local electromagnetic environment, this can result in variable types of noise results from the output stream.
In the previous comments, I suggested trying some type of cup electrodes with paste. Have you been able to do this? That is the most obvious solution. Since both the old and new Ganglion boards exhibit strange noise with your dry electrodes, this seems to be the only explanation.
Another test you can do with any EEG system, is to use alligator or other clip leads to connect all three pins together: channel, reference, and ground. This will result in a signal very close to zero microvolts.