Troubleshooting & Clarification (Beginner)
Hello. I've been using the Ultracortex MarkIV with the Daisy + Cyton boards and v5.2.1 of the GUI. The main problem I have is that the waveforms in the time series graphs appear identical in most of the 16 channels. I've attached some screenshots to show what I mean. Other common problems I've been having include intermittent packet loss warnings/data streaming errors, railed time series channels, and poor impedance checks on certain channels even though all the electrodes are screwed in all the way. Hardware settings and filters are left at default, hopefully that's optimal. I also did the FTDI fix that was described in the documentation. The dongle is connected straight to the PC port since I do not have an extended cable, which I believe improves the signal slightly. The headset is fairly new and all the electrodes look grey so they shouldn't be worn down or anything. I do tests in a tech lab so maybe there's some electronic interference but I'm usually the only one in the room and the other PCs are typically in sleep mode.
Other things I'd like to clarify:
- Railed means the signal is too large for the vertical scale correct? To get clean data, the channel should say not railed or at least near railed?
- On average, what range of values should the uVrms be per channel?
- On average, what should the head plot look like? pale colors? darker colors in areas where activity occurs? blue corresponds to negative voltages while red = positive right?
- Should I change the hardware settings/filters?
- For impedance checks, would values below 2000 kOhms be good enough for decent readings?
(I believe the data streaming error that appears on these screenshots was due to low battery whereupon it failed to generate any waveforms, but I've usually had it to where the dialog box appears despite waveforms appearing. Maybe the latter happens due to delay between when I press start streaming and when data is received.)
Comments
Hello again,
I've posted more screenshots of a session down below, this time the lithium ion battery was charged. As you can see I am still having the consistent issue of identical waveforms throughout all channels, which I doubt is accurate. Also the head plot is entirely a deep red which I also think is wrong. Maybe there's something wrong with the hardware? What steps could I take to remedy these errors?
Thank you for sharing the screencaps. I see identical noise on all channels. Let's try to address this with fine-tuning your set-up and reducing environmental noise.
First, see general tips for reducing noise (Cyton and Ganglion boards):
Moving away from the desk helps reduce the noise seen in the Time Series widget. The 20Hz noise is a harmonic of 60Hz, and shows up especially where the electrode does not make good contact, or where there are simply too many cables and electrical devices around. This is a very common artifact.
In the GUI's hardware settings, check that the SRB2 is ON for all channels.
Check reference and ground electrode connections:
Lastly, please see the recommended use note for the Ultracortex electrodes:
https://docs.openbci.com/AddOns/Headwear/MarkIV/#the-complete-ultracortex
"before adjusting/twisting the blue electrode mounts, you must disconnect them from their colored cable. it is NECESSARY to ensure that the white cable end of the electrode mount is not connected to anything. Otherwise, a broken wire is probable after repeated turn adjustments."
Zach, hi.
Apologies for not responding earlier. Your second set of images show very similar (nearly identical) noise on ALL CHANNELS at once. The only electrode shared with all channels is the Reference (SRB2) ear clip lead. This would be the first thing to suspect.
Have you tried swapping or replacing the ear clips?
Independent of your Ultracortex setup, have you tried some simple tests, perhaps using another setup? For example using cups and paste, or the adhesive type electrodes? As described in the tutorials.
https://docs.openbci.com/GettingStarted/Boards/CytonGS/
https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/
The noise shown is also extremely high amplitude for EEG, nearing 1000 uV. Normal EEG is generally below 100 uV. So you might expect the noise could be related to environmental EMF electromagnetic field noise from nearby equipment, such as: power supplies, extension cords, wall wart transformers, conduits in walls floors ceilings, wireless equipment, etc.
To eliminate the environmental aspect, you might try setting up in another room, again away from the sources mentioned.
Has this OpenBCI gear EVER worked in a more normal fashion? When purchased? Have you interacted with customer support (contact at openbci.com) ?
Regards, William
Thank you all for your responses!
As per your recommendation, I did some tests with a different setup. I still used the Ultracortex Mark IV, earclips, and spiky electrodes, and recorded in the same room, but this time with a Ganglion board rather than a Cyton+Daisy. I posted some screenshots of the results below. Looks like the readings are more accurate and stable: <1000 kOhm impedances, <100 uVrms, lighter colors on head plot. Based on these findings, it seems that the Cyton+Daisy was either setup incorrectly or is altogether faulty. I will try setting up the Cyton+Daisy again and see if I still have data issues, but I'm fairly confident that all wires were connected as specified in the documentation (Y-splitter cable to both boards' SRB2 and everything).
Although the data was better, I ran into some issues with the streaming process. Sometimes it streamed fine, other times it was slow and halted via data streaming error. I'm wondering if that was because I had a google chrome window open (I was writing in google docs as I streamed data)? The accelerometer also wouldn't produce any readings sometimes.
Thank you again for your help!
Are you able to see alpha brainwave (~ 10Hz) generation with eyes closed? This is stronger in occipital / parietal area, than in frontal / temporal.
As far as lost packets, try using a usb extension cord for the dongle, and drape the dongle / cord over a nearby monitor to get it away from the table surface / laptop / desktop etc. This might fix your accelerometer issue as well.
Hi,
I had a partner look at the FFT plot while I was recording data with my eyes closed and the occipital channels seemed to spike. Appears that the ganglion is accurately reading brain signals! I hadn't used a USB hub yet but the software worked ok after some rebooting.
As for the cyton, I tried removing the daisy extension, swapping the pins from the bottom to the top row, and using cups and paste, but I was still getting the same issue of identical channels and high voltages. This leads me to believe that this is an issue with the electronics on the board. I posted pictures of the board we have (w/ daisy, ear clips removed). I don't see any egregious damage or anything but perhaps you can spot an issue? Maybe we should send back our board to get it fixed/replaced or hold a zoom meeting with someone to find a solution? The next time I do testing I'll get a USB hub set up but would that alone be the cause of such odd issues?
Thanks again
To be more specific, you are looking for an increase in ALPHA band with eyes closed.
We said nothing about using the top row on the Cyton mainboard. You MUST use the row closest to the board (bottom row), because the SRB2 pin there acts as a common reference.
We asked previously:
[ Has this OpenBCI gear EVER worked in a more normal fashion? When purchased? Have you interacted with customer support (contact at openbci.com) ? ]
PLEASE, next time try your testing with the cups exactly how it is described in the tutorials. Just use the ONE channel as shown, with electrode at the Oz (rear of scalp) position. This will enable seeing the alpha with eyes closed.
https://docs.openbci.com/GettingStarted/Boards/CytonGS/
https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/
Let us know the results of your alpha test on Cyton.
William
Hello,
Firstly I'd like to clarify a few things in my last post. I mentioned there were spikes in the occipital channels on the FFT plot using the ganglion; these spikes did in fact occur in the 10 Hz area whenever I closed my eyes, thus alpha waves were detected. Also, in the majority of my testing, I used the bottom row of pins on the cyton to wire all my electrodes; I had just wanted to try the top row for one test because it seemed like each channel shared two pins (including SRB).
A colleague and I went through the tutorial exactly as outlined in https://docs.openbci.com/GettingStarted/Biosensing-Setups/EEGSetup/ (except I used a fully charged lithium ion battery rather than AAs) but still we got the same erroneous readings, therefore no alpha waves were detected in channel 7 (occipital). I've posted the pictures of the session below. As you can see, the channels are still identical and too high in voltage. I noticed an LED turn red on the dongle, but it seems its just indicative of data streaming. At this point I am fairly confident that there is something wrong with the cyton board, as I've gotten the same results out of the cyton using multiple setups. What would you recommend we do to have the board fixed?
Thank you.
Asked this TWICE already:
Because you are getting identical noise on BOTH channels, it means that something is wrong with your reference (SRB2, left ear lobe) connection. Either too little paste, bad electrode, bad cable, bad connector, etc. Or possibly could be related to the subject MOVING during the session. Try different possibly known good electrodes. Gold cup electrodes can go bad if the plating wears off or chips.
It is hard to read the max scale you have set on your time series, too fuzzy to see in the photo. Is it 1000 uV. That is WAY more than you should be seeing with EEG. EEG is normally below 100 uV.
I believe the cyton board and spiky electrodes were purchased in April of this year so it is fairly new. The Ganglion and gold cups were purchased before then, about a year ago. The ganglion seems to be functional still as I was able to get appropriate readings from it.
We were planning on contacting customer support but wanted to verify our findings on this forum first. My advisor has just emailed [email protected].
The uV readings from our most recent test were consistently higher than the expected range of 0-100 uV, at times getting as high as 2000 uV, and these do not seem to reflect any sort of behavior from the subject. Sorry for the low resolution, I was unable to access the screenshots at that time so we resorted to regular pictures.
I had used a good amount of paste for each cup used, the subject sat fairly still, and as you can see in the pictures the wires were fully plugged into the pins. As I had mentioned in the post I quoted below, the same spiky electrodes and earclips we had used with the cyton produced accurate readings with the ganglion, suggesting that the wiring/electrodes is not the issue. The common denominator throughout these tests seems to be the cyton board, perhaps there is an issue with the SRB pin/port?
And no the cyton has not worked normally since we had purchased it, it consistently produces readings of high voltages and identical waveforms throughout each setup.