Cyton impedance check, freeze issues
Natty_noodles
Australia
Hey openBCI community!
Hopefully someone is able to offer me a hand here.
When I check my impedance on the openBCI GUI by clicking on the omega symbol below tha channel number, it flatlines all my channels and won’t return to normal function until I turn it off and on again and reconnect the cyton.
I’m wondering if anyone had a fix for this or an idea of what could be causing it?
I’m thinking it could be an issue with the way it’s communicating with the microcontroller?
Halp!
Thank you!!!
Comments
Nat, hi.
When did you purchase this Cyton? Did it ever work correctly, EEG, EMG, etc., impedance, etc. Are your channel, reference and Bias wires properly connected to skin? What type of electrodes and electrolyte? Are you using the latest GUI version?
William
Hey, thanks for the quick reply!
Are there any extra considrations we should keep in mind when taking impedance measures that we have potentially negleted?
For example, we usally find that the first channel we test for impedance is okay, but the second or third channel we checl is when in glitches and flatlines or railes on all channels.
Thanks!
Nat
Yes, give it a try with gold cups and paste throughout. Technically it's not wise to mix electrode metals. But sometimes an exception can be made with silver chloride.
Can you elaborate? The black ear clips normally are used with the Ultracortex headset. Do you also use that?
It does sound likely that your ADS1299 analog to digital converter chip on the Cyton may be defective. Since it requires a power cycle to return to 'normal'. If the all-gold-cups also fails, then email contact at openbci.com and mention that your impedance function has never worked, and in fact crashes one of the components.
Yes, we have the ultracortex headset with the dry comb electrodes but we were getting lots of noise and high impedance (and the same freezing issues) so we decided not to use the headset and just to use the gold cups instead. Will try put bias and gnd on the gold cups and keep you posted and if that doesnt work, email the address you provided.
Thanks William
so this is what we're getting (pic attached).
What concerns me is that the major frequecy spike is around 30, even though we're not looking at impedance in that image, so maybe its leaking current at that freq? This is with the gnd and bias with gold cup electrodes too.
Any thoughts?
Let's see what Richard @retiutut has to say. The 31 Hz injection for impedance is normal, but the RAILING etc. looks unusual. The 31 Hz injection should be happening on the channel you have pressed the 'ohm' button on. But instead you see RAILED.
The all-gold-cups test should be helpful to see. It's also possible you have a flakey reference electrode. Sometimes it helps to swap Reference and Bias to possibly locate a bad cable / connection.
William
The screenshot above does not have enough information and context to determine what is happening.
Please:
FYI many first-time users get poor connections and signal quality because electrodes are not properly attached and pick up EMF noise. Blaming faulty hardware is actually the last thing we should do, unless there is overwhelming evidence (e.g. dongle blue light does not come on when plugged into computer).
Another thought, maybe try toggling impedance while the data is Not Streaming. I see in the console log you shared that there is packet loss after this command is sent.
In the current development copy of the GUI, the GUI requires users to Stop Streaming in order to access Hardware Settings for the ADS1299.... though the channel on/off and impedance buttons do not require this. Maybe this needs to change so that all streaming must be stopped to toggle impedance. We can also move Cyton impedance check user interface to an existing widget (used by Ganglion).
Hi Richard and William,
1. Yes, our channels get railed very easily. We swapped the bias and ground, it made no difference. However for some reason today we are not railed, but the last week of trying the device, we were. We dont know why and havent changed anything in our setup.
We followed your steps using the ultracortex headset with dry electrodes this time and with the black ag/cl ground and bias that come with the headset. We normally use 50hz filter and gain = body uV but left it out since they were not in the steps advised. The screenshot is from the first set of instructions pre-impredance check and post-impedance check on FP2. I want to make a note, that the issues with checking impedance does not occur on the first check, its usually when we are checking 3 or 4 different channel's impedance that it freezes.
We tried toggling impedance while data is not streaming, the screenshot for that is attached.
Thanks,
Nat
Nat, thanks.
So I'm trying to summarize your situation: impedance check usually works fine if you are only checking one channel. But if you turn on the impedance check for several channels at once, after some number of checks are turned on, everything freezes. And you have to power cycle to clear that situation. Is that correct?
If you have a freeze issue like that, I do believe you have an issue with the Cyton mainboard failing in some way, such as a marginal ADS1299. Even turning on ALL channels for impedance check, should not cause a freeze.
You always want to use your mains notch filter. So the screenshots you posted are a bit hard to interpret. But it does look like you were not seeing any RAILED conditions; which were showing up on your gold cup test when many channels were not connected.
So basically everything is working except using impedance check on more than 1 channel, will causes hard freezes, requiring a power cycle. This is a hardware failure, so go ahead and email OpenBCI support as previously mentioned. @openbci @retutut.
Regards, William
I'm not thoroughly convinced this is a hardware failure. I see 50Hz noise, this really should have been removed, as it's difficult to create 1 set of instructions that works for every situation. Also, it would have helped to see the channels on "Auto" vertical scale or Body uV.
@Natty_noodles If it is indeed "freezing", could you please create and share a screen recording/video along with the full Console Log file?
Please configure the GUI to remove mains noise and set TimeSeries Vertical Scale to Auto. Try Impedance check on 1 channel only, preferably the forehead since this usually has the best impedance.
Richard, she has stated MULTIPLE times, that her Cyton completely FREEZES up, after impedance checking more than a couple channels. This is abnormal, is it not? She has to power cycle to clear. Looks like flaky ADS1299 to me.
I'd like to see that the GUI data stream freezes, as opposed to getting "stuck" and returning bogus values when it receives certain commands. A video recording and full console log will also provide conclusive evidence as to what causes this, if that's what happens. Very hard to see what's actually happening with just pictures.
@Natty_noodles, can you describe, EXACTLY what happens during the freeze condition? Also provide log snapshots.
I'd feel more comfortable with a video recording and full console log file. The existing descriptions don't seem to clarify exactly what is happening and there is a breakdown in semantics.
There are no signs that the GUI has crashed. If it does seem to "freeze", please do not stop the video right after. Please try to start data stream or other commands. Leaving out information will only prolong the time it takes to diagnose what is happening.
3000kOhm Impedance on the forehead w/ MarkIV using Gain Mode Classic seems pretty good. I am getting similar readings. So, impedance check is working. Let's just see about the "crashing" or "freezing" with a video.
Also, I'm not having any issues running impedance check on more than one channel with GUI 5.0.1. Granted, the visual data is not meaningful, only the number in the top right of each channelBar.
Okay, thanks for the response guys. Thats right William, the impedance problems come from chekcing more than one channel. This is because for the experiment we wanat to conduct, we need to record the impedance of each channel. I've got work now but ill get the video and full console file to you a little later.
Yes it seems to rail less with the dry electrodes than with the gold cup, but when the gold ones are good, they are better than the dry.
Would it be helpful to send the same images as before but with body uV on? and a notch filter
Thanks for the help!!
We're mainly interested in the EXACT circumstances of the freeze: what do you do, how does the GUI react,, what do you do in response, etc. Is there any way to proceed without power cycle? Etc.
Thanks for clearing that up. Okay, there has been no way to proceed without power cycle. The GUI freezes and does not respond. Even stopping streaming and restarting does not fix it, we need to close everything, turn off the board, and restart. This occurs usually at the 3rd or 4th channel that we look at impedance for. We never leave impedance on for a channel, we check a channels impedance, then return it to regular recording, and they repeat that on each channel, usually by 3 or 4 it freezes. However ill record it for you this evening (its morning now for me in aus) and lets go from there.
OK, this is great, thanks. I think Richard @retiutut would especially like to see the logs when you do this. I don't know that the screen recording of the GUI user interface actions, would add anything to what you've already said. Richard?
If you do make a large video file, please upload to Google Drive.
I really do need to see both a video and a full console log file. There is enough qualitative data here already. Together, this will tell the full story of what happens. Also, this helps me try to reproduce the exact sequence of events and compare.
When you share the video and log, please share what channels are connected and the electrode type, or any other information that will not show in a screen recording.
In the remote possibility that the freeze could be related to operating system or usb issues, it may be helpful for Nat to try a test on another laptop or OS. Though that seems like a long shot, since freeze only occurs with multiple impedance checks. My guess is that something in the ADS1299 freezes up, then that wedges the PIC32. USB traffic during impedance check, is the same as a data stream, but with 31 Hz current injection at the electrode.