Gold cup electrodes for EEG-SMT

[michaloblastni]

After much effort, I figured out how to use professional electrodes with your EEG-SMT device.



You'll need:

• EEG-SMT https://www.olimex.com/Products/EEG/OpenEEG/EEG-SMT/open-source-hardware
• 5x 3.5" stereo jack, 3 pole https://www.aliexpress.com/item/1005006946002972.html
• 5x gold plated cup electrode  https://www.aliexpress.com/item/1005008999662099.html

Attach the single wire from the electrode to the long leg of the jack. The middle part of the jack is a ring. That's where the signal is.

If you have a digital multimeter, attach one electrode to the ring and the other to the gold cup electrode. It should show there is a working contact. Make sure there is not a contact between the gold cup and the tip, or the sleeve (the remaining two parts of the jack).



Now, you can use EEG-SMT with your new electrodes.

Gold cup electrodes require a conductive paste on your skin:
https://www.youtube.com/watch?v=FCX9oP8fpU8&ab_channel=CNSACMedShop

Remember to buy a conductive paste for EEG. Without it, the electrodes will not attach properly to your skin and the signal will be noisy or flat.

You can try water, but the signal will have a lot of noise without a proper paste.

Conductive paste for EEG: https://www.aliexpress.com/item/1005005209117678.html
You may also need a prep creme to scrub your skin as on the video above.

This solution is much better than the original passive electrodes. Gold cup electrodes are small and easy to attach properly. They stay in place. I also like that there is only one wire, not three.

I briefly evaluated this with water and it worked, however there was quite a lot of noise. After my conductive paste arrives, I will reevaluate it.

Coming up next is an experiment with EEG Cap from OpenBCI and OpenBCI electrodes. I plan to connect them to EEG-SMT and see how it performs.

What are your thoughts? Have you already tried using better electrodes?

[LubOlimex]

Nice. Thanks for sharing.

The original project had a lot of people construct passive electrodes and they used silver contacts for better contact. Check them in the links from here:

https://openeeg.sourceforge.net/doc/hw/electrodes/passive/

But in my opinion active electrodes are better if you don't want to chase perfect skin contact. They can measure through hair and without gel. We sell active electrodes but you can also check the DIY original considerations here:

https://openeeg.sourceforge.net/doc/hw/ae.html

[michaloblastni]

I really appreciate you sharing these design options. Thanks for that, Lubolimex.

With professional electrodes connected, the EEG-SMT can be a powerful device for anyone who wants to run professional-level applications. You can get high-quality electrodes for just a few bucks, and the best ones give you maximum signal with minimal noise. Plus, they're designed so you can easily attach them to your head without making a mess, and it's simple to remove and reattach them in the exact same spots as before. That's important, because you don't want any unwanted movement that could mess with your machine learning models that were trained on previous data. For the DRL, an EEG ear clip works better than using an electrode in your hand. Another important design factor is minimizing pressure on the scalp and head, since long-term wearing needs to be comfortable and safe.

I've already figured out how to use professional passive electrodes with the EEG-SMT, but now I'm working on connecting professional active electrodes.

What you'll need:

• EEG-SMT https://www.olimex.com/Products/EEG/OpenEEG/EEG-SMT/open-source-hardware
• OpenBCI kit (this includes active electrodes) https://www.aliexpress.com/item/1005002683152100.html
• EEG ear clip (for DRL) https://www.aliexpress.com/item/1005007480172861.html
• 3D printed Mark III Nova (you can find cheap 3D printing services in your city. 1kg PLA filament costs only $15) https://github.com/OpenBCI/Ultracortex/tree/master/Mark_III_Nova
• Loctite super glue with Cyanoacrylate https://www.aliexpress.com/item/1005009296409832.html
• 5x 3.5" stereo jack https://www.aliexpress.com/item/1005006946002972.html

I'm still waiting for my electrode kit to arrive. However, these are DRY electrodes that don't require any EEG gel. I will try them with EEG-SMT once they arrive.

My 3d printed EEG cap is here:


After printing and attaching 8 bolts, it looks like this:


Only four electrodes will be used with the EEG-SMT, which means I'll have just two channels, since the device unfortunately requires two electrodes per channel.

I'm planning to experiment with a different wiring setup that uses only one electrode per channel, taking the differential against the DRL. I'll let you know if that works.

I'm also not sure if the DRL will function correctly when I use an ear clip instead of the same dry electrodes, so that's another test I want to try. The OpenBCI electrode kit comes with enough electrodes to use one as the DRL instead of the ear clip, so I'll be able to compare both options.

After that, I'll compare the amount of noise I get with professional gold cup electrodes and real EEG gel versus the noise with OpenBCI electrodes in a 3D-printed EEG cup.

Finally, I want to compare the noise from the EEG-SMT with the noise from the OpenBCI 32 8-channel board, using the same OpenBCI electrodes and the same EEG cap for both. All the parts should arrive within a week or two, so I'm looking forward to running these tests and sharing the results.

By the way, Lubolimex, have you, or anyone else, tried connecting professional active electrodes to the EEG-SMT yet?

[LubOlimex]

We haven't tested any professional electrodes.

This cap is looking nice it would be nice if you can share the 3D design, I often get questions on how to attach the electrodes in place to the head!

I am also curious whether using USB isolator between the PC and the boards would reduce noises induced by the PC. Something like USB-ISO-HS or similar:

https://www.olimex.com/Products/USB-Modules/USB-ISO-HS/

[michaloblastni]

The design is at https://github.com/OpenBCI/Ultracortex/tree/master/Mark_III_Nova It's Open Source.

I 3d printed:
FRAME_FRONT (for me it's medium size) https://github.com/OpenBCI/Ultracortex/raw/master/Mark_III_Nova/STLs/FRAME/medium/FRAME_FRONT_medium.stl
FRAME_BACK https://github.com/OpenBCI/Ultracortex/raw/master/Mark_III_Nova/STLs/FRAME/medium/FRAME_BACK_medium.stl
OCTANUT https://github.com/OpenBCI/Ultracortex/raw/master/Mark_III_Nova/STLs/MECH_PARTS/OCTANUT.stl
OCTABOLT https://github.com/OpenBCI/Ultracortex/blob/master/Mark_III_Nova/STLs/MECH_PARTS/OCTABOLT.stl
OCTARING https://github.com/OpenBCI/Ultracortex/blob/master/Mark_III_Nova/STLs/MECH_PARTS/OCTARING.stl
E_HOLDER https://github.com/OpenBCI/Ultracortex/blob/master/Mark_III_Nova/STLs/MECH_PARTS/E_HOLDER.stl
COMFY_INSERT https://github.com/OpenBCI/Ultracortex/blob/master/Mark_III_Nova/STLs/MECH_PARTS/Comfy_Insert.stl


SUGGESTED PRINT SETTINGS

FRAME_FRONT & FRAME_BACK

    Material: PLA
    Supports: YES
    Raft: hopefully NO (but if supports aren't sticking, try the raft)
    Infill: 20%
    Layer Height: 0.2mm
    Number of Shells: 3
    Speed while extruding: 50-70% (slow it down if possible; these parts are detailed)

MECH_PARTS (OCTANUT / BOLT / SPRING_CASING / ELECTRODE_HOLDER)

    Material: PLA
    Supports: NO
    Raft: NO
    Infill: 20%
    Layer Height: 0.2mm
    Number of Shells: 3
    Speed while extruding: 50-70% (slow it down if possible; these parts are detailed)


Yes, I'm using a cheap USB isolator based on ADUM3160:


https://www.aliexpress.com/item/1005006860967668.html
Data sheet for the isolator: https://www.analog.com/media/en/technical-documentation/data-sheets/adum3160.pdf

Once my OpenBCI electrode kit arrives, I'll test the difference in noise between an isolated vs. non-isolated port with ADUM3160. If the isolator will be doing a poor job, I'll have to buy https://www.olimex.com/Products/USB-Modules/USB-ISO-HS/

[LubOlimex]

Nice. It is probably good enough, we have similar isolator with ADUM4160 (main difference between ADUM4160 and ADUM3160 is the isolation rating, 4160 is rated 5kV, while 3160 only 2.5kV). Ours is called USB-ISO. The difference between USB-ISO and USB-ISO-HS is the throughput - the ADUMs are good for up to 12MB/s (Full-speed USB) while the USB-ISO-HS can work also at 480MB/s High-speed USB. This shouldn't matter for EEG-SMT, but is important for audio processing.