[Update 6/29] – TComp has generously shared additional instructions for assembling these active electrodes which can be found via this link
In a previous post (https://openbci.com/community/posts/active-electrodes-for-openbci) I had described my solution on how to build low-cost active electrodes for the OpenBCI headset. The DYI active electrodes based on previous works from Joerg Hansmann and Jarek Foltinsky at the OpenEEG project (http://openeeg.sourceforge.net/doc/hw/ae.html) worked very well and significantly reduced the impedance of the circuit.
The goal of the second part of the project was to find a way of attaching the active electrodes to the existing Ultracortex Mark III headset without any modification of the components. To achieve this, I built a simple assembly by using an M2 long screw that was passed through a hole in the upper side of the 3D printed enclosure and then through the center hole of the electrode holder. All components are then tightened with a double nut and spring washer on the upper side of the electrode holder. A 3D printed pin with a blind hole at the bottom was pressed on the nuts; it served as a guiding pin going through the upper hole of the octabolt.
The system is spring loaded in the same way as the original passive electrodes assembly suggested by the OpenBCI team. The octabolt will need to be unscrewed out of the headset in order to make space for the active electrode assembly.
The 3 wire shielded cables from the active electrode assemblies is then connected through a special connector blocks to the Cyton board. The 5V power comes from the AVDD (+2.5V) and AVSS (-2.5V) pinouts on Cyton. The signal wire is goes to the signal channels in the same way as regular passive electrodes, while the shield is connected to the ground. The small internal Li-Ion battery (as recommended in the OpenBCI setup) was replaced by a set of larger bateries since the active electrode setup will require more power than just the board alone.
All electrodes in the system present significant lower impedance than the original passive electrodes and performed very well when tested with EMG signals such as eye blinks or teeth grinding.