I soldered silver plates to active electrodes and now everything works perfect. No "jumping" at all, signal is very stable and of very good quality. There is no 60Hz interference, the tests were done with dry active electrodes.
See below three eyes blinks (active electrodes are on forehead), and then teeth grinding.
Active electrodes works very well. It is amazing expecially considering the fact, that the design is very simple.
Most analisis were performed with using BioEra, a few tests with ElectricGuru were done for verification.
Active electrodes add another level to safety. So high skin-electrodes impedance (megaoghms) means also much higher isolation.
Only DRL works with this prototype. I do not know why, but touching VGND has no effect and doesn't allow eeg reading.
I suspect, that the "jumping" effect (desciribed here) can be the result of dynamic changes in input electrode-skin impedance. This can occur easily with dry, non-adhesive electrodes. My silver plates are very polished, almost like adhesive and that could also be an important factor.
Below is the scheme of one active electrode in this prototype. The bottom opamp is actually not needed. TLC272 that I used has two op amps, and I wanted to show precisely what I did. Improved version of this design scheme proposed by Joerg is available here.
Cables in my prototype are very short now, about 70cm (2feets). I wanted to extend them. See what happened. Extender was 2 meters (6 feets) long and shielded!
The above result shows that DRL may not be able to compensate interferences between active electrode and modularEEG. Therefore either the connection must be short and probably shielded, or another active electrode design with positive gain should be used, see new design here.
Here is picture of modified schematic proposed by Joerg, it has added 100ohm coupling resistor that should help with the above problem, it also adds HF rejection.
All previous tests were done with active electrodes placed on forehead, this was due to flat shape of electrodes.
To try active electrodes on another parts of my head, I created pin electrodes and connected them instead of plates.
There were made of pure silver as the plates were.
I was able to measure eeg on back of my head through the hair.
One of the measurement results can be seen here.
This was done with dry silver pin electrodes without any skin preparation.
I should admit that due to poor skin-electrode connection, the stability was not very reliable.
Sometimes I had to move electrodes a few times, until good spot was found and proper results were shown.
Additionally because of the shape of electrodes it was also not very comfortable.
The measurements were possible, but the skin-electrode connection would have to be checked continously to guarantee good data.
Since something like that is now possible in BioEra, I want to come back to those experiments at other time.
Links to another active electrodes
Here is description and picures of active electrodes built by Joe Street.