Another side note: while working with the ESP8266 I found big differences between the pins in terms of measured discharge times. I'm not sure what causes this (internal resistance or stray capacitance), what is interesting is that this is very constant across hardware, and not affecting the accuracy of the measurements, after calibration of course. Especially the TX/RX pins (GPIO1 & 3) are very different. All the measured times are roughly doubled.
I haven't done any testing of the same using an ATmega processor. In a week or two I hope to start working on that. Going to be a fun learning curve, as this method is pushing the limits of the processor - it sometimes even feels like it's actually USING the limits of the processor and the IO pins, while that shouldn't be the case. The 330 Ohm resistors limit the current to 10 mA (or about 15 mA on the 5V ATmega), which should be comfortably under the port limit.
The thing I've been struggling with the most is that all the different flavors of electrodes Im testing seem to suffer from bubble buildup on the surfaces. So tapping the electrode on the side of the container will create a new "set point" which will hold stable until I stir/tap it again. They all fall within a range determined by the solution, but the bubble buildup variability is enough to prevent my calibrations from getting below 10% error.
I think I may have to go below 0.4v on the driver pins to address this.
No, you have to switch to AC.
Using a DC will mess up your measurements even if there's no electrolysis taking place, as the ions in the liquid will move to the electrodes.
Getting to 10% error with your setup is very good already.
I am depolarizing the electrodes by reversing the polarity and doing a "forward" read & a "backward" read, and each take exactly the same amount of time. Doing this above 10kHz for less capacitance in Z.
Something more than simply reversing polarities is needed in high concentration solutions (ie seawater). Currently experimenting with "relaxation" phases in the mix with electrodes grounded to each other. That is improving the situation.