pH probe accurate until adding plants and oxygen

I recently started a project to build a hydroponics monitoring system. I have everything working correctly, though, I'm noticing some issues with how my pH probe is functioning. It measures the buffer solutions perfectly, but when I place it into a deep water culture reservoir with a well rooted plant and an air stone, the measurements are significantly off. I'm getting a fluctuating reading around 7.8-8.1. When I take some water out of the reservoir and measure that in a separate cup, I get a reading of 3.95. This matches the readings of my store bought sensor. It's not a deep reservoir and if I hold the probe near the top of the water it still reads high. Has anyone run into this? Does anyone know if it's the oxygen, the roots, or possibly something else?

Yes, my water should not be at 3.95 pH. I was too trusting of the probe because of how accurate it was with the buffer solution that I went crazy with pH down solution...

The presence of nearby active electrical circuits (pumps, lights, heaters, etc.) could profoundly affect such sensitive circuitry. Of which you have told us absolutely nothing.

I'm a moron. I was just about to say I had a tds probe in the water...

Ground loop issue.

You have to electrically separate the two - and that's not easy.

I'm using two capacitors to isolate my TDS probes, that helps a lot (on the TDS side as well).

Yes, the measurement circuit must be totally isolated from the water tank, to a very high
standard (even tiny leakage currents will swamp the result).

I've had great results from a pH probe that I just plunked inside a stainless steel tank; with motor running inside the same tank (brushless submersible 12V DC pump) and EC probe outside the tank in the hose. Probably I've just been lucky. Or I'm just very good, you never know :slight_smile:

Ground loops are a major issue with pH probes. Very hard to find and get rid of. The EC probe is a major risk of causing ground loops, and a potential victim of the same.

I'm trying to design a circuit that can completely isolate the pH probe: NME0505 to completely separate the power supply; ATtiny reading output from precision OpAmp; optocouplers for serial interface (100 kHz I2C is possible through optocouplers but quite expensive, 9600 bps serial should be enough). That should absolutely get the ground loops out of the way without adding too much cost, though those NME0505s cost more than the rest of the components and the PCB together...