The capacitance of the tank is really low and thus I've been using a 10M resistor to make the charge up take long enough for the Arduino to be able to measure it. I would suspect that using a long wire run would be very susceptible to noise. As such I'm going to keep the run from the circuit to the plate as short as possible (and the ground wire from the bottom probe to the circuit as short as possible as well).
On our Axis all of this is very accessible:
That is the front of the compartment where the black tank is (as you can see the probes there). I'll just be mounting the circuitry somewhere in that space.
Been experimenting with putting the ATMega to sleep. When running the circuit is consuming 25ma. Putting the ATMega to sleep reduces the current draw down to 1.4ma. I may just use a wake up circuit on the rising edge of the sense wire.