By Charles E. Kinzer
We had designed a new direct-contact electrode conductivity meter at a water instrumentation company to measure conductivity in water. The theory is actually simple, where an oscillator driver signal is connected to one of two electrodes in the water and a sensitive amplifier connected to the other. This makes a glorified ohmmeter with the water path between the electrodes being the resistance that is measured.
Oscillator excitation is used because electrode plating from materials in the water occurs with DC but not with AC. Also, results display for conductance in MHOS instead of resistance in OHMS.
It had very bad performance and would not read a low conductance. Even in distilled water or just in the air, the conductance read very high when it should have been zero. We could find no excessive capacitive or magnetic coupling anywhere. This left direct signal injection as to the signal amplifier as the probable reason, but how?
Finally, we removed the oscillator drive wires from the standard black terminal block where they resided along with the signal wires and connected them separately from the block. It then worked perfectly. If the driver signal were merely touched to the terminal block with the signal wires, the signal jumped dramatically confirming the problem. Our highly sensitive instrument was doing an excellent job reading the resistance of the black phenolic plastic of the terminal block. A simple circuit board revision to give the driver signal wires an independent terminal block was all that was needed. While we seemed to know that terminal block insulation resistance was a concern with extremely high voltages, we had never considered it could also matter with extremely low voltages.