Measuring conductivity in water

I want to be able to measure the output on a RODI unit. Pure water measures 18.2 megaohms / cm and goes down from there if the water contains more ions.

I am sure that I'll be building (or buying) a probe of some type with electrodes 1mm apart or something and measuring resistance values by setting up a potential divider. I think I can handle all of the maths involved, but I could use some help setting up something that not just works but works halfway decently well and putting some realistic numbers to what I can do. How would I get the best resolution out of it? Do I make a divider with a resistor larger than what I want to measure or smaller?

Im not sure on the values but I think a wheatstone bridge is what you are looking for, check wikipedia out for how it works

You might also need some sort of depolarizer (AC excitation ?) since DC applied to a pair of probes is bound to set up an electrolysis reaction.

Perhaps if the voltage is low enough you can prevent electrolysis, and make sure the probes don't react witht the water and change the ions

Perhaps if the voltage is low enough you can prevent electrolysis

No, if the voltage is sufficient to cause conduction to measure it is sufficient to cause electrolysis.

@Delta_G
This is a lot more difficult that I think you imagine. 18M is a very high resistance to try and measure and impossible without proper amplification. Also you are going to have a very wide dynamic range which means a high precision measurement. That is before you start an try and deal with the electrolysis problem. It is not something a beginner can tackle. It would be a major project for me and maybe I am a tad more experienced than you.

Grumpy_Mike:

Perhaps if the voltage is low enough you can prevent electrolysis

No, if the voltage is sufficient to cause conduction to measure it is sufficient to cause electrolysis.

@Delta_G
This is a lot more difficult that I think you imagine. 18M is a very high resistance to try and measure and impossible without proper amplification. Also you are going to have a very wide dynamic range which means a high precision measurement. That is before you start an try and deal with the electrolysis problem. It is not something a beginner can tackle. It would be a major project for me and maybe I am a tad more experienced than you.

Thanks Mike. That's exactly the type of info I was looking for. I had a feeling it might be harder than I thought.

So if I want to monitor water quality, I might be better off looking into an off the shelf solution that I can use Arduino to monitor as opposed to a DIY probe. TDS meters aren't very expensive, so I would assume that the parts can't be that expensive. Worst case scenario I'll be hacking an off the shelf meter.

If the RODI unit you are referring to does in fact provide a chemically pure DI water output a simple ac powered conductivity measurement device will indicate "end of Life" for the RO filters as the filtrate (output) will begin to conduct progressively more current. There are several methods that might well be applied, 1. measure the conductivity with an AC excitation and set a limit where an indication of a progressively lower value of conductivity would indicate EOL for the filter. 2. Measure the dielectric constant of the water, any contaminants would change the dielectric constant slightly (much the same as Solution #1) this could be done with a pair of insulated plates (a capacitor) and the water is the dielectric material 3. This is for me a real stretch but I think it is possible to use a non porous membrane containing an electrolyte material (a dissolved salt in water) and measure the voltage developed when the ionic content of the measured sample changes. This will create a very small emf differential between the sample and the electrolyte cell. NONE of the above are simple methods all will require a great deal of electronics experience and likely a lot of experimentation. #'s 1 and 3 are subject to the influence of any other metal present as the combination would in of itself be an offsetting factor for the measurement. #2 because it would/must be electrically isolated from direct dc contact with the electrodes might be the easiest to do BUT as Mr Grumpy Mike said it is not a trivial topic and even though I spent nearly 20 years in the field I cannot think of an easy way to do it without a lot of investigation as mike said "

It is not something a beginner can tackle. It would be a major project for me and maybe I am a tad more experienced than you.

" and I most whole heartedly agree with him.... I wouldn't like to try doing that. Did you look to see if the Mfr of the RODI system had a device for sale?

Doc

Might work

http://webpages.charter.net/tdsmeter/

Do I make a divider with a resistor larger than what I want to measure or smaller?

A few possibilities:

  1. measure resistance: the simplest is to use a divider. Pick a resistor value that provides you the best resolution in the range. If for example your high end is 2Mohm, I would pick something that's 10% of that -> 200k. This approach favors the lower end of the measurement. Connect that resistor to a pin to energize the divider only during measurement phase. If the water's resistance doesn't change fast - likely the case, put a capacitor there to help out.

  2. measure frequency: you can use the resistor to form a relaxation oscillator and you can measure the frequency of the oscillator to get a sense of the sensor's resistance.

  3. treat the whole sensor as a lossy capacitor: this approach will require fancier math / electronics but you get to detect if the water is there.