Capacitive Water Sensing

[Noggin]

I need to detect the presence of water in a tube and I am wondering what the theory is behind why this works.

I have made an electrode out of a 1" wide strip of aluminum foil (just the standard stuff in your kitchen) and wrapped it around a PVC hose. I then connected it to a constant current source (a few 10's of micro amps). I turn the current source on for a short while and then take an ADC reading. If the ADC reading drops, then that means the capacitance increased and that water is present.

**broken link removed**

You can see my test fixture in the image above. Without water in the hose, the ADC reading is about 800-850. When water is in the hose and near the electrode, the ADC reading drops to about 600-650. If I grab the hose where the water is, then the ADC reading drops 100-200 counts.

I understand that the dielectric strength of the water is much higher than of air and that is why it increases the capacitance.... but where is the other "plate" of my capacitor?

 

[BrownOut]

Your attachment doesn't seem to work.

EDIT: Never mind. I see it now.

 

[mneary]

The other "plate" is your board and everything attached to it.

 

[Noggin]

It just seems that it is too far away for that to be it...

I'm trying a few other electrode configurations that I think will work better. Such as two "C" shaped electrodes on opposite sides of the hose, one being ground the other being my sensing electrode. I used a shielded cable to get to it, which makes the cable nearly immune from me touching it. The little 30 gauge unshielded wire is very sensitive to me touching it. However, with the other electrode being so close to it, I get only about 1/4 of the change in ADC as with my single electrode sensor. But, the shielded wire adds a large amount of parasitic capacitance to the circuit which is very likely the major reason the sensitivity is reduced. But, the reduced sensitivity also brings with it stability in readings

 

[Noggin]

Does anyone know of a source for an electrode similar to what I have above? Soldering to aluminum foil is rather tedious and I'll need a part number to put on a BOM. I came up empty at Digikey and Mouser, but I wasn't sure what to search for other than "electrode". I considered wrapping a bare wire around the hose and then putting a piece of foil tape over it but I don't have high hopes that it would be a long term reliable connection (20+ years is the desire), not to mention whether or not the foil tape would even make electrical contact to the wire through the adhesive.

 

[Boncuk]

Why not try brass sheets?

They are available down to 0.06mm thickness and easy to solder.

They are easy to work with - you just need a pair of strong scissors to cut them to shape.

I used them to make wing tanks for a model air plane for long endurance flight (> 16 hours flying time with no worries about travel of center of gravity.)

Boncuk

 

[Noggin]

I was hoping for something that could be ordered from digikey, this is going to be a mass produced item.

I think my next plan is to place an electrode on the PCB and sandwich the hose between the PCB and a metal grounded bracket. I think I may get acceptable results.

 

[Boncuk]

Digikey is the wrong address for brass sheets.

Try at a metal store, where iron, copper, aluminum and brass sheets are sold.

 

[KMoffett]

Noggin,

Since you're going commercial, you're probably well down the road on your design, but this is a liquid detection circuit that I've used. The grounded foils on both sides of the detection foil reduce the sensitivity to external objects. For the electrodes, you could use brass tubing, available at hobby stores, as solder-able crimp ferrules. But that brings up two questions:
1. Will brass be OK in the environment where you are intending to use them them?
2. Can the electrodes be slipped down the tubing before it is connected to the water supply, or do they need to be attached to already installed tubing?

Ken

 

[Noggin]

Thanks KMoffett. I worked with the mechanical guy and had him modify the design to bring the tubing close enough to the board so that we can actually zip tie it to the board and hold it up against an electrode on the board. We're getting the boards in today so I should be able to test it out. With my prototype testing I think it'll work great.

 

[KMoffett]

If?..When?...it works, would you post a photo?

Ken

And, that avatar still creeps me out.

 

[Boncuk]

Ken

And, that avatar still creeps me out.

Hi Ken,

I've no objections against Urangutans in the forum.

Boncuk

 

[Noggin]

I wouldn't feel comfortable posting a picture as it is a customer's board and we're under an NDA. It isn't significantly different from the photo in my first post though. You can make out the electrodes on the board and it'll be zip tied to it.

And I love my avatar It's so cute and happy!