Making water conductive with out salt

December 1, 2006 7:00 PM

Options other then salt to make water conductive to low voltage (5 - 9v) DC current.

I'm working on a water fountain project and I need to determine when a drop of water passes through what I'm calling a 'water gate'. The 'water gate' is a circuit that will be completed when the drop of water is between two small wire probes, this will then cause some code to be run on an Arduino board.
I have access to +5 or +9 volts DC through the board and I'd like to keep from using another power supply thus the need for the low voltage.
The other problem is because this is a fountain the water will be running though a pump, and I don't want salt to build up on the pump and case it to die before its time. Also the salt build on the out spout that forms that drips could get rather large I would think.
Are there any other options to make water conductive to this low of a voltage with out the build up that would/could occur with salt?
I've tried tests with all kinds of water (normal tap, distilled, RO) but still I've had to add salt. It's not the end of the world to add the salt, but I would rather another option if it's out there.

posted by blackout to science & nature (24 comments total)

Any ionization in the water should work (I'm trying to remember chemistry...), but I can't think of something other than salt that you could add.

The reason salt does it is because it dissolves into Na+ and Cl- ions in the water, and those allow electricity to move through. You just need something else that dissolves that is LESS bad for everything than salt is.

Sorry I cant give you a concrete answer, but maybe this can at least help point you in the right direction.
posted by cschneid at 7:03 PM on December 1, 2006

How is this possible, if you're looking to dissolve something in the water that you don't want to be there? Anything but distilled/RO water will eventually cause buildup somewhere.
posted by rxrfrx at 7:07 PM on December 1, 2006

Perhaps the sensor could measure thermal conductivity instead - the difference between water and air is orders of magnitude. One approach is to measure the current needed to keep a resistive heater at a constant temperature. A sudden change in required current would indicate the presence of a drop of water (as long as you've shielded against breezes!)
posted by Mapes at 7:10 PM on December 1, 2006

rxrfrx has a very good point.
Mapes, I'm not sure if it would work with the rate of drops there will be. I still need to do some fine tuning but it seems it's going to be anywhere between 1 - 5 drops per second. And with all that water you would think the heat loss from evaporation would cause false positives. Although I've never worked with what you suggest.
posted by blackout at 7:17 PM on December 1, 2006

Acetic acid (vinegar) or ammonia will make a weak electrolyte, likely conductive enough for your purposes.
posted by Wet Spot at 7:26 PM on December 1, 2006

An acetate solution, in the right conditions, will grow a wicked cool crystal.
posted by rxrfrx at 7:28 PM on December 1, 2006

Can you change your sensors at all? There are lots of water detecting flood alarms out there that run on a 9V battery, that's all you're essentially making.
posted by Science! at 7:33 PM on December 1, 2006

IANEE (I am not an electrical engineer) yet, but...

I wouldn't go the water contamination route. how about a laser pointer and a photoresistor? Arranged correctly, the refraction of the light upon entering the water will deflect the beam to hit or not hit the sensor. Or the diffusive properties of the water will alter the resistance measurably.

Depending on how quick a reaction you need, how about a Thermistor?

I take it this is for art of some sort? Please describe the project further...
posted by phrontist at 7:43 PM on December 1, 2006

Distilled water is still somewhat conductive, I think you need to introduce some current gain.

Here's a few links that I really haven't looked over.

• 1
  
• 2
  
• 3
  
• 4

posted by substrate at 7:45 PM on December 1, 2006

the drop of water

Didn't see that before. Go with the photogate described above.
posted by phrontist at 7:46 PM on December 1, 2006

What I'm working on is a clone of a time fountain (see here and here) I need to detect the drop in order to determine when to strobe the UV leds.
Thanks for all the ideas guys, I'll try them all out, I really like the idea of the photogate though, that should be just as easy to rig up.
posted by blackout at 8:09 PM on December 1, 2006 [1 favorite]

YES! Timefountain! I want to do this now! OOOOH!
posted by phrontist at 8:11 PM on December 1, 2006

I plan on writing a how to as well as releasing any source code I write.
I'm waiting on my shipment of UV Leds at the moment though, should arrive early next week.
posted by blackout at 8:18 PM on December 1, 2006

did you try it with the marker dye already?

i ask because lots of fluorescent dyes are actually ionic salts that will also make water conductive.
posted by sergeant sandwich at 8:25 PM on December 1, 2006

Not yet, I will when I get the shipment of LEDs in though.
All great input guys, thanks so much.
posted by blackout at 8:26 PM on December 1, 2006

If you want to go back to the original plan of increasing conductivity without adding something that will crystallize out, consider adding some alcohol, like rubbing alcohol (isopropanol).
posted by ikkyu2 at 8:31 PM on December 1, 2006

I'll take tomorrow and test all the ideas out that I can with what I have now. And let you guys know what works and what doesn't.
posted by blackout at 8:32 PM on December 1, 2006

I couldn't find the conductivity of 3% hydrogen peroxide, but it wouldn't clog at least (would probably bleach your dyes, though).
posted by jamjam at 8:40 PM on December 1, 2006

How about working it the other way around? Start the drops at a certain rate, and tune the strobes until it looks right.

_{This sounds like a job for LeakFrog!}
posted by Mr. Gunn at 6:40 PM on December 2, 2006

That would work if you could control a few other things, the main being voltage. In the area where I am the mean voltage out of the wall when you plug something in is 124 Volts AC, although this changes all the time I've seen it as low as 109 and as high as 130 volts sometimes. Even those few volts are going to change the rate at which the pump motor works so it will change the rate at which the drops fall. So it's much more easy to determine when the water passes a point, because the rate at witch it falls is not going to change, within reason, the rate will change with the wind in the room, etc, but it will not be much. That small change verses the constant change in voltage that controls the pump, I would much more prefer knowing when the water passes a point.
At current the area where the fountain will be displaying (i.e. the are where the UV LEDs are to strobe) is just a mere 15 inches thus the water drop will not reach terminal velocity before it hits the pool of water below. Yes it will be speeding up the entire fall, but I think the increase in speed will be small enough to not really matter. Again I've not got my UV LEDs to test this out with, they should arrive sometime early next week.
Really there is no way to have a constant flow of drops at a set rate, thus the need to determine when the drop pases a given point to control timing and other events.

To those that care, I didn't get a chance to test all the ideas out today (rather yesterday as it's 1:58 am local time) I will indeed try them all out as I said I would and I will post the results in this question. I ought not to give a time frame though because I'm sure I would miss it.
posted by blackout at 1:00 AM on December 3, 2006

Just looked up leakfrog and that sounds like a great product for it's intended use, but the use I need would require me to rip it apart and make it work as a sensor to the arduino board. And although that would not be the end of the world I already have a laser point that I ripped apart and photoresistors are very cheap.
Also to get more into the design of the fountain, in the original design (see here) it looks to me that he has all of his LEDs flashing at the same time. I intended to have just a row (that is a pair on the same level) of LEDs flash at the same time. My current design has a total of 10 rows of LEDs (for a total of 20 LEDs) to cover the 15 inch drop area. With each row being controlled independent of the other rows. I think (again I have not the LEDs to test yet) this will allows my to show just one drop at a time whereas you can see (in the video in the above link) the original maker has a few drops at any given time shown, due to the fact that all of the LEDs [in that design] flash at one time. There is not video of the second design so I don't know if he fixed that or still uses the old method.
Once I get my LEDs I will be able to play around much more and determine what works and what doesn't. Who knows my whole design may have to be redone.
posted by blackout at 1:18 AM on December 3, 2006

Now that I've followed your link, which I should have done at first, of course, I see that the dye used is soluble fluorescein, a disodium salt, which sgt. sandwich apparently, impressively, knew just off the top of his head. Your solution will almost certainly be conductive enough without further additives.

But I worry your scheme may have a flaw in the physics that renders that moot: if your wire probes are at the nozzle or before, the current will start to flow as soon as the solution bridges the wires, but that will not tell you when the droplet leaves the area of the probes, and that interval is likely to be significant and variable. But if you then decide to base your timing on when the current stops flowing, that moment (if it occurs at all) may not always coincide with the departure of the drop because of wetting effects at the probe, which I believe are likely to be enhanced by current flow. If you put the probe somewhere in the free-fall path of the drop, the drop will splash off the probes, and how it behaves after that is anyone's guess.

Not to mention that the electrolysis produced by the current will degrade the fluorescein.

Phrontist's optical method looks like the way to go to me.
posted by jamjam at 10:48 AM on December 3, 2006

well, now that i think about it, there's probably some kind of surfactant or something similar in the marker ink that keeps the fluorescein solvated when the ink dries out.

dried fluorescein has a kind of dark orange color and it doesn't fluoresce very well until it's in solution, but clearly when a marker dries on paper it's still pretty bright. so i think there's some extra stuff happening with marker ink that might interfere with conduction in solution, but i don't know.

it's certainly true though that passing current will cause the same kind of deposit build-up/electrochemistry that rxrfrx alluded to early on and that jamjam points out. this might be a big problem, or it might just mean you'd need to de-gunk and refill the fountain every so often. i agree that the optical method sounds most promising.
posted by sergeant sandwich at 11:29 AM on December 3, 2006

Last summer I made a bubble counter out of a RadioShack matched IR LED/phototransistor, a 2N2222, a comparator, a potentiometer, and some resistors. Though it detected bubbles rather than drops, I ripped the idea off from detectors that are used for IV's and titrations. I'm sure it would work just as well for detecting drops.

The design is fairly simple, and I got it to count anywhere from <1 to hundreds of bubbles per minute across a 1.5 o.d. plexiglass tube. you can also get matched leds/phototransistors that are housed in a small plastic enclosure with a small gap for something to pass through. i'm sure a href="http://www.digikey.com">Digikey has them.

Feel free to email me if you want the design.
posted by dsword at 6:37 PM on December 3, 2006