MRE Electronic Pond Level Control
For many years I have used a float based switch and relays to automatically fill my half acre duck pond. In typical "toilet tank" control fashion it opened an electric valve when the pond got low and shut it off when the level reached the top setting. This system proved to be very reliable with the only problems being that all the moving parts failed infrequently. I have replaced the float, the switch, and the relays a couple of times over 8 years.
While working with the LM393 voltage comparator chip used in my 12V UPS circuit I noticed just how sensitive the chip was and noted the tiny currents needed to "flip" the output state. This gave me the idea of using the LM393 to sense the water level in my pond and operate the valve. I tested this idea by extracting a sample of water from the pond and placing wires in the container to see if the small currents would trigger the chip. It worked better that I expected!
Feeling confident the LM393 could be configured to operate a small relay and control the 24VAC valve, I set out to design a sensor that would endure the not-so-clean pond water for many years. I knew that even with tiny currents metals immersed in dirty water could form oxides and corrode. Remembering my 7th grade science class and experiments with hydrolysis of water, I decided to use carbon electrodes. I had on hand a box of 1/4 inch copper clad Arcair electrodes which worked out well. I was able to build a sensor with nylon supports and slide it inside a 1-1/4 inch PVC pipe. The copper cladding on the carbon rods allowed me to solder the wires on all three rods such that the joints would always be above the highest water level. One long rod serves as the common while the other long rod and the short rod serve as the bottom and top contacts. The nylon disks are drilled with three holes on 90 degree increments. The two long rods are glued to the bottom disk with JB Weld but not glued to the top disk which is allowed to slide on them to adjust the differential levels. The short rod is glued to the top disk. The disks are held in place inside the PVC pipe with a stainless steel screw in each. I set the differential on my application at 6 inches which means the distance from the bottom tip of the long electrodes to the bottom tip of the short electrode is 6 inches. This determines the cut on / cut off points on the control. When the pond level falls below the long rods the valve turns on. When the pond fills and the water touches the top rod the valve turns off. I should note here that the ends of the long rods are positioned about 1-1/4 inches above the lower end of the pipe. By so doing the waves and ripples on the water surface have little effect on the sensor since the bottom of the pipe always remains below the water surface. I drilled six 1/8 inch holes around the bottom edge of a cap placed on the end of the 1 1/4" pipe that allows water to enter the sensor but screens out trash and fish. At the top of the pipe the hole where the power and valve control wires exit serves as a vent that allows air to enter and escape as the pond level changes.
I built the circuit on a universal circuit board and mounted it inside a short length of 2 inch PVC pipe. I used my lathe and boring bar to cut offset slots on the inside of the pipe such that the circuit board would slide neatly inside. A hole was drilled and a clear Plexiglas plug was inserted to allow light from the bi-color LED to shine through. This can be seen in the photo where the white lead curves up to the base of the LED. By using an 8 pin plug on the board, it can easily be removed from the pipe for servicing. A pipe cap and a reducer complete the top assembly and make it weatherproof. I did not glue any of the PVC joints as they were quite secure by just pressing together. I used some cotton like plastic packing material in the space between to circuit board and the lower pipe (in the tapered section of the reducer) to help deter any insects that might enter the lower tube and invade the circuit board space.
The control circuit operates directly from the 24V AC supply needed for the 1" sprinkler valve. I elected to go with a Zener diode controlled TIP41C transistor to provide the 12V DC current for the LM393 chip. Since the voltage on these 24V transformers can run well over 30 volts with no load a common regulator will not handle the rectified DC voltage. The heavy duty transistor dissipates the power easily which is good because the high voltage drop, even with small current, will generate some heat. The sensor voltage is fed through a 1K resistor (R2) to limit the current in the event of a sensor short circuit such as a marine animal getting across the line and it also allows the momentary contact switch SW1 to serve as a manual trigger/test switch. SW1 can be used to start the fill cycle when the water level is between sensors. Those of you familiar with industrial controls will recognize that the voltage comparators and transistors serve the same function as a start/stop pushbutton station on a magnetic starter. When the top sensor (S1) is dry, Q2 conducts (stop button released). When the bottom sensor (S2) is dry, Q3 conducts (start button pressed) and pulls in the relay which latches through the lower set of contacts and holds the relay energized as the pond fills and S2 again becomes wet (start button released). The relay remains energized as the pool fills until S1 becomes wet and turns off Q2 (stop button pressed) which breaks the holding circuit. The other set of relay contacts are used to operate the 24VAC valve and the LEDs which are reverse voltage protected through D3 and D4. One should include a power switch on the 120V side of the 24VAC transformer. This main power switch is useful to manually terminate the fill cycle and also to kill the circuit in the event of a malfunction. I also strongly suggest a 0.5A inline fuse on the 24V side since I have discovered over the years that these little solenoid valve coils have a nasty habit of failing with a shorted coil. The fuse will save your wiring, transformer, relay, etc. when this happens.
Some other thoughts... (1.) The hydrolysis of water by passing a DC current through it splits the water molecule into hydrogen and oxygen. I suppose it is possible that over time and even with the tiny currents involved, some hydrogen could accumulate in the upper part of the tube around the circuit board. Thus far the top of my device has not exploded and I don't really expect it to, but it is a factor worth mentioning. A rain proof vent might be something to think about! (2.) I sprayed my completed circuit board with clear acrylic to protect it from the elements. It is after all living just a foot or so above the water. (3.) Even though this is a 24V circuit and the transformer provides a level of isolation from the power line it is always wise to power systems used outdoors with ground fault breakers. This goes double when working around wet areas. A GFCI outlet or breaker should be used to provide the 120VAC power for this or any other outdoor tool or project.
I hope you find this information useful. As usual I end by noting the fact that I am not an electrical engineer, just a hack tinkerer ham radio guy so take that into account when considering this design. If you damage your equipment or burn down the block trying to use something you find on this site, don't come crying to me. Heck, how do you think my junk box gets supplied? Enjoy!
Email me with any comments or corrections by clicking here.
After several months of operation two problems have been encountered with the initial design. (1.) Lightning has damaged the LM393 chip and transistors twice. I have recently installed three 15V transient-voltage-suppression (TVS) diodes (P6KE15CA) across the input lines of the chip to ground. That would be across S1 and from pins 3 and 5 to ground. (2.) The lower disk supporting the sensor rods became coated with algae and gunk. The contamination was conductive and resulted in the circuit failing to turn on at low water level. I removed the bottom disk and glued all three rods to the top disk. Now the two longer rods are suspended inside the tube and there is no submerged support disk to contaminate.
It remains to be seen if these design changes will improve the reliability of the system. I will note that the carbon rods appear to be holding up well. The positive rod was still coated with copper whereas the more negative rod was showing a much thinner layer remaining. I expect that over time the more negative rod will become bare carbon. In my mind this confirms that the use of carbon, at least for the low level sensor rod, is necessary as a metal rod would eventually be corroded away by the tiny currents.
Almost another year has passed and the thunderstorms are back. I again had issues with the electronics getting zapped. As of now I have abandoned the electronics and reverted to a float controlled electromechanical system using 2 micro switches and a relay. This time around I did enclose the float and switches inside a 2" PVC pipe to protect them from the elements and the sensor retains the vertical profile.
I am leaving the system design on the site because it actually works great but it appears to not be suitable for my particular outdoor application. It was a nice experiment in electronics and hopefully will inspire other ideas...
THE ABOVE CIRCUIT AND TECHNIQUE FOR IMPLEMENTING AN ELECTRONIC POND LEVEL CONTROL SHOULD BE DONE AT YOUR OWN RISK. THE AUTHOR, THIS WEBSITE, NOR THE WEB HOST ASSUMES ANY RESPONSIBILITY FOR THE SAFETY OR ACCURACY OF THIS INFORMATION. THERE IS NO WARRANTY EITHER EXPRESSED OR IMPLIED THAT THIS METHOD AND PROCEDURE IS SAFE AND NO LIABILITY WILL BE BORNE BY ANY PARTY INVOLVED IN PREPARING OR PUBLISHING THIS DOCUMENT. ELECTRONIC POND LEVEL CONTROL FAILURE MAY RESULT IN DAMAGE TO THE PROPERTY ON WHICH IT IS IMPLEMENTED AND OR PERSONAL INJURY MAY RESULT. THIS CIRCUIT AND SYSTEM SHOULD BE CONSIDERED EXPERIMENTAL AND IS OFFERED HERE FOR INFORMATIONAL PURPOSES ONLY.
USE AT YOUR OWN RISK!