Outdoor Pond Water Level Control / Valve

[usafltg]

Hi guys -

I'm trying to find a solution, that hopefully might already be available to solve this problem. I have 2 waterfall / ponds (they're pretty elaborate), that use commercial grade pumps to flow the water (10,000 gallons per hour). One is in the front, and one is in the back. Unfortunately, on a daily basis, due to a leak or something else going on that would virtually cost a fortune to rip and replace... both ponds slowly lose water throughout the day, and thus have to have water added to them at various points.

I know they make "valves" that utilize a 2 wire connection solenoid, and potentially a wireless one as well, that could be soldered into the copper line that could turn the valve on or off. I'm trying to find a solution using a Pi, a sensor, and valve, such that when the water level drops a few inches, it would trigger the valve to open and flow water back into the pond, and when the water level reaches the full mark, it will shut the valve off. I'd love it if there were a wireless solution, but my gut tells me that's probably not likely for the distance between where the sensor would need to be installed in the pond to the raspberry pi / valve.

I don't think a "float" type trigger would work, as pond retention areas aren't that deep, just wide, therefore only an inch or 2 in drop.

Does anyone know if there's a viable solution out there for this? I've been trying to wrap my head around for about a year now.

[Andyroo]

You could possibly use an ultrasonic sensor or even a simple pair of wires where the contact is made by the water.

Are waves a problem?

[pfletch101]

If you are not using something like an ultrasound depth gauge that gives an 'analog' depth measurement, you will need two sensors, one to signal the 'full' point, and one for your chosen 'refill now' point. I would probably use a small float on a (also small) hinged arm (a miniature version of the setup that controls a toilet's fill valve) and attach two tilt switches to the arm in a way that allows you to adjust their 'attitude' so that they switch appropriately at the two levels of interest. Since each switch is going to be either closed (near zero impedance) or open (infinite impedance), you should be able to run thin, low-voltage, wiring (e.g. 3-core thermostat wiring) from the 'sensor' to your Pi, without worrying too much about false triggering, and I would do this, rather than looking for a wireless solution.

Coding should be fairly straightforward, though you should account for any likely messy failure modes - e.g. cut off the flow if it takes more than (e.g.) 110% of the usual time to get from 'refill' to 'full'.

[LTolledo]

If using a float switch, I also recommend putting it in a sort of enclosure that can "attenuate" the waves, to remove errors in activation...

[MarkR]

Whatever you use, it needs to work reliably outside and get a bit wet.

I wouldn't recommend sticking electrodes in the water, that will probably fail quite quickly (corrosion, bioslime etc). A float switch sounds fairly plausible (although why wouldn't you just use a toilet fill valve?).

Ultrasonics can probably fail in some quite exciting ways. Remember that if this fails, you'll get a very damp garden and a big water bill

So probably one advantage of using a Pi is that you can built into the software, some data logger, monitoring, and rudimentary fault detection (e.g. if the tap is on for 10 minutes and the water level is still too low, something's failed)

A float switch sounds most likely to keep working (to me).

[pfletch101]

…. (although why wouldn't you just use a toilet fill valve?).

The OP specified that there wasn't a great deal of 'drop' between the 'full' and 'refill' levels, so you need a float to be on a short arm to get enough rotation to overcome the hysteresis of a typical level switch. Thinking about it, however, if you mean 'solve the entire problem using a toilet fill valve', that might even work! Making sure that the valve continued to function well over time in the less than ideal conditions of an outside pond might not be trivial, however.

[omegaman477]

Hi guys -

I'm trying to find a solution, that hopefully might already be available to solve this problem. I have 2 waterfall / ponds (they're pretty elaborate), that use commercial grade pumps to flow the water (10,000 gallons per hour). One is in the front, and one is in the back. Unfortunately, on a daily basis, due to a leak or something else going on that would virtually cost a fortune to rip and replace... both ponds slowly lose water throughout the day, and thus have to have water added to them at various points.

I know they make "valves" that utilize a 2 wire connection solenoid, and potentially a wireless one as well, that could be soldered into the copper line that could turn the valve on or off. I'm trying to find a solution using a Pi, a sensor, and valve, such that when the water level drops a few inches, it would trigger the valve to open and flow water back into the pond, and when the water level reaches the full mark, it will shut the valve off. I'd love it if there were a wireless solution, but my gut tells me that's probably not likely for the distance between where the sensor would need to be installed in the pond to the raspberry pi / valve.

I don't think a "float" type trigger would work, as pond retention areas aren't that deep, just wide, therefore only an inch or 2 in drop.

Does anyone know if there's a viable solution out there for this? I've been trying to wrap my head around for about a year now.

Keep it simple. There are many cheap water tank level detectors available. Most are either Ultrasonic or conductive. A good quality conductive sensor in a freshwater pond, with gold contacts will survive for many years, and they are cheap. Look for an AC detector, rather than a DC (DC will encourage electrolysis, and corrosion of the electrodes) But be real, is it so bad to have to replace a $5 sensor every 4 years?

There are also capacitive sensors, usually as cheap.

But the super simple approach is two float switches. Cheap, rugged and reliable. Most float switches are based on mercury contact switches, and sometimes magnet and reed. Mercury tilt switches are accurate (reproducible) to the mm. Magnet /Reed switches tend to gain hysteresis as they age. As to the problem of 'Waves' this can be easily resolved via software 'hysteresis' or averaging. My only advice with float switches in the 'real world' is to ensure they don't get fouled with leaves and twigs. Place them in a riser tube with a screen at the bottom.

[pfletch101]

But the super simple approach is two float switches. Cheap, rugged and reliable. Most float switches are based on mercury contact switches, and sometimes magnet and reed. Mercury tilt switches are accurate (reproducible) to the mm. Magnet /Reed switches tend to gain hysteresis as they age. As to the problem of 'Waves' this can be easily resolved via software 'hysteresis' or averaging. My only advice with float switches in the 'real world' is to ensure they don't get fouled with leaves and twigs. Place them in a riser tube with a screen at the bottom.

The sale and/or use of mercury switches are prohibited by law in an increasing number of jurisdictions - at least in the northern hemisphere - because of the high toxicity of metallic mercury, if it escapes.

[omegaman477]

But the super simple approach is two float switches. Cheap, rugged and reliable. Most float switches are based on mercury contact switches, and sometimes magnet and reed. Mercury tilt switches are accurate (reproducible) to the mm. Magnet /Reed switches tend to gain hysteresis as they age. As to the problem of 'Waves' this can be easily resolved via software 'hysteresis' or averaging. My only advice with float switches in the 'real world' is to ensure they don't get fouled with leaves and twigs. Place them in a riser tube with a screen at the bottom.

The sale and/or use of mercury switches are prohibited by law in an increasing number of jurisdictions - at least in the northern hemisphere - because of the high toxicity of metallic mercury, if it escapes.

They are being rolled off everywhere. Hence my inclusion of magnet & reed. Most low cost float switches are usually M&R.

[ptimlin]

While I can appreciate if the idea is to learn to use the Pi for something like this then I say have at it. But if the idea is simply to get an auto top off for the ponds, any reason why you can not use plain old mechanic float switches that simply open or close a water valve? Search for something like "livestock water tank float valves" to get an idea of what is available.