Most of the homes in the area where [Raikut] lives have tanks on the roof to hold water. Each is filled from a well using a pump, with gravity serving as a way to pressurize the home’s water supply. The system isn’t automatic and requires the home owner to manually switch the pump on and off. [Raikut] made this process a lot easier by designing an LED bar indicator to monitor the water level.
The sensor is very simple. Each LED is basically its own circuit controlled by a transistor and a few resistors. A 5V signal is fed from 7805 linear regulator into the tank. The base of each transitor is connected to an insulated wire, each extending different depths in the tank. As the water rises it completes the circuit, illuminating the LED.
[Raikut] is conservation minded and built a buzzer circuit which is activated by the LED indicating the highest water level. If someone walks away from the pump switch while it’s filling the alarm will sound as it gets to the top and they can turn it off before it wastes water.
[via Reddit]
why bother when all you need is a simple PID controller?
Why not take it 1 step further, add a relay to the pump
Yeah, automagical it. Circuit at (or near) bottom starts pump, circuit at top turns off.
Hell, make the package pretty and start selling them.
I am sure that turning the pump on automatically might not always be desirable.
But I do agree that implementing relay switch for automatic turning the pump off seems pretty obvious. Especially when author figured out to use the water itself as electrically conductive element.
Why not close the loop and have it enable/disable the pump?
Using the drum itself as the +ve contact will slowly cause rust. Better to reverse the sense and have the drum be the -ve contact.
If his schematic is correct, the transistors are wired with collector and emitter reversed, so will have lower gain (hFE≈10 instead of hFE≈150). I’m pretty certain the transistors are being operated in saturation anyway, and even if they’re not it’ll only make the LEDs dimmer than they would be otherwise.
I believe you are correct, assuming his schematic matches the implementation.
Maybe he’s not allowed to modify the pump, it might belong to the municipality and he just owns the tank, which is why he used this method, but if he does own the pump I agree automate the filling process, he certainly has the skill.
There are a few ways around that:
1. They may own the pump, but they can’t own how it’s connected to the home’s mains. Even if they own everything up to the pump’s power connector, one could simply add a relay at the connector.
2. If somehow the company has the pump wired to power separate from the home owner’s own connection (unlikely), It would be trivial to activate the switch using a servo or solenoid. I’d hope that if this were the method used, the owner would install some weight based failsafe that turns the switch *off* in the event of a failure. Perhaps a mechanical power float switch (like a sump pump’s) as another backup to cut power to the switch-flipper.
I made one of these for a water tank once. After a few months the copper wire descentegrated. I assume it was some sort of electrolysys type chemical reaction.
The professional sensors use a float with a direction sensor inside. When the float hangs down, it is open. When the water lifts the float up, it turns over and the circuit is closed.
Or a capacitive level sensor that has no wetted parts.
Or a piezo sounder/mike forming an ultrasonic version of the pinger on Das Boot. Fire it upwards from outside the tank bottom, or downwards from the tank top, or both, using the time delay for the main reflection at the water’s surface. Periscope and depth charges optional.
Hack Man has a point… you can go for a float based system, but the moving parts can lead to complications such as getting jammed, or giving false readings. The capacitive alternative may lead to better results. One of such that I have found to work fairly well is the water level sensor here: http://vegetronix.com/Products/AquaPlumb/
Pumps are expensive. Especially if they break from over-use. Once a day is most likely more than adequate to fill the tanks.
It’s also important to keep a person in the control loop. If something should go wrong with the ‘automatic’ version, then the house could go without water (which could be a bigger issue than having to manually run a pump every day).
Yea… the HUGE risk of overflowing… like everyone else in his hood doing it manually…
It’s a water pump… couple it with a timer so it only check level at night if you are that scared of water… (if you need to push the derp so far in order to be able to make a statement… maybe it’s time to put the keyboard away…)
My guess is that you have never had water damage in your home. It might be a small risk, but the consequence is that everything could be ruined.
So use multiple failsafes in descending order of complexity. If you’re truly paranoid, install 2 sump-pump style level switches that are turned “off” by high water level. Put in series they will still turn the power off even if one fails.
So use a microcontroller and set it to fill once a day automatically at 8am or something. And have an override that if it goes empty it fills again (all the way to the top), if this happens more then once a day it sets off the buzzer.
This will work well…until the wires corrode.
What’s wrong with a simple float switch to shut off the pump? Whole lot simpler/cheaper.
Does your toilet buzz when it’s ready to flush? Float on vertical rod, two switches large range between top and bottom.
We had these when I lived in south central Mexico. When and how much we filled the tank would also gauge our water temperature. you would never want to fill it full in the morning unless you wanted boiling hot water in the evening. The well is nice and cool, black rotoplas is nice and hot.
You could make some money selling white canvas covers for them.
These tanks need to become hot periodically or they’ll develop legionella and you become sick from the water.
“The best is the enemy of the good.”
This is a great project as it solves an immediate problem. Yes, features can be added, but for the problem at hand it’s perfect. Maybe it’s a stepping stone to something better.
But, having said that, those sensor wires will corrode, unless he uses an ac signal. A better way would be an ultrasonic distance sensor mounted in the top of the tank, pointing downwards. Then an Arduino (or if you want to be very chic, a Raspberry Pi) to drive the sensor and calculate the water level.
Or a series of float switches. The floats could push rods upwards out of the water to activate a switch, so that the switches and wiring are away from the water.
Given the cost of what he built and being in India, I’m not certain there’s any budget for a microcontroller at all. Wire, battery, LEDs, resistors, transistors? A time source is probably actually the hardest part here.
Honestly, if the markets where he lives are anything like the ones in Old Delhi, I’m sure sourcing cheap parts won’t be too difficult, however I’d have to agree. Simplicity wins here.
Not really clear the details of the community’s water system is it? Sounds like community well with each residence having water tank on the roof. Sounds like mess to me, if each residence can control the well pump dependently. The Jones family turns off the pump because their tank is full. The Smith family needs water in their tank so they turn on the pump Now the Jones tanks overflows. In the event each home has their own well & pump, time for each home consider using a volume tank to provide the pressure with a pressure switch controlling the pump. As a community well & pump time for a water tower for real water pressure. I believe a column of water exerts .044 PSI per foot. water tank full level of 10′ the pressure would only be .44 PSI at ground level, even less at the sink levels.
I think you’re off by an order of magnitude:
$ units
You have: 1 foot / 1 mm * mmH2O
You want: psi
* 0.4335275
It’s a nice hack. I am also thinking of implementing some system like this. But is it safe to do this? Also how much cost is involved in making such a circuit?
I really like the user interface.
if he wants to monitor it because he wants to know if its empty or full it’s ok… but if he’s monitoring it to know when to turn the pump on and off… this is a total overkill!!! he could buy a “float switch” (http://en.wikipedia.org/wiki/Float_switch) to turn the pump on and off, “what if he doesn’t own the pump but just the tank” some may ask… well… then connect it to a solenoid valve and thats it!! thats one of the most common setups in mexico at least
Float switch? Leds? Transistors?
Ball (float) valve.
Designed for the job. Simple. Cheap. Low maintenance. Been around since the Victorian’s discovered the joy of a flushing toilet.
Please forgive my ignorance but, I don’t understand the purpose of the pump and barrel. Is there something wrong with the city/county water pressure? Or is it because the person in question gets there water from a well?
can i know the name of the sensor used for sense the level of water….