How do you build a water level sensor


Thread Starter


I'm working on a project where i need some kind of level to indicate how much water i have in the tank. I have done some research on the net to find such a sensor, but all the ones i have found are very expensive and i need to build this sensor from scratch. can anybody give me some hints or schematic i can look at to design this?

Many ways to make a sensor, depends on high / low limit sensor or constant level sensing. The simplest constant level may be to take a float, connect to arm, connect to potentiometer with small voltage and measure wiper.
Hi You may get some ideas from the web sight:
For low cost I would suggest a simple plastic model located in - products - level sensors - single point. You could then attach a rod to it a stick it in the tank to any level. Continuous model - it would be cheaper to buy one - than
make one. If you want to buy section parts to make or assemble a multi point unit - email a request. Email Almeg for a quote. Thanks Bob Hogg
Expensive is relative. If this is an industrial project you're better off to bite the bullet and buy one rather than construct one, you'll save youself some hassle and money down the road. If however this is a school/home/farm project there are a number of cheap alternatives.

Remember that when it comes to CHEAP -vs- GOOD -vs- FAST, you can have any two, but never all three. Cheap and fast won't be good, good and fast won't be cheap, cheap and good won't be fast.

1) Use load cells to weigh the tank. If the tank is symmetrical and evenly distributed on its supports, you can get a reasonable approximation with just one load cell under one support and multiply the result by the number of supports. Accuracy can be improved by calibrating the load cell output to the actual level as measured with a dip stick or other simple device. Inexpensive
load cells are readily available in most industrialized areas.

2) Mount a multiturn potentiometer with a tooth belt pulley on it at the tank top. Using a long tooth belt cut in half, attach one end to a float and the other to a counterweight that hangs ouside the tank. You will have to tinker with the counterweight so that is just balances the float. If your tank is very turbulent this is going to be a troublesome approach, but for a calm tank its cheap if not a bit kludgey. You can also use the counterweight as a pointer to a (inverted) visual scale on the tank exterior.

3) Use your imagination and create a float/slidewire/variable resistor that fits your tank. Remember that with counterbalancing weights or springs you can put the resistor almost anywhere outside the tank. Get creative.

4) If you happen to have a low pressure transducer you can just measure the fluid pressure at the bottom of the tank and calibrate that to level. Good pressure transducers can be obtained for $200 and up.

5) If you only need point level, you can get float switches from "": or "": for about $15.00 each.

6) Extend two parallel conductor rods into the tank from the top to bottom. The resistance will change with the ammount of rod sumberged in the water. Make sure the rods are insulated from the tank if it is metallic. Its cheap but its also quite finiky and subject to changes in your water (dissolved minerals, etc) and doesn't offer much resoultion without expensive electronics.

7) How cheap do you want to go? Is this a high school project where you want to impress an easily impressed teacher? Is your tank fairly small? If so, look up LVDTs on the net and see how they are constructed. You can make a crude one with a piece of thin wall PVC pipe, an iron rod, and some varnished copper wire. Attach the rod to a float large enough to support the rod. The electronics are failry simple and you can find this info on the net. This approach has a high "cool factor" for a school project, but I wouldn't recommend it anywhere else, unless you already have an industrial LVDT, but your range of change in level will be somewhat limited.

8) You can take a temposonics transducer and mount the magnet to a dough-nut float. A bit more expensive, but a very good fluid level measurement method for a few hundred dollars.

9) I once got a free-sample of a micro-machined air pressure sensor chip that was advertized as being sold for around $10.00 or something like that. Its been so long that I forgot who made it, but it measured low air pressure levels in the range of a few inches of water column. If your tank is not too tall, you can try and find one of these and set up a bubbler system. Entend a small hose to the bottom of your tank. The pressure in Inches-H2O to push air bubbles out of the hose is equal to the height of the water in your tank in inches. Keep the bubble rate low so your hose doesn't become significant source of air flow resistance thus creating a pressure drop in the hose. As the water level changes, the pressure in the bubbler hose will change.

The most important thing is: Imagination is the key to creating your own sensors. There are literally hundreds of ways to measure a fluid level.

Try these links as well:
"": and type in level
measurement and click search.

Hello friend,
You can use a simple float with pulley arrangement as a sensor or you can use a simple transistor circuit i.e. you have to sense water level. When that level is reached then your electronic circuit completes & gives water level indication.

Robert Scott

If you could build from scratch, in quantity one, a sensor that was as good as what is commercially available, and for less money, then it stands to reason that you should quit whatever job you have now and go into business competing against these suppliers. Otherwise, buy an off-the-shelf solution. The only exception I can think of is if your particular application allows a cost-reduction that is not available to most other applications.

Curt Wuollet

There are some rugged yet inexpensive pressure transducers available that would seem well matched to the task. With the pricing I've seen even I, the DIY advocate, would probably just use them with a lookup table. The trick is to look in the general electronics catalogs rather than the "industrial" catalogs. They come in atmospheric and relative versions for pressure vessels as well. I'll find the place where I've seen them when I get a chance and post it. You can build around the basic transducer.


There are several variables to consider.

>> some kind of level Does it need to be a continuous level signal, or can it be derived from discrete points. If discrete, how many points need to be measured (high and low - high, low, and control - high fault, high warn, control, low warn, low fault - or some other arrangement)? If continuous, what accuracy and repeatability are desired?

Is the water clean, or contaminated in some way? If clean, is it to be used for human consumption?

What is the temperature range?

How large is the tank? Where are the access ports, and what are their sizes and orientations?

How will the level information be used? What type of controller will be used? (hardwired conventional relay, "smart" relay, PLC, other computer type).

One might make a cheap analog level measurement system by first building a floating deck. This could be made with a couple of empty 2 litre soda bottles mounted in a frame. The frame could be centered in the tank using a pair of 1/2" PVC pipes, or a square or octagonal rod spanning the height of the tank, and fitted to prevent rotation of the floating deck.

On the deck connect a length of 3/8" wide timing belt long enough to span the tank, and then some. Build a pulley structure above the tank to loop the floating deck timing belt through. On one of the pulleys connect a 10 turn pot, or an incremental encoder. Connect a small counterweight to the free end of the timing belt (just enough to keep tension on the belt). If a pot is used then calculate the diameter of the pulley such it will turn less than 10 turns from tank empty to tank full (or, gear down the pot to stay below 10 turns).

Empty the tank, and set the pot a 1/2 turn from zero (which gives a 'live' zero than can be discriminated from, for instance, an open pot wire or similar failure) then mesh it with the belt. Note the distance from the floating deck to the top of the tank, and wiper voltage. Fill the tank and note the wiper voltage, and distance from the floating deck to top of tank.. Subtract the distances to learn the zero to span level change, and wiper voltages. For instance, if the 'empty' distance is 60" and wiper voltage is 0.5V, and 'full' distance is 10" with a wiper voltage of 9.5V then water level increase is 50", and gives a 9V control signal. Do the calculations, and find the volts per inch (or inches per volt), then use that information as needed.

If the parts are sourced from surplus outlets the parts costs could be contained, but there's a fair amount of labor and design involved, so the cost difference between a 'roll-your-own' and an off-the-shelf sensor may turn out to be quite small.

Rather than build one I would purchase a capactive proximity switch to be used outside of the tank. If the walls of tank are plastic, a good adjustable switch will allow you to "see" thru the wall and only look at the water inside,
without need to contact the tank or the fluid. If the tank walls are metal, a teflon well can be inserted into the tank and the switch threaded into the well, again, not contacting the water. Those type of switches cost less than $100. Unless you're dirt cheap or bored, I wouldn't bother.