I want to make a water level indicator but, instead of all LEDs lighting at 100% (for example) I want the top one only to light and the rest turn off
I tried this circuit but it did not work with water.
An optical prism water sensor is 'on' by default (dry) and 'off' when immersed. Sounds like exactly what you're looking for, and you avoid the hassle of passing electricity through water.
More about those here: https://www.electroschematics.com/optical-liquid-level-sensor/
You could literally just wire LEDs to the phototransistors and get the effect you want.
The basic problem with the original circuit (besides a schematic that is difficult to decipher) is that water resistance is too high. The schematic doesn't say what the operating voltage is, but the water resistance could be a few hundred K ohms. This will not allow enough current through to saturate two transistors with collector currents in the 10's of mA.
One possible solution is to change all of the transistors to small-signal n-channel MOSFETs. Add a 1 M resistor from each gate to GND. While you're at it, I think you can replace three of the transistors with 6 diodes.
Here is a first draft of a solution using N-MOSFETS. Each one turning off the previous one. SW1 to 4 are float switches. Only the float switches are in the water.
Capacitors should be between 0.1uF and 1uF (test before deciding). Capacitor voltage no less than 25V.
Schottky Diodes prevent current dropped by the LED's to reach the previous MOSFET gate. (update)
simulate this circuit – Schematic created using CircuitLab
I know this thread is three years old, but it has some recent activity, so . . . here.
As a public service for future generations, here is the OP schematic redrawn for clarity. The structure of the circuit is fine. It is not clear why the LED resistors are split into two resistors with the bypass transistor connected in the middle. The circuit would work fine and have much less power dissipation if the bypass transistor were connected directly to the LED anode, and the two resistors combined into a single 300 ohm part.
Operation is straightforward. Each switch closes successively as the water level rises. As each switch closes, one transistor lights its own LED and another transistor bypasses the previous stage's LED, turning it off.
Each switch represents two probe wires maybe 1/2" apart. When water rises to their level, it appears as a resistance between the probe wires. The only real problem with this circuit is that water resistance is too high to drive a bipolar transistor, let alone two in parallel. To barely saturate both transistors would require a water resistance of 18K or less, while real water could be several hundred K.
Using a second transistor to inhibit the previous stage LED rather than the previous stage driver works well, and is in fact less complex that using steering diodes.
As above, the only real problem is the inadequate drive current caused by the high resistance water. A solution is to change the transistors to small-signal n-channel MOSFETs. Here is a reworked schematic with FETs. This version also replaces the bypass transistors with diodes. A diode is able to bypass an LED because while the LED's forward voltage (Vf) is around 2 V or more, a signal diode Vf is around 0.7 V. The diode clamps the LED's anode at a voltage too low for it to conduct and illuminate.
With this approach, even a water resistance of 1 M would still present a gate voltage of 6 V, more than enough to "saturate" (fully enhance) the FET.
This can be achieved fairly easily with the help of a microcontroller and a bit of code.
You can use an Arduino which is quite easy to start with.
Connect each Led to a pin of the Arduino and use what is called a PWM to vary their frequency.
Connect your sensors likewise to other input pins.
There are tons of example on Google on how to achieve that, they keyword is LED PWM control Arduino.
The best way you can do it is by a digital circuit. Create the truth table , do the K-maps required and find the boolean expression of the circuit and use digital logic gates.
It so much easier to do it using digital circuits than analog circuit because in the end if you try to do it with a analog circuit you will end up building digital logic gates , dont reinvent the wheel!
