Whenever I switch on the circuit, the LEDs glow fully for a second and gets off. After 2-3 seconds they start glowing slowly.

What might be the reason for the above?

 

Your wiring between the Probes and the Meter have probably created an Oscillator of some sort.
They can also act as Antennas and pick-up Electrical-Noise.
Try putting a 100nf Capacitor on each Input to Ground.
It would also be a good idea to use a Shielded-Cable.
The Input-Resistors are probably unnecessary.
.
.
.

 

Thanks for the reply. Can you elaborate where to put the capacitors?

 

Also, I'm using a 9v battery.

 

Install the caps from pins 1 thru 5 to ground on the ULN2003A
BTW what does that little push button switch do? It's not on the schematic.

 

Install the caps from pins 1 thru 5 to ground on the ULN2003A
BTW what does that little push button switch do? It's not on the schematic.

Thanks for the reply. Little push button switch is for kept for testing purpose for other circuit.

 

Your sense electrodes may be getting coated with hydrogen and oxygen bubbles, causing changes in the resistance between the electrodes and the water or whatever you are sensing. Lowering the current through the electrodes might help.

 

Also, I'm using a 9v battery.

Perhaps the battery is failing.
Make sure it's a good quality, new one. When all the LEDs are on the current draw is about 25mA, so a little PP3 battery won't last long.

 

Your sense electrodes may be getting coated with hydrogen and oxygen bubbles, causing changes in the resistance between the electrodes and the water or whatever you are sensing. Lowering the current through the electrodes might help.

It turns out, thanks to another thread here, I learned from some research the minimum voltage for electrolysis of water is 1.228V. I found this surprisingly low, but the chemistry demands it so who am I to argue?

 

Perhaps the battery is failing.
Make sure it's a good quality, new one. When all the LEDs are on the current draw is about 25mA, so a little PP3 battery won't last long.

Thanks. If the battery fails then the LEDs shouldn't glow after few seconds.

 

Our motorhome had a similar arrangement, but used a moving coil meter. It was set up with a momentary on switch to minimise electrolysis but didn't work for long and was never very accurate. When I removed it the stainless steel anode in the tank was coated. I reused the wires and fitted a float switch 25% up the tank that brings a red LED on, and that is very reliable.

 

I suppose some enterprising person could detect water level more consistently using thermistors, which could be electrically isolated from the liquid.

 

There are a number of ways to measure water level while isolating the electrical components from the conductive fluids. The method used will not be dependable due to corrosion and as others pointed out electrolysis. A float system with variable resistance as output, ultrasonic sensor, or a load cell are a few possible solutions.

 

IF the tank is non-metallic, there is a novel method for using a capacitive sensor on the outside to detect the level continuously, though it is somewhat complicated.

Since the current instrumentation is only detecting discrete levels anyway, using a set of single capacitive sensors at the desired intervals would also work with no need to immerse antything.

This sort of thing can be attached to the outside of the (non-metallic) tank: (chosen at random for cheapness)
https://www.amazon.com/DollaTek-Non-Contact-capacitive-External-Monitoring/dp/B099DF6DHM/

​

 

Install the caps from pins 1 thru 5 to ground on the ULN2003A
BTW what does that little push button switch do? It's not on the schematic.

After putting the 100nf capacitors from pins 1 to 5 the same has repeated. It didn't work out.

 

How could I correct it?

 

Your circuit uses DC which causes electrolysis and plating/corrosion on the probes. That is why most soil moisture detection circuits use AC at the probes, not DC.

 

Years back when playing with (forbidden topic) I used a float with a magnet inside it. When the water level dropped so low the magnet activated an exterior reed switch which turned on a pump. When the magnet was far enough away from the switch it opened and shut off the pump. Something similar could be done with a series of reed switches at differing levels. Give me some time and I'll bang out a picture for you.

 

With this system you can have as many LED's as you like. No need for logic chips or any kind of submersed sensors. Just a simple float with a strong magnet internally will trigger each reed switch. As the level drops the level of LED's will drop accordingly. In this arrangement there are diodes that provide a current pathway to each lower LED while blocking the upper LED unless the level is at the top. Each diode will have an associated forward voltage (Vf) drop, so I'd recommend a higher voltage over a lower voltage. With the lower voltages you'll have too much Vf which will cause each lower LED to be dimmer. But with super bright LED's even the lower brightness will be minimized. 12 volts DC would be better than 9VDC. Better still, power it from a 12 volt wall wart. You don't need much amperage. One amp should be sufficient to light all the LED's simultaneously.

 

I like the idea but wouldn't the magnet need to be always in contact with a reed switch or none of the LEDs would be lit.
Depending on the size of the magnet and tank seems like it would need a lot of reed switches.