water leak sensor

[karl222]

Nearly any small N-FET should work (BS270, BSS138, 2N7000, 2N7002.....)

Just one point:
These FETs are somewhat sensitive to electrostatic discharge.
So a serial resistor between gate and sensor and a small ceramic capacitor 10nF .. 47nF between gate and ground
gives at least some protection against static voltage.
Remember, this makes the sensor a bit slow (3..5 Seconds).

Standard NPN transistors won't work reliable with such simple circuit, it needs
more electronics around to make that work.
The reason: Standard Transistors are driven by current, FET's are driven by (nearly pure) voltage.
For a standard transistor it needs a low resistance of water on the sensor, about 10..100KOhm.
It needs a big sensor and a lot of water, some drops of pure water won't do.

Hi,
I don't understand what you mean exactly
it is clear that there should be a small capacitor between gate and ground
you mean gate --- resistor 10k --- water --- +

but it doesn't work. esp vdd goes to 3,0V only and consumes 120mA instead of 70mA.

any ideas how to get the MOSFET powering up the ESP properly. via the VDD pin..

thanks
karl

[kimot]

But do not forgot, that if you wake up your sensor only when water leak, you will never know, if your sensor is still alive after some time.
Better way is wake up sensor periodically and send some data to controller ( with battery voltage for example ).
And if controller do not receive any data for some period, it will send alert to you.
Then you have assurance, that your sensor is ready detect water leakage all the time.

You can use simply internal timer and wake up cca every hour with 20 mikroAmp. consumption in this time.
Or build something sophisticated and wake up every 2 hours with consumption 1 mikroAmps between wake ups.
Look here for inspiration:
http://www.kevindarrah.com/wiki/index.php?title=Pusher

[karl222]

You are right, probably the better way...

Thx

[papperone]

But do not forgot, that if you wake up your sensor only when water leak, you will never know, if your sensor is still alive after some time.
Better way is wake up sensor periodically and send some data to controller ( with battery voltage for example ).
And if controller do not receive any data for some period, it will send alert to you.
Then you have assurance, that your sensor is ready detect water leakage all the time.

You can use simply internal timer and wake up cca every hour with 20 mikroAmp. consumption in this time.
Or build something sophisticated and wake up every 2 hours with consumption 1 mikroAmps between wake ups.
Look here for inspiration:
http://www.kevindarrah.com/wiki/index.php?title=Pusher

I agree this page bring a lot of inspiration but still... I don't see this guy using deepsleep for waking up on a regular interval! Am I missing somethign so obvious that I can't see in his schematic?
I will brainstorm with myself (the best way to come up with bright or sh***y ideas ) as I'd like to build such setup for several sensors (including water leak which is in my high priorities when few months ago I had my ground floor flooded after a broken water pipe while nobody was at home... Murphy's law) and the "must" are (1) running on batteries (2) last at least 6 months between recharging cycle (3) send out battery level status to prevent sensor dying without alert

[kimot]

He do not use deep sleep because high consumption 20 mikroAmp and wake up interval max 74 minutes.
His solution uses timer to wake up possibility every max 7 200 seconds ( not wake up really - switch on ESP and through sw action switch off then ) or through external event.
Consumption 1 mikroAmp.
Will bi nicer "wake up" one time per day - maybe with some RTC chips with alarm pulse possibility or another long interval timer.

[papperone]

He do not use deep sleep because high consumption 20 mikroAmp and wake up interval max 74 minutes.
His solution uses timer to wake up possibility every max 7 200 seconds ( not wake up really - switch on ESP and through sw action switch off then ) or through external event.
Consumption 1 mikroAmp.
Will bi nicer "wake up" one time per day - maybe with some RTC chips with alarm pulse possibility or another long interval timer.

Yes got it, I was before looking the only IC I didn't know (TLP5111) and it looks perfect for such a job.
I've already a latch circuit I designed to allow micro to turn off completely itself, but it was for a project where the turn on was to be manually activated then the micro performs itw own cycle of actions and turn itself off completely once done using a GPIO as a trigger.
I think I will try to add this IC to my latching switch circuit so to have the timely wake up and as well the sensor wake up, thanks a lot for the link!!

[kimot]

Some inspiration:

http://www.esp8266.com/viewtopic.php?f=13&t=8315

[Finnarduino]

hi! is it possible to create water flow sensor by using sound detection?

How this works?
streamlabswater.com

[ziancolins]

Finnarduino, I think you should go to the youtube, and watch videos for understanding the process or work of water flow sensor.