I have an automatic irrigation system that uses water from a bucket, and I installed it in a way that I would be notified if the bucket level went low. It stays in a hard to reach place, so that’s why there’s a sensor there.
However… this is the second sensor that seems to be faulty. I know, I know, most generic AliExpress / Tuya devices are not trustworthy, but today when I was facing the “your device is dry” false warning again (which is on for, idk, more than a week already?)… I was thinking: should those devices work fine when the probe is submerged in water for weeks in a row?
AFAIK, all such devices (like mine) work with two or three probes, and when it gets in contact with water, those probes get shortened and then the device detects that as “I found water”. But could it cause the sensor to go haywire if the situation is the opposite and the probes are constantly shortened, and only rarely go dry?
I’ve seen some water-level sensors already, but they’re too expensive for something I could do mostly fine with a backup alarm to check water every 10 days or so. Also, while I know ESPHome could help me getting custom stuff for my home, I’m not yet ready to learn electronics and spend time with such projects. That said, I’m open to other simpler suggestions for this scenario!
Don’t worry, you’re not doing anything wrong now that you’ve clarified.
You shouldn’t be getting false readings if your probe is submerged. If your device is showing such readings intermittently, then maybe set up an automation to only warn you if it’s in that state for x amount of time.
It’s not intermitent, but I wasn’t able to pinpoint the frequency either. It simply looks like the device starts to think it’s dry and that’s it, it stays that way until I actually have to refill and try to reposition it.
I’ll try further investigations and start annotating when it happens!
With any form of measurement and electronics, the question is
do I measure a thing, or do I measure a change in the thing?
With water, finger, presence, PIR, heat, smoke etc… we can measure something (which usually ends up looking at current through or voltage across a resistor) or we can detect a (significantly) rapid change in a measurement.
Since measuring ‘things’ is fraught with accuracy, calibration, stability and battery-draining issues, looking for ‘change’ is always going to be much easier (and less demand on the battery).
So, plunge two wires into water and yes the conductivity changes.
Plunge two wires into water, and wait, and yes nothing much changes.
Smoke detectors don’t detect smoke - they detect a change in optical translucence, or heat, or particulate scatter of radiation.
Dry to dry - there is no leak
Dry to wet - there is a leak
Wet to wet - same as dry to dry to dry, so must be no leak
When I was younger, and we did not have fancy electronics, we used floats on long string with a micro-switch. That worked.
I am using a similar device for my pool but having it constantly showing ‘water’ seemed to drain the battery quickly. I attached this, instead of the two pins, so now it only shows when the level reaches a lowest point. There are other solutions which can detect the level, this is binary
@Biscuit yep, I do remember water “sensors” my dad used in the building we lived when I was a kid - a float which toggled off the street pump. That was smart tech! haha
But yeah, besides that, I’m still unsure what might be going on. I get it’s not supposed to change when nothing much changes, but what gives if it reports the wrong state for days in a row?
@vingerha I DID see those devices on Ali, when looking for some cheap-ish and more reliable option to measure water level! I had no clue what they were or how to use it, figured out it was just a component for DIYers. And I still have little clue haha… How did you adapt the usual sensor I posted with that gadget?
PS: just while I write this, I got another faulty notification to fill up the irrigation bottle
I only connected the Ali floater to the already available wires, just removed the 2-pin thing. When it reaches the low end, the sensor will trigger, that is the only thing I need. As it is in a pool and the water moves with people swimming, I set alarm only if it stays ‘on’ for 1 minute…you should not need this I guess.
EDIT: battery change in 1y+, with the two pins in less than 2 mnths
EDIT2: after a second coffee, I may have soldered it (was 2+y ago) and can show you if you have issues. I already had the leak-sensor and do not need ‘level’ so this was the cheapest option at the time, I bought it with the risk that it would not work but that risk was 5EUR…easy decision
I suspect water leakage sensors are meant to be dry almost all the time and only let you know when there is a “black swan event”. I expect its electrodes might corrode if persistently in water. A capacitance sensor can be made of a thin metalic strip and laminated in plastic and left dipped in the water and last longer and even give idea of how deep the fluid level is.
I have tried two different tuya ZigBee water sensors submerged in water in a pail and they initially show ‘wet’ but after a few minutes they show dry even when still submerged. I was able to repeatedly recreate this. They somehow reset to dry on their own. This was not good for my use case and I stopped using them.
