The lion’s share of soil moisture monitors we see are meant as add-ons for a microcontroller. So we’re glad that [Miceuz] tipped us off about this soil moisture alarm he built with analog parts. It’s really not hard to take the concept and build it in the analog world. That’s because you’re just measuring a resistance value. But for those of us who never really got started with analog parts this is a great project to learn from.
A high-efficiency op-amp is doing the brunt of the work. When the soil is moist the resistance is rather low compared to a reference voltage provided by a separate resistive divider. But when the plant gets thirsty and the soil dries out the resistance increases, triggering the op-amp to illuminate an LED and create some noise on the buzzer (we’re a bit confused on how that buzzer works).
Unfortunately this isn’t a viable long-term solution as the battery calculations show it lasting only about four months. That’s where a microcontroller-based circuit really shines, as it can put it self in low-power sleep and wake infrequently to take readings.
19 thoughts on “Analog Soil Moisture Alarm”
With a row of plants a 12v 1A wall adapter would power a whole heap of these things.
Needs a setting to adjust how wet/dry the soil gets before the alarm activates. Some plants like it wet, and some dryer.
Slather the whole thing in silicone sealant and its set for outdoor use.
Instead of slathering the whole thing in silicone (though that can be fun) why not keep the unit indoors and put the probes on the end of a cable?
That’s exactly what i was thinking about… Build one unit and use multiplexer (digital or even analog one) to take measures from multiple electrodes…
I bet the analogue ones you buy at the garden store wouldn’t run for four months on a set of batteries if you left one on!
I bet you could use a trimmer pot to adjust the reference voltage to change the setting for how wet or dry it is.
Monocrystaline PV solar panel perhaps.
There are some long term difficulties with this solution:
Different pH-values, chemical changes on the surface of electrode material, density of soil etc..
The result is a drifting impedance value.
I’m not sure about the electrode materials. I would guess, that copper for instance, brings an inhibition of microbial activity.
Maybe there’s a better way by using capacity changes of the soil (not tested yet).
Some “analog” moisture readers are a zinc and a copper probe attached to an an analog meter movement. (maybe they have a shunt resistor and/or resistive divider. Moisture in the soil will cause a voltage differential to develop across the probes and the meter deflects to the amount of moisture. If the soil is dry, no voltage, meter reads low/dry.
I guess copper and zinc electrodes would corrode quite quickly and readings would drift with time
Yeah, the one I had needed to be removed from the soil on occasion and the probe cleaned/polished.
I believe that biasing your sense electrodes with DC is a bad idea.
A better solution, I think, would be to use an oscillator to provide a small AC bias. You can sense the return with an op amp, then rectify it to provide you with a DC value that can be measured or used in a threshold detector.
I’m thinking of a circuit similar to those people have homebrewed to measure ESR in electrolytic caps.
The result is a current flow that won’t favor the erosion or deposition of material from probe over another, and the extremely low current means little electrolytic action in any event. Use stainless steel probes.
What about using a 555 timer and a power MOSFET to emulate sleep by switching the whole system on and off intermittently? Set a sufficiently long period and a relatively low duty cycle, and include a way to override it or at least keep the LED/buzzer/notifier-of-dryness on once moisture drops too far. Granted, I don’t know whether that would save much power over the existing arrangement, but it seems like it has potential to help.
I thought about 555 in the beginning, but it’s a huge power hog compared to all those nanoamp op-amps.
Another way could be to set up one low power op-amp as relaxation oscillator that switches power on and off, but it gets really hairy, power consumed would be somewhat similar to a low power attiny with watchdog running, and complexity increases with no appearent benefit, except c00l hax0r pride of course ;) hm, maybe i try that ;)
Oh, okay. Thanks for the comparison, and for sharing this neat hack!
Pick the right design and analog beats all. First there are much lower power opamps available. Second by dropping the 12V supply significantly you reduce power, cost and possibly gain some electrochemical benefits which prolong sensor life.
And then you can consider a relaxation oscillator, but you won’t need it, the nails will be gone by the time the battery dies. Digital has significant costs and is not always the way to go. Adding a timer though is not a big complexity increase.
BTW, buzzer does not sound, it only clicks loud enough when it gets powered and when power is turned off.
Ahh, that makes more sense since it’s just a piezo element. Cool!
Another problem with this design is that the constant low current in the same direction will corrode the probes and cause the measurement to drift over time.
Much better design at: http://sbolt.home.xs4all.nl/e-groeneVinger.html
It is cheap, ultra low power (runs nicely from a little solar panel), and uses a balanced square wave to probe the soil.
I’m looking to build automatic vegetable garden control soon and have been thinking about a number of home made moisture sensors. Pity though a non-contact method couldn’t be used. Something that can be completely sealed against water. The electrical properties of the soil should change with moisture content, so I was thinking of trying to play with capacitance and inductance.
I have made one with a simple opamp (a 741 that I had lying around) but also with a attiny85. In that one I just feed the electrodes from one of the output pins that I switch on just a few times a day to take a reading. saves a lot of corrosion.
Currently thinking of maybe using 2 output pins to feed it (one low, other high) and then from time to time go through a polarity reverse cycle.
I dont think capacitance measuring is the answer as that also would have metal plates in the soil, unless you’d want to keep them isolated as part of yr dielectrium.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)