Moisture-Duck Sensor: Measure Soil Moisture with ATtiny13A and Duck-like Precision

In this video, I'll explore several different methods for measuring soil moisture, including gravimetric, neutron probe, tensiometer, time domain transmissometry, time domain reflectometry, capacitive, and resistive methods. I'll explain how each method works and discuss their advantages and limitations.
I'll also show you how to make your own DIY soil moisture sensor PCBs for the capacitive and resistive methods. For the capacitive sensor, I'll be using an MSP430 microcontroller and for the resistive sensor, I'll be using an ATtiny13A microcontroller. I'll go over the design and assembly process step-by-step, and provide all the necessary code and schematic files.
Whether you're a gardener looking to optimize your watering schedule, a farmer managing large-scale irrigation systems, or a student or researcher studying soil moisture dynamics, this video will provide valuable insights and practical knowledge for measuring soil moisture using different methods and building your own custom sensors.
Don't forget to subscribe to our channel for more DIY electronics and sensor projects!
Link to the project files:
github.com/Mak...
▬ Contents of this video ▬▬▬▬▬▬▬▬▬▬
0:00 - Intro & summary
0:20 - Gravimetric Method
1:13 - Neutron Probe Method
2:03 - Tensiometer Method
3:43 - Time Domain Reflectometry (TDR)
4:57- Time Domain Transmissometry (TDT)
5:22 - Capacitance Sensors
9:09 - Resistive Sensors
9:46- Test of Corrosion on Resistive Sensors
12:14 - Making the Schematics & PCB
14:15 - Summary

[@Aleksei_R]

I used resistive fork from aliexpress like in your video long time ago. Beat the corrosion was very easy. I cut off the pcb fork with corroded copper and soldered stainless steel electrodes to the remaining part. I took stainless steel from used car wiper.(rubber part is held by two ss plates)Cut the Stainless to the same lenghts as pcb was and insert to the pot, so the moisture and soil only contacts stainless plates and not solder or pcb.

[@makersfunduck]

Hi @Aleksei_R, your idea is really interesting. It never occurred to me to use stainless steel. indeed stainless steel will corrode with a slower rate than copper or tin. What about the solder? For example I tried aluminum rods, but they do not hold the solder very well. Is it easy to solder stainless steel?

[@Aleksei_R]

@@makersfunduck after about a year of usage there was not remarkable corrosion if there was any. And it worked in analogue circuit, where current runs all the time. Soldered relatively easy. But i can not give any flux recommendations , because i used self made flux from rosin and soap boiled in vinegar and water🤣🤣🤣

[@AdityaMehendale]

I am planning to make s similar device, with some differences: 1) small solar cell instead of battery; flashing LED instead of buzzer 2) Preferably no MCU - just CMOS/digital; ATTiny, MSP430 STM32 as a last resort. 3) It is true that DC corrodes the electrodes. Also correct that capacitance method in side-by-side electrodes if iffy. This can be countered by having face-to-face electrodes with a water-permeable membrane (e.g. a sheet of microfiber "paper") sandwiched between coated electrodes. Permittivity-changes should IMO become obvious and clear in this configuration. The paper may eventually get contaminated by stuff in your potting soil. Resistance-measurement is also possible, but IMO "AC" measurement (i.e. with an AC biasing stimulus, with each electrode having a series-capacitor) would eliminate galvanic corrosion even if you did not run the device intermittently. The problem with DC interfaces is that even very small leakage currents, over extended periods, shall eat away the electrodes (and cause electropotentials) whereas the DC-insulation offered by a plastic-film capacitor is outstanding, robust and cheap. 4) about your "GHz" problem: (out of academic interest) --> An ESP32 has a 2.4 GHz pin, and RSSI indicator ;) - I wonder if this can be repurposed for humidity measurement. (self-attenuation due to moist soil)

[@makersfunduck]

i think your solution work, but if you can use at least a timer, it would definitely increase the life of the electrodes. considering more or less a cheap mcu more or less costs same with a 555 IC, MCU would make more sense. But your ESP idea is really worth investigating and I find it really interesting. but their power consumption is questionable even during deep sleep. Still I will look at the datasheet for a bit deeper:)

[@AdityaMehendale]

@@makersfunduck I was thinking lower tech - sth like a single-gate 74HC14 and a 16-bit (Divide-by-65536) counter. Slow flash --> all ok; fast flash --> water!

[@makersfunduck]

@@AdityaMehendale Sounds really interesting. I would love to see your design, if you make it open sourced.

[@neworginallogin]

I wonder if you could use graphite probes (like pencil cores) to reduce corosion even more

[@makersfunduck]

It will also corrode. and you need to replace the probes as well. but it is an interesting question. I will try how long it will take. Also I am wondering if they provide reliable results.

[@hanyanglee9018]

what about rf energy, like 50mhz, does water block it ? Or if water changes Z0 of a coax, it should be possible to test it.

[@makersfunduck]

I tried capacitive design on different frequencies. around 10mhz, you start getting feedback once you put it in water. 50 mhz is also better. however once I put it inside a real soil, it did not work, and academic papers requested that if there are gaps inside soil, those frequencies does not provide a reliable results, as I observed. But around 1Ghz , it was working fine. however I couldnt find a cheap MCU which can work on these frequencies.

[@hanyanglee9018]

@@makersfunduck So, how about heating up the soil a bit and test the difference of temperature? Let's say 5v on a 100ohm for 30 sec and test the temp before, after and 2 min after? Should be way cheaper.

[@makersfunduck]

@@hanyanglee9018 I have found most relevant these 2 papers discussing your proposed solution. sci-hub.se/doi.org/10.1093/jxb/20.1.46 www.proquest.com/openview/497729083bc894de658f1591fac75101/1?pq-origsite=gscholar&cbl=52938 What my take from these (and other papers) is, it can be done, if you calibrate and get the real soil content for the type of the soil that you have. it is different for different type of soils However it wont be a low power design, since to get a decent result, you need to apply considerable amount of current (in the paper they used around 350mA to get a decent results). Also I am not sure if it would be cheaper, since you are adding another component (temperature sensor) to the BOM. But I think it is indeed an interesting concept that worth writing another scientific paper :)

[@zoenagy9458]

There wasnt any conclusion, are you saying resistive is better than captive?

[@makersfunduck]

There isn't a conclusion about which one is better. My conclusion is that both of them have ups and downs, and users should know their advantages and disadvantages. The resistive version provides reliable results, but corrosion is the biggest issue. Therefore, you should do the measurement sparingly. The capacitive version doesn't have this issue; however, it is not suitable for all soil types unless you can go to GHz-level frequencies. So you should have certain types of soil without air gaps and insert the probe very carefully to get a trustworthy result.

[@zoenagy9458]

@@makersfunduck thank you that was useful! we might not need high frequency, just output a square wave and measure local edge vs return edge with 32bit hw timer, the difference would be the propogation delay, and repeat multiple times.

[@danheidel]

@@makersfunduck One technique that is sometimes used with resistive probes is to encase them in a known porous matrix such as plaster of paris. This reduces the effect of different soil types since the moisture measurement is always done on the same material. There might still be issues with variation as different ions migrate into the matrix but it should help. This is an older method and so I'm not aware of any testing done with it and capacitive probes but I imagine that it would also work for the as well.