Soil Moisture Monitoring: Capacitance vs. Temperature

This video summarizes the first seven weeks of soil moisture monitoring in a shallow rainwater harvesting basin using both soil capacitance and temperature sensors wired to Feather MO Datalogger.
Results suggest temperature may be more effective at capturing rainfall infiltration and soil moisture dynamics when compared to the ubiquitous soil moisture capacitance sensors sold on eBay.
Please see this playlist for further background and details: • Soil Monitoring Experi...
Please see this video for details regarding instrument setup and links to code: • Long-Term Soil Monitor...

[@Blueberry8737]

Awesome video as always. I have plans on trying to monitor my yards soil moisture content to know when to water and this research is super useful. I can't remember if you mentioned the large quality control issues with the capacitive sensor that many other TH-camrs have shown? Also it might be beneficial to invest in a high end soil moisture sensor to see if you can draw a correlation between temperature and moisture content like you are theorizing.

[@ModestMaker]

Thanks for your feedback. Yes-- I am aware of the quality control issues on those cap sensors and selected them accordingly as summarized in this video: th-cam.com/video/lb4HdcpYWCQ/w-d-xo.html. The ones that were installed all met QA criteria and demonstrated good repeatability prior to installation. Bottom line, I really wanted these sensors to work, but I don't trust them for all the reasons outlined in prior videos. Sadly, they just don't respond well to a drying profile among other issues.

[@donaldhoudek2889]

Great video! You may want to enclose the chip/connector end of the capacitance sensors in either a Heat Shrink material and or an acrylic coating to insure that the moisture is not somehow having an affect on the sensor. Did you seal the sides of the sensor PCB? Keep the videos coming

[@ModestMaker]

Agreed. Yes-- I thoroughly covered the electronics on the cap-sensors with electrical-grade silicon given the potential of moisture impacting the same. Another possibility is the wire connection to the analog pins on the Feather MO needs to be tightened. I will check that soon but frankly, those sensors are not something I feel comfortable relying on outside of detecting an initial pulse of infiltration. The experiments I've tried in the past demonstrate the response curve is not linear and thus are not likely to give you a good indication of soil moisture as things dry out. It would be great if we could make these cheap sensors work but I don't have much confidence.

[@J0stik]

hello, i am using those soil capacitive moisture sensors in clay soil and in coco coir/perlite mix .... clay soil, response from sensor is not as good, used two point calibration, from permanent wilting point to field capacity (i basically let my peppers wilt, measured response from sensor and then fully saturated soil, waited one day to let excess water be drained and then measured response from sensor) using sensors with tlc555, so far i needed to replace 2 of 3 sensors, all of them were coated with epoxy and insulated circuitry, and response from them was good for only few days ( soil was dry and sensors were showing moisture above field capacity) @@ModestMaker

[@J0stik]

other thing is coco coir /perlite mix ..... response from sensors fantastic, using two sensors in one planting pot .... difference in those two measurements is max 2percent, validating data with weighting of dry pot / weighting watered pot / reading sensor data and it is +/- 3 percent, and that is something i can live with :) @@ModestMaker

[@ModestMaker]

@@J0stikThat’s great that you got the sensor to work in your mix of soil backed by data rather than just observation. I considered soil texture as a factor which is why I always qualify my results as being specific to a silty clay. It intuitively makes sense to me that these might work better in lighter soils.

[@hardwareful]

Didn't I comment under an earlier video that these capacitive sensors are garbage? Well, there you go. Some slap concrete or clay onto the cheap capacitive sensor PCBs in an effort to create a more stable measurement zone, but once it flakes off the PCB, it's game over again.
All in all, I doubt they're worth out time in any regard. Better methods are RF absorption, time-domain reflectometry, ...

[@ModestMaker]

Sure, but now there's hard data that demonstrates the challenges rather than anecdotal claims. Having said that, there may be ways to redesign the RC constant to be more sensitive to soil moisture conditions (eternal optimist). I agree there are better methods, but not everyone can afford tensiometers, time-domain reflectometry, etc. The idea is to develop and tests alternatives to expensive venues that may not be accessible to broader public. In this regards, there may be something we can salvage in temperature by looking at the change over time (working on it) :)

[@hardwareful]

@@ModestMaker I still have a bag with 15 sensors of said type. We used them in a little indoor farming experiment, and logged their readings over time. Sadly that failure was not documented well, so points go to you for doing the science :) ironically, you might be able to bury an ESP8266 in a water tight enclosure and measure RSSI at the top, or at least between two sensor units spaced 10-20cm apart.
One of my experiments involved the use of a 2.4 GHz directional coupler. An RF antenna is detuned when the soil around it gets wet. When you step the frequency, you can sample the S parameters (think CC1101, CC2511 and the likes) - this is more of a volumetric measurement than the capacitive one (which is surface-selective), but a lot more sophisticated to design.

[@ModestMaker]

That RF design sounds pretty cool and very innovative. Is there a site or video where I might learn a little more?