Thermal Soil Moisture Sensor

Design by Node257 | Licensed Under CERN-OHL v2

Summary:

The thermal SMS uses the high specific heat of water in the soil to measure moisture content. The current design achieves this using a combination of a heating resistor and a thermistor. A separate (or on-board) microcontroller does the following:
1) Heats a resistor buried in the soil, at a precise voltage, for a set period.
2) Measures the temperature change in the soil, and calculates the moisture content.
3) Outputs a value in any desired unit, for example: 5% (Dry) – 95% (Saturated), or plant-specific actions: ‘Needs Water’, ‘Too Wet’, Etc.

After calibration for a particular soil composition, the signal should not drift due to corrosion. Since both the thermistor and heating resistor/pcb heater can be insulated from the soil, the sensor itself does not interact with the soil electrically at all, exclusively using heat generation and temperature change to calculate water content. The system only needs to heat a small amount of soil by a fraction of a degree to obtain useful data from the thermistor, so it doesn't need a large battery. The soil will very quickly return to a nominal temperature, and any moisture evaporated from around the heating resistor is replaced through diffusion/wicking. As long as the unit only took measurements a few times a day, there is no reason the thermal aspect of it would have any measurable effect on the soil around it, such as chronically drying it out. With a small solar panel and battery attached, the unit could broadcast sensor data over RF for extended periods with no maintenance.

This could also be achieved inexpensively by interfacing with a cheap solar garden light. They include a solar panel, battery, and microcontroller; for less than $5 each. Simply tapping off of the battery to power your own microcontroller would work in that situation and solve the power issue. Those lights are small, but mostly empty. You could put your own stuff in them. My objective with this project was to make the first version as low cost as possible so it could be used in schools or agriculture in large numbers. However, it could also be useful to design a high accuracy version with a thermocouple amplifier (such as the Adafruit MAX6675), and a waterproof probe, for more demanding applications.

Theory of Operation:

Amount of Heat / Change in Temp = Quantity of Water
It takes 1 Calorie of energy to increase the temperature of 1 Cubic Centimeter of Water by 1 degree Celsius.

Examples:
A sensor in the field may have a 100 Ohm, 1/4 Watt heating resistor that it heats at 5 volts for 30 seconds, which will produce about 7.5 Joules or 1.8 Calories of heat. If the soil between the heating resistor and the thermistor has an approximate volume of 2 cubic centimeters - that 1.8 Calories has a measureable effect of temperature, maybe 0.5 degrees Celsius, depending on soil composition and moisture content.

E = P x t
0.25W x 30s = 7.5J
1.8 kCal / 0.5°C ≈ 50% H2O Saturation
(For a particular Soil Composition)

About Me:

Hello! I'm an Electrical Engineering student from Denver. This has been a pet project of mine for years, but it's time to just release it. I developed another design using different physics that you can find at the following link: Infrared Soil Moisture Sensor. I have listed both projects as 'work in progress', since the PCB designs still needs to be validated. I will continue to work on these, though work and school keep me very busy. Any donations will go directly towards building more prototypes, and are greatly appreciated! I am publishing the designs under an Open Hardware License to encourage others to refine and diversify the technology. I think it could be very useful for water conservation to have accurate, affordable sensors that don't drift or corrode. Thank you for checking out my project!