Desk of Ladyada - Desk plant health monitoring with SproutSense

Exploring soil moisture sensing with our prototype, SproutSense, to monitor plant health. After neglecting a plant, we're motivated to improve and use adafruit.io for data tracking. We're puzzled by rising capacitance over days and exploring various moisture measurement methods. Plus, we're seeking a slim, low-cost SMT 20mm coin holder for a sensor update.

[@mloik1]

Plant scientist and major fan (and customer) of Adafruit here. State of the art soil moisture these days is done with time domain reflectometery. Perhaps if you get fast speeds with the square wave PWMs you could approach that? Similar sensors, such as that from METER, have the separation gap up the middle of the two prongs like you mentioned. An ESP32 version would be awesome, lose all the cabling. The cable memory with small sensors is a giant pain as you noted. That coin cell would have to be well-sealed to avoid dirt and water ingress; greenhouses, fields and the even the surface of individual pots are incredibly dirty, wet places for electronics. The light sensor might not tell you much about light in the open but sometimes light under the canopy is of great interest. In fact, comparison of light above vs below the canopy tells us how much light was absorbed by leaves; using the Beer-Lambert Law we can infer the Leaf Area Index of plants (LAI = area of leaves per unit area of ground covered.) As that changes over time, it is a good indicator of plant growth.

[@BrummBilt]

Seal the edges of the PCB with conformal coating. The light sensor may not tell you much under the plant canopy. Temp and humidity is useful if you want to see how those affect soil moisture (science experiment), but would be a microevironment that close to the plant (which is fine too).

[@john_hind]

NFC would be good for the plant sensor. It could be a hi-tech version of one of those plant label stakes. Wave your phone at it and it brings up an info page for the specific plant, shows the logged sensor data and maybe even gives you some care advice based on the sensor info plus the plant type.

[@JehovahsaysNetworth]

Try using a portable solar light to interpret plant cell biology synapses and merge the Computer Hardware and use ai to communicate with the plant. I like your educational video.

[@KarlMiller]

I have seen several designs for these small PCB soil sensors but I could never quite understand two issues:
1. Why is it so important to conserve board space for a soil sensor? Are people only using these to monitor soil for tiny plants? Who cares if the pcb is slightly larger - a larger PCB could fit more/better sensors and even a bigger battery and antenna for wireless versions.
2. Why such a short sensor when soil sensing is most importantly useful for the plants roots? Roots can go deep for some plants and the location for sensing can change based upon watering technique.
Also, the capacitance traces could be farther from the edges of the PCB to limit seepage or even coat the base in an UV-cured epoxy.

[@john_hind]

Have you considered a software solution for I2C address changing rather than the board space, cost and GPIO disadvantages of a DIP switch? SEESAW could have a standard I2C transaction to set the address in non-volatile memory and then it comes up at the new address after power cycle. There could be a maintenance script which hunted for the current address and changed it to the desired address with just one device on the bus. Or I believe the AT-Tiny chips can respond on two addresses, so there could be a reserved address for config and again you plug the boards in one at a time for configuration. There could be other config settings as well, such as setting the power-on state of GPIOs.