Final update on this topic. I’ve been fiddling for a week with the sensors, doing tests, posting data to Gemini, analysing the data and the conclusion is the following:
- The sensors work fine
- Erratic behaviour is due to my misunderstanding how the sensors are supposed to work and how plant biology works with different species, different soil types and different pots
- The way you pour water in the pot makes a difference
- The sensor should not be treated as a thermometer style of sensor. Values should be interpreted in broader context, not as an instant reading of the current state
According to Google Gemini, this chart, together with some photos of the plans, the soil, and the environment, shows normal and healthy behaviour of both the plants and the sensors.
Started with removing the sensors to record 0%. I tested this previously and they always record 0% outside the pot and 100% submerged in water. When re-seating the sensors in the pot there are some considerations:
- The blade should be placed vertically in the soil
- The sensor should be positioned midway between the plant root and the edge of the pot
- One must ensure firm contact between the sensor blade and the soil, pressing the soil to the blade by hand, if needed
Once properly seated, the sensors need some time to “settle in”. It takes time for capillary action to re-distribute moisture into the disturbed soil around the sensor. It usually takes 24 - 48 hours for this to happen. This is why frequent sensor removal or physical disturbance should be avoided. Capacitive sensors create an electromagnetic field in the soil immediately surrounding the blades. Moving them breaks the “bond” with the soil structure and resets the electrical baseline. If removal/disturbance is necessary, once should take note of this and interpret the data accordingly.
After about a day i watered the plants, that is the recorded peak on the chart. Depending on the actual water flow in the pot, the type of plant, soil and pot, the content of the “water”, whether it is clean water or has fertilizer in it (mine had some aquarium water for some flowers), the readings would be different. Some will peak, and start do decline, as the plat is sucking in the moisture, with or without a small climb (like the red Snake Plant or the yellow Yucca here). Heavy fertilization may oversaturate the sensor. The Yucca’s soil is good, not “swampy” as the numbers might indicate. Some will rise sharp and continue to climb. In some cases, the sensor will record water movement inside the plant as increased soil moisture (the Ficus and the Peace Lilly). As long as the rise stops and transforms into a decline, all should be good.
Changes in soil moisture should be observed in tandem with the changes in the local air moisture in and around the pot (that is why most devices like this have those sensors as well). The plant feeding process include a transpiration phase where the plant releases excess water through the leaves.
One final note. All sensor data should be verified in person. Inspect the soil, try to estimate using fingers or a stick. Also, have in mind that these processes are slow. Waiting a day to confirm the moisture trend should not kill the plant. Overwatering it, on the other hand, will.
Happy gardening!
). Probably off by 2% from my other humidity sensors in my room.
