Also the way you are powering the sensor may be an issue. Did you try powering the sensor directly from V instead of through a MCU pin?
I use a Kalman Filter for smoothing, see post#13.
I saw the earlier post with the Kalman Filter. Very interesting! I spent some time last night researching it. I plan on incorporating it in my program. For me smoothing isn’t an issue; I get smooth, consistent readings. It’s that the readings don’t change as the soil dries. It’s almost as if these probes are great at detecting a “dry to wet” scenario but I’ve been unable to get them to show the gradual “wet to dry” scenario that naturally occurs as soil dries. Could it be my power configuration? Currently, I have 3 capacitive soil moisture probes powered at once using the 3.3 volt pin of an Esp32. (Potted plants, one probe per pot.) I’m using a transistor as a switch to save power during deep sleep. I give the transistor 2 full seconds to “warm up” like Aggertroll pointed out. Take one reading per probe, turn the transistor off, then go back to sleep for an hour. I never touch or move the probes. Seems to me like I should see a graph like the one Aggertroll showed in Post 6. My graph after a week looks like a straight horizontal line, even though the soil has dried considerably. Then if I were to take one probe out and reinsert it, that reading drops instantly and shows the correct level of dryness. Like I said I’m still testing. I’ve now put the probes in sand, in a wet sponge etc. Thanks so much for your help guys, a lot of us are benefiting from this.
I think powering the sensors from the GPIO pins of a ESP32 is not the way to do the thing.
I have the same sensors in a cactus plant using soil I got from the Idaho Desert. When the soil gets down to 22% the motors fire off, the sensors work well and show good drying trend. I have, from time to time, after an initial insertion of the sensor, had the sensor not respond to dropping moisture levels. I reseat the sensor, there is a sudden drop, the pump kicks on and the sensor does not need messing with again. I have the sensor set where the water from the pump flows to and being its pooling. I push the sensor deeply into the soil, I wiggle the sensor to loosen the soil around the sensor and let the sensor settle into its new role. I give a spot of water to the sensor spot and watch it over the next few days for a drying trend. Once the sensor is set in the soil correctly sensing drying, I leave it alone.
Ok, excellent. Thank you. Right now I am using the 3.3 volt pin, which is basically the output of the Esp32 onboard voltage regulator which can supply approx 230 mA (each of these probes need 5 mA), So think I am good with my power configuration. I am going to try and create some space around the sensor by wiggling it in the soil like you mentioned. One thing about the peat/coco coir mix that I am using is that while it is an “airy” medium it also also sticks to the probes, so the moisture around the sensor may not be evaporating. Thanks for your input. I’ll keep this topic informed of my findings.
This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.