-
Notifications
You must be signed in to change notification settings - Fork 151
What to Expect from the Sensor
Some users experience an initial drift in soil moisture readings when b-parasite is first deployed (see #2, #19, #50, #104).
It make take a couple of watering cycles for the sensor to fully accommodate itself in the soil and for the readings to stabilize. While it happens, it is not unusual to see upward drifts in the sensor readings.
The calibration has been done by taking readings while the sensor is submerged in water and while completely dry. This was done while varying the input voltage, in order to account for the natural discharge of the battery. For each scenario, a second-degree polynomial was fit over the collected data.
This simple model does not account for other factors that can impact the readings - for example: air humidity, temperature, different types of soil, proximity to plants roots. You may experience drifts in readings due these unaccounted variables. This also means that a sensor out in the air may read values that are greater than 0%
.
While b-parasite may output readings with arbitrary precision in the [0%, 100%]
range, the issues discussed above increase the uncertainty and variance of its values. In practice, a good mental model is to treat its output in ranges of 10%
to 20%
, effectively reducing its output space to 3
or 4
bits.