OK, an update about the “strange” Battery behave for MiFlora Sensors…
Here is the history of my three plant-sensors and their battery state over the period of March…
I would not expect an increasing of their battery state - but again, I guess, there’s an issue in the firmware.
The Battery state will probably be calculated out of the voltage of the battery.
And I do think, that this value might increase when the sensor will be hit by sunlight and warming up.
Therefore, the resistance is changing and the voltage information changes → then, the sensor will report a higher battery state …
Why do I beleive this?
→ the light blue device (bogenhanf) is placed on the floor, while the others are near a window and therefore will be hit when the sun is shining through the window and warming the sensors up.
Let me ask the question, do you really need to know the battery state every 6 hours?
Pulling the battery state is an expensive operation for the miflora sensor. It requires an active connection be made to the sensor, that is accomplished over multiple operations.
A Miflora will passively beacon their state down to about ~20% of the battery, sometimes even lower. The active connection required to obtain battery status requires a battery which is closer to 40%; the act of requesting the battery status will sometimes cause the miflora device to go offline/do-a-hard-restart.
I currently check the battery state every two weeks; a miflora sensor will run for months, the battery declines along a graceful curve.
My Mi Flora’s previously had great battery life but in recent months I’ve noticed them getting drained extremely fast. I previously used the ESPhome component to manage them but swapped over to having the ESP’s act purely as Bluetooth Proxies and letting HA manage the Mi Flora’s… It seems that the battery life took a dive after making this change.
How fast is “extremely fast”? I track my battery changes with Grocy and using the same setup as you (ESP proxies) and I get around 8-10 months between changes.
In my experience the percentage values that the sensors report are not reliable. Most of my sensors (16) can report 100% from the day they have a fresh battery until they die.
In HA, Bluetooth>settings you can change to be only passive scanning. I personally did not like it to be directly in HA because I have many of temp/humidity Bluetooth sensors and this mi plant sensor and they are sending a lot of readings to HA with decimals in up and downs. now you in HA you can remove decimals but still a lot of data. So I prefer to still use esphome in esp32 to get data in passive mode and filter for my needs. I removed battery level to be sure nothing depletes battery faster. Buying often button battery it is more important to me that battery level.
I haven’t had any issues with the battery life of the MiFlora Sensors so far… they are still all working fine.
BUT:
The information given by the sensor is just poor…
I have one device, which is reporting 0% State of Charge for quite a while now… yet, the sensor is still working fine
I’ve added template sensors that detect when a plant sensor has stale data, I found that more reliable than the battery sensors. I am using the illuminance senors with a time frame of 12 hours.
Edit: I had a closer look at my local hardware shop (Bunnings). There were only a couple that were 3.2V. They tend to be the higher lumen, higher cost ones (so not so tempting to tinker). Most of the cheap ones are 1.2V. Quite a few use 3.7V (especially in the higher lumen range).
The Plant Integration is creating a “device” for the plant, with the ability to add sensors from different sources (FlowerCare Sensors, Humidity Sensors, and whatever).
In Combination with the OpenPlantBook integration, you can also pull the specific parameters for your plant from OpenPlantBook.
All will then be included in a Card like this:
And this is how my dashboard is now looking (there are two plants with no images available at the moment)