Have you figured out some filter combination that gets rid of enough noise while not adding a significant lag, and that is a nice balance between amount of data and precision?
Indoor temperatures do not change very fast so I use this:
sensor:
- platform: dht
pin: D5
model: DHT22
update_interval: 15s
temperature:
name: "Bathroom Temperature"
filters:
- sliding_window_moving_average:
window_size: 15
send_every: 15
send_first_at: 15
Updates every 3m45s with the average over that interval.
If you want more responsive but without flooding the network with messages you could use something like this:
update_interval: 1s
filters:
- sliding_window_moving_average:
window_size: 10
send_every: 1
- or:
- throttle: 120s
- delta: 0.2
This sends a 10 second average every 2 minutes unless the temperature average changes by 0.2°, if that happens the average is sent immediately.
6 Likes