I am currently building a sensor with this component and so far it mostly works pretty great.
But when I try to utilize the deep sleep component of esphome no attribute updates seem to be sent to the coordinator for quite a while after reset even though there are no errors visible in the logs.
It seems like it needs at least 30s after boot to even send out a single attribute update. I find this extremely inefficient.
I assume there must be a better way since it is explicitly mentioned in the readme right?
My working ESPHome config
external_components:
- source: github://luar123/zigbee_esphome
components: [zigbee]
esphome:
name: zigbeetest
friendly_name: ZigbeeTest
esp32:
board: esp32-c6-devkitc-1
partitions: partitions_zb.csv
framework:
type: esp-idf
logger:
hardware_uart: UART0
i2c:
sda: GPIO22
scl: GPIO23
id: bus_a
sensor:
- platform: scd4x
co2:
name: "CO2"
id: co2
temperature:
name: "Temperature"
id: temp
humidity:
name: "Humidity"
id: humid
update_interval: 10s
deep_sleep:
id: deep_sleep_1
sleep_duration: 10min
run_duration: 40s
zigbee:
id: "zb"
on_join:
then:
- logger.log: "Joined network"
endpoints:
- device_type: TEMPERATURE_SENSOR
num: 1
clusters:
- id: REL_HUMIDITY_MEASUREMENT
attributes:
- attribute_id: 0
id: hum_attr
type: U16
report: true
device: humid
scale: 100
- id: TEMP_MEASUREMENT
attributes:
- attribute_id: 0x0
id: temp_attr
type: S16
report: true
device: temp
scale: 100
- id: CARBON_DIOXIDE_MEASUREMENT
attributes:
- attribute_id: 0x0
id: co2_attr
type: single
report: true
device: co2
scale: 0.000001
- attribute_id: 0x2
type: single
value: 10000Another issue with this configuration is that the sensor expects to be polled every 10min as per manufacturer spec as it does a re-calibration routine after a prefixed number of polls and with this 10s polling interval (which is needed since I need to try a few times for the attribute updates to actually make it through, even though I see no errors in the logs of the device) the assumption about the number of polls before the sensor runs it's calibration is wrong.
So in the end what I am wondering are the following points:
- Is there a way to reliably know if a zigbee attribute update was in fact sent out, which ultimately could be used to trigger the device to go into deep sleep?
- Is there a way to make the startup faster so that I don't need to be wasting power for ~40s just to send a single status update?
- Is there maybe some example configuration which could illustrate how to set this up? I assume that I am missing something here
I am currently building a sensor with this component and so far it mostly works pretty great.
But when I try to utilize the deep sleep component of esphome no attribute updates seem to be sent to the coordinator for quite a while after reset even though there are no errors visible in the logs.
It seems like it needs at least 30s after boot to even send out a single attribute update. I find this extremely inefficient.
I assume there must be a better way since it is explicitly mentioned in the readme right?
My working ESPHome config
Another issue with this configuration is that the sensor expects to be polled every 10min as per manufacturer spec as it does a re-calibration routine after a prefixed number of polls and with this 10s polling interval (which is needed since I need to try a few times for the attribute updates to actually make it through, even though I see no errors in the logs of the device) the assumption about the number of polls before the sensor runs it's calibration is wrong.
So in the end what I am wondering are the following points: