All triggers being send by the device directly after the event occured, except for the vibration strength
The vibration strength is reported ~300 seconds after the first vibration event occured, and can be considered to be the maximum vibration amplitude in the past 300 seconds
Tilt-, Vibration-, Drop-alarm triggers originate out of the reported motion and are reset by Homey based on the duration as defined in the settings; the device does not provide an alarm trigger, nor a alarm reset (aka Homey special)
For the tilt data, I’m using the RAW force vector as provided by the device, not the directly reported angle. (aka Homey special). This enables a distinction in tilt direction (X, Y). I’ve added several additional features to the driver to
be able to be able to compare the tilt to it’s “reference position” and “previous position” (capability and trigger cards)
set the reference force vector upon next tilt report (settings > mark “Set tilt reference vector” > press save > tilt the sensor to it’s new reference position ie. roof window closed)):
determine if the angle provided is the absolute or singed value (settings)
Keep in mind that the tilt angles are determined based on a change of the force vector (as result of the gravitational force). A tilt will be reported when the orientation of the force vector will change. Rotating the sensor, while the sensor is lying flat on a table top, will not change the orientation of the force vector, therefore no change in tilt is reported
Due to a limitation of Homey’s Zigbee core, it is not possible to adjust the vibration sensor sensitivity. I’ll further explore options to still add this setting to the driver.
Battery capabilities have been implemented, but will not show any data, awating a fix in Homey’s Zigbee core
I’m curious on your experience with this sensor and use cases: please share them
Plugin v0.4.2 installed via cli.
Sometimes switch (QBKG03LM) control works well, but time to time plugin unable to turn it on or off.
And after several attempts it starts working again.
Other xiaomi-zigbee-devices not affected at all. Working as expected.
I created issue on github, but it seemed necessarily to me to write here too.
@FantomNotaBene thanks for reporting.
v0.4.2 did not change anything in the implementation of the Aqara Smart Light Wall Switch (L) Double (QBKG03LM). Considering the error code (coming back from the Homey core), it points towards a mesh issue / disturbance. Let’s follow-up on the GitHub issue, to prevent parallel discussions.
Indeed a known issue, but I only have seen it when using the beta Homey app…
Can you check if it also occurs when using the old app / webinterface?
You right i don’t get error messages from old app and webinterface but the behavior is same. Sometimes it not reacting sometimes it likes that motor has own life
For example if i set manually curtains to 50% it closes and after few seconds goes back to open and then goes to my setpoint. I will upload video later
I am very curious how you were able to implement this.
In my testing with two different sensors, the accelerometer force values output have not been consistent enough to guarantee an accurately calculation of the sensors current resting tilt angle. Also, there is an imbalance in the two ends of the output range, particularly in the values I’ve seen for the X-axis.
Although I’m not a Homey user, I also just want to say how much I respect the great work you’ve been doing in creating compatibility for Xiaomi sensors on the platform!
I’m not able to pair my aqara vibration sensor in v0.4.2
The blue led of the sensor blinks 3x but nothing happend. I have tried this with the preview app and via my.athom.com.
I have restarted my homey, but this changed nothing. Am I the only one with pairing issues?
@Emile86 please check with a defined sequence… it could be that the first command gets blocked, creating a queue of commands that are still to be send…
if the first command is “released”, that the remaining queue is still send to the curtain controller
@veeceeoh thanks for the compliment… one day, you’ll buy a Homey
The source code is available on GitHub for sharing knowledge. Which shows the calculation methods I used for the vibration sensors.
On the accuracy… I’m aware of the imbalance you mentioned on the other GitHub site, but looking at the usability options, I think most users will look at the relative tilt where absolute accuracy is not that relevant.
Perhaps with some calibration, it would be possible to increase the accuracy (further)…
@Wdl I always refer to the by Xiaomi provided inclusion instructions…
Can you try if the sensor will include when tripple clicking the button (instead of press and holding the button for 5 seconds)?
@Marius_Muresan “no pressure”
Just received the involved sensor from @Mustang… and did some preliminary tests…
Also here different capabilities (1x, 2x click, hold and release) compared to the “other” Aqara Wireless Switch.
Goal is to have an update early next week that will add support for these same devices with different firmware and capabilities…