I think I’ve isolated an issue I originally thought was Rule Machine intermittently missing contact events.
I have two Zooz ZSE41 Open/Close Sensors, both using Hubitat’s built-in ZSE41 driver and both on firmware 2.30:
- Garage to House Contact
- Wine Room Door
Both sensors have now shown open in Hubitat while the physical doors were definitely closed.
The interesting part is the timing. On the Garage sensor I captured:
Sep 7 21:46:07 battery = 100%
Sep 7 21:46:08 contact = open
The door was not opened — I was asleep at the time.
I saw the same pattern again the next morning:
Sep 8 09:41:53 battery = 100%
Sep 8 09:41:53 contact = open
Again, the door had not been opened.
also - here are some logs from the other sensor / door:
The battery reports appear to occur on roughly a 12-hour schedule. After one of these false opens, Hubitat remains showing the sensor as open even though the door is physically closed. Physically opening and closing the door restores the correct closed state.
I have now also seen my second ZSE41 (Wine Room Door) sitting in an incorrect open state while that door was physically closed, so this does not appear isolated to one sensor.
This may also explain what I originally thought was an intermittent Rule Machine/Simple Automation failure. If a periodic report has already left Hubitat believing the contact is open, physically opening the door later can produce another device “is open” log entry without generating a new state-change event for the automation. Consequently the automation does not trigger.
Things already tried/confirmed:
- Hubitat C-8 Pro
- Platform 2.5.1.181
- Two separate Zooz ZSE41 sensors
- Both firmware 2.30
- Both using Hubitat built-in ZSE41 driver
- Hub rebooted
- Database rebuilt
- Physical open/close immediately restores correct state
Has anyone else seen a ZSE41 generate an open report during its periodic wake/battery reporting? I’m particularly interested in whether this could be a ZSE41 firmware issue, something in the built-in driver’s parsing of the periodic report, or a regression in 2.5.1.x.
thanks!
