Weird issue with LZW36 fan turning off randomly in the night

I haven’t figured this one out yet, but figured if I’m having the issue, someone else may be as well.

I have several LZW36 switches in our home. Since switching to using the zwavejs2mqtt docker, we’ve had an issue where our ceiling fans will turn off randomly during the night while we’re all asleep.

Problem is, in Home Assistant, this appears to be a local control toggle. :ghost:?


It’s happened almost every night, usually between 11:30PM and 2:30AM, and on multiple fans. I don’t see much relevant information in the zwavejs2mqtt logs either. This is the only thing that I found that matched the timestamp:

  2021-03-06 01:25:18.182 DEBUG MQTT: Publishing to zwave/4/38/1/currentValue: {
  time: 1615015518181,
  value: 0,
  nodeName: 'mbdrm_ceiling_fan_light',
  nodeLocation: 'master_bedroom'
} with options { qos: 1, retain: true }
2021-03-06 01:25:18.183 INFO ZWAVE: Node 4: value updated: 38-1-currentValue 0 => 0
2021-03-06 01:25:18.287 DEBUG MQTT: Publishing to zwave/4/38/2/currentValue: {
  time: 1615015518286,
  value: 0,
  nodeName: 'mbdrm_ceiling_fan_light',
  nodeLocation: 'master_bedroom'
} with options { qos: 1, retain: true }
2021-03-06 01:25:18.287 INFO ZWAVE: Node 4: value updated: 38-2-currentValue 45 => 0

The actual zwavejs log doesn’t show anything at the timestamp, so all this really shows is that this particular node has been updated.

At least the ghost is smart enough to tap once. Beats most sig others here.

I’ve been having exactly the same thing happening with my LZW36 using the ZwaveJS integration. Glad it’s not just me going crazy.

It could be this issue:

Fan turns off about 2 hrs and 10 minutes after being turned on via zwavejs. It might be a firmware bug. There appears to be a temporary config file fix in an upcoming zwavejs release.

Yep, that seems to be it. I just observed the device RGB me, and then the fan shut off. It was 2 hours and 8 minutes after turning it on from HA. Good catch!

Marking this as solved for now as we’ve identified the issue per @jtronicus’s post. I’ve also tagged it as a firmware bug. I believe we’re working on this with the HA team right now.