I would like to bring a recurring SONOFF ZBMINIL2 Zigbee issue to the attention of the eWeLink team, as I have already reported it to SONOFF support but, unfortunately, they seem rather reluctant to investigate the actual technical problem.
The device is used with Home Assistant ZHA. It is an End Device and is located approximately one meter from a Zigbee router. Other End Devices in the same network do not exhibit this behaviour.
The problem has now occurred several times.
The pattern is remarkably consistent:
- There is a relatively minor change in the Zigbee network topology.
- The ZBMINIL2 continues to work normally for several hours.
- Several hours later — sometimes 10+ hours after the topology change — the device suddenly becomes completely unreachable from ZHA.
- Power-cycling the ZBMINIL2 does not restore communication.
- Restarting ZHA does not restore communication.
- Restarting the Zigbee coordinator does not restore communication.
- Reconfiguring the device in ZHA may complete successfully, but the device remains offline.
- The relay itself still works locally.
- The only reliable way to restore the device is to pair it with the Zigbee network again.
This does not look like a relay malfunction or a simple range problem.
It looks much more like a problem with the ZBMINIL2’s handling of parent loss / Zigbee rejoin as an End Device.
I have already provided SONOFF support with a detailed description of the behaviour. Unfortunately, their response so far has focused largely on obtaining the serial number, proof of purchase, etc., and on asking me to verify whether the relay itself can still be controlled.
That misses the point.
The relay works. The problem is that the Zigbee stack stops being able to recover the device’s connection to the network.
SONOFF also stated:
“We haven’t received similar feedback from other users.”
I would be very cautious about treating that as evidence that the problem does not exist. An occasional Zigbee device going offline is very easily attributed by users to their coordinator, network topology, Home Assistant or Zigbee in general. Most users will simply pair the device again and move on rather than spend hours investigating why it disappeared several hours after an apparently unrelated topology change.
There are already reports from other ZHA users describing very similar ZBMINIL2 rejoin/drop-off behaviour, including cases where the device appears to perform a rejoin and subsequently leaves the network again.
I am therefore asking the eWeLink team to help escalate this to the appropriate SONOFF firmware/engineering team.
The questions that need answering are fairly simple:
- Is there a known issue with ZBMINIL2 firmware 1.0.14 (
0x0000100e) concerning parent loss or rejoin? - How is the ZBMINIL2 expected to recover when its Zigbee parent becomes unavailable?
- Is there a known limitation in its End Device rejoin implementation?
- Is there a newer firmware which addresses Zigbee rejoin/parent recovery?
- If not, can SONOFF investigate this behaviour?
I am not asking eWeLink to fix the firmware. I understand that eWeLink is not responsible for the Zigbee firmware itself.
I am asking eWeLink to help get this issue in front of the people at SONOFF who actually are responsible for the firmware.
At this point, repeatedly pairing the device again is not a solution. It is merely a workaround for a device that should be able to recover its Zigbee connection on its own.
I would really appreciate it if someone from eWeLink could help push this investigation forward rather than letting the case become another “device offline” ticket.