Issue was resolved by deleting and uninstalling zigbee2MQTT, hard reboot of Home Assistant os, re-install of zigbee2MQTT then re-pair 90+ devices.
Over night only had three devices and one non functioning device.
CORRECTION on investigation two router devices were offline due to loss of power, so one device was ‘lost’ overnight, the other never seemed to have paired properly in that it had never populated its fields, a quick re-pair resolved both.
So still a better night than I have been experiencing over the last 5 months of zigbee2MQTT wackamole and the whole purpose of switching to smzb06.
The low lqi was resolved by wiping zigbee2MQTT and completely restarting.
The first 40 devices took about an hour to do, all routers.
The next 10 routers were all a pain to get into pairing.
The next 30 end devices/ batteries went in in about and hour, and the remaining 10 battery devices were playing hard to get and took multiple pairing or just would go into pairing modes, looking hard at you Aqara leak detectors……
I have three devices, 2 in the same room, line of sight to the SMLight smzb06, which a little troubling.
Also have one device, battery that has not been receiving data over night, in fact has not reported since pairing 15 hours ago, so will re-pair that now.
All the Node Red automations came straight back up very easily without me doing anything, I have a standardised naming convention and if I doubt was able to check the old device pairing in MQTT.
Home Assistant automations did not appear to survive the rebuild, so have some work there.
I could see zigbee2MQTT connecting to the smzb06 on the smzb06 dashboard so it was not a networking issue.
Also, the complete delete and wipe of zigbee2MQTT resolved the issue using the same zigbee2MQTT config data copied and pasted from the smzb06 config page.
So not a finger problem.
The issue seems to be related to migrating from the sonoff P usb coordinator to the Ethernet smzb06.
I complete zigbee2MQTT wipe and rebuild resolved all potential network/aerial issues, so suspect the problem is in zigbee2MQTT not the smzb06.
Yea pairing can be a hit and miss but usually routers go in quickly, its the battery stuff that drives me mad sometimes.
I got another UZG-01 that I am currently testing as well, will compare to the SMLIGHT when it finally arrives. My goal is to try and create more smaller networks to get the best stability.
No surprise the UZG-01 worked straight out of the box
From what I understand, the only problem with wifi intefearence is if your device is next to a wifi transmitter, or worst still if your coordinator is next to your wifi router. If it is sufficiently far away or your neighbour’s wifi, it should not cause problems. Still though is better to have a channel that is not occupied by wifi.
I try my best to not have too many wifi devices as well high powered wifi devices like soundbars and TVs, if I can get them wired, I have them wired.
Just to rule out any weirdness of the hardware, I actually replaced my UZG-01 with another UZG-01 (It arrived quicker than the replacement SMLIGHT). Wanted to do this before I split my network and go down that route. So far seems good but the problem usually only starts after 24-28 hours, so will have to wait before coming to conclusions.
Ok I think I have a handle on the issue, where when migrating between coordinators you need to power off all router devices, ie power down the house.
Here is what SMLight support sent:
“Device does not start with a connected antenna, because some Zigbee routers can “hold” old Zigbee network, and at the same time coordinator cannot operate 2 Zigbee networks on the same channel with the same ID, so the coordinator does not start the network if it finds an already existing one
Migration between coordinators requires a mandatory reboot of all Zigbee routers. Usually, the best option is to turn off the power in the house for a couple of seconds after disconnecting the old coordinator and connecting the new one, then all zb routers will lose the old network and nothing will interfere with the new one.”
Makes sense I suppose but I never encountered that yet fingers crossed. I just swapped yesterday, granted from the same zigbee chipset so maybe that is why. I just shut down z2m, swapped the coordinator and restarted, everything just came back.
Hello, those of you connected via ethernet, please check if you restart the router, the connection between the zigbee2mqtt addon and the coordinator is restored, for me after restarting the router the addon does not reconnect to the coordinator and the whole zigbee network drops…
For what it’s worth, I’m running 2 zigbee networks (ZHA on a “not recommended” Sonoff bridge and Z2MQTT on a TubesZB ethernet dongle).
Both coordinators have been happily chugging along with next to no issues for around 3 years, within 1.5m range of not one, but two wifi access points.
Obviously, your mileage may vary and I might just be lucky, but I wanted to share my own experience.
Yea its very subjective, but I guess just best practice to have it just away from wifi routers or any other kind of EM radiation.
I think also a lot of Zigbee issues seem to be due to bad devices that cause the network to unstable. Very hard to debug and not a lot of definitive information out there.
I got my replacement, they sent me the wrong one :D, the 06M version. I notified them when it was shipped so wasnt expecting it to arrive.
Anyhow I tested it and it worked out of the box. Got an IP and I even re flashed the zigbee firmware to the new ember for Z2M. Will try to do more tests but unlikely to use this for my production Z2M network.
Anyhow for everyone that was following, finally got the 06 version delivered today. It works good at last.
So not sure what the issue was, surely something with the LAN chip itself. I would recommend everyone having an issue to contact support to get it replaced
I received my SLZB-06 earlier this week and started testing last night - worked straight out of the box with POE to a FortiSwitch S124E no dramas at all.
One observation, it did auto negotiate speed to 100Mbps, haven’t tried to force the port to 1Gbps no need really. POE drawing a report 1.7 Watts
First went with DCHP address and move that to a reserved DHCP address that required a reboot of the 06 and again no issues with that.
Only a small Zigbee network for now, with 8 devices (2 routers & 6 End Devices) for testing.
With regards to the RSSI being unknow, SMLite confirmed this is a bug jn ZHA during migration. The solution for now is to start from scratch. I took the opportunity to switch to Zigbee2mqtt and re-paired all my devices.
and not a lot of definitive information out there.