HomeAssistant Yellow Devices Dropping Off

Configuration
1

I have close to 85 zigbee devices running off of a homeassistant yellow. I have updated the zigbee firmware to the latest, but I still have major device drop issues. The majority of my devices are light bulbs and I can’t go a day without a bulb dropping off. I have to manually power cycle them and they come back online. I’m out of town and already like 10 bulbs are offline. Anything I can do? Is the yellow known to have issues like this? I get errors like the below when they drop it seems:

  • NWK conflict is reported for 0xb191
  • Found b0:ce:18:14:03:6c:14:28 device for 0xb191 NWK conflict: sengled E1C-NB7
  • NWK conflict is reported for 0x0a1b
  • Found cc:cc:cc:ff:fe:c6:11:2b device for 0x0a1b NWK conflict: The Home Depot Ecosmart-> ZBT-A19-CCT-Bulb
  • Watchdog heartbeat timeout: TimeoutError()
  • Received relays from an unknown device: 0x5f78
  • Unknown device AddrModeAddress(addr_mode=<AddrMode.NWK: 2>, address=0x5F78)

I didn’t have this issue before upgrading to the yellow, my network IS congested but it’s normally 70-80% which isn’t terrible. Ive tried changing channels and no luck.

Edit. I’m going to try moving my HA Yellow hub elsewhere, if that doesn’t fix, I’ll be getting rid of it for a better solution. I think the integrated ZHA is worse than an external solution you can move wherever, I guess I could try getting a third party dongle and moving everything onto it, I’ll see.

2

What did you have HA on before you migrated to the yellow? If running zigbee are you using the on-board radio, while at the same time running WiFi?

It would probably be better to hook up the yellow to the house network via the wired ethernet port (RJ45 connector). Then use the radio on board just for zigbee. The other is to get a zigbee stick and attach it with a 3 to 6 foot usb cable.
This will eliminate the channels of zigbee and Wifi from beating on each other.

3

I have it hardwired, all my access points are manually set to channel 1 for 2.4ghz and low power setting, and the one closest to the ap has 2.4 disabled completely. WiFi is disabled on the yellow as I have the Zwave module installed. rssi looks pretty good, plus bulbs should act as routers…

4

Something I want to do is remove everything BUT lights and see if maybe it’s routing through a battery powered device. I have a couple climate sensors and some motion sensors, I guess if it’s using those and they drop something that could be an issue. Hmm. Further troubleshooting to come.

5

Found b0:ce:18:14:03:6c:14:28 device for 0xfb87 NWK conflict: sengled E1C-NB7
NWK conflict is reported for 0x1110
Found cc:cc:cc:ff:fe:c3:16:d0 device for 0x1110 NWK conflict: The Home Depot Ecosmart-ZBT-A19-CCT-Bulb
NWK conflict is reported for 0x38d5
Found b0:ce:18:14:03:6c:2e:bb device for 0x38d5 NWK conflict: sengled E1C-NB7

So many of these all the time, doesn’t seem to care what brand of bulb.

6

I’m getting the exact same thing, and it’s infuriating. I’m on a generic x86 install. Did you ever figure out how to solve this issue?

My log entry:

Logger: zigpy.application
Source: runner.py:188
First occurred: 2:48:29 AM (13 occurrences)
Last logged: 2:49:04 AM

    Received relays from an unknown device: 0xFF93
    Unknown device AddrModeAddress(addr_mode=<AddrMode.NWK: 2>, address=0xFF93)
    Unknown device AddrModeAddress(addr_mode=<AddrMode.NWK: 2>, address=0x9D70)
    Received relays from an unknown device: 0xA995
    Unknown device AddrModeAddress(addr_mode=<AddrMode.NWK: 2>, address=0xA995)
7

Moved yellow hub to a different location, away from any aps and my rack, same. Had a bulb drop off within 12 hours. May need to rethink my move from hubitat as I had no drops with hubitat. It could 100% just be the ecosmart bulbs aren’t routing randomly.

8

Ok, I have gone an entire day with 0 issues. The changes I made:

I threw away 4 crappy sonoff climate sensors that were always running out of battery, maybe they were dropping routes. They were cheap.

I removed multi-protocol, flashed the the standard firmware to the radio, and migrated the radio.

I changed back to channel 25 since the only devices with issues were the sonoff stuff.

Will report back to confirm devices stay on longer. Usually I’ll get 3-4 bulbs dropping off a day so it’s already better.

9

Ok the issue seems to be resolved, I have changed channels before and enabled/disabled multi protocol before so I am verily sure that’s not it meaning it was those dumb sonoff sensors.

TLDR I think I had a some cheapo devices that weren’t routing correctly and would periodically screw up the network. Deleted and burned said devices in the fireplace and network has been stable for 2 days which is 2 days longer than it’s ever been stable on HA, success.