ZHA Zigbee Tested Devices...Please add your device results

@photo64 Hrm… not what I expected. I had stability and pairing issues before but I’m up to almost 30 devices paired to a hubzb1. I have no idea what could be causing issues with only 4 paired devices… unless it’s a timeout thing. You could try adding the following 2 lines to the configuration area of ControllerApplication in bellows:

yield from self._cfg(c.CONFIG_END_DEVICE_POLL_TIMEOUT, 60)
yield from self._cfg(c.CONFIG_END_DEVICE_POLL_TIMEOUT_SHIFT, 10)

Sorry for the formatting… on my phone. You can see the actual code for these here: https://github.com/Yoda-x/bellows/blob/2a660abd1af2095cdc025580ebefbded1c6ec158/bellows/zigbee/application.py

@dshokouhi these are the settings that I use to manage a large network: https://github.com/dmulcahey/bellows/blob/dev/bellows/zigbee/application.py

The config block for the table sizes is what helped my network stay stable as it grew.

1 Like


I’ll take a look at that and try it out. I saw there was a bellows update and I was hoping that would fix the disconnect issue but was concerned about going to 0.62 and 0.63 due to a a number of other issues (waiting on 0.64 now). Not sure what happened between 0.57 and 0.61 but now it just loses connection too often and I have to restart HA to get it back. I rely heavily on zwave and starting to add more and more zigbee. Was hoping the Visonic motion sensors would be added but as far as I know they still are not working in HA. A great cost saver if they ever do get added.

@dmulcahey thats what I noticed as well. Looks like a similar commit was made earlier:


But from what I can see your changes are slightly different lol


Not comfortable enough guessing the correct placement, should these lines
yield from self._cfg(c.CONFIG_END_DEVICE_POLL_TIMEOUT, 60)
yield from self._cfg(c.CONFIG_END_DEVICE_POLL_TIMEOUT_SHIFT, 10)

be added after this section at line 51 of the code?

    yield from self._cfg(c.CONFIG_APPLICATION_ZDO_FLAGS, zdo)
    yield from self._cfg(c.CONFIG_TRUST_CENTER_ADDRESS_CACHE_SIZE, 2)
    yield from self._cfg(c.CONFIG_ADDRESS_TABLE_SIZE, 16)
    yield from self._cfg(c.CONFIG_SOURCE_ROUTE_TABLE_SIZE, 8)
    yield from self._cfg(c.CONFIG_MAX_END_DEVICE_CHILDREN, 32)
    yield from self._cfg(c.CONFIG_PACKET_BUFFER_COUNT, 0xff)
    yield from self._cfg(c.CONFIG_KEY_TABLE_SIZE, 1)
    yield from self._cfg(c.CONFIG_TRANSIENT_KEY_TIMEOUT_S, 180, True)
    yield from self._cfg(c.CONFIG_END_DEVICE_POLL_TIMEOUT, 60)
    yield from self._cfg(c.CONFIG_END_DEVICE_POLL_TIMEOUT_SHIFT, 6)

@photo64 if that’s from your file, you already have them. It’s the last 2 lines you pasted there.

@dshokouhi yeah, I’m not sure what is the best way to make these generic enough for everyone. I’ll start a conversation in discord to see what everyone thinks.

1 Like

Would changing the 6 to a 10 make any difference?

@photo64 not positive but I doubt it…


I just added those 2 lines to my application.py file and restarted HA. What I posted above was from your link not my original file. I’ll keep my fingers crossed and let you know if this keeps me connected. Thanks for the suggestion

Is there some documentation from SiLabs or elsewhere on these variables and what they are actually doing? I have about 30 devices as well connected to my HUSBZB-1 and would like to get some more information on tweaking the right values.

I’ve been running Yoda-X’s fork of bellows in test, and I believe his larger table size defaults used all the buffers on the stick:

@photo64 did that help at all?


I believe it did, I had it connected for 3 straight days before I had to restart HA yesterday for another reason. It wasn’t getting past a day and a half before not working so I am hopeful that was it. I also noticed less errors in the logs although I do still get these every 3 minutes -> WARNING (MainThread) [bellows.zigbee.application] Unexpected message send notification. Any solution for those?


Nope, those are just warnings though… nothing to worry about.

all of my “unexpected message send notification” errors went away after upgrading to 0.63


I may have spoke too soon, after running straight since I last commented it seems ZHA has once again disconnected from HA and I am not getting my automations to fire off. Any thing else I might be able to try?

ZHA is broken for me in HA 0.64.0. I went back to 0.63.3 and everything started working again.

see this issue for a potential fix: https://github.com/home-assistant/home-assistant/issues/12675

@photo64 we’re gonna need to look at full logs to see what’s happening. I would suggest logging at debug for zigpy, zha and bellows or just bellows and zha if your on a version from before the zigpy split. Then we can dig in more.


Before I do that I would like to upgrade first since I am still at 0.61.1 to see if that changes anything. If I still run into problems I can create an issue with log info. I am waiting to see if 0.64.1 takes care of the broken ZHA in 0.64