Help with an Unreliable Hub - Unstable Zigbee Mesh

Hi @hubitrep . I don't want to rule anything out. I've been running some sort of Deco mesh system before and after installing hubitat. The only change was upgrading to wifi 6 APs. I don't recall there being any immediate collrelation between installing the wifi 6 points and the problems. My hue bridge is rock solid and the Ikea one never fails either, nor does the hive home heating system. I can't change the channels manually on Deco, but I have checked and the 2.4ghz band is on channel 9 with the 5Ghz on channel 36.

I don't suppose you can control the channel width either. If you can afford to, try turning off the 2.4 band on the deco and see if it makes a difference.

What @hubitrep wrote is on the mark. TP Deco meshes are zigbee killers. You cannot select the channel and you cannot select the channel width, which defaults to 40 MHz.

Do yourself a long-term favor and a get better mesh WiFi system. One relatively inexpensive and easy to install system is a UCG Ultra/Max router along with UX mesh points.

1 Like

Also you have a lot going on on the 2.4 ghz band, setting a high power level on the hub to have it “scream” over the others (while the zigbee devices can’t increase theirs in response) isn’t a great solution imho

1 Like

Yup. Because it also increases noise!

Thanks aaiyar. I appreciate the feedback, but a router and 4 APs would be a very expensive solution, even with the more reasonabiltiy priced ubiquity gear. I hoping it wont come to that.

@aaiyar @hubitrep all valid points, but I am wondering would interference from Wifi actually cause the Zigbee radio to go offline? This is more than a problem of just the devices not responding, but the hub is actually showing that Zigbee is offline.

Just want to make sure we are not chasing down a path which would not actually cause the main problem at hand.

2 Likes

Hi All. Yeah, I hope its not a case of trying to find an open window when the door is wide open lol. I reset the hub this morning, and the zigbee network is online and working flawlessly. Everything is connected and my automations are running. The hub doesnt seem stressed. My hue integration has no problems and my home wifi is working perfectly.

Hi All. I've messaged Bobby and I'm waiting for a reply, but by way of a sit-rep, I've had two network failures over the last 4 days or so. I've attached the logs below.

The radio appears to be going on and off line. I've been keeping an eye on the hub load and it isn't stressed. I've had long standing issues with geofencing too and the hub not making my better half or I as away when we're outside the geofence. I have changed phones and its still not working. I'm starting to think there may be something wrong with my hub.

I’m not super familiar with the Zigbee protocol stack or the SiLabs chipset but yeah if interference is causing retry loops, it’s something one would want to rule out. Certainly I’ve seen a wifi router knock itself out because of interference (a simple change of channel resolved the issue).

Thanks @hubitrep . Hub is on channel 25 atm having been changed from 20 in the past. The hue bridge is on 20 and has no issues. I also live in a rural area, so unless its an in-house issue (which I hope it isnt) there are no other wireless signals near me. I'm half hoping its a hardware issue at this stage.

1 Like

most likely as said noty a hardware issue but bad devices.. hub load is not a good indication..

post your settings/zigbee page sorted by number of messages.,., that may help diagnose device issues.

aLSO what version of the firmware as there were some fixes for this recently

Thanks @kahn-hubitat and @hubitrep

My zigbee settings sorted by messages is below, both from top to bottom and bottom to top.
There is one water leak sensor under the bath that I need to reset. Aside from that, everything else connects when the rub is reset. The most busy items are the light switches that are actually sonoff zigbee relays (ZBMINI Extreme) using the inbuilt generic zigbee switch driver. The temperature sensors are sonoff temp and humidity sensors (SNZB-02D) using the Tuya driver developed by kkossev and a third realy temp and humidity sensor using the inbuilt driver.

Appreciate your insights.

Those stats reset when you reboot, so need to know how long hub was running since last reboot for the numbers to have meaning. Logs > Device Stats shows uptime at the top.

Thanks @jtp10181
Device stats below.

The Eddi and Zappi are API calls to the MyEnergi Cloud and the two VR Thermo are virtual thermostats. The rest are bulbs connected through the hue bridge.
The first non-hue bridge device is a govee light strip over LAN.

Hi All

Other outage last night. Logs below.
The hub is in a cloak room next to my network rack with the switch and beside the hue and ikea bridges. As a last resort, I'm going to move the hub to somewhere else with an ethernet connection as a last resort.

Turn on debug logging and see what these are spamming out constantly. Do they do power reporting by chance? I wonder if that can be disabled.

What other devices are you using with custom drivers besides the ones you mentioned above from kkossev? Any Aquara or Xiaomi devices?

Hi @jtp10181 , I'm not using any other custom drivers. I have no Aquara or Xiaomi devices. I've avoided them because I didn't want another hub.
I've turned on debug logging and I don't see too many errors. They don't have power reporting and I have no power reporting plugs.
I moved the hub to the living room where another ethernet cable is and no change. My zigbee radio is still going on and off. See below.

I may have to throw in the towel here unless @bobbyD can determine if it is a hardware issue or not.

Thanks again for all your help. The community is truely amazing.

Those two devices that are just spamming out off commands, can you check the event tab on those and verify that nothing is calling refresh on them constantly? It would show as command-refresh in the events. If that is not there then the devices themselves are spamming on their own.

It is not too horrible, it is minutes apart each time, but they are definitely very chatty. Maybe a combination of multiples of those chatty devices is overwhelming the radio?

You could send a PM to @bobbyD or @support_team with your Hub ID (found in settings) so they can check the engineering logs. That way when they do find this thread they already have it from you.