I need help with an unreliable hub and an unstable zigbee mesh. I've tried all I can think of and I'm at my wits end.
I'm running a C8 hub that's hard wired to gigabit switch, that is in turn being serviced by a TP Link wifi 6 mesh system where all nodes are directly wired to the switch. I have a philips hue hub with several hue and innr lights and switches connected, and also an IKEA dirigera hub with one motion sensor and one lightbulb. I also have several Google Nest Minis, Hubs (1st and 2nd Gen) and two apple homepods. My SO and I both have iPhones.
The problems I have been having are getting worse. These are:
At least every second night, my zigbee radio is going offline. When I wake up, I have to reset the hub.
For the past several months, I have several zigbee devices becomming unresposive or simply failing to report. These are a mixture of hue motions, third reality contact and motion sensors, and several sonoff contact, motion and temperature sensors. I have constantly having to repair them and readd them to the network.
I have tried adding signal repeaters (Tradfri) and at least one zigbee plug (usually Ikea Tradfri and Tretakt) in each room.
Room lighting apps fail to turn on the lights through the hue bridge, dispite saying that the app is active. I have to go in an manually press the activate button to get the lights to turn on.
Room lighting apps failing to reset after restrictions are turned off, and then not activing again in the morning.
I'm relatively learned when it comes to technology, but I simply cannot seem to find a source(s) of all these issues).
For completeness, my hub it up to date, running 2.3.9.201 and my zigbee channel is set to 20 with the power level at 12. I have checked, and the hue hub is also using channel 20 but I'm not sure what the Dirigera hub is using.
Thanks in advance to anyone who can help. I love the platform but I'm getting more and more frustrated with it.
My recommendation would be for you to try another Zigbee channel and see if it stabilizes things. Try 25 for now. Also, in most cases, a lower signal strength can perform better. Try 8 to see if it helps.
Hi Sebastien and thanks for the advice. I changed the zigbee channel to 25 and reduced the signal strength to 25. I gave it a day and most of the devices found the new channel and I repaired the ones that didnt. Things seemed more stable for 12 hours or so and since then, I have again been experiencing significant issues. These are things like plugs, contact sensors and temp and humidity sensors becoming unresponsive and the zigbee radio going off line. I've pinned an extract from the hub events log, and also the network graph view.
By way of additional info:
I have 51 zigbee devices with 7 of those being plugs (5 tradfri and 2 tretakt) acting as repeaters. I also have 2 IKEA signal repeaters.
I don't have any z-wave devices.
I forgot to mention in my first post that I also have a Hive Heating system (UK version of something like EcoBee and Nest) that also uses zigbee. It is connected to one main thermostat, a receiver and two TRVs.
The 2.4 GHz radio space is super-saturated, without even thinking about any possible BT devices that are present in your home. And if I had to bet, the major culprit is your TP-Link mesh WiFi system.
Please read this link for details:
Here's what I would do:
Put all your TP-Link mesh points on WiFi channels 1, 6, or 11 for the 2.4 GHz band. Further, ensure that the channel width is 20 MHz and not 40 MHz. If your TP-Link system doesn't let you pick the WiFi channel or set the channel width, consider replacing it - their Deco mesh systems are an anathema to a stable zigbee system.
Use the following zigbee channels for your Hubitat, Hue, and Dirigera - 15, 20, 25. Don't go crazy on radio strength - that also increases noise.
I think Hive uses low channel zigbee (i.e. zigbee 0-10) (just checked, it doesn't). So, configure your Hive system to use zigbee channel 11.
Thanks @aaiyar .
I've checked the Deco system and the 2.4GHz is on channel 5 with the 5GHz on channel 36. It doesn't look like you can manually change them.
The Hue system is using channel 20 and the HE set to 25. I can't change Hive or Dirigera manually, but it looks like they both use 11, as you say.
I've ran a channel scan on HE and its showing the following.
I don't have many bluetooth devices in the house. I was looking at some switchbot things, but I'll put those off pending getting this resolved.
It'd be a punch in the gut to replace the Deco system. It's not long installed and it wasn't cheap.
If it helps with diagnoisis, the Tretakt plugs are connected to HE and are using Dandanache's drivers. They have worked flawlessly and never go offline.
My advice to anyone who wants a good zigbee experience is to avoid TP-Link Deco mesh systems. If you can find a way to force the Deco mesh to use WiFi channel 2 or 3, then even with a 40 MHz channel width, you should get some stability in zigbee channels 15, 20, and 25.
Hi Sebastien. Thank you for the advice. Switchen to chanel 25 has helped but it wasn't stable with a power level of 8. Things are alot more stable with a power level of 12, but my zigbee network is still dropping offline at least once every two days. When I reset the hub, most of the zigbee devices reconnect, which they weren't doing before. Any advice?
Thanks in advance.
Any chance you have Zigbee bulbs on your network? Some of these non-Zigbee 3.0 will not repeat correctly and cause major mesh issues.
Also, some older Zigbee devices won’t play well with each other. I use to have some thermostats that would kick my Aqara contact sensors off the mesh all the time. They repeat well for everything else though…
It is also likely that there is a lot of interference or too much « space » between devices, causing them to drop. I use to have those issues as well. I no longer have any repeating bulbs (only Sengled which do not repeat) on my hubs. I have also added quite a few repeaters.
I think its also possible that the hub is stressed either CPU or memory which can then cause the zigbee radio to shut down or flip flop on and off. Should look at the Logs > Device and App state to rule that out. Also next time you are having issues check the hubs notifications to see if you have any load or memory warnings.
Hi @Sebastien and @jtp10181 , thank you so much for your reply. Now for a sit-rep, I've replaced all my old Tradfri outlets with newer Tretakt ones and things are alot alot alot more stable. Automations are happening quicker and lights are comming on in sync as opposed to the 'popcorn' effect. I'm still using Chanel 25 and I've upped the power level to 16. I've had one instance of the zigbee network going offline and that was today, but there was no evidence of the hub being overloaded. I'm going to keep an eye on it for the next few days and see how things go. Thanks again for your help.
Hi @Sebastien and @jtp10181. I've two more instances of the Zigbee network going offline. When it is online, all is fine. All divices are connected and automations are running fine. Is there anything I can look at that may give me an inidcation as to why the network keeps going offline? There are no load or memory warnings. Thanks in advance for your help with this.
Sorry I am far from a Zigbee expert, I only had that one piece of advice from reading other posts in the past. Just to be clear though when you say it is going offline, when you go to Zigbee details it says it is OFFLINE next to status? Or do the devices just stop working and you assume it is offline?
Hi @jtp10181 , no need to be sorry. You're very kidn to help. I am getting a notification saying that the ZigBee Network is Offline and it is showing as 'Offline' next to status. I've been keeping an eye on the hub load, and I can't see anything that would indicate why this is happening.
It is less likely a hub load issue, and more likely a mix of devices problem. One or more devices are overwhelming the radio. The first thing to do is to use tested devices from the Compatible List. If you have any non compatible devices, and or using non built-in drivers, you may want to remove those devices and/or switch to built-in drivers. If that doesn't apply to you, please send me your hub's MAC address in a private message, as you may be dealing with a hardware malfunction.
Hi @bobbyD , thank you so much for the message. I have a mixture of devices, with the majority using in-build drivers. I have a few temp and humidity sensors using kkossev's drivers as the in-built hubitat one wasn't reporting right, and a number of Ikea Tretakt smart plugs and two Tradfri buttons using dandanache's Ikea drivers. The other user drivers I have are for wifi devices. I'll send a private message with a print out of the zigbee details showing whats more active on the network.
Thanks again and happy new year.
What Wi-Fi devices do you have and what drivers are being used? Some of the community wifi integrations may cause issues, especially if the device goes offline.
What polling interval do you have set on the Wi-Fi devices?
I have a Govee LED strip behind one TV that I've had installed for nearly 2-years, and two tasmota energy monitoring plugs that I installed last week. Aside from that, I have two thread-based contact sensors (Eve) integrated through Matter from a HomePod. The only other devices connected are an EV car charger and water heater, both MyEnergi, but they are only integrated through a cloud API call.
The Govee light strip is set to poll every 300-seconds and the plugs every 60-seconds when they're in use.
Looking at the logs, the hue bridge is eating up alot of busy time, but I have it on a separate zigbee channel. It is set to poll every 1 minute but I've moved to the V2 api so I'll move that to 3 minutes.
Again, apprecaite your help.
Mark these as "disabled" in the device list for now, to rule it out as an issue. Power monitoring stuff can be very chatty sometimes.
That is only relevant if you compare it to the total busy time at the top of the stats page. For example if total busy time is 2% and Hue Bridge shows 90% of total busy it just means it was busy for 90% OF the 2%, so not very much at all in the grand scheme of things. The percent of busy just ranks things as most busy to least.
For the Tasmota plugs, I'm using Garyjmilne's drivers. I only installed these last week and this problem has been on-going for a few months before that. Before I changed the Zigbee channel, it was going offline once a day.
I see now re. the busy time. Its showing total is 11.3% with the hue bridge taking up 75% of that.
If changing the channel has helped don’t rule out interference just yet.
You already received all the advice I had. @kkossev and @dandanache drivers are fine. Your TP-Link Deco mesh probably should be your main suspect at this point. You wrote it was not installed long ago, does its installation coincide somewhat with your problems ?
