i’ve 6 TRV-ZBT and 3 of theme are connected to 3 SNZB-02P as external sensors located in the same rooms as the thermostatic valves.
Everything worked fine for about two weeks but, for the past five days, a message has been appearing that says something like “TRV now uses the external sensor as the room temperature reference again”. The strange thing is that message appears cyclically about every hour. I disconnected the external sensor from TRV’s configuration and,after a while,i reconnected it and everything seemed fine, but now another TRV is sending the same message!
The TRV have a RSSI of about -60db while the corresponding external sensor have -57db.
Since you’re using a Sonoff Bridge Max, I’d be interested to know whether all three TRVs are running the same firmware version. Also, do all three SNZB-02P sensors show the same behaviour, or does the issue move between TRVs? This might help determine whether the problem is related to the TRV firmware or the Bridge Max.
Yes, that’s correct. In the Sonoff ecosystem, the external temperature readings are relayed through the Zigbee bridge, rather than being sent directly from the SNZB-02P to the TRV.
Since the issue is now affecting almost all your TRVs and seems to move from one device to another, I’d be interested to see whether the Bridge Max firmware or its handling of external temperature updates could be involved. Have you tried restarting the Bridge Max since the issue started?
A new TRV-ZBT firmware version 1.1.3 (0x00001103) has been released. It includes improvements to Zigbee reconnection after prolonged disconnections and error recovery to enhance device stability.
I’d suggest updating your TRVs to this version and checking whether the issue persists. I’ve installed it myself and so far haven’t noticed any regressions.
It might be worth giving it a try before making any further changes to your sensor configuration.
i think it was a problem related to the firmware update because, on one of the TRV,once it was disconnected, notification about the availability of new firmware appeared for a few seconds. When I restarted Bridge Max, the firmware update notification appeared on all the TRV thermostats.
Since the firmware update didn’t fully resolve the issue, I wouldn’t assume it was the root cause. At this point, I’d be more interested in the communication between Bridge Max and the TRVs, rather than the firmware itself.
Could you clarify whether the TRV actually loses its connection to the Zigbee network, or whether it simply stops receiving temperature updates from the external sensor? Also, does this happen at regular intervals?
These details could help narrow down where the problem really lies.
Hi, I’m having the same problem—10 TRV-ZBT units (SW 1.1.3) + SNZB-02DR2 and SNZB-02P + ZBBridge-U (SW 2.10.0):
When installing with the previous version software of the TRV-ZBT and ZBBridge-U software (SW 2.8), only one thermostatic Valve kept disconnecting repeatedly. I tried pairing it repeatedly, resetting it to factory settings, and using a different type of temperature sensor, but nothing worked. After an initial period of 24–36 hours during which the thermostatic Valve functioned, it would repeatedly disconnect from the sensor.
Eventually, I replaced this thermostatic head with another unit, and everything worked stably for about 3 weeks
Now, in my naivety, I installed the new software version 1.1.3 on the TRV-ZBT and version 2.10.0 on the ZBBridge-U. Since then, my eWeLink app has been literally flooded with messages from various thermostatic heads and various thermometers—see the attachment
My observations:
It’s not a signal strength issue; the signal is strong
The network topology consists of a ZBBridge-U + approximately 10 switches + 10 MINI-ZBRBS units
If I reset the ZBBridge-U, the messages stop for a while
I haven’t been able to identify any direct correlation, but it seems to me as though it somehow depends on the time synchronization (time offset) between the software running on the thermostatic heads and the ZBBridge-U software.
That’s why resetting the ZBBridge-U always has a different effect and “fixes” the error message at a different time and for different thermostatic heads.
I’d really appreciate it if the eWeLink and Sonoff teams could fix this, because I have to admit I’m already frustrated by the number of different issues—among other things, I’m also waiting for a fix for the MINI-ZBRBS.
I’m running several TRV-ZBT units with firmware 1.1.3 as well, but I’m using ZHA rather than the Sonoff ZBBridge-U / eWeLink ecosystem. I haven’t observed anything similar - no repeated notifications about external sensors going offline and coming back online.
This makes me wonder whether the issue is actually related to the TRV firmware itself, or rather to how the ZBBridge-U / eWeLink handles communication and synchronization with external temperature sensors.
Questions:
When these notifications appear, does the TRV actually stop using the external temperature sensor, or does it continue working normally despite the notifications?
Do you see any corresponding loss of availability of the temperature sensors, or are these messages generated only by the TRV?
Did you notice whether the problem started immediately after updating the ZBBridge-U to firmware 2.10.0?
Since both the TRV firmware and the bridge firmware were updated, it’s difficult to identify which change might be responsible.
I wouldn’t rule out a firmware regression, but I would be cautious about blaming TRV-ZBT firmware 1.1.3 at this point. The fact that the same TRV firmware works without these symptoms in ZHA suggests that the bridge or the eWeLink integration might be worth investigating as well.
Hopefully someone from Sonoff will take a closer look at this, because repeated notifications across multiple devices certainly don’t look like normal behaviour.
Hi Jam3, thank you for your response. Here are my answers:
I happened to notice that the TRV does indeed stop using the external temperature sensor, but the outage seems to be very brief. If you look at the messages, you’ll see only messages about the sensor reconnecting—no messages about the sensor disconnecting. Actually, I’m not sure if this message comes from the TRV or from ZBBridge-U.
The thermometers are still online. In most cases, I have two TRVs connected to a single thermometer, and the malfunction affects only one TRV.
The problems began immediately after updating the TRV software on the relevant valve. Since I also had a problem with one valve head before the upgrade, I think the issue isn’t related to the new software but to the TRV restart, and there’s some time-related connection between the TRV booting up (completion of the upgrade or insertion of batteries) and the ZBBridge-U restart. That’s why the problem affects different valve heads each time after the ZBBridge-U restarts.
The fact that the temperature sensors remain online, and that only one of two TRVs assigned to the same sensor is affected, is particularly important. This makes a general Zigbee connectivity issue or a problem with the thermometer itself much less likely.
Your observation about the TRV reboot is also interesting. Both a firmware update and battery replacement involve restarting the TRV, so I think it’s worth investigating what happens when the device reconnects to the bridge. One possibility is a synchronization or initialization issue between the TRV and ZBBridge-U, rather than an actual loss of communication with the external sensor.
I would be particularly interested in one test: have you tried restarting only an affected TRV (by removing and reinserting its batteries), without restarting the ZBBridge-U? Does the same problem reappear? If so, it would help establish whether the TRV reboot alone is sufficient to trigger the issue, or whether the interaction between the TRV and the bridge is also required.
I think this is something Sonoff should investigate at the firmware level. The fact that the thermometer remains online while an individual TRV temporarily stops using its temperature is a rather important clue.
Hi Jam3, thank you for your valuable feedback. I was finally able to get the system to a state where it doesn’t send any thermometer disconnection messages for 12 hours:
By repeatedly restarting ZBBridge-U, I managed to get the system to the point where the thermometer disconnected only from one TRV-ZBT
For the affected thermostatic radiator valve, I selected a thermometer in another room as the external sensor. When no disconnection message arrived for 1 hour, I switched the external sensor back to the thermometer in the room where the TRV is located. Since then, no sensor disconnection messages have appeared.
It’s a paradox—the entry in the TRV-ZBT regarding the external sensor must have been “corrupted” in some way, but not completely, since it worked most of the time. I’m not a Zigbee specialist, so I don’t know exactly how the topology works, but one theory occurred to me. Normally, communication proceeds as follows: SNZB-02P ↔ Router (MINI-ZBRBS) ↔ ZBBridge-U ↔ Router (MINI-ZBRBS) ↔ TRV-ZBT, which would explain the strong signal from both the thermometer and the valve. I don’t know if this is possible, but if a different path were used to transmit control data from the external thermometer to the TRV-ZBT—for example, SNZB-02P ↔ Router (MINI-ZBRBS) ↔ ZBBridge-U ↔ TRV-ZBT—that would explain the strong signal displayed for the SNZB-02P and the TRV-ZBT, but also the signal dropouts from the thermometer acting as an external sensor for the TRV-ZBT.
It would be really great if someone from the eWeLink and Sonoff teams could comment on this theory and get back to us.
Thanks for sharing the results of your tests. This is actually a very interesting finding. The fact that changing the external temperature sensor to another one, and then assigning the original sensor again, seems to have resolved the problem is particularly noteworthy.
I would be a little cautious about the Zigbee routing explanation, though. Zigbee routing and the mechanism used to provide an external temperature reading to a TRV are two different things. The signal strength reported by the thermometer and the TRV doesn’t necessarily tell us anything about how the external temperature value is actually delivered to the valve. It simply doesn’t work exactly that way
I’m increasingly wondering whether this is related to how ZBBridge-U manages external temperature sensor assignments. Perhaps, after a TRV restart or a bridge restart, some internal state associated with the sensor assignment isn’t properly synchronized. Reassigning the sensor might then force the bridge and the TRV to establish that relationship again. Of course, this is only a hypothesis, but your latest experiment makes it worth investigating.
It would be particularly useful if Sonoff could clarify how external temperature sensor assignments are implemented in the ZBBridge-U ecosystem, and whether assigning one sensor to multiple TRVs is officially supported. Hopefully someone from the Sonoff team will take notice, because your testing has provided some very useful clues.
thanks for your support. In these days i made some tests:
i moved the external sensor near the bridge but nothing changed
i swapped the external sensor on this TRV with another one but nothing changed
i replaced this TRV with a new one and it worked correctly
i reassigned the correct external sensor and was fine.
finally, i repositioned the external sensor in the correct place and now it has been working fine for 24 hours.
From these test i deduced that the problem is the TRV so i’m trying to refund.
Replacing the TRV with a new unit and getting everything working again certainly points towards the original valve being involved. However, I wouldn’t be completely certain that this is a hardware fault. Replacing the TRV also means going through the pairing and external sensor configuration process again. So it’s possible that the issue was related to the internal configuration or the way the original TRV handled its external sensor assignment.
The fact that reassigning the sensor also appeared to resolve the problem is worth noting. In any case, I think you’ve done a pretty thorough job of troubleshooting. If the replacement unit continues to work reliably, requesting a refund for the original one is perfectly reasonable.
It would be interesting to see whether the same problem could be reproduced with the original TRV after a factory reset and fresh pairing, but I certainly wouldn’t expect you to spend even more time troubleshooting a device that should simply work.