Same for me
1 Like
The web app is very buggy! Use mobile app to update your devices.
I have five TRVZB valves but no update found, I’m still on 1.4.1 with ns Panel pro 120 at 4.3.3.
1 Like
Based on my experience with Sonoff firmware upgrades, I decided not to update the TRVZB before leaving the weekend house… Smart move! 
1 Like
It offered me an update for 8 out of 10, and the next day for the remaining 2 heads. This time I didn’t notice any serious problems, but the update of just one device, which took about 45-50 minutes, is terrible and the new adaptive mode doesn’t seem to work as it should.
1 Like
I find the update using NSPanel Pro, but the valves are not taking the update. No error message, just “success”
1 Like
How did you find the update via nspanel pro?
Oh I meant I could see the available update on trvzb devices connected to the NSPanel
I also have trvzb connected to a bridge, these could be upgraded without any issues
Yesterday, the Adaptive Mode option suddenly appeared on 1 of 8 updated TRVZBs (fw1.4.4), but the functionality is strange.
1 Like
Judging by the tactical silence around the TRVZB firmware update, it looks like the dev team got absolutely steamrolled by the issue. And it doesn’t seem to be just the download or installation part — something else clearly refused to cooperate 
Looks like bailing out from the TRVZB ecosystem into Home Assistant was the right call after all — the thing has been running smoothly for days now with schedule automations, mode switching, and external sensors. Quite a contrast to the rather… underwhelming cooperation with the NSPP, where the thermostats kept periodically breaking up and getting back together like a dysfunctional couple.
Sonoff even cooked up a so‑called quirk for TRVZB in ZHA. It’s oddly structured, the code is half‑baked and a bit twisted, but hey — it works. When I get some time, I’ll dig into it and maybe polish a few things.
Addendum
Well, it no longer works. After updating HA to 2026.2.0, many entities stopped being available. It turned out that the working quirk is now built into ZHA, and simply removing the Sonoff‑made one was enough to regain control. Let’s say it’s not full control, because you still have to account for the peculiarities of the TRVZB firmware.
For example, there is absolutely no way to convince the TRVZB to cooperate with an external temperature sensor. But there is a workaround. The Better Thermostat integration handles this elegantly by taking over the control logic and bypassing the buggy TRVZB firmware. It treats the TRVZB as a not‑so‑smart device. And because it offers extensive configuration that doesn’t get into deep conflicts with the device’s twisted firmware logic, even features equivalent to Adaptive Mode work correctly — and in several flavors. Also with firmware 1.4.4, which finally downloaded onto all of my TRVZBs.
Finally, everything behaves according to common sense and expectations in ZHA. If anyone is tired of wrestling with the TRVZB in eWeLink, I highly recommend going in this direction.
1 Like
I updated software for
ZBBridge-P => 3.0.0
TRVZB => 1.4.4
but there is no adaptive mode …
I also noticed invisible “boost” mode on the history chart of TRVZB 
I used this mode several times last week but it was never shown on the history chart…
For example this night at 1a.m. boost mode was turned on… temperature rised but set value didn`t change at all… it is confusing… I know the setting is just temporary (depending on the timer set by the user - 7min in my case) but there should be at least some peak on the chart noticeable…
An update to adaptive mode appeared on all thermostatic heads. After three days of testing, I turned off adaptive mode on all of them. The temperature constantly exceeded the set temperature and was not constant at all. The heating was unnecessarily turned on almost the entire time I was testing the new adaptive mode.
4 Likes
Unsurprising really.
If adaptive mode reduces the water (heat) flowing into the radiator when the temperature is approaching the setpoint value, it has to keep the boiler and pump running. (Or at least the pump).
Otherwise the flow is either on or off, so there would be no point in having the variable valve opening/closing.
This is why I said elsewhere in this forum, I don’t think PID control is really suitable for Gas / Oil fired boiler, in domestic central heating systems.
I think you are just confirming it…
1 Like
@Ramsdale Yes, it is difficult to control the temperature of homes with gas/oil boilers using only thermostatic heads. Here, even more emphasis should be placed on boiler control.
Actually proportional control works absolutely fine for gas boilers and radiators - normal TRV heads have a temperature proportional response, and do not just switch on and off with temperature. As the room temperature comes up, the valve gradually closes down until it’s fully off, but it will then gradually open to allow water through slowly to maintain the temperature.
They are actually quite good at holding a temperature once they have stabilised, but they don’t have such precise or consistent control over the actual temperature, which is why they have numbers eg 1-5 instead of temperatures.
The question is one of tuning rather than unsuitability. Here is a room which I have updated the firmware on and enabled the adaptive control:
You can see that the temperature overshoots and then shuts off while the temperature drops, at which point it does gently come back on to demand heat and maintain a quite flat temperature.
The problem is with the initial response, not longer term control. Once the temperature has stabilised it does quite a good job of keeping it at the set point. Perhaps the PID coefficients need tweaking a bit to reduce that overshoot. The difficulty is that radiators and heating systems vary so much it’s difficult to find a one size fits all solution.
Here is a TRVZB being controlled by Versatile Thermostat using direct valve control:
The valve position is modulated directly to control temperature, so it’s possible to get very good control if things are set up right.
There are two ways to control a boiler with smart TRVs - You can control the boiler with its own thermostat independent of the TRVs, in which case the TRVs are acting mainly as temperature limiters for the room. You can also control firing of the boiler by demand indicated from the TRVs - how you do that will have a big impact on how well the overall system works.
A single radiator calling for heat might not be a very efficient way to do it, but if you can restrict boiler firing to only when a certain number of TRVs want heat it will probably be more efficient.
The problem here is one of tuning, not that the concept is fundamentally flawed.
I think you’re oversimplifying a system that is far more complex in practice than in theory. Yes, classical mechanical TRVs behave approximately proportionally — but that doesn’t automatically mean that electronic TRVs with PID‑like logic will behave well in a real water heating system. The two mechanisms are fundamentally different:
- a wax/bimetal TRV reacts slowly and locally,
- an electronic TRV reacts quickly, digitally and with sampling delays,
- the heating system has huge thermal inertia,
- the boiler has a minimum firing power that often exceeds the demand of a single radiator,
- flow characteristics are nonlinear and installation‑dependent.
This is why “proportional control works fine” is only half true. It works fine within the physical constraints of a mechanical TRV, not necessarily when you try to apply a textbook PID to a system with minutes of delay and kilograms of water to heat.
Your graphs actually illustrate the core issue: the overshoot is not just “a tuning problem” — it’s a symptom of the system’s inherent inertia and the fact that the actuator (the TRV) cannot influence the boiler’s minimum output or the thermal mass of the radiator. No amount of PID tweaking will remove that completely, because the physics don’t change.
Direct valve modulation can work well, but only in ideal conditions:
- large radiators,
- stable flow,
- well‑balanced system,
- boiler with wide modulation range,
- minimal cross‑room interference.
Most real installations don’t meet those assumptions.
And regarding boiler control: whether TRVs call for heat directly or the boiler runs independently has a massive impact on stability. A single radiator calling for heat from a boiler with a 6–10 kW minimum output is almost guaranteed to cause oscillations, regardless of how clever the TRV algorithm is.
So yes — the concept isn’t “fundamentally flawed”, but it’s also not as universally applicable as you suggest. The limitations come from the system’s physical properties, not just from imperfect tuning.
The point is that it depends entirely on how the supposed ‘smart’ algorithm has been implemented, and what factors have been taken into account.
If it is a basic PID controller with fixed gains, then the developers will have had to make an educated guess as to what coefficients to use and there is such a large variety of heating systems that it is likely that it will not be satisfactory in all use cases.
I disagree here though - the overshoot in the first graph is due to a poorly tuned controller, not a problem with modulating the opening of the valve. The graph shows that the valve remained open for much longer than necessary, I would guess because of integral wind up or something like that. I have no idea if the developers made any attempt to account for the flow curve of valves or if they treat it purely linearly (which would not apply to most valves).
To prove the point that is is possible to get good results, here is a different TRVZB being controlled by a PI controller:
The temperature was maintained with no more than 0.1ºC from the set point throughout the day, which I would say is quite acceptable for a domestic heating system.
To explain the graphs, the red at the top indicates whether the boiler was firing - in this case it was set to fire whenever any radiator requested heat.
The next graph is the room temperature and set point. The orange heating indicators are when the controller was calling for heat. Any power below 5% is regarded as off, so that the boiler does have he opportunity to switch off at very low power levels.
Below that is the actual valve opening position, and at the bottom is the power level commanded by the controller.
I use an automation to interpret the power output of the PI controller in order to linearise the response of the valve, since most radiator valves are the quick opening type. This effectively compresses the 0-100% output of the controller into the linear portion of the valve movement, eg 20% valve opening gives something like 80-85% flow.
The other function it has is to force the valve to slam closed rather than gently close from a low opening percent - the TRVZB quickly loses its calibration if that is done repeatedly. Whenever the controller commands 0% opening, the automation fully opens, then closes the valve. So far I have not had any valve using this go out of calibration.
Just for the sake of comparison here is the same valve using the original on/off logic and accuracy set to 0.2º:
Adaptive mode is not suitable for all heating systems.
Adaptive mode is not a “magic button” that will work for everyone.
It is only recommended for buildings with radiators that are larger than the heat demand of the room.
A suitable heat buffer is also required.
1. Oversized radiators (Low-temperature ready)
Adaptive mode aims to find the lowest possible supply temperature that will maintain thermal comfort.
If the radiators are “on the same level,” the system will have to operate at high parameters, which precludes subtle adaptive adjustments.
A large heat transfer surface allows for operation at a low delta, which is an ideal environment for algorithms that learn from the building’s inertia.
2. A heat buffer as the “lung” of the system
Without a buffer, in systems with low water flow (e.g., radiators only with thermostatic valves), timing issues arise.
Cycle time is the number one enemy of the lifespan of compressors in heat pumps and heat exchangers in gas boilers.
The buffer allows the heating device to operate in long, stable cycles, even when heat demand is momentarily minimal.
In buildings with low inertia or with a poorly sized heat pump, adaptive mode can actually worsen comfort, causing the device to frequently turn on and off instead of providing smooth modulation.
In summary: Proper heating system design (stable hydraulics and an appropriate heating surface) is essential. Only then can intelligent automation be implemented.