Repository navigation
TRVZB : Can't retrieve constant temperature and "warm" radiator that I had with my dumb TRV #1564
Replies: 2 comments 16 replies
Yeah this is an issue with valves and radiators. They are not inherently linear devices, whereas the tpi algorithm is written with linear power output change in mind. There are a few things you can try. Firstly, if your radiator does not start to get hot until the valve is open at 19%, then set the minimum opening degree to 19%. That means that as soon as the VTherm calculates it needs power, it will move the valve to the 19% position. Note that this is separate from the opening threshold, which acts like a cut-off for the minimum calculated power - eg below that threshold any calculated power is considered to be zero, so VTherm calculates it wants 4% power and the opening threshold is 5%, then nothing happens until the power required increases to 5% or above. There is no mechanism to set the maximum valve opening at the moment, but there is a feature request open for it here: #1554 Assuming that gets added, this would scale the 0-100% power request to the functional range of the valve, but that still doesn't mean its heat output will be linear with that range because the heat output depends on the difference between the average temperature of the radiator and the room, it's not fixed like an electric heater is. As the room temperature gets closer to the water temperature the power output of the radiator drops. If you have heating with a low flow temperature (eg a heat pump at 20-30ºC) the this change could be considerable. If you have an older style boiler with a high flow temperature, it won't be as significant. It's also worth looking at what the two TPI coefficients do - internal sets the responsiveness to internal temperature error, eg how far away is the room from the set temperature, and external sets the amount of output needed to offset the heat loss to the outside. That is directly proportional to the temperature difference between inside and outside. The steady state heat loss requires a lot less power to replace than actually heating the room does, so you want the output from the internal coefficient to be much more significant than the external. Start with your internal coefficient at something like 0.5, and the external at either 0 or 0.001. See how it behaves, and only then increase the external coefficient by small amounts - 0.001 or less. If you find the temperature constantly overshoots, you have gone too far and should reduce the external coefficient a bit. Personally I think the TPI algorithm is not great for controlling radiator valves - it expects a linear response, but radiators do not give that so it's hard to get a satisfactory result. I have switched to using the internal logic of the TRVZB with the dead band set to 0.2º There is a proportional/integral algorithm in development which should work better as it can inherently compensate for the non-linear response. |






Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Hello,
Before installing Versatile Thermostat and the TRVZB I manage to had constant temperature (at least with few variation) with a dumb graduated TRV.
e.g. during a typical day :
The drops are just because I opened the window. The radiator were always warm, not very hot (except to compensate the windows opening).
Now I have installed a VTherm to control a climate of the TRVZB. They are installed on my radiators with no adapter, and the screw is well tighten. I am also using external temp sensors
With following parameters
max_closing_degree = 10min_opening_degree = 15opening_threshold = 10I am having hard time maintaining constant temperature since. I observe a lot of oscilliations : Next picture is showing temperature variation for two days + the valve opening degree. During the days the valve is closed (at 90%) so obviously the temperature is dropping, but then it is hard to maintain the 19 degrees like before. (there is a windows opening around 8 an, hence the temperature drop)
The thing that scared me is the sensitivity of the
valve_opening_degree.I witness the followings
I have the same behavior in the other rooms. It seems that the playground is only between 20 and 30%. So I am not sure the work I am doing with the
coeff_intwill have an influence at all. Basically I have the impression to have a on/off radiator.For the record here is the full configuration
Any help appreciated :)
Thanks in advance,
Best !
All reactions