Out of interest: Why use temperature based control (introducing the need for OVP calibration)

[mjkl-gh]

Hi there,

First of all thanks again for this integration. I've been trying it out for a week now and after some initial hurdles it is now working fine.

However, since I don't have a thermostat other than SAT, I'm now relying on home assistant to heat my home. Something i'm not sure i'm completely comfortable with.

Before trying sat I was busy building my own thermostat based on ESPHome, however this wasn't working that great as I had yet to learn "low load control" existed. So my boiler tended to start oscillating. Having learned the term "low load control" I came across your integration and decided to try it out first before building something myself again, but since I'm not sure I like the dependency on home assistant I am doubting to switch back again.

One question I had was: Why use temperature based control. From what I understand doing that introduces the need for establishing the OVP value to decide when low load control needs to be active.

My idea was to use the output of the pid to set the relative modulation level directly. Then to enable low load control, take the reported minimum modulation level (LB of opentherm message id 15) to determine if we need to run in low load control. For example, My boiler has a reported minimum modulation level of 25%. That means that any PID output >25% would set the relative modulation level to (pid-output - 25%). Any output of <25% would set the relativel level to 0 and cycle the requested temperature according to the duty cycle required (pid_output/25%)

This would eliminate the need for determining OVP, as the OVP is effectively retrieved from the boiler.

I was wondering on your thoughts on this

[sergeantd83]

#95 will clarify things about how SAT is working against other approaches. SAT performs complicated calculations in order to strategically control the boiler without cycling while maintaining the room temperature at setpoint.

[mjkl-gh]

I've read #95, maybe I misunderstood, but it felt like a comparison to other systems that don't do any low load control at all.

To dig in deeper on the "without cycling part" and why I thought power based control (including low load control) would be better:

When I calibrated sat on the first try, I fully opened the TRV of the floor heating in the attic, as I thought that was adviced in the guide. It therefore determined my OVP to be at 38 degrees. Since then I closed the TRV on the attic as that floor is under construction.

This meant that at a point sat was sending a temperature of 38 degrees continously to the boiler. However, since closing the TRV my actual OVP had changed to 45 degrees. This meant the temperature of the boiler would continue to rise to 45 degrees. At 38+5 degrees my boiler passed an internal threshold and turned off due to overtemperature and started cycling. This was made worse by my model of boiler missing a return temperature probe, meaning it would oscillate/cycle every 30sec or so.

In my mind power based control would have prevented this, as in this case the thermostat instead would have sent a max_modulation_level of 0 with a temperature target of something unreachable, let's say 80 degrees. This means that after the attic TRV is closed the OVP would have automatically shifted to 45 degrees, as nothing in the system is looking at the temperature of the water anymore, only the requested, and actual power.
If the house would heat up too quickly due to more power going to the living room than expected, the PID would correct that and request power would quickly go lower than the minimum of 25%. This in turn would activate low load control, only turning on the boiler on the duty cycle. As it eventually would have using the current implementation.

However, I now fully ignore the other inner workings of sat or the tuning of the pid values. Therefore I was hoping to pick your brains on why temperature based control was chosen over power based control.

[sergeantd83]

If you have a multi-room installation, you can update to the latest develop version, enable the Dynamic Minimum Setpoint (Experimental) option, and then select your boiler’s flow water temperature in the General Settings. This mode “senses” the flow water temperature and adjusts the control setpoint accordingly in multi-room environments.

Regarding power control:

A strict power-based control strategy relies on accurate and continuous feedback of the boiler’s relative modulation value ( ID 17 ). However, in practice, many OpenTherm boilers either:

Do not respect Maximum Relative Modulation management, or

Report fixed modulation values (e.g., always 0% or frozen data).

In such cases, any thermostat that bases its control loop on the reported relative modulation will fail, because the feedback signal is unreliable. SAT avoids this limitation by not relying solely on ID 17. Instead, SAT uses a hybrid control model with its own modulation-inference layer and multiple fallback mechanisms. This allows SAT to maintain stable and safe control even when the boiler does not provide valid modulation feedback.

As a result, SAT is currently the only OpenTherm thermostat capable of robust and predictable control across all OpenTherm boilers, including those with partially implemented OpenTherm behaviour.