max radiator valve algorithm

[malc_b]

Hi,

Is there anything around on the max thermostat algorithm?  I'm playing with one on the bench and what I'm seeing I think is:

1. set 0.5C under room temperature - no change in valve position

I'm guessing it has a deadband, maybe +/- 0.5C to stop lots of motor movements.

2. Set 1.5C under room temperature - valve closes a bit but then no change (at for 30 min)

I'm guessing that the control is just linear, with maybe 50% being spot on the set temperature and then higher or lower if temperature is lower or higher than set point.  I think the band is 1.5C so something like this I think:

Code Auswählen Erweitern


If ((Tdesired-Tset)>0.5) {
  Valve = 50 - (Tdesired-Tset-0.5)*50;
  if (Valve < 0) {
    Valve = 0;
  }
}
If ((Tset-Tdesired)>0.5) {
  Valve = 50 + (Tset-Tdesired-0.5)*50;
  if (Valve > 100) {
    Valve = 100;
  }
}


I'm not seeing any sign of integration to reduce the error.

Update:  Maybe I am seeing some integration.  If I change the set point to the room temperature I would expect nothing to happen but it does.  But I can't make any sense of it.  Room is 21.5C, set 19C, valve goes to 0%.  Set 21.5C, valve goes to 61%.  Hmm.  I've just noticed I have set -1C offset so I'll need to try this again.

[chapelhill]

The valve position will be based on a pid (proportional, integral and differential) algorithm which learns ove time.

Easiest way to see what it is doing is to log the temperatures and valve positions on an active radiator.

Note the design of these valves is for european installation position which tends to be at the top level of the radiator and horizontal, but UK tends to be vertical and at the bottom of the radiator.

Don't get excited about valve temperature but room temperature as that is what we are controlling the valve temperature is just a comprimise to achieve control and works fine in my opinion.

[malc_b]

Do you have linked wall thermostats?  On just the rad valve alone I don't think I see any integration but maybe I'm wrong.  I don't have the valves in system so I'm just manually setting the desired temperature relative to the current temperature and seeing what happens.

Do you have any typical logs you could share?  For a room with just a radiator valve, no wall stat.

[chapelhill]

I will get some graphs over the weekend. I have two thermostats as well as radiator valves but the most unreliable part of configuration is the association between the devices which seems to get forgotten and is then an extra device to set the schedule on and to switch into different modes.. Running with a thermostat did not seem to make temperature control noticably different.

[ggaljoen]

Was just thinking; can you do something with the boost function?
There you define the valve % and the time, if needed you could deactivate also.

[chapelhill]

Attached is an image of a graph from earlier this week.
It is for a single radiator in a large room with the valve set for 15 degrees during night, 19 in the morning, 15 during the day and 17 in early evening.
The blue shows the desired temperature. Missing bits are at 15 degrees.

You have the boiler function overlay in black labelled boiler it shows its valve position as it is a max 240v switch.

The separate thermostat that is not linked to anything is labelled Actual temp in a browny red colour. The thermostat is about 2 meters from the radiator at about 1.5 meters from the ground resting on top of a dresser cupboard.

The weather about 3-4 degrees below zero overnight and warmed up after lunch to about 4 degrees which it maintained overnight.

I hope that is of some help.

[malc_b]

Thanks

That's an interesting graph.  The valve temperature measure is significantly affected by the heat from the radiator.  The actual temperature isn't changing as much as the valve temperature.  Not surprising and I think eq3 have allowed for that as this would explain the dead band the rad valve has.  It's also interesting that the valve temperature doesn't match the desired temp setting.  If it was integrating and correcting for error then you would expect this to bracket the desired temperature.

[chapelhill]

These valves are designrd for eurpopean plumbing installtions which tend to have the valve at the top of the radiator and in a horizontal orientation. This will tend to reduce the net difference between radiator valve temperature and perceived room temperature and I suspect the horizontal orientation being horizontal and pinting away from the radiator would reduce the direct heating affect of the radiator on the valve.
With all the shortcomings it still performs pretty reasonably. I did consider getting the radiator connections changed but I don't think it wouod make much difference, perhaps just a change to teperature offsets, and now the credit limit is manageable I am happy with my system and looking to expand it.

[malc_b]

I don't think top or bottom makes a lot of difference.  If anything lower would be colder air.  Horizontal or vertical I suspect makes a fair bit of difference because when vertical it is above the hot pipe to the radiator.

FYI the new max basic thermostat looks a better design to me, as well as being a lot quieter than its bigger cousin.  The basic has a 45 deg top.  If you mount that vertically the 45 would point up towards the radiator.  The air vents in the basic are under the 45 deg slope so the natural air flow up the radiator would pull air through the valve from away from the radiator.  The valve would still get the IR from the radiator but better to have a good airflow design than not.