Page 17 of 18 FirstFirst ... 712131415161718 LastLast
Results 161 to 170 of 176

Thread: OpenTherm control behaviour

  1. #161
    Automated Home Legend paulockenden's Avatar
    Join Date
    Apr 2015
    Location
    South Coast
    Posts
    1,719

    Default

    I think this could be solved in the controller.

    Right now it appears to simply to do a max() of all of the zone demands to calculate the value it sends to the BDR91 or OT bridge. If, instead, it did an average I think this situation would be avoided.

    When a new zone kicked in, wanting full heat the boiler would be turned up a bit, but not to maximum.

  2. #162
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,361

    Default

    Quote Originally Posted by paulockenden View Post
    I think this could be solved in the controller.

    Right now it appears to simply to do a max() of all of the zone demands to calculate the value it sends to the BDR91 or OT bridge. If, instead, it did an average I think this situation would be avoided.

    When a new zone kicked in, wanting full heat the boiler would be turned up a bit, but not to maximum.
    Nice idea but not workable unfortunately.

    The controller has to consider that this "new" zone may in fact need full boiler flow temperature to achieve its set point - as both the set point and the thermal characteristics of the radiator and room could be very different from other rooms.

    I think most houses have a room or two with undersized (or obscured or blocked) radiators that struggle to meet their target in winter. In our house our dining room, (partially blocked 2m long rad - colder at the bottom - needs replacing, have already tried to flush it out on the lawn..) living room (3 old style double panel radiators without convection fins) and hallway (not sure why that one struggles as it is quite large and is a convector, but it does, possibly the hallway and front door are draughty) and in the depth of winter they all need full flow temperature just to reach a comfy 20/21 degrees.

    On the other hand some of our radiators are oversized - in particular both upstairs bedrooms, and overshoot is more of a problem there. In these circumstances if the controller were just to average the demands of the different rooms the ones that struggled would simply never reach their set points. So doing a max() of all the demands is actually necessary if you expect all rooms to achieve their set points if there is a significant variance in the abilities of the radiators from one room to another...

    I talked about a possible solution in another thread as you may recall, something that could actually be implemented purely in controller software and which I think would help minimise overshoot -

    1) When there is a sudden jump in the combined heat demand of all the rooms (after the max() calculation is made) for example one cold zone suddenly being scheduled to be on, don't just suddenly turn the boiler up to maximum, ramp up the heat demand sent to the boiler progressively over a 10-15 minute period so that other zones have time to sense a small amount overshoot and start closing their valves a bit. Because you are only ramping it up not limiting the maximum you still ensure that rooms that struggle will get there - they may just take about 10-15 minutes longer to get there, but I think that is an acceptable compromise.

    In any kind of feedback system with a long time constant like this, responding well to a "step change" is the hardest thing to get right. So don't cause a step change in the flow temperature in the first place. Ramp it up smoothly over a reasonable amount of time.

    2) Increase the wake-up frequency of HR92's for a while when a step change in demand and therefore flow temperature has occurred - it looks from the protocol that the sleep time of HR92's is controlled by a message from the controller, potentially dynamically, although I don't think they are varied at the moment. Having the HR92 respond more quickly to sensed temperature changes will lessen the degree of overshoot when another zone has called for a large increase in flow temperature. Once conditions have stabilised again and all zones are back in the proportional band you can go back to less frequent wake-ups to conserve battery life.

    Combining 1 and 2 together should in theory greatly improve the temperature control of zones at equilibrium when a new zone comes online and calls for a lot of heat, but won't prevent a weak zone from reaching its target, at most delay it slightly.
    Last edited by DBMandrake; 2nd December 2017 at 08:28 PM.

  3. #163
    Automated Home Legend paulockenden's Avatar
    Join Date
    Apr 2015
    Location
    South Coast
    Posts
    1,719

    Default

    Quote Originally Posted by DBMandrake View Post
    Nice idea but not workable unfortunately.

    The controller has to consider that this "new" zone may in fact need full boiler flow temperature to achieve its set point - as both the set point and the thermal characteristics of the radiator and room could be very different from other rooms.
    The temperature of the water in the heating loop will be higher than the temperature in the zone in question, so the new zone WILL heat up. Just not as quickly as with the boiler going full tilt.

    Also, if we're using an average (or perhaps a weighted average) to calculate the boiler demand the heating loop water will get a bit hotter for a while - not not the stupidly max-temp hotter that currently happens.

    P.

  4. #164
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,361

    Default

    Quote Originally Posted by paulockenden View Post
    The temperature of the water in the heating loop will be higher than the temperature in the zone in question, so the new zone WILL heat up. Just not as quickly as with the boiler going full tilt.
    You're not quite getting what I'm saying. For any given radiator/room/outside temperature combination there will be a minimum flow temperature with the radiator valve fully open and flowing as much as possible that will be able to reach the room set point.

    If the flow temperature is lower than this the room wont get up to the set point. Period. Not more slowly, it just won't make it at all. If most of your rooms only need a flow temperature of 50 degrees under certain conditions to meet their set point but one room comes online that needs 65 degrees but we just take the average of the different zone demands and end up at 52 degrees, that room wont ever reach the set point target...it will sit there forever with the HR92 fully open and the room below the target.

    The flow temperature needs to be at least as high as the most demanding room's requirements - the radiator valves in the other room can then close to reduce how much heat they put out at that boiler flow temperature.
    Also, if we're using an average (or perhaps a weighted average) to calculate the boiler demand the heating loop water will get a bit hotter for a while - not not the stupidly max-temp hotter that currently happens.

    P.
    A time weighted average could actually be a good way to do it - but a time weighted average of the combined (max()) demand of all rooms. A time weighted average is just a low pass filter function after all.

  5. #165
    Site Sponsor The EVOHOME Shop's Avatar
    Join Date
    Dec 2014
    Location
    Powys, Mid Wales
    Posts
    483

    Default

    Quote Originally Posted by paulockenden View Post
    I think it was Richard who said he'd managed to measure it using an OpenTherm monitor.

    Unless I've remembered incorrectly.

    P.
    It was and nothing hearsay about it...

    The control bandwidth for OpenTherm is 1 degree either side of temperature setpoint of the zone.

  6. #166
    Automated Home Legend
    Join Date
    Jul 2014
    Posts
    2,307

    Default

    I am waiting for the Heat demand implementation on the firmware to compare what the OT Bridge is sent. At the moment I dont think Domoticz is tracking the correct value, largely driven by the fact that Dan doesn't have an OT Bridge. My biggest worry is that there doesn't appear to be a great deal of RF chatter between the controller and the OT Bridge. It almost appears that the OT bridge reads all the messages flying around and decides what to do without the controller telling it. But am not entirely sure too.

  7. #167
    Automated Home Ninja
    Join Date
    Aug 2016
    Posts
    489

    Default

    I very much doubt that's happening. Given the relative location of a couple of my HR92s and my OT Bridge, I don't think they'll be able to hear each other.

    When you say "there doesn't appear to be a great deal of RF chatter", how are you gauging that? Remember that the firmware we're running on the radios doesn't parse OT messages very well, so you won't see them at all in the Evohome logs unless you're looking for parsing errors. (Oh, unless you're using an HGI80? I can't remember.)

  8. #168
    Automated Home Legend
    Join Date
    Jul 2014
    Posts
    2,307

    Default

    I have an HGI80.

  9. #169
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,361

    Default

    Quote Originally Posted by bruce_miranda View Post
    I have an HGI80.
    If it works like a BDR91 then you'll only see a heat demand message sent to it every 20 minutes, or when the heat demand changes significantly. If your house is at a nice steady equilibrium there may be no need to send changes to the heat demand - in which case you'd see one message every 20 minutes repeating the same or very similar heat demand.

    Is this what you see ?

  10. #170
    Automated Home Legend paulockenden's Avatar
    Join Date
    Apr 2015
    Location
    South Coast
    Posts
    1,719

    Default

    I suspect that a BDR91 boiler relay and an OT Bridge are effectively working off the same signal from the controller.

    The only difference is how they deal with anything less than 100% heat demand. With the BDR91 it's by pulsing the boiler on and off, whilst with the OT bridge it's by passing the relative heat demand through to the boiler. The controller doesn't micro-manage either device.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •