Page 3 of 13 FirstFirst 12345678 ... LastLast
Results 21 to 30 of 125

Thread: Evohome Opentherm Boiler decision and set up

  1. #21
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    1,788

    Default

    Now that we have the heat demand display in the new firmware, and I can also measure exact temperatures (not rounded and biased towards the set point as displayed by the controller) using simple Python scripts, (I can call the script manually to get up to date readings that are valid as recently as 10 seconds ago, not relying on 5 minute data points from Domoticz which will be too out of date) I'm going to do a little testing to find out for certain exactly what the proportional band of the Evohome (HR92 actually) really is...

    Anyone want to place any bets ? +/- 1.0 degrees ? +/- 1.5 degrees ?

  2. #22
    Automated Home Guru
    Join Date
    Nov 2015
    Posts
    143

    Default

    Quote Originally Posted by HenGus View Post

    My wife and I were looking around a new home over the weekend which has a Valliant 18 system boiler. What struck me was the minute size of all the radiators in the property. They were all single panel. I came away thinking what, if any, thought has been given to boiler condensing efficiency? Not a lot, I suspect.
    I would say None whatsoever

    I agree that most of your savings were down to the New Boiler Like a lot of things we get into Diminishing returns here the majority of savings come early as we upgrade poor sysyems to modern efficent ones But after that Depending on Fuel Costs etc it can be hard to justify the spend on extra controls etc ok we might like them but do they really save money I ran my House on a BDR for 2 years with a WB Boiler now have opentherm and the savings are minimal as the BDR Honeywell Evo and a well insulated home is very efficent I might be saving a little but hardly enough to justify the Opentherm Bridge conversion especially if I had to pay someone to do it ROI would be pretty long I suspect

  3. #23
    Automated Home Legend
    Join Date
    May 2014
    Posts
    988

    Default

    Quote Originally Posted by StephenC View Post
    To be fair, the radiator point I was making is part of my own longer term strategy towards a heat pump based heat source due to not being satisfied with gas/fossil fuel usage, and will be taking place when other large works will occur.
    I wasn’t suggesting it was an answer to anyone, merely my thoughts are now thinking beyond gas and OpenTherm as the “destination” solution as that is clearly out of my control at this point with the current kit/investment.

    The selection of EvoHome and afterwards after looking at it OpenTherm, are part of reducing peak load shifting, and a reduction gas/fossil fuel usage (that’s important to me) as well as the usual cost and comfort reasons to be clear. Maybe I hadn’t stated that as a driver.
    I accept all of that. It rather depends on how long you expect to live in a particular house etc. If I was starting from scratch, then I would look probably look at alternative heating solutions.

  4. #24
    Automated Home Legend paulockenden's Avatar
    Join Date
    Apr 2015
    Location
    South Coast
    Posts
    1,561

    Default

    Quote Originally Posted by DBMandrake View Post
    to find out for certain exactly what the proportional band of the Evohome (HR92 actually) really is...

    Anyone want to place any bets ?
    Despite what we've always been told, I think it's +/- 1.0 degree.

    P.

  5. #25
    Automated Home Legend
    Join Date
    May 2014
    Posts
    988

    Default

    Quote Originally Posted by paulockenden View Post
    Despite what we've always been told, I think it's +/- 1.0 degree.

    P.
    If that is true (and I have no reason to doubt you), then H'Well really needs to review its controller BDR/OT command logic. Intuitively, the range seems to be too small for a multi-zoned home. Conversely, it would probably work well for a single TPI thermostat.

  6. #26
    Automated Home Guru
    Join Date
    Dec 2016
    Posts
    121

    Default

    Quote Originally Posted by DBMandrake View Post
    Now that we have the heat demand display in the new firmware.......Anyone want to place any bets ? +/- 1.0 degrees ? +/- 1.5 degrees ?
    I am curious about the Evohome proportional band and whether that is conditioned by the 'learning' that Evohome does. I have no HR92s but my house is zoned. One zone has under sized radiators and very high heat capacity (i.e. it takes a lot to shift the temperature). With the new demand display, I see that that zone is never anything other than 0% or 100% and that switches when the zone temperature is (apparently) at the set point, never above or below. Of course, I can only see that to the nearest 0.5oC so I guess it is actually switching at finer granularity than that. But to me it looks like a proportionality band of zero.

    My other zones show a proportionality band of +/- 1.5oC. I see heat demands between 0% and 100% for temperatures in that range. But what is the variable that Evohome 'learns' for each zone? Does it vary the proportionality band based on the learned rate of change of temperature? And if so, isn't that differential control (rate of change) rather than TPI?

  7. #27
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    1,788

    Default

    Quote Originally Posted by Edinburgh2000 View Post
    I am curious about the Evohome proportional band and whether that is conditioned by the 'learning' that Evohome does. I have no HR92s but my house is zoned. One zone has under sized radiators and very high heat capacity (i.e. it takes a lot to shift the temperature). With the new demand display, I see that that zone is never anything other than 0% or 100% and that switches when the zone temperature is (apparently) at the set point, never above or below. Of course, I can only see that to the nearest 0.5oC so I guess it is actually switching at finer granularity than that. But to me it looks like a proportionality band of zero.
    If it's not an HR92 what kind of zone is it ? Is it a "zone valve zone" that is controlling a zone valve ? Do you have manual TRV's on those radiators as well ?

    When you say it drops suddenly from 100% to 0% do you mean as the room warms up, or is that only happening when the room is scheduled to drop in temperature ? If the radiators are undersized and struggle to reach the target (or only just reach it) it could well be that they are calling for a 100% heat demand all the time, which would only go away when the room gets scheduled to a lower temperature, which would result in 0% demand if it was more than a couple of degrees below the current room temperature.
    My other zones show a proportionality band of +/- 1.5oC. I see heat demands between 0% and 100% for temperatures in that range. But what is the variable that Evohome 'learns' for each zone? Does it vary the proportionality band based on the learned rate of change of temperature? And if so, isn't that differential control (rate of change) rather than TPI?
    Before everyone starts trying to independently figure out the width of the proportional band, it's important to make clear that the measured temperatures displayed by the Evohome (and HR92 and wall stats) is biased up to 0.5 degrees towards the set point and then rounded to the nearest 0.5 degrees. This makes it impossible to measure what the proportional band is just by looking at the figures reported on the controller.

    For example, if the real temperature is 19.2 degrees and the set point is set to 10, it will bias it down -0.5 to 18.7 and then round it back up to 19 for display. However if you change the set point to 21, it will bias it up +0.5 to 19.7 and then round it up to 20. So the displayed temperature can be significantly different from the real temperature and always biased towards the setpoint. This is a questionable choice IMHO and is probably done to make the temperature look more stable than it really is when it hovers either side of the set point.

    So you need to get the raw unbiased, unrounded temperatures to measure the proportional band. Which you can get either with Domoticz and an HGI80 or via the V1 WebAPI - which kindly provides two decimal places with no rounding or biasing. This is the one I do my graphing from.

    As for whether the proportional band changes - generally no, the width of a proportional band in a PID controller system is fixed, but the proportional, integral and differential terms can be tuned to change the response within in that band.

    The main one of interest is the integral term - this compensates for offset error over time and is the one that allows the controller to converge on the correct heat demand to meet the target set point no matter what that heat demand is. (Unlike a mechanical TRV, which will always have an offset error)

    So pretend the proportional band is +/- 1 degree and the set point is 20 degrees the way it would work is:

    room temp <19 degrees - heat demand is always 100%
    room temp >21 degrees - heat demand is always 0%

    room temp 19-21 degrees - heat demand can theoretically be anywhere between 0% and 100% anywhere within this temperature range, although in practice the learning process will usually mean that heat demand slopes up towards 100% at the bottom end of the range and down towards 0% at the top end of the range.

    Under conditions where the room is slowly warming up in response to other heat sources (say a TV, lamps and people) and no longer needs the help of the radiator, you may find that the heat demand at the set point keeps dropping until it reaches 0% even though you're within the proportional band at or just above the set point - in this scenario the whole top half of the proportional band will have a head demand of 0% because even with no heat from the radiator the room temperature is being maintained or is gradually being exceeded. (Another example would be a room that was heated in the morning but can maintain or exceed its set point with sunshine through the window later in the day. It's still at the right set point (or a little above) but the radiator has gone off completely.

    Conversely you might have a room that is struggling to reach it's target and is persistently staying in the bottom half of the proportional band - the heat demand will keep creeping up thanks to the integral term and may potentially hit 100% while still within the proportional band just below the set point if it just can't meet it's target.

    The "learning" process of the HR92 is going to be mainly self tuning of the differential term of the PID equation - this is the one that helps prevent overshoot by making predictive changes based on what is going to happen rather than what has already happened. For example it allows the heat demand to be throttled back in advance of reaching the target. The integral adaption also probably looks like "learning" but it's just the normal process of integrating a long term error and applying a correction for it.

    While everyone says the proportional band should be wider (wider than what though ?) there is a negative for having too wide a proportional band, and that is integral windup where if you are a long way from the set point for a long time, then you keep adding and adding more integral offset trying to get there faster than you can, so by the time you get there you have a large offset that then causes an overshoot so that the temperature goes right past the target and keeps going for quite some time - this overshoot is like the overshoot you get from thermal mass but it isn't physical in nature its a quirk of the control algorithm. Essentially your control algorithm has a lot of undesirable inertia.

    The usual solution to this is to avoid modifying the integral while you're outside the proportional band. So the integral adaption is basically "put on hold" until the temperature gets within the proportional band, and then it is allowed to start acting and adapting. So if you make the proportional band really wide you'll get problems with integral windup, and you may also slow the heating process a lot as it will start to taper back the heat demand far too early.

    So the choice of proportional band width in degrees is a trade off between these and other factors.
    Last edited by DBMandrake; 29th January 2018 at 01:18 PM.

  8. #28
    Automated Home Guru
    Join Date
    Dec 2016
    Posts
    121

    Default

    Quote Originally Posted by DBMandrake View Post
    If it's not an HR92 what kind of zone is it ? Is it a "zone valve zone" that is controlling a zone valve ? Do you have manual TRV's on those radiators as well ?
    It's a whole floor controlled from a wall thermostat in the hall. It is a 'zone valve zone'. The TPI control acts on the BDR91 controlling the zone valve, which I have set to 3 cycles per hour (20 minute periods). I have manual TRVs on all radiators.

    Quote Originally Posted by DBMandrake View Post
    When you say it drops suddenly from 100% to 0% do you mean as the room warms up, or is that only happening when the room is scheduled to drop in temperature ?
    I mean as the room warms up. I think it is performing exactly as intended and I have no complaints about it. The heat demand in the zone is only ever 100% or 0%. If the zone does reach its setpoint (which is only when the outside temperature is not too cold), the heat demand goes from 100% to 0% with nothing in between. And the controller shows the zone temperature to be at the setpoint before and after that change. (Of course, when the schedule reduces the setpoint below the actual temperature, the heat demand goes to zero.)

    But I still think I have a zone with a proportional band of zero and I am curious what variable Evohome 'learned' to determine that.

  9. #29
    Automated Home Sr Member
    Join Date
    Oct 2017
    Location
    France & UK
    Posts
    74

    Default

    Quote Originally Posted by mylesm View Post
    This is First Winter with Intergas Boiler I had a worcester was about 10 years old so changed it for Intergas As you are probably Aware Intergas Boilers can be setup as Combi,System,Open Vent,simply by changing a parameter in the Menu Mine is currently open vent Heating Only no Stored Hot Water All Rads have HR92 Valves except Bathroom which acts as bypass rad I intend to connect the Combi Side for Hot Water in Kitchen later this year so far Opentherm is working well I see flow Temps between 30 - 65c I have 65c set as Max Ch Temp

    TO be honest in any well insulated house the savings using opentherm over BDR TPI is minimal
    There is even evidence to suggest that even TPI versus an old standard thermostat does not provide any energy savings - see the report in the link below. There is a lot of hype about Opentherm but I have yet to see any real studies to show the actual savings when applied to the real world, whether this be a near passive, average or poorly insulated property. Possibly there are arguments in favour of comfort. Certainly our house is much more comfortable with Evohome using a BDR91 switch compared to our old set up and of course the ease of control via a central controller , apps etc is a bonus, but I doubt that we would notice any change on switching from this to Opentherm.

    https://www.gov.uk/government/upload...ng-boilers.pdf
    Last edited by rvb99; 29th January 2018 at 05:11 PM.

  10. #30
    Automated Home Legend paulockenden's Avatar
    Join Date
    Apr 2015
    Location
    South Coast
    Posts
    1,561

    Default

    Any savings with Evohome are more likely to come from not heating unused rooms.

    TPI is, to my mind, more about comfort than money.

    P.

Posting Permissions

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