Page 4 of 13 FirstFirst 123456789 ... LastLast
Results 31 to 40 of 126

Thread: Evohome Opentherm Boiler decision and set up

  1. #31
    Automated Home Guru
    Join Date
    Dec 2016
    Posts
    124

    Default

    Quote Originally Posted by DBMandrake View Post
    The "learning" process of the HR92 is going to be mainly self tuning of the differential term of the PID equation....... 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.
    Is there a differential term? I thought Evohome uses TPI (i.e. Proportional and Integral only). Or is it just the Evohome controller that applies TPI control and the HR92s do something different?

  2. #32
    Automated Home Ninja Dan_Robinson's Avatar
    Join Date
    Jun 2012
    Posts
    347

    Default

    Evohome is tpi, until you plug an OT bridge in.
    Kind Regards - Dan Robinson (Jennings Heating Ltd)

  3. #33
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,193

    Default

    Quote Originally Posted by Edinburgh2000 View Post
    Is there a differential term? I thought Evohome uses TPI (i.e. Proportional and Integral only). Or is it just the Evohome controller that applies TPI control and the HR92s do something different?
    Don't confuse two completely different things.

    The HR92 is a PID controller which is a feedback system where measured temperature and set point are the inputs and valve position is the output, which is then also modified by a transfer function to produce the heat demand percentage sent to the controller. The aim of it is to adjust the valve position and heat demand to reach the target set point as quickly and as accurately as possible without overshoot or oscillation. (Quite a challenging task, as it happens)

    TPI (time proportional integral) is something completely different. It's really just a fancy name for duty cycle modulation or pulse width modulation where there is a constant cycle time (default 10 minutes on the BDR91) and the on time is proportional to the desired heat demand - hence the "proportional" bit. Integral is referring to the total heat output of the boiler averaged over time over multiple cycles. (Since integration calculates the area under a graph)

    So, not related at all. The PID controller is what measures the temperature and compares it to the set point to vary the heat output from the radiator, TPI is just a method of modulating the power output of a boiler that doesn't normally have any means to do that. (Fixed user set flow temperature) If you are using OpenTherm you are not using TPI but the HR92 is still behaving as a PID controller.
    Last edited by DBMandrake; 29th January 2018 at 09:32 PM.

  4. #34
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,193

    Default

    Quote Originally Posted by paulockenden View Post
    Any savings with Evohome are more likely to come from not heating unused rooms.
    Or only heating rooms at times of the day that they are needed based on usage patterns, if you have predictable patterns.

    For example we don't heat our bedrooms during the day, and at night the bedrooms are on at a low temperature but the downstairs is all off. In the early evening most of the downstairs is on at 20'ish (except for a spare room which is 15) but in the later evening all the downstairs except the living room drops a couple of degrees. (I've compared gas usage from not doing this - the two degree drop from 8pm to 11pm in 4 rooms makes a big difference!)
    TPI is, to my mind, more about comfort than money.
    I agree, TPI as a means to save gas is overrated or perhaps even miss-sold, what it is is a superior way to modulate the output of a boiler for accurate control which was not originally designed to be modulated in a way that can avoid long cycle times and large room temperature oscillation.

    If you compare it to the normal thermostats that came before (first mechanical, then electronic) they typically had a 2 degree hysteresis. Set it to 20 and the boiler would run at full set flow temperature until the room reached 21 degrees then the boiler went completely off. With thermal inertia the room might gain another degree and end up at 22 degrees before it started to drop back down again.

    The boiler would remain off, potentially for an hour or two until it reached 19 degrees then turn on again. Rinse and repeat. You end up with a very long period of time where no other rooms can get heat because the boiler is off for so long each cycle and the room temperature is fluctuating 2-3 degrees every hour or two as the boiler cycles on and off. Not very comfortable.

    TPI means that in the proportional band the boiler is cycling on and off at a regular rate - for example every 10 minutes, and by adjusting the on period the average heat into the room can be very precisely controlled (theoretically infinite resolution) and thus the exact balance of heat into the room can be established and the room temperature held within a fraction of a degree.

    No more repetitive 2-3 degrees swings, no more hour or more period where the boiler isn't on where rooms with manual TRV's can't get any heat - as long as there is at least some heat demand the boiler will fire for at least some time in every 10 minute cycle which means rooms with manual TRV's get a chance to get some heat in every 10 minute period.

    OpenTherm takes it one step further of course and directly controls the flow temperature which leads to even greater precision in heat output control from the boiler, but perhaps more importantly is far quicker and more responsive. As TPI is based on average heat output over a (by default) 10 minute cycle when the heat demand changes it is on average 10 minutes before that change has any tangible effect on radiator flow temperatures, whereas a change from say 45 to 55 degrees via OpenTherm will be actioned by the boiler immediately and it may reach the target in under a minute. This faster response time can help give better overall control as the more delays you add to the system response the more prone it will be to overshoot and oscillate.
    Last edited by DBMandrake; 29th January 2018 at 09:52 PM.

  5. #35
    Automated Home Legend paulockenden's Avatar
    Join Date
    Apr 2015
    Location
    South Coast
    Posts
    1,693

    Default

    Quote Originally Posted by DBMandrake View Post
    Or only heating rooms at times of the day that they are needed based on usage patterns, if you have predictable patterns.
    That's what I meant. Don't heat rooms when they are unused!

  6. #36
    Automated Home Guru
    Join Date
    Dec 2016
    Posts
    124

    Default

    Quote Originally Posted by Edinburgh2000
    Is there a differential term?

    Quote Originally Posted by DBMandrake View Post
    Don't confuse two completely different things. The HR92 is a PID controller ....TPI (time proportional integral) is something completely different.
    Is that a "Yes"? (The control method of determining the % heat demand applied to zone valves (rather than HR92s) by the controller including a differential term.) Does the controller react to the rate of change of zone temperature?

    (And thanks for the best explanation of "TPI" that I have seen on this forum. That penny has now dropped with me.)
    Last edited by Edinburgh2000; 29th January 2018 at 11:23 PM.

  7. #37
    Automated Home Legend
    Join Date
    Jan 2015
    Location
    NE UK
    Posts
    1,176

    Default

    Quote Originally Posted by blowlamp View Post
    No contest in my opinion.

    Intergas are built like no other -
    Does Veissmann still have rubber hoses to and from the heat exchanger?
    The intergas would be simple to descale if ever necessary (are you in a hard water zone?) and don't forget that you'll be paying for the ten year warranty of the Veissmann in the purchase price.



    Martin.
    I am wary of any warranty. Many long term warranties, if you read the fine print, are only covering the bits that are unlikely to fail. Then you have the likes of one well installer offering wonderful warranties, and expensive installations, but you have to use them for servicing every year and they are not cheap. There is probably enough said in these pages to persuade you to go Intergas!

  8. #38
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,193

    Default

    Quote Originally Posted by Edinburgh2000 View Post
    Is that a "Yes"? (The control method of determining the % heat demand applied to zone valves (rather than HR92s) by the controller including a differential term.) Does the controller react to the rate of change of zone temperature?
    Yes. The D term of a PID controller reacts to the rate of change of temperature and applies a moderating effect. In essence if the temperature is rising quickly towards the set point and would otherwise cause the system to overshoot the target, the high rate of change will apply a "retarding" effect to the valve position/heat demand to reduce it sooner than would be the case if the temperature was rising slowly.

  9. #39
    Automated Home Legend
    Join Date
    Sep 2014
    Location
    Scotland
    Posts
    2,193

    Default

    Quote Originally Posted by paulockenden View Post
    Despite what we've always been told, I think it's +/- 1.0 degree.
    I did a little careful testing tonight and I can say with pretty high confidence that it is more than 1 degree. I checked 4 different HR92 zones, including ones known to "struggle" to reach the set point, as well as ones that tend to overshoot, as the two are likely to have very different learnt characteristics for the integral term. I tested in the heating side of the differential zone - eg set point set near equal to the current measured temperature and increased in 0.5 degree steps above the measured temperature.

    I then subtracted the (raw) measured temperature from the set point and compared it to reported heat demand percentage. The results were:

    Code:
    Temperature difference - Heat Demand
    
    Living Room
    
    1.32C - 67%
    1.82C - 100%
    
    Bathroom
    
    0.51C - 7%
    0.78C - 21%
    1.28C - 100%
    
    Kitchen
    
    0.71C - 7%
    1.21C - 35%
    1.71C - 100%
    
    Study
    
    0.78C - 14%
    1.24C - 65%
    1.54C - 100%
    The contrast between the kitchen and bathroom is quite interesting and will be a result of the very different thermal characteristics of the two rooms and the different adaptions made by the two HR92's.

    In summary, the proportional band (at least in the positive, set point higher than measured direction) must be at least 1.32C and is probably less than 1.54C. (More samples needed though to confirm the latter) I suspect it will be the same in the negative direction but to test that I'd need to redo the test taking the valve position reading directly from each HR92 as the valve position readings below 30% are truncated and appear as a 0% heat demand all the way from 30 down to 0.

    A bit more sampling (with some luck with the precise temperature readings) would narrow it down a bit more but its definitely more than 1C, and 1.5C is a nice round number, just saying.
    Last edited by DBMandrake; 30th January 2018 at 12:22 AM.

  10. #40
    Automated Home Legend
    Join Date
    May 2014
    Posts
    1,000

    Default

    @DBMandrake

    Simon - your detailed knowledge and willingness to pass on the information in a clear and reading style is much appreciated. Noting the above, what conclusions have you come to read the boiler demand setting?

    IMG_0944.jpg

    What does the boiler control of 37% actually mean in practical terms.(NB: the percentage is still higher than the highest zone demand; that said, the particular zone comprises three small single panel radiators)

Posting Permissions

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