Evohome - poor temperature regulation

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • greyhound1234
    Automated Home Jr Member
    • Mar 2015
    • 33

    Evohome - poor temperature regulation

    I have purchased a Honeywell Evohome and am running with radiator zoning valves. I am having issues with the way this system is regulating the temperature.
    My experience very much echoes that in this thread

    however I thought I wouldn't hijack it as it is long enough

    I have come from a Honeywell CM907 digital stat, which was very good at maintaining the temperature, to within +/- 0.2 degree. It used to do pulsing of the boiler when just under the set point, and stop when it reached it. I was hoping get that sort of control in all my rooms using the zone valves (where suitably thermally disconnected from each other).

    I have found that the ability of the Evohome to hold a set temperature is very poor. It consistently over-shoots by 1.5 degrees before the heat demand stops.

    System:
    Honeywell Evohome with BDR-91 relay and HR92 zone valves. Combi boiler.
    Firmware: v25, dated 16 Jan 2014

    I will describe the results of some observations.
    ----
    I used Evohome internal sensor and in single zone/thermostat mode. I set the temperature to 18.0. The boiler was fired continuously until 17.5 was reached. Then it went into the proportional band, and commenced 1 minute pulses every 10 minutes. This continued until the room temperature - as measured by Evohome itself - was 19.5 degrees.

    I then installed the zone valves, in the hope this behaviour would go away.

    I did a long constant temperature test on the lounge zone. From a current room temperature of 16.0, I set it to a constant 18.0 overnight. Initially, it kept pulsing and got to 19.5 degrees before closing the valve and stopping calling for heat. Overnight it gradually fell, and by 3 in the morning it was approaching 18.0. On arrival at 18.0 it did the same thing and got back to 19.5 deg C.

    Conclusion: this 1.5 degree overshoot cannot be resolved by adjusting my temperatures down by 1.5 degrees, since it does eventually approach the right temperature for a while, then when triggered overshoots again. This is a serious bug in the way the control system has been implemented.

    Also this overshoot behaviour is not related to a high radiator temperature, as I repeated this with cool radiators (boiler down low) and a slower slow temperature rise still resulted in a deliberate overshoot to 1.5 degrees more than requested.

    ---

    This next observation was independent heating of other zones, which also doesn't work unless the other zones are set down cold.
    The lounge was set constant at 18.5 degrees, and lounge room temp was indeed stable at 18.5 degrees (presumably having peaked earlier in the day and come down).
    Bedroom 1 and bedroom 2 were set very low, and were both at 16.0 degrees.
    Bedroom 1 and 2 zones were then set to 18.5 degrees to trigger the heating.

    The heating came on, and the bedrooms started heating up. But the lounge radiator was also allowed to run, even though no heat was needed.
    The bedrooms only started to reduce the calls for heat at 19.0, and (no surprise) stopped calling for heat at 19.5 (both of them).
    The lounge also kept heating up, and the lounge rad valve gradually started to close at lounge temp of 19.5, and closed down fully at 20.0 (set point of lounge + 1.5 degrees).

    Conclusion: Any zone calling for heat, will keep going until it gets to 1.5 degrees over the set point. Regardless of which zone is calling for heat, all zones will run until they are 1.5 degrees more than the set point, until their zone valves are instructed to close.
    -----------
    This is offering very poor control - since all zones will go up by 1.5 degrees even if only one calls for heat. Indeed it defeats the object of zoning, might as well just use normal TRVs!!
    and yes, I have also tried removing the battery and doing a cold boot and rebinding all my valves 'slowly'. No change.

    -----------
    Some interesting quotes from other thread...
    Quotes:
    "When using Evohome in thermostat-modus (only Evohome+BDR91, no radiator valves), it will continue slowly heating up the room untill the temperature reaches 1.5 degree above the requested temperature. After that, it will finally stop requesting heat. This process of slowly heating up above requested temperature can take 2 hours easily. It's not a short accidental overshoot. It's a continuous, well orchestrated slow build up of overshoot on overshoot on overshoot. It's a huge bug in my opinion.

    Some people have suggested to put my requested temperature a degree lower in Evohome, to compensate for this Evohome behaviour. This is not a solution. Yes, it will prevent Evohome from heating up the room too much, but it won't prevent Evohome from letting it cool down too far afterwards."

    "I don't think TPI is the problem. The old Honeywell Round does TPI as well, I think, but works fine. The problem is a bug in the implementation in Evohome is my guess. It never skips a TPI cycle. But it should. But it doesn't. If there's a bug that big in the system, they will HAVE to make new firmware available. The product is too expensivee not to. This would mean they would need to give everyone their money back for a non-functioning system."
    ------------

    I am attempting to get an answer from Honeywell engineers via their support email. What I hope would happen, is that they admit this is not intended behaviour; that they find the bug; that they fix their firmware; and that they release the new firmware with PC bootloading software and instructions for users to update it. That would cost them, and what I actually think will happen is probably the opposite of all that, and I will return the product to the supplier as not fit for purpose.

    1 degree is pretty much spans too cold - ok - and too hot. That was about the accuracy of the old analogue clicky thermostats. So if Honeywell consider that 1.5 degree overshoot as a design feature, it is definitely not fit for purpose.
  • top brake
    Automated Home Legend
    • Feb 2015
    • 837

    #2
    Originally posted by greyhound1234 View Post
    I have purchased a Honeywell Evohome and am running with radiator zoning valves. I am having issues with the way this system is regulating the temperature.
    My experience very much echoes that in this thread

    however I thought I wouldn't hijack it as it is long enough

    I have come from a Honeywell CM907 digital stat, which was very good at maintaining the temperature, to within +/- 0.2 degree. It used to do pulsing of the boiler when just under the set point, and stop when it reached it. I was hoping get that sort of control in all my rooms using the zone valves (where suitably thermally disconnected from each other).

    I have found that the ability of the Evohome to hold a set temperature is very poor. It consistently over-shoots by 1.5 degrees before the heat demand stops.

    System:
    Honeywell Evohome with BDR-91 relay and HR92 zone valves. Combi boiler.
    Firmware: v25, dated 16 Jan 2014

    I will describe the results of some observations.
    ----
    I used Evohome internal sensor and in single zone/thermostat mode. I set the temperature to 18.0. The boiler was fired continuously until 17.5 was reached. Then it went into the proportional band, and commenced 1 minute pulses every 10 minutes. This continued until the room temperature - as measured by Evohome itself - was 19.5 degrees.

    I then installed the zone valves, in the hope this behaviour would go away.

    I did a long constant temperature test on the lounge zone. From a current room temperature of 16.0, I set it to a constant 18.0 overnight. Initially, it kept pulsing and got to 19.5 degrees before closing the valve and stopping calling for heat. Overnight it gradually fell, and by 3 in the morning it was approaching 18.0. On arrival at 18.0 it did the same thing and got back to 19.5 deg C.

    Conclusion: this 1.5 degree overshoot cannot be resolved by adjusting my temperatures down by 1.5 degrees, since it does eventually approach the right temperature for a while, then when triggered overshoots again. This is a serious bug in the way the control system has been implemented.

    Also this overshoot behaviour is not related to a high radiator temperature, as I repeated this with cool radiators (boiler down low) and a slower slow temperature rise still resulted in a deliberate overshoot to 1.5 degrees more than requested.

    ---

    This next observation was independent heating of other zones, which also doesn't work unless the other zones are set down cold.
    The lounge was set constant at 18.5 degrees, and lounge room temp was indeed stable at 18.5 degrees (presumably having peaked earlier in the day and come down).
    Bedroom 1 and bedroom 2 were set very low, and were both at 16.0 degrees.
    Bedroom 1 and 2 zones were then set to 18.5 degrees to trigger the heating.

    The heating came on, and the bedrooms started heating up. But the lounge radiator was also allowed to run, even though no heat was needed.
    The bedrooms only started to reduce the calls for heat at 19.0, and (no surprise) stopped calling for heat at 19.5 (both of them).
    The lounge also kept heating up, and the lounge rad valve gradually started to close at lounge temp of 19.5, and closed down fully at 20.0 (set point of lounge + 1.5 degrees).

    Conclusion: Any zone calling for heat, will keep going until it gets to 1.5 degrees over the set point. Regardless of which zone is calling for heat, all zones will run until they are 1.5 degrees more than the set point, until their zone valves are instructed to close.
    -----------
    This is offering very poor control - since all zones will go up by 1.5 degrees even if only one calls for heat. Indeed it defeats the object of zoning, might as well just use normal TRVs!!
    and yes, I have also tried removing the battery and doing a cold boot and rebinding all my valves 'slowly'. No change.

    -----------
    Some interesting quotes from other thread...
    Quotes:
    "When using Evohome in thermostat-modus (only Evohome+BDR91, no radiator valves), it will continue slowly heating up the room untill the temperature reaches 1.5 degree above the requested temperature. After that, it will finally stop requesting heat. This process of slowly heating up above requested temperature can take 2 hours easily. It's not a short accidental overshoot. It's a continuous, well orchestrated slow build up of overshoot on overshoot on overshoot. It's a huge bug in my opinion.

    Some people have suggested to put my requested temperature a degree lower in Evohome, to compensate for this Evohome behaviour. This is not a solution. Yes, it will prevent Evohome from heating up the room too much, but it won't prevent Evohome from letting it cool down too far afterwards."

    "I don't think TPI is the problem. The old Honeywell Round does TPI as well, I think, but works fine. The problem is a bug in the implementation in Evohome is my guess. It never skips a TPI cycle. But it should. But it doesn't. If there's a bug that big in the system, they will HAVE to make new firmware available. The product is too expensivee not to. This would mean they would need to give everyone their money back for a non-functioning system."
    ------------

    I am attempting to get an answer from Honeywell engineers via their support email. What I hope would happen, is that they admit this is not intended behaviour; that they find the bug; that they fix their firmware; and that they release the new firmware with PC bootloading software and instructions for users to update it. That would cost them, and what I actually think will happen is probably the opposite of all that, and I will return the product to the supplier as not fit for purpose.

    1 degree is pretty much spans too cold - ok - and too hot. That was about the accuracy of the old analogue clicky thermostats. So if Honeywell consider that 1.5 degree overshoot as a design feature, it is definitely not fit for purpose.

    Hi and welcome

    If you could share some photographs of your installation that would help the community respond

    Particularly the location of the relay box and the evohome controller

    Did you self install or get an installer in?

    thanks
    I work for Resideo, posts are personal and my own views.

    Comment

    • Rameses
      Industry Expert
      • Nov 2014
      • 446

      #3
      What you are experiencing is the evohome system trying to modulate the room temperature based on what it is learning.

      Depending on the boiler (age, make, type, etc), hotwater and how you determine the temp - based on what information you have given we feel the system operating within normal parameters. Closing the circuit and changing the heat/flow etc all will impact the 'system' and not just the room being 'watched'

      An overshoot is typical, especially in the learning phase or when the room has experienced changes which affect the original learning - but these a quickly adapted during the following cycles.

      The evohome system is trying to achieve the best 'average' temperature nearest to your setpoint, for the amount of time entered into the schedule. The best metaphor is driving a formula one car with only the pedal full pressed or fully off. The heat is either falling or rising, never static. A room overshoot of 2 degrees for 10 mins is probably acceptable if the room temp was to be maintained for 1+ hours, considering the curve of decline. (some systems do this by undershooting, sacrificing comfort for the learning)

      Our advice is put the system to your desired schedule (not just one room etc), leave it to 'settle' , with all the settings on default. If you still feel the system is 'wrong' then please come back to us.

      Just as a side mention - we are creating a 'white paper' which will give our users more insight into the behaviour and answer common questions. We will communicate these out shortly.
      getconnected.honeywell.com | I work for Honeywell. Any posts I make are purely to help if I can. Any personal views expressed are my own

      Comment

      • orange
        Automated Home Guru
        • Dec 2014
        • 149

        #4
        Originally posted by greyhound1234 View Post
        commenced 1 minute pulses every 10 minutes
        but why would evohome ever switch the boiler on for 1 minute every 10 minutes ? How can this ever be correct behaviour ? To continue the 'driving analogy' which has been used before by customer support (still never got a reply on that - several weeks later) - I would consider that terrible driving.

        greyhound1234 - what is your "minimum on time" - in your settings ?

        Comment

        • G4RHL
          Automated Home Legend
          • Jan 2015
          • 1580

          #5
          Whilst I remain very happy with the system I do note I also initially get an overshoot of 1.5 degrees but I do not see that as a big issue, the system cuts out and in time a balance of the set temperature is maintained, i.e. initially it overshoots by 1.5c but then settles back. Never a constant overshoot.

          What I am not 100% convinced about is the learning process. Yes, optimisation does work to an extent but for example after 7 weeks I would not expect optimisation to assume it takes 60 minutes to get a zone up to temp when in reality it is taking 40 minutes, if not less. It does not seem to matter what the start temperature is, it comes on 60 minutes before.

          Comment

          • erik
            Automated Home Guru
            • Feb 2015
            • 244

            #6
            Trust me, Greyhound. Evohome is not going to learn. Not after 1 week, not after 2, not after 3. The implementation is flawed. It will always go ON at least 1 minute per 10 minutes. It never skips a cycle. Logic dictates that, if your room heats up from 1 minute of heat per 10 minutes, it will ALWAYS build up a huge overshoot. Untill it finally decides to stop asking heat (1.5 above setpoint) and then it will stay off for a few hours and, very uncomfortably, cool down to setpoint. Only to carefuly build up the same huge overshoot afterwards again.

            And Honeywell seems to have a hard time believing it, even though there's clear proof, from multiple users. They'll try to convince you on how complicated everything is and that the behaviour is 'normal', that it's still learning, or that you're doing something wrong. It's plain hoggwash! They're really convinced their system is perfect, it seems.

            However, there's NO, and I repeat NO reason for a thermostat to ask for even more heat when the room temperature is already 1 degree above setpoint and steadily rising. Especially when it takes 2 hours or more without heating to cool down 1 degree. It's not smart. It's dumb, very dumb.

            Or to use the car analogy: let's say you want to do 100km/h. You're doing 90km/h. So you push the throttle for 1 second and then release it for 1 second. It takes you to 92km/h. You push the throttle 1 second again, 94. Again, 96. Again, 98. Again, 100. Again, 102. Again, 104. Again, 106. Again, 108. Again, 110. Now... we're already 10 above the wanted speed. We have learnt from past that if we don't push the throttle for 10 seconds, the speed decreases by about 2km/h. What do you think? Should we push the throttle again? Anyone with common sense would say NO. Let's wait 10, 20 or 30 seconds and MAYBE then push the throttle again. But what does Evohome do? It just pushes the throttle AGAIN. Probably untill you're doing 120 or so. And then FINALLY it will stop pushing the throttle and let the car slow down to 100. Only to start slowly throttling up to 120 again afterwards.

            Greyhound, your old Honeywell thermostat worked fine (0.2 precision). So did mine (0.3 precision). Evohome works badly (1.5 'precision'). It's really not that complicated.

            No whitepaper is going to change this. What we need is recognition of the fact that there's a problem. And then a fix. Honeywell is currently testing the Evohome behavior during low load conditions, I will wait and see what the outcome will be. If the testing confirms the problem and if they create a fix, I will be very thankful and I would happily advise the system to others and support them during their installation etc. If Honeywell continues to deny and ignore the problem... I'll make sure to warn everyone who's thinking about getting the system... Making mistakes is human. It happens. What's important is how you deal with those mistakes and how you fix them...
            Last edited by erik; 12 March 2015, 09:17 PM.

            Comment

            • orange
              Automated Home Guru
              • Dec 2014
              • 149

              #7
              Originally posted by erik View Post
              Making mistakes is human. It happens. What's important is how you deal with those mistakes and how you fix them...

              agree 100% - let's see if Honeywell do fix these issues - although it's already been over a year so I'm not at all hopeful.

              Comment

              • greyhound1234
                Automated Home Jr Member
                • Mar 2015
                • 33

                #8
                Thank you for all your responses. I hope my previous post didn't come across to harshly, but to be honest I was getting frustrated with having the system get whatever temperature it felt like, having been used to decent thermal comfort from my trusty CM907.
                If anyone has any suggestions I'm happy to try them.

                Re: Installation. Yes it was a self-install.

                Re: position of things. Small modern house, thin walls. BDR-91 boiler relay is on the lounge wall, boiler is in the adjacent kitchen (i.e. it is not too close). Evohome controller in the lounge about 3 m from the BDR-91. All valves 'know' what zone they belong to.

                Re: learning. This is after a month of use. It did this the same from day 1 and it does it now on day 30.

                "No whitepaper is going to change this. What we need is recognition of the fact that there's a problem. And then a fix. Honeywell is currently testing the Evohome behavior during low load conditions, I will wait and see what the outcome will be. If the testing confirms the problem and if they create a fix, I will be very thankful and I would happily advise the system to others and support them during their installation etc"

                I quite agree. My testing has clearly illustrated this is not desirable behaviour. Certainly not sensible programming. It is not a low-load problem; it is just flawed logic. I have turned my rad temp right down and if it was a low-load vs high boiler temp problem it would improve, but the behaviour is the same.

                Logically the system should work something like this:
                - Radiator valves should be fully open at Tset - 1.0 degrees, and fully closed at Tset (or maybe Tset + 0.5, but certainly not T+1.5). And varying in-between.
                - Proportional band should be Tset -1.0 to Tset exactly for each zone. (perhaps open this up as a customer parameter)
                - Heat demand should be 'full' if T < proportional lower temp on any zone.
                - Heat demand should be off if T > Tset on all zones.
                - If neither of those conditions are met, at least one zone is in the proportional band and then do some magic and work out an optimum pulsing boiler schedule.
                Simples!

                "Whilst I remain very happy with the system I do note I also initially get an overshoot of 1.5 degrees but I do not see that as a big issue, the system cuts out and in time a balance of the set temperature is maintained, i.e. initially it overshoots by 1.5c but then settles back. Never a constant overshoot."

                I agree after a few hours it settles back - once every zone is at the temperature and there is no heat demand. But being too hot for a few hours is not ideal, and for me this is not acceptable having experienced much better control on the single zone thermostat, and now just having spent 5x as much on this system.

                I think if Honeywell is not willing to admit they have a bug and roll out a fix, they should at least either A. open up more advanced parameters so the user can set both ends of the proportional band with respect to the setpoint, or B: open up their source code and boot loaders so the community can sort of the problems for them.

                I really hope Honeywell can find this issue and sort it out quickly. But based on their responses so far, they appear to be in denial so I am not holding my breath.

                Comment

                • Rameses
                  Industry Expert
                  • Nov 2014
                  • 446

                  #9
                  Where are you taking your temp readings from?

                  And you mentioned in a previous post you are using the internal sensor on the controller, is this correct still?
                  getconnected.honeywell.com | I work for Honeywell. Any posts I make are purely to help if I can. Any personal views expressed are my own

                  Comment

                  • G4RHL
                    Automated Home Legend
                    • Jan 2015
                    • 1580

                    #10
                    Originally posted by greyhound1234 View Post

                    I agree after a few hours it settles back - once every zone is at the temperature and there is no heat demand. But being too hot for a few hours is not ideal, and for me this is not acceptable having experienced much better control on the single zone thermostat, and now just having spent 5x as much on this system.
                    But I am never too hot for a few hours. The boiler cuts out shortly before the set point is reached, the pump continues on overrun and as a result the temperature rises further to usually no more than 1.5C over the set point if that and thereafter the set temperature is more or less maintained within 0.5 degrees. To me that is what the system should be doing. Currently I am in my study, it is set to 21C, the TRV reads 21.5C, the boiler is off, it is not repeatedly recycling. The temperature will drop and in due course the boiler will come back on to maintain the set temperature. That is what I expected the system to do and it does. Where I find it does not do what we are told it does is properly optimise. It does not learn as much as we are lead to believe. That is not an issue for me.

                    Comment

                    • erik
                      Automated Home Guru
                      • Feb 2015
                      • 244

                      #11
                      Greyhound, the reason they're testing it under low load, is that if there's high load (for example -5 outside and +20 inside and poor isolation), that 1 minute of heat per 10 minutes will probably not raise temperature in the room. In that case, you won't notice the problem. It will seem like it's keeping temperature perfectly.

                      There's logic behind requesting heat even when 0.1 above setpoint. It prevents drops below the setpoint. But there's no logic anymore when it keeps getting warmer and warmer in the course of a few hours and it then still requesting heat 1.0 above setpoint.

                      Rameses: the problem occuris with ANY temperature sensor. HR80, HR92, Evohome and Round, located at GOOD spots. Also on the same spot as the old thermostat which does fine. Also, my own thermometers tell me the same thing (except that Evohome of course is always lie'ing a bit about the measured temperature, making it seem closer to setpoint than it is).

                      G4RHL, Evohome overshoots even further if you use it in thermostat-modus (no radiator valves involved), so it's more noticable in that case. Also, Evohome lies a bit about measured temperature. If it says 21.5 during setpoint 21, it might very well actually be 21.9. Have you confirmed with your own thermometers that it 'only' goes up and down 0.5? I expect it to be around 0.7-0.8 when using radiator valves.
                      Last edited by erik; 13 March 2015, 07:00 AM.

                      Comment

                      • greyhound1234
                        Automated Home Jr Member
                        • Mar 2015
                        • 33

                        #12
                        Originally posted by Rameses View Post
                        Where are you taking your temp readings from?

                        And you mentioned in a previous post you are using the internal sensor on the controller, is this correct still?
                        When I put the rad zone valves on, I set all zones to use remote device temperature sensor, not evohome controller sensor.

                        Comment

                        • greyhound1234
                          Automated Home Jr Member
                          • Mar 2015
                          • 33

                          #13
                          Ignoring arguments over how the TPI algorithms should behave, the behaviour of other zones is clearly not correct. Can anyone else replicate this?
                          e.g. Zone A and B both previously set to 17.0 and both holding 17.0. Zone A is then set to e.g. 20. Zone A will heat up. Zone B will heat up. Zone B will only stop heating up at 18.5 (1.5 degrees above it's setpoint, the zone valve closes off).

                          Comment

                          • greyhound1234
                            Automated Home Jr Member
                            • Mar 2015
                            • 33

                            #14
                            Perhaps we should have a poll of what firmware we have, and whether we are happy with it.
                            v25 16Jan2014, and not happy...
                            Rameses - humour me, please tell me what firmware version the Evohome controller units are currently leaving your factory with, right now.

                            Comment

                            • orange
                              Automated Home Guru
                              • Dec 2014
                              • 149

                              #15
                              Originally posted by greyhound1234 View Post
                              Perhaps we should have a poll of what firmware we have, and whether we are happy with it.
                              v25 16Jan2014, and not happy...
                              Rameses - humour me, please tell me what firmware version the Evohome controller units are currently leaving your factory with, right now.

                              there is only that initial release firmware - AFAIK - as of today no further updates are planned to existing units - bugs or not.

                              Comment

                              Working...
                              X