Evohome Zone Going Over Setpoint Every 3Hrs

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • WiteWulf
    Automated Home Jr Member
    • Mar 2016
    • 40

    Evohome Zone Going Over Setpoint Every 3Hrs

    Evening all, I've been graphing my new evohome install for a few days now since it was installed and have noticed something a little off with one zone. The house is unoccupied at the moment, so all zones are set to 15degC. With the exception of obvious solar heating as the sun moves around the house during the day I can see that all zones but one are sticking close to the setpoint. The odd one out, however, is rising up to 18degC and cooling back down again on what seems like a 3hr cycle. I've included a screenshot of the graph below:

    3hr_periodocity.jpg

    I'm going to move the HR92 to another room and see if the problem follows it tomorrow but thought I'd ask on here first in case anyone has seen a similar issue before. FYI the system does not currently have any optimisation enabled.
  • paulockenden
    Automated Home Legend
    • Apr 2015
    • 1719

    #2
    Yes = I posted about something similar. Everyone told me the room was too small or the radiator too big, and that I should wind down the lockshield.

    I did. It made naff-all difference.

    P.

    Comment

    • rotor
      Automated Home Guru
      • Aug 2015
      • 124

      #3
      Is it a small room, with a possibly over-sized radiator? I ask, because I was one of those that suggested to Paul. I have a tiny room with a normal-room-sized radiator, and it is the room that gives me the worst spikes.

      Comment

      • DBMandrake
        Automated Home Legend
        • Sep 2014
        • 2361

        #4
        Originally posted by WiteWulf View Post
        Evening all, I've been graphing my new evohome install for a few days now since it was installed and have noticed something a little off with one zone. The house is unoccupied at the moment, so all zones are set to 15degC. With the exception of obvious solar heating as the sun moves around the house during the day I can see that all zones but one are sticking close to the setpoint. The odd one out, however, is rising up to 18degC and cooling back down again on what seems like a 3hr cycle. I've included a screenshot of the graph below:

        [ATTACH=CONFIG]771[/ATTACH]

        I'm going to move the HR92 to another room and see if the problem follows it tomorrow but thought I'd ask on here first in case anyone has seen a similar issue before. FYI the system does not currently have any optimisation enabled.
        A couple of possibilities:

        1) The radiator is oversized for the room, the room is really well insulated or the flow temperature is set too high for the size of radiator, room and weather conditions, (warm weather) or a combination of factors.

        Describe the room and in particular the size of the room and the radiator.

        Basically what can happen is if the minimum amount of heat the HR92 can put into the radiator (by opening the valve slightly for a minimum of 4 minutes) is too much for the room then as soon as the radiator panel heats up even if you closed the valve then and there the temperature of the room would continue to rise and overshoot the set point due to the thermal mass of the radiator staying hot for quite some time.

        2) The system has not calibrated itself to the characteristics of your radiator valve, radiator and room yet. This takes time and you say you have only had the system a few days - that is not long enough. You need to give it a couple of weeks.

        Over time it adapts to both the opening characteristic of the TRV valve body itself, and the thermal response time of the radiator and the room.

        Every TRV valve body is a bit different even for the same model of valve (mine all differ noticeably) and the HR92 has to learn where the "opening point" of the valve is in terms of pin travel, which initially it has no way of knowing. EG does the water start to flow at 20% pin travel ? 30% ? 50% ? Likewise the operating range of valves differs - once they start to open how far does the pin have to travel until the water is flowing without restriction ?

        It learns this empirically by making adjustments to the valve position and monitoring the temperature changes over time. Eventually it figures out with a great degree of precision just at what point the valve opens and once it has learnt this its ability to regulate the temperature without overshoots improves.

        The other thing it has to learn is the thermal response of the radiator and room - eg how quickly does the room heat up and how much thermal overshoot is there after the valve is closed ?

        Eg if your target is 20 and it waited until it actually got to 20 before closing the valve it will overshoot the set point, possibly by several degrees because the radiator stays hot long after the water flow is stopped, still putting heat into the room. So what it does is measure this initial overshoot and when the same set point change comes around again the next day it will start to close the valve earlier than the set point by the right amount calculated to "coast" to the target temperature.

        The degree of accuracy that it can do this to in most cases is amazing, and makes it look much easier than it really is. However any change in the room characteristics can temporarily fool its prediction of when to start turning the valve off causing undershoot or overshoot until it adapts again.

        Some rooms are just naturally problematic and "unstable" and are difficult if not impossible to prevent overshoot, which can be caused by a combination of factors like:

        Radiator oversized for the size of the room, room very well insulated, high outside temperature, (low heat loss) flow temperature too high, radiator valve sticky and difficult to control accurately.

        I would give it another week or two and see if it has made any improvements.

        By the way all this learning and adaption is always enabled, and is completely separate to "optimisation" which is actually optimal start and optimal stop. Optimal start and optimal stop simply bring the set point changes ahead in time from their normal schedule, but aren't related to the process of learning the thermal overshoot of the room or valve control.
        Last edited by DBMandrake; 23 April 2016, 11:27 PM.

        Comment

        • DBMandrake
          Automated Home Legend
          • Sep 2014
          • 2361

          #5
          Originally posted by paulockenden View Post
          Yes = I posted about something similar. Everyone told me the room was too small or the radiator too big, and that I should wind down the lockshield.

          I did. It made naff-all difference.

          P.
          Adjusting the lockshield isn't ideal and would only have an effect if it was nearly closed, which would then affect full flow. I'm not sure if its a bathroom but have you tried leaving a towel over it ?

          It takes a day or so to adapt to a change like a towel over the radiator so try leaving it over for a couple of days.

          In my bedroom if I throw a towel over the radiator mostly but not completely covering the radiator (which is a fairly large 2.3kW double convector, a bit oversized for the room) then the room takes significantly longer to heat but initially it will overshoot by a couple of degrees. (Since the towel slows the heat transfer into the room increasing the thermal lag)

          However the very next day the system has adapted and learns to start closing the valve even earlier so that it doesn't overshoot. If I then remove the towel then for a couple of days it undershoots the target until it adapts again.

          With something like a towel over your radiator I suspect after a couple of days you may see the overshoot disappear. (I'm just suggesting it as an interesting experiment, not a solution)

          Comment

          • WiteWulf
            Automated Home Jr Member
            • Mar 2016
            • 40

            #6
            Thanks guys, executive summary: stop worrying and leave the system to settle down for a while

            All the rooms in the house are empty at the moment, so the thermal profiles will all change dramatically once we've moved furniture and bodies in.

            Thanks again...

            Comment

            • DBMandrake
              Automated Home Legend
              • Sep 2014
              • 2361

              #7
              I just wanted to report that now that heat demand is so low due to warm weather (the system just runs a little bit in the mornings and the bedroom over night to keep the newborn happy...) I am seeing sporadic overshoot in some zones that should not be and I have now pinned down the cause.

              Sticky/Stiff TRV valve bodies! This was one of the suggestions that I gave in this thread but I have now seen it first hand in three different zones.

              The symptom is that the zone generally controls the temperature well, but then somewhat intermittently I'll find the zone 1 to 1.5 degrees over temperature and the radiator still flowing quite audibly and red hot. This should not happen.

              Overshoot is often just a matter of thermal mass of the radiator - the valve was closed and water stopped flowing completely when the set point was neared but because the radiator holds its heat for a long time it keeps radiating heat into the room well after the water flow stops, until the radiator cools, causing an overshoot.

              But that is not what is happening here - the water is still clearly and audibly flowing when it should have long since stopped as the temperature is past the set point and still rising. Holding the menu button and going to option 10 Valve Position reported 18% - which is normally well into the fully closed region. So clearly the HR92 is trying to close the valve and thinks it has closed it, because generally the closing point is about 30-40% indicated after the initial calibration.

              The interesting thing is that going through a valve calibration temporarily restores the correct threshold, and the valve seems to work better for a few days but then the problem recurs. I have done this several times in one zone only to have it overshooting in the same way a few days later. For some reason it gets out of calibration again. Here is how to test for this happening:

              Remove the HR92 from the radiator and turn the black wheel fully anti-clockwise. Fit it back on and click the latch. Turn the temperature up to at least 3 degrees above the current room temperature. After about a minute the zero calibration process will start where it winds the valve right down, then because you have the temperature set high it will then wind it fully open again. So both ends of the valve pin travel are calibrated correctly to 0 and 100%.

              Now set the temperature to about indicated temperature and listen for water flow. Make 0.5 degree adjustments until the water flow through the radiator just stops. This is the closing point of the valve. Go into menu option 10 and find the valve opening percentage. I find it tends to be between about 30-50%. Note this figure down. A few days later do the same test where you adjust the temperature by 0.5 degree intervals (waiting each time) until it just closes. See if the percentage figure has dropped dramatically over what it was. If it has you may have the same problem as me.

              In my case the valve closed calibration creeps down from about 40% to under 20% in a few days in the problem zones. In fact, if I repeatedly fully close and open the valve (by setting to a very low temp, waiting for it to close, setting to a very high temp, waiting for it to open etc) then each time it goes through an open/close cycle it seems to creep down a bit.

              So is the valve "sticking", for example a swelling o-ring that is grabbing the pin or just metal to metal friction ? Could be - the valve pin seems fairly stiff on these valves which are Pegler Bulldogs, which seem to me to be of rather mediocre quality. (They are only just over one year old too)

              Another possibility is that if the valve is stiff/sticky it may be causing the HR92 to prematurely think it has reached the fully closed (pin down) position before it really has, thus causing it to keep "recalibrating" the closed position a little bit each time you go from open to fully closed. When it winds the valve down it has a memorised closed position from previous calibration cycle, but if the motor finds it "too hard" to wind down to that point it will actually stop before it gets there and recalibrate that new, not quite as closed as before point as the zero point. So a valve position that used to be 40% might now be represented as 30% because the 0% point was recalibrated.

              So a stiff/sticky valve could well cause calibration creep that eventually results in the valve only closing properly when nearly 0%. (My offending valve was only fully closing at an indicated 10% after this "calibration creep" had occurred) To confuse matters its possible that a sticky valve may vary in stickiness depending on temperature or other factors like pressure. (How many other radiators are open) I've spoken to Richard at the evohomeshop about this and he agrees with my theory that it is probably sticky valves and that the Bulldog valves are not the best quality... (I won't repeat his exact words )

              So I think my plan is to bite the bullet and replace the Bulldog TRV valve bodies with the Honeywell Valencia ones during my summer maintenance in a few months...fortunately I "only" have 9 radiator valves, so whilst expensive (even with just the one valve body by itself per radiator) its not prohibitively expensive. I've never been that happy with the Bulldog valves (the original TRV heads are crap and inaccurate, with one even failing due to the wax pellet exploding...) and now that I've seen its causing me intermittent temperature overshoot issues the perfectionist in me will not be happy until the problem is resolved...

              If/when I replace the valves I'll report back the results.
              Last edited by DBMandrake; 24 May 2016, 01:09 PM.

              Comment

              • paulockenden
                Automated Home Legend
                • Apr 2015
                • 1719

                #8
                The Honeywell Valencia valves are very good, and of course play well with HR92s.

                Not particularly expensive, either.

                Comment

                • DBMandrake
                  Automated Home Legend
                  • Sep 2014
                  • 2361

                  #9
                  Originally posted by paulockenden View Post
                  The Honeywell Valencia valves are very good, and of course play well with HR92s.

                  Not particularly expensive, either.
                  That was my thinking - if anything is going to play well with the HR92 then it will be the Honeywell Valencia valves. They're about twice the price of the Bulldog valves if you buy a full kit (inc unwanted lockshield and TRV head or hand wheel) but being able to buy just the TRV valve body without the rest brings the price down a bit.

                  7 of my 9 valves are 8mm microbore, but I should be able to use standard 15mm to 8mm internal reducers which can be had for about 80p each.

                  Anyone thinking about using an HR92 with the Bulldog valves - they're crap. They worked OK for about 6 months but started getting sticky after that - first one valve in one room, then a few months later two other valves followed suit. Whilst they're technically on the HR92 compatibility list I wouldn't recommend these valves knowing what I do now.
                  Last edited by DBMandrake; 24 May 2016, 01:56 PM.

                  Comment

                  Working...
                  X