Strange bug

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • paulockenden
    Automated Home Legend
    • Apr 2015
    • 1719

    Strange bug

    First time I've ever seen this one, but just before I went to bed last night I set economy mode for three hours. This normally sets the system to 'quiet' mode, so I don't get kept awake by the boiler/pump/pipes noise.

    Only the boiler kept kicking in for the next hour or so. So I checked as although most of my zones were at eco mode temperatures, a couple were showing mysterious over-rides. Sensible temperatures and times, but nothing whatsoever to do with me. And on two completely separate zones.

    Ghost overrides, if you like. Very strange. Can't think of any logical way that this might happen. Will keep an eye open (or an ear) in case the same thing happens again.
  • bruce_miranda
    Automated Home Legend
    • Jul 2014
    • 2307

    #2
    I am sure I have encountered this before and on Eco too, this was before the last firmware version. Couldn't repeat the issue.

    Comment

    • DBMandrake
      Automated Home Legend
      • Sep 2014
      • 2361

      #3
      Were the ghost overrides valid temperatures from your schedule around the time you switched to eco ?

      If so you've hit an issue that has been discussed before, and which I have experienced about half a dozen times now - HR92's send out a periodic set point change message (once an hour from the time of boot-up of the HR92 from memory) to what is already the current set point. It's some sort of set point "refresh" programmed into the HR92 firmware.

      If the Evotouch receives a set point change to what it believes is the current set point, nothing happens, no override is displayed and it is basically ignored.

      However this leaves a small window of opportunity of up to 4 minutes where trouble can happen. If your schedule happens to change a zone set point within the 4 minute window before the HR92 sends its hourly set point refresh, (or you manually change the override at the controller or use eco, etc) then the controller changes its opinion of the set point but does not send it to the HR92 until the next communication, which may be up to 4 minutes away. Meanwhile the HR92 is still at the original set point.

      If the HR92 happens to send its hourly set point refresh to the controller before the controller sends the set point change to the HR92, because the temperatures don't match the controller interprets this as if you had made a local override on the HR92 and shows it as a local override with the pocket watch icon, as well as reverting back to the previous temperature.

      The end result is that your scheduled or manual change made at the controller is ignored and it appears as if someone manually overrode the zone back to what it was before the change was made using the dial on the HR92. Which could potentially last for many hours depending on your schedule!

      As I discussed in the previous thread pertaining to this bug, it could be a very tricky problem to completely solve if the HR92 firmware is non-upgradable and the hourly set point refresh can't be distinguished by the controller from an actual user initiated manual override on the HR92... (potentially impossible to fix 100% if the set point change messages are identical)

      This issue might actually be why local override display was not available for so long - as the issue has only manifested since the local override display feature was added. (previously the controller would ignore ALL set point changes sent from HR92's)

      If the set point was a seemingly random temperature and not even one from your schedule then that is a separate issue, and one which I have also had happen twice! (And discussed with graphs in another thread) That issue I think just required a reboot of the offending HR92.

      Question for DanD if he is reading this - does the current Domoticz code monitor for set point change messages from HR92's ? If so, does it pick up the automatic hourly set point updates ? And if so, do the messages look identical to a set point change made by turning the dial on the HR92 ? That would be very interesting to know...
      Last edited by DBMandrake; 30 April 2017, 10:23 AM.

      Comment

      • paulockenden
        Automated Home Legend
        • Apr 2015
        • 1719

        #4
        You're right - the temps and end-points for the override all looked sensible (although I wasn't awake enough to remember the exact details).

        Sounds like Honeywell could easily code around this bug with a firmware update for the controller. Simply set a timer so that you know when the hourly refresh messages will arrive for each zone (to distinguish them from other messages), and if one arrives within five minutes of a (non-HR92) zone over-ride or mode change then just ignore it.

        P.
        Last edited by paulockenden; 30 April 2017, 11:13 AM.

        Comment

        • DBMandrake
          Automated Home Legend
          • Sep 2014
          • 2361

          #5
          Originally posted by paulockenden View Post
          You're right - the temps and end-points for the override all looked sensible (although I wasn't awake enough to remember the exact details).

          Sounds like Honeywell could easily code around this bug with a firmware update for the controller. Simply set a timer so that you know when the hourly refresh messages will arrive for each zone (to distinguish them from other messages), and if one arrives within five minutes of a (non-HR92) zone over-ride or mode change then just ignore it.

          P.
          That would be a hack.

          Because now an actual user initiated set point change on an HR92 made within that time period would also be ignored. This could lead to a manual override not being reported on the controller and thus not being cancelled correctly at the next set point, if the next set point was the same temperature.

          If there is some difference in the automatic set point message vs a user initiated one (like a different flag) the fix should be easy - if so why is it not fixed yet. On the other hand if the messages are indistinguishable, then it would be impossible to fix it 100% without a firmware update on the HR92 (which I'm assuming is off the table) and any hacks like the one you describe are only going to introduce other possibly more common and more noticeable side effects.

          That's why I was curious to know whether Domoticz logs these messages or not. If it does, they could be compared to see if they can be distinguished on something a bit more certain than timing...
          Last edited by DBMandrake; 30 April 2017, 12:51 PM.

          Comment

          • paulockenden
            Automated Home Legend
            • Apr 2015
            • 1719

            #6
            Another controller-level fix would be to send schedule/mode/over-ride changes immediately.

            P.

            Comment

            • DBMandrake
              Automated Home Legend
              • Sep 2014
              • 2361

              #7
              Originally posted by paulockenden View Post
              Another controller-level fix would be to send schedule/mode/over-ride changes immediately.

              P.
              Can't do that unfortunately. The HR92 will be asleep and won't receive the set point change! An HR92 only wakes up approximately every 4 minutes to receive communications from the controller, so you can't send it a message at an arbitrary time. If you could we wouldn't be waiting several minutes for HR92's to react every time we make adjustments from the controller.

              Interestingly the sleep time is not hard coded in the HR92 but is sent in a synchronisation message from the controller each time the controller communicates. In other words the controller says "Here is the data I have for you now, I will send another communication in 300 seconds", the HR92 will sleep for that amount of time before waking up and listening again.

              So, not a solution, and I can't see how even using a variable sleep time would help for user initiated changes - for a scheduled set point change the wake up time for the HR92 for the period just before the change could be shortened to coincide with the set point change and that would probably help with scheduled changes (I'm surprised this isn't done already) however a user initiated change could occur at any time so can't be predicted in this way.

              Even timing the communication window correctly only reduces the chance of it happening, it doesn't prevent it so still falls under the "hack" category in my mind.

              Comment

              • paulockenden
                Automated Home Legend
                • Apr 2015
                • 1719

                #8
                Surely a dirty hack that reduces the chances is better than the current situation?

                Comment

                • DBMandrake
                  Automated Home Legend
                  • Sep 2014
                  • 2361

                  #9
                  Originally posted by paulockenden View Post
                  Surely a dirty hack that reduces the chances is better than the current situation?
                  I'm not sure that the suggested hack would actually work though, and is more likely to cause trouble than help.

                  By the way, it must be Karma, but I just had it happen to me tonight!

                  The heating has been in heating OFF mode most of the day due to warm weather, a few minutes ago I cancelled the off action and all the zones reverted to their normal schedule on the controller. Just a few minutes later I walked past the controller and noticed that both my dining room and living room were blue instead of the expected orange that they had been moments before - both were now set to 5 degrees as a timed override!

                  So the exact same issue you reported, now conclusively shown to still occur on the latest firmware version. It's many months since I can remember this problem last affecting me.
                  Last edited by DBMandrake; 30 April 2017, 09:44 PM.

                  Comment

                  • paulockenden
                    Automated Home Legend
                    • Apr 2015
                    • 1719

                    #10
                    OK Simon, see if you can pick holes in this fix! ;-)

                    - IF local overrides are disabled for a zone then ignore any incoming messages which differ from the controller managed setpoint.

                    I know if it won't fix this for everyone, but it would in my case.

                    In fact, I'd say the fact that the controller does take notice of these phantom local overrides, even when overrides are disabled, is actually a bug in itself.

                    Comment

                    • bruce_miranda
                      Automated Home Legend
                      • Jul 2014
                      • 2307

                      #11
                      Another variant of the same bug? I have a Quick Action that sets all zones to 23C (used to ensure full system circulation). When I switched it on the controller showed me that the zones had a Setpoint of 23C. But shortly after that, 2 zones went back to 5C, which they were set to prior to me switching the Quick Action on.

                      Comment

                      • paulockenden
                        Automated Home Legend
                        • Apr 2015
                        • 1719

                        #12
                        Sounds the same.

                        Simon, you never replied whether my latest fix might work, at least in some instances.

                        Comment

                        • g6ejd
                          Automated Home Guru
                          • Oct 2016
                          • 153

                          #13
                          I've just started to use Economy mode to stop rooms being heated (correctly) when it's relatively warm outside and I notice the CS92 link keeps failing to report water temperature (--) on the display. When the schedule starts the next day it works again and if I don't use Eco mode the system works OK. I can't say if it was OK before the latest update, but it's behaviour that is wrong.

                          Comment

                          • paulockenden
                            Automated Home Legend
                            • Apr 2015
                            • 1719

                            #14
                            I can't see any way that the HW temp link could be connected to the Economy quick action. Doesn't make sense. I suspect the occurrence is coincidental.

                            Comment

                            • DBMandrake
                              Automated Home Legend
                              • Sep 2014
                              • 2361

                              #15
                              Originally posted by paulockenden View Post
                              Simon, you never replied whether my latest fix might work, at least in some instances.
                              Ignoring the set point updates from an HR92 that is supposed to have local overrides disabled ? Yes, I think that would avoid the problems on zones set to local override disabled.

                              Doesn't do anything for zones that do have local override enabled though - which would be the majority I would think, since its the default ? I've only disabled local override in two rooms, because I have a 1 year old that thinks that every button and knob that can be found in the entire house within his reach should be turned and pressed until it wears out...

                              Originally posted by paulockenden View Post
                              I can't see any way that the HW temp link could be connected to the Economy quick action. Doesn't make sense. I suspect the occurrence is coincidental.
                              I was going to reply and say the same thing - I can't see any connection between using the Economy action and problems with the hot water sensor, as the economy action doesn't touch anything to do with hot water.

                              If the temperature reading goes to -- it means that the Evotouch hasn't received any temperature updates from the CS92A for a long time. I don't remember how long it is but its something on the order of 2 hours.

                              If the temperature in the cylinder isn't changing at all, or is well below the bottom differential temperature and is not changing much (middle of the night with the hot water scheduled off and not being drained) the CS92A only sends an update about once an hour. If a couple of those very infrequent updates go missing in a row because of comms issues (collisions, noise, etc) then it could easily change to --. I've actually seen this happen once myself before I moved my CS92A.

                              The reason it doesn't happen when the hot water is on and in use is because the CS92A sends much more frequent updates when the temperature is changing, especially when within the differential band. So if one of those updates was missed due to comms problems then another update will be along soon enough and it won't cause it to time out.

                              So unfortunately my thinking is that the underlying issue here is comms problems between the CS92A and the Evotouch, which is only made obvious when the CS92A falls back to infrequent updates when the system is idle and the temperature is not changing.
                              Last edited by DBMandrake; 10 May 2017, 09:26 AM.

                              Comment

                              Working...
                              X