Evohome HR92 valves reporting wrong temperature
Hi all,
I just got a evohome installed (by some professional as I needed some boiler work, installing some valves on radiators that didn't have them, etc...).
My setup includes a condensing combi boiler for heating and hot water, an evohome controller + relay, 6 radiators with 6 HR92 valves, 3 radiators with normal valves (two will be
replaced by heated towel rails, one will be removed when I redo the kitchen in a few weeks). The radiators and their valves are all far from furniture and not covered by curtains.
I have a few problems/questions with the temperature controls and I haven't found the answers I need in the old posts.
The temperature detected by the valves when the radiator is on or still warm is several degrees (4 or 5) higher than the one in the middle of the room. On the other hand, the temperature is correct then the radiator is cold, so I cannot just change the temperature offset.
In the last two days I had to keep setting the room temperature to 24C until the centre of the room is warm (about 20C) and then decrease the set temperature to 21 or 22C so it will stay at about 20C.
Is this normal? I expected some difference, but not this much as it basically makes the system useless and impossible to control.
Could some DTS92 thermostats help? If I understand correctly, I can use those to pick up the temperature and use the HR92 valves only as actuators.
Is it possible to fix the DTS92 to the wall instead of just using the desk stand?
My other question seems to be about a potential bug in the software. Let's say that the valve (and controller and app) think that a room is a 20C, but it's actually colder, so I set the temperature manually (with the app or controller) to 22C. Immediately after doing that, the app/controller will report an increase of 0.5C in the detected room temperature, even before actually turning the heating on. This seems to happen about 50% of times. Does this happen only to me?
HI mbf82,
I've had the system since November and I too have noticed the temp instantly reports a 0.5C increase. Must be a flaw in the App as I have not spotted this on the controller. I tend to be now using the App miore than the controller but would also appreciate some insight as to when users will receive the new firmware (I accept that beta trialling has to complete first).
This isn't a bug/fault in the HR92, but a fundamental limitation of measuring the temperature of a room right beside a hot radiator. It doesn't matter whether you have a conventional wax pellet mechanical TRV or an electronic sensor like the HR92. On a conventional manual TRV you just never knew how inaccurate this measurement location was because it has no readout. Ever wonder why conventional TRV's are labelled with 1 through 5 instead of a temperature in Celsius ? Now you know.Originally Posted by mbf82
As the room heats up from cold the middle or far side of the room could initially be as much as 3-4 degrees or so below the temperature measured by the HR92 as you have noted. About 1 degree of this error is due to the the close proximity of the HR92 to the side of the radiator - which can be compensated for by setting the calibration to -1 degree. I have all my HR92's set to -1 after comparison with a more distant room thermometer after the room (with closed doors/windows) has reached final equilibrium. Of course this means that when the radiator is cold the temperature reading will be about 1 degree below the "true" reading, but there is not much you can do about that unless you use a remote sensor that is not right beside the radiator.
The rest of the initial error that you describe (up to about 2-3 degrees I have seen) is not actually an "error" but a reflection of the fact that the side of your room near to the radiator is genuinely 2 or 3 degrees hotter than further parts in the room when the room is initially heating up. I think people generally assume that a room only has one temperature but this is not the case - any time the room is heating up in response to a radiator on one side there will be a temperature gradient across the room that causes heat to gradually flow from the hot side of the room to the colder side. There is a "thermal lag" on the far side of the room, and depending on where you measure in the room you will get quite a different result.
The HR92 will maintain a steady set temperature in the vicinity of the radiator during this time and will not report any change in temperature however heat will flow from the hotter side of the room to the cooler side of the room and the rest of the room will eventually "catch up" with the temperature of radiator side of the room reaching an equilibrium - so you would see this temperature gradually climb on a thermometer located on the other side of the room, but it can take a long time (an hour or more) depending on the construction of the room. We have solid brick with plaster for all interior walls downstairs and I definitely notice this effect when I have checked with another thermometer.
As this is not an error in the reading but simply a reflection of the fact that you're measuring the temperature of the hot side of a room that initially started off cold and therefore has a heat gradient across the room, you can't try to correct for this with the calibration setting. Another thing to keep in mind is that only a closed, reasonably well insulated room will reach a final equilibrium where the entire room is about the same temperature. If you have a room where the radiator is at one end and an open door into a colder or drafty hallway is left open there will always be a heat gradient across the room where the radiator side of the room is warmer than the door side of the room, probably by 2-3 degrres - no heating control system can do anything about this it's a fundamental fact of physics and the way heat flows.
If you would like the "middle" of the room to heat up quicker to your target temperature than in this scenario you could use a remote stat like a DTS92 either wall or table mounted at least 1.5 metres from the radiators. Now it will try to optimise the temperature where the stat is located instead of in the vicinity of the radiator. However assuming you put the stat on a wall near the "middle" of the room, by doing so the temperature of the room near the radiator will now overshoot the set target temporarily as the room heats up while heat is still flowing from the hot to the cold side of the room. When eventual equilibrium is reached (with closed doors/windows) the temperature of the radiator side of the room will drop again and the room temperature will be relatively even throughout the room again.
So it depends whether you prefer the cold side of the room to warm up quicker or whether you don't want the hot side of the room to overshoot. Personally I find overshoot more uncomfortable than just taking a bit longer to warm up.
In something like a living room an acceptable compromise might be a wall stat near the seating position. A remotely located stat is always going to be more accurate because it is not influenced by the direct heat from the radiator and it can be placed in the part of the room that you're interested in, such as the seating position. Just keep in mind that the whole room will still not have one temperature when heating up, the side near the radiators will be warmer than the other side of the room until it reaches thermal equilibrium. (Actually there will always be a small gradient)
No, it's a known behaviour of the evohome system - apparently it rounds the displayed temperature towards the set point by up to a maximum of 0.5 degrees. Why they chose to do this, nobody knows but it has been confirmed to behave this way. It doesn't affect the functioning of the system but personally I don't like this behaviour.My other question seems to be about a potential bug in the software. Let's say that the valve (and controller and app) think that a room is a 20C, but it's actually colder, so I set the temperature manually (with the app or controller) to 22C. Immediately after doing that, the app/controller will report an increase of 0.5C in the detected room temperature, even before actually turning the heating on. This seems to happen about 50% of times. Does this happen only to me?
It does however mean that if you are trying to do a temperature comparison between the evohome and another thermostat, the comparison is only valid if the room is at the set point - if the set point is a long way above or below the current temperature there will be a bias of up to half a degree in the direction of the set point.
Another thing to consider when comparing readings to another thermostat is that different thermostats have different response times to changes - the evohome sensors are quite quick responding compared to many other typical stand mounted room thermostats, such as the weather station one I use. (Which includes outdoor temperature, humidity etc) So if you are trying to do a comparison between the two not only should they be near each other you should let the room temperature stabilise for some time so that differences in the response time of the thermostats doesn't introduce another error in your comparison.
Last edited by DBMandrake; 15th January 2016 at 12:55 PM.
This behaviour is still there on the new firmware. Unfortunately I suspect this is considered to be a "feature not a bug" by honeywell.
I understand that one side is warmer than the other. I just expected the gradient to be less “sudden” as the thermometer was less than one metre from the living room radiator and about 80cm from the small bedroom one.This isn't a bug/fault in the HR92, but a fundamental limitation of measuring the temperature of a room right beside a hot radiator. It doesn't matter whether you have a conventional wax pellet mechanical TRV or an electronic sensor like the HR92. On a conventional manual TRV you just never knew how inaccurate this measurement location was because it has no readout. Ever wonder why conventional TRV's are labelled with 1 through 5 instead of a temperature in Celsius ? Now you know.
I think I will get some DTS92s for the living room and bedrooms and try to find a good position to fix them on the wall.
I've looked at some manuals I found on google, but I haven't found any documentation on how to associate a DTS92 to an existing valve and controller. Does anybody have any pointer?
I see. It vaguely makes sense, but seems like an odd and confusing behaviour.
Ok I didn't realise you were measuring that close.
In that case I would suspect that the thermometer you're measuring with is quite slow to react, thus is lagging behind the real temperature by a significant amount of time. As I mentioned, I've found the temperature sensors in both the HR92 and the Evotouch itself to be very quick responding and they respond far quicker than my table top thermometer, so if the room is warming up quickly the table top thermometer lags behind by quite a lot until the room temperature is steady.
Try positioning your thermometer propped up right beside the HR92 and try it again to see if there is a discrepancy or not. If there is still a bit discrepancy when the room is warming up it's likely that it's just a slow thermometer.
You don't associate it to the valve, you associate it to a zone on the controller. It's pretty easy to add to an existing zone - make sure the DTS92 is cleared of any bindings - it shouldn't have any when it's new but you never know.I think I will get some DTS92s for the living room and bedrooms and try to find a good position to fix them on the wall.
I've looked at some manuals I found on google, but I haven't found any documentation on how to associate a DTS92 to an existing valve and controller. Does anybody have any pointer?
Instructions are here:
http://www.honeywelluk.com/documents...stallation.pdf
In the evotouch you would hold down on Settings for 5 seconds to go into the Installer menu, choose to edit the existing zone and there is an option that says sensor, which will currently say radiator. Press this then choose remote sensor, it will ask you to bind the remote sensor - press the buttons on the DTS92 described in the manual above to bind and that's it! To go back to the HR92 as a sensor follow the same procedure but press the binding button on the HR92 rather than the DTS92.
BTW the Evotouch itself has a built in temperature sensor so for testing purposes you could try using the evotouch in place of where you are thinking of putting a DTS92 to get an idea of what effect it might have on the temperature control. This assumes your evotouch isn't already using it's sensor for the zone its in. The cavet if you do this is it needs to stay in the same location and it does of course need to remain powered.
It's quite fast. The sensor seems to update ones every minute. If the difference of temperature is big it can take a couple of updates to stabilise.
Oh, that's interesting. Thanks!You don't associate it to the valve, you associate it to a zone on the controller. It's pretty easy to add to an existing zone - make sure the DTS92 is cleared of any bindings - it shouldn't have any when it's new but you never know.
Oh, that's cool as I'm considering wall mounting the controller in the living room.BTW the Evotouch itself has a built in temperature sensor so for testing purposes you could try using the evotouch in place of where you are thinking of putting a DTS92 to get an idea of what effect it might have on the temperature control. This assumes your evotouch isn't already using it's sensor for the zone its in. The cavet if you do this is it needs to stay in the same location and it does of course need to remain powered.
How do I associate the controller itself to a zone?
By fast I don't mean how often it polls the sensor to update the displayed temperature, I mean how quickly the sensor reacts to a large change in temperature.
So for example if you moved a thermometer from a 15 degree room to a 20 degree room, even if the reading is polled and changes slightly every minute, it may still take a long time until the reading approaches 20 degrees.
This is mainly dependent on the thermal mass of the sensing element and also how well air can flow through the casing. My thermometer reads down to 0.1 degree increments and I think it updates once every minute like yours, but it is still very slow at adapting to changes in room temperature. I would say it would take it half an hour probably more to go from 15 to 20 if I moved it between two rooms maintained at those temperatures.
The same way as above - you go into the installer menu, edit the zone you want to change, and then for sensor you choose evotouch instead of remote sensor. In the system settings section you can also adjust the calibration (offset) of the built in sensor in 0.1 degree increments. (Unlike the HR92 where you can only change it in full degree steps)Oh, that's cool as I'm considering wall mounting the controller in the living room.
How do I associate the controller itself to a zone?
I understood what was happening.
The heating was off all day during some of the coldest days of the year, so the rooms cooled down quite a bit. When the heating turned on the radiators got very hot (and the room a bit warmer but still too cold). Then the radiators turned off because the sensor thought it was much warmer. As the radiators take several hours to cool down, the sensor kept thinking it was quite warm.
A few hours later the cycle happened again and the room got closer to the temperature I want and so on.
Over the weekend I managed to keep the temperature I want everywhere because the heating was on all day, so the rooms couldn't cool down enough to show this problem (and I set the target temperature to 1C more).
In the living room things worked better as I associated the controller as sensor for the zone.
After this experiment I decided to get some DTS92s for the various rooms (and not bother with hallways and landing as they don't matter much and they are surrounded by warm rooms).
I have used Evohome now for over 18 months and, yes, it is true that from time to time the odd zone can get above target temperature. In my experience, this is because the controller does not - and cannot - regulate the temperature of the CH water coming out of the boiler. In the 'good old days', the timer clicked ON and the hall thermostat demanded heat. Water circulated around the system and TRVs closed as the required temperatures were reached. Zoning allows heat to come on whenever and where ever it is demanded. In some zones, the CH water will be warming up: in some zones the water reaches the radiator at max boiler temperature (or as new as is possible). It is in the latter situation, when most of my zones are effectively 'Off', that I see HR92 over-reads. It doesn't happen as much if the HR92 is just maintaining a zone temperature: in such situations, the usual over temperature is 0.5C. Personally, whilst I love gadgets, I am not sure that spending even more on Evohome is worth it from a comfort or energy saving/cost saving perspective.
Posting Permissions
Reply With Quote