Summer Setting

[paulockenden]

Within the first few months of using Evohome I learned not get obsessed by numbers.

I just tweaked my schedules until the house felt comfortable, almost ignoring the temps recorded and set.

Even where I graph my system (using Demoticz and Initial State) I rarely look at the actual values. It's more about graph shapes and trends.

P.

P.

[DBMandrake]

As a slight aside, I'm pretty sure I see a similar behaviour in all zones when the setpoint drops in the winter. The displayed temperature does fall pretty quickly, despite that fact that when I put thermometers in these rooms, they show a much slower drop.

Keep in mind that the indicated temperature is rounded towards the set point, (why, nobody outside Honeywell knows) so for example if the set point and actual temperature were both 20 degrees then 20 degrees would be indicated, however as soon as the set point changes to 5 degrees at the HR92 (which can take a few minutes after the schedule changes) the reading will usually drop to 19.5 even though no temperature change has occurred at the HR92 yet. This doesn't have any effect on the control of the radiator, its purely a cosmetic issue. It does also affect wall stats like the DT92.

However, yes, on top of this rounding effect I do see what you do in all my zones where an HR92 is the sensor - as the radiator cools there is a drop of at least 1 degree in the reading whilst the room temperature (measured elsewhere with a digital thermometer) has not dropped at all.

So the indicated change can look like 1.5 degrees or more a short time after a set back, although the part of the temperature "drop" caused by rounding towards the set point only affects the displayed temperature, not the behaviour of the radiator valve in terms of whether it will come back on or not, so can be ignored.

I don't see this quick drop behaviour at all in rooms with DT92 sensors (well, I only have one at the moment, but I have tried it in two different rooms) nor my hall where the Evotouch is used as the sensor - when the radiator cools down because the room has reached the set point the reading stays the same and agrees with a 3rd party thermometer in the room, and there is no tendency for the radiator to cycle back on again when the room is still warm enough.

Unfortunately I'm pretty sure this is just an inherent limitation of trying to measure room temperature beside a radiator that neither Honeywell or anyone else can solve. You can't beat the physics - its just a really bad place to get a representative measurement of room temperature... The only real solution to eliminate all sources of measurement error is to measure the temperature away from the radiators out in a good location in the room, such as:

* On a wall about 1.2 metres high (half typical ceiling height) so that you're not measuring too close to the cold floor or warm ceiling. (And so that the controls are at a usable height)
* At least 1.5 metres from the nearest radiator to avoid any direct heating from the radiators
* Not near any other sources of heat like a TV, AV Cabinet etc
* Not in a location that would have direct sunlight shining on it at any time of the year
* Not under a window
* Not next to a door or open hallway.
* Preferably somewhere near the occupants of the room, if the room is large, such as near the sofa in a living room if it is a large room.

Finding a location that meets all those requirements at once can be pretty tricky to nearly impossible in a smaller room. In our living room due to furniture placement the only two feasible locations are near the door either next to the light switch on the door knob side, or on the wall that the door opens against, just past the door opens. (So it is not obstructed by a fully open door) I ended up choosing the latter after some testing with a standalone thermometer.

I found that when the door was open even a small amount, the light switch location on the door knob side would register a temperature drop fairly quickly if the hall way was cold, whereas the other location on the wall that the door opens against would not register a spurious drop, but gave the same reading as with the door closed. Since installing the DT92 I've been very happy with the temperature measurement and control of the room in changing conditions, and have monitored the room with an independent thermometer as a reference.

Before the weather got as warm as it is now (where the room goes well over-temp even with no radiator on - as high as 25 degrees lately) I found that the temperature it was maintaining was within about +/- 0.2 degrees of the set point according to my 3rd party thermometer that was placed on the arm of a sofa near the DT92, and subjectively the room temperature felt comfortable as well. Can't complain!

[top brake]

Eco mode works for me.

[guyank]

Within the first few months of using Evohome I learned not get obsessed by numbers.

I just tweaked my schedules until the house felt comfortable, almost ignoring the temps recorded and set.

Even where I graph my system (using Demoticz and Initial State) I rarely look at the actual values. It's more about graph shapes and trends.

P.

P.

I agree. I don't pay any attention to what the numbers actually say and have the setpoints made to make the rooms comfortable. However, it is a bit of a problem when the weather heats up and the room is still comfortable, but because the HR92s are now sensing a lower temperature, they demand heat. I tired turning it back on last night to see what would happen, and sure enough, almost every room in the house demanded heat this morning. Unfortunately I'm not there to see whether this demand stops after a while.

Keep in mind that the indicated temperature is rounded towards the set point, (why, nobody outside Honeywell knows) so for example if the set point and actual temperature were both 20 degrees then 20 degrees would be indicated, however as soon as the set point changes to 5 degrees at the HR92 (which can take a few minutes after the schedule changes) the reading will usually drop to 19.5 even though no temperature change has occurred at the HR92 yet. This doesn't have any effect on the control of the radiator, its purely a cosmetic issue. It does also affect wall stats like the DT92.

I am aware of the setpoint rounding issue, but as you see, the temperature does drop by more than the 0.5C rounding 'error'. I have considered external temperatures, but the DT92 is not the best looking item of kit and the T87RF2033 is too expensive to justify. Plus, when I suggested the having a thermostat fixed to the wall in the lounge, the look that I got told me it wasn't going to happen

It's just something that I'll have to live with, but its another reason why I would not use this system if I was doing a house refurbishment and was making a mess of the walls anyway.

[DBMandrake]

I have considered external temperatures, but the DT92 is not the best looking item of kit and the T87RF2033 is too expensive to justify. Plus, when I suggested the having a thermostat fixed to the wall in the lounge, the look that I got told me it wasn't going to happen

I agree that both DT92 and T87RF are too expensive, certainly to be considered a standard piece of kit to get the best performance out of the system in every zone, and not all rooms need a sensor that also has a display and buttons. (Although I find it convenient in a heavily used room like the living room, as the primary way to adjust and see the temperature closer to eye level than the HR92)

There is the HCF82 of course which is a sensor with no buttons or display, but IMHO it's rather angular and ugly (I much prefer the look of the DTS92) and bizarrely is even more expensive than a DT92 at £58 versus £55. What gives Honeywell ?

There is a real need for a really small, unobtrusive, WAF compatible, anonymous looking temperature sensor that performs the function of the HCF82 but with no buttons, fins, branding or anything else visible, that sells for no more than £20. The binding button could be on the back so that it has to be lifted off a slide in bracket to perform the binding then put back on place. If such a thing existed for no more than £20 you could make a case for deploying it in most/all zones that don't already have a full DT92 or T87RF. But at £55-£71 for the existing temperature sensors this is simply not an option for most people.

It's just something that I'll have to live with, but its another reason why I would not use this system if I was doing a house refurbishment and was making a mess of the walls anyway.

As this is a fundamental limitation of measuring the temperature next to the radiator, what other system do you propose to use that wouldn't have the same problem ?

Any system where all zones have the ability to call for heat that measure temperatures beside the radiator are going to suffer from the same issue as its a matter of physics and how heat distributes in a room. Even manual TRV's that can't call for heat from the boiler will still come on when the radiator is cold when the room is up to temperature if the boiler is turned on - they suffer from the same issue, the difference is that they don't have a temperature readout so that you know what temperature they are sensing...

At least with the Honeywell system there is an easy solution to the problem - a remote wireless temperature sensor that can be located anywhere suitable in the room to more accurately sense the temperature and control the radiator. A lot of other systems simply don't give you that option.

The only issue is that the sensors are too expensive and may not have sufficient WAF but a much cheaper and unobtrusive sensor could easily mitigate this, so go to it Honeywell!

Or just use Eco or turn your heating off manually.

[guyank]

I'm not disagreeing that this is a fundamental physics issue with this type of installation and for a retrofit installation I believe that it is still the best option. I'm sure that now Honeywell have set a standard, other manufacturers must be looking at alternatives that may be better implemented but will still not get over this particular issue unless they also produce small, cheap sensors.

The point I was making was that if I were doing a new build or a whole house refurbishment, Evohome would not be my system of choice because of the limitations induced by putting a sensor next to the radiator. I haven't looked at this in detail as it is not something I am looking at doing in the near future, but if you used the Heatmiser system or something similar and a heating manifold, I believe that you would achieve a much more accurate level of control. However, it would probably be considerably more expensive.

[top brake]

I'm not disagreeing that this is a fundamental physics issue with this type of installation and for a retrofit installation I believe that it is still the best option. I'm sure that now Honeywell have set a standard, other manufacturers must be looking at alternatives that may be better implemented but will still not get over this particular issue unless they also produce small, cheap sensors.

The point I was making was that if I were doing a new build or a whole house refurbishment, Evohome would not be my system of choice because of the limitations induced by putting a sensor next to the radiator. I haven't looked at this in detail as it is not something I am looking at doing in the near future, but if you used the Heatmiser system or something similar and a heating manifold, I believe that you would achieve a much more accurate level of control. However, it would probably be considerably more expensive.

Not sure why so many of you feel that measuring temp at the HR92 is not good. Look around all houses and you will find TRVs. The only issue is if the HR92 is covered by a rad cover, curtains etc

Majority of evohome installations just use HR92 rather than the other way round.

[G4RHL]

I have no issues with the HR92 sensor being at radiator level. It is designed to take account of where it is and whilst as a room warms up it may read a degree or so out from a thermometer elsewhere in the room it seems to eventually balance. Plus I don't see it all that important so long as there is something ensuring the room is at a comfortable temperature. When the system is new one can't help playing and fiddling, but leave it alone and it does its stuff. You soon find if you generally need a slightly different setting. Have to say the concerns over the position of the HR92 sensor etc. is a little bit of navel gazing. The HR92 sensor is far better and more accurate than other TRVs I have had.

[DBMandrake]

I have no issues with the HR92 sensor being at radiator level. It is designed to take account of where it is and whilst as a room warms up it may read a degree or so out from a thermometer elsewhere in the room it seems to eventually balance.
[...]
Have to say the concerns over the position of the HR92 sensor etc. is a little bit of navel gazing. The HR92 sensor is far better and more accurate than other TRVs I have had.

Every room is different though - if you have internal walls that are solid brick/plaster like we do, there is a lot of thermal mass in the room to heat so it can take hours for the bulk of the room to reach an equilibrium at the desired temperature if the temperature is measured at the radiator.

For example in our living room with the temperature measured at the HR92, say the room is scheduled to 21 degrees at 8am on a Saturday morning. In cold weather the radiator will come on at about 6:30am and "reach" it's target of 21 by 8am. However when we get up soon after this we find the room subjectively feels a lot colder than 21 degrees - a thermometer sitting on the coffee table in front of the sofa confirms it - and only reads about 17-18 degrees even though the heating has been "on" for an hour and a half. It takes about a further 3-4 hours for that coffee table reading that is more representative of the room to creep slowly up to 21 degrees and the room to start to feel comfortable and actually feel like it is 21 degrees.

It takes so many hours to eventually get there because when the HR92 measures 21 degrees at the radiator it throttles the radiator right back even though the rest of the room is nowhere up to temperature yet. In reality we don't wait, we just turn up the temperature to 22-23 for an hour or so to speed up the process but then need to turn it back down again to 20-21 otherwise the room will eventually end up too hot.

The remote DT92 completely solves this - because it measures the temperature well out into the room the radiator keeps going full tilt until the room itself nears 21 degrees and only then does it throttle the radiator back. The warm-up time of the true room temperature is vastly faster when measured remotely like this as it avoids the radiator being prematurely throttled back. We find that when it claims the room has reached it's target 21 degree temperature as reported by the remote sensor it does indeed feel like it has, and there is no need to make a manual adjustment and no chance of the room ending up too hot later due to forgetting to turn the temporary temperature boost back down again.

In short, it becomes set and forget. Up to temperature quickly and efficiently and at a comfortable temperature no matter what the prevailing outside conditions are. I'm no longer having to make any corrections to set points with changes in weather conditions outdoors. Call it navel gazing if you like, but my decision to use a remote sensor in the living room is not just based on theory, its based on months of temperature monitoring at various different locations in the room and comparing those to both the HR92's internal sensor and how we subjectively feel the temperature, and the verdict is clear - the remote sensor gives a reading that tracks much more closely with the feeling of comfort in spite of weather changes while the HR92's measurement does not, and it also optimises warmup times.

I won't argue with you that the HR92's sensor is far more accurate than a conventional TRV - this kind of goes without saying as manual TRV's are horrifically inaccurate for multiple different reasons that I won't go into, (why do you think they're only labelled 1-5...) and under fixed conditions like consistent prevailing outdoor temperatures the HR92 can maintain a very steady and consistent temperature in the room, once that long equalisation period has happened through the room. (which can take many hours however)

However make a radical change to the conditions such as a sudden cold snap in the weather and you will still find yourself having to make changes to your set point to maintain subjective "comfort" because the temperature offset between the HR92's sensor and the real room temperature has now changed.

People are so conditioned to "turning up the TRV" in colder weather that they don't stop to realise that if they have to do this it means the temperature measurement and control system is flawed and is not actually maintaining the temperature at the part of the room where the humans reside... You do not need a higher temperature in the room just because the outdoors is colder during a cold snap, what's really happened is that the room is actually colder than it was previously, despite the same set point.

This is especially so with a manual TRV for multiple reasons. An HR92 eliminates some of those errors (such as elimination of proportional offset error) so the need to make adjustments to the set point with changes in weather are significantly reduced with an HR92 vs manual TRV but not completely eliminated. A well chosen remote temperature measurement point eliminates the remaining sources of error so that measured temperature corresponds closely to subjective comfort no matter what the prevailing conditions are outside.

BTW for those already using domoticz to measure their room temperatures with independent (non honeywell) temperature sensors, it is apparently now possible to bind those temperature sensors to the evohome via domoticz, and thus utilise them in place of a DT92 etc...

http://www.domoticz.com/forum/viewto...t=10555#p86178

It still looks fairly experimental but for anyone already using domoticz this could be a solution to the "remote sensors are too expensive and/or obtrusive" issue, and while I don't use domoticz at the moment it is an interesting development.

[Bazinga]

We've gone down the external sensor route for one of our bedrooms that never warmed up properly, and the DTS92 made a huge difference.
Our main room has the evotouch controller in it, and this room is usually OK, but it's got a very high ceiling and it's south facing, so I've been using the HR92s' temperature offset to try to keep the room under control. This method does work (-2 in winter, -1 in Spring /Autumn, 0 in Summer), but it's hardly "set and forget". Can I use the internal sensor of the evotouch as the remote sensor for this zone, or do I need another DTS92 or a T87RF, etc. for this zone?