So I've finally got this house I've been working on up and running. Extremely tight construction so very low infiltration plus excellent insulation in a moderate climate (Portland Oregon) should equal moderate to low heat load. Construction is plates over the subfloor covered with either wood flooring or radiant friendly carpet and pad so relatively low thermal mass.

Because the system uses air sourced heat pumps there is a tremendous advantage to keeping the water temperature low (better efficiency). Current set point is 95F.

Which brings me to my question. What differences might I expect between dumb and smart thermostats Aube or Tekmar?

The reason I ask is that there are some really elegant thermostats made by a German company called Jung. Stainless steel, real small, very simple but the problem is they only appear to come using something akin to the old bimetal design.

I can understand how in a high mass design it would be crucial to have something smart an predictive to avoid over or undershooting but what about my set up? Are most of you guys running these smart puppies or something plain Jane like what the forced air folks use and if so how does it work for you?

What I'm trying to get at is if the radiant thermostats are a necessity to avoid problems or a nice upgrade that makes thing work a bit better?

with low mass they are less critical for sure, though I do still prefer, with any radiant system, to use a thermostat that can take a floor sensor input in addition to the onboard air sensor, simply for comfort sake.

this generally means we're using only smart thermostats. I think it's silly to do radiant without the floor sensor guarantee these days, but if you aren't going to do that, then the additional difference between smart and dumb thermostats is incremental, not massive.

Rob,

Walk me through the logic of the floor sensor if you don't mind as I'm having trouble understanding why one really cares what the floor temperature is. I know lots of people use them so that sounds like heresy but it just isn't clear to me what that information gains you.

in a constant circulation system, nothing.

but in most systems installed in america today, they cycle on and off on an air sensor. trouble is, the air temp is not the only thing that determines your comfort. imagine you are over a cold crawlspace or a garage (to take the extreme example), air temp is satisfied, but the floor gets quite cold in between heating demands. you can get a "cold 70" that way pretty easy.

basically, it gives you direct control over one more aspect of your comfort. it can also, in high mass situations, eliminate lag coming out of solar gain periods.

Nothing is worse than having a cold, heated floor. If you're going to heat the floor, it's small dollars to guarantee that it's never cold unless you want it to be.. why not?

So Rob, the room's 70 but the floor's not warm. How does the logic here work? You run heat 'til the floor warms to 70? 100? Won't that overheat the room? Is this a 'smart' Tekmar dohickey that decides all this? And what does its flowchart look like - simplified for us laymen.

set the minimum to room temp, and then you can't overheat the room by more than a degree or so. it's simply to avoid letting the floor get too cold... not so it feels noticeably warm. just comfortable.

there are a few thermostats out there that can take additional floor sensors. many tekmar stats can do it, but I think Aube has one and Uponor as well... others probably do as well.

its' very simple. the thermostat calls for heat if EITHER the air OR the floor sensor is not satisfied.

This only works if the house isn't particularly low loss. In a very well insulated and low infiltration house it would probably result in overheating.

I ran some tests a couple of years back to get a handle on what temperature felt neutral (neither warm nor cold) for various material. I don't have the number at hand but as you would expect stone/tile was the highest, followed by wood, then carpet. This is because the stone conducted the heat away from whatever its in contact with, like your feet, much more quickly than the wood, and the wood than carpet.

But consider a tile floor in a extremely tight house. If the load is less than temperature required for the floor to feel neutral (which could easily happen) one has to choose between letting the floor feel cold and overheating. I don't see anything wrong with that but it wasn't my understanding of what the smarts in a smart thermostat were for, but maybe I'm wrong.

So, putting the concern about the floor getting too cold for comfort aside for the moment (because I made sure there was only wood and carpet flooring in the building I'm working to eliminate this exact problem) what if anything would a smart thermostat by me? I'm focused on the predictive ability.

Any ideas?

not true. if you set the floor temp to the desired room temp, you won't put much heat into the space. the upward range is a degree or two, perhaps, at most, due to thermal mass shift.

You are speculating that the floor temp is set to a higher temp for noticeable comfort or truly neutral temps: I don't recommend that at all.

I am saying you set it to avoid very unnoticeable DIScomfort... that is, you are limiting how chilly the floors can get. A 70 degree floor might feel cool, but it's not uncomfortable. a 65 degree floor, more uncomfortable, and some floors can get much chillier than that too.

floor sensors cut out that discomfort zone entirely.

Other than floor sensing, in a low mass situation the only thing smart thermostats really do is help adjust for poorly set reset curves and narrow the temperature differential of the space to +/- 1 degree instead of 2 or 3. If you have a good reset curve set, you can use a simpler thermostat.

Do these 'smart' thermostats, then, let you set a floor and an air temperature or does it go with a single setting for both? Rob, what's the "minimum" you refered to?

they are separate. I was referring to the floor minimum temperature setting. they come with a maximum setting as well (which should almost always be turned off) just in case you are trying to protect a fancy wood floor or something, and your water temps won't do it.

Got it. Thanks

I'm still having trouble understanding the floor sensor. If you set it low, then I agree you're less likely to overheat the room, but its less likely the system will come on due to a call for heat. Conversely if you set it higher then the floor will be warmer but you are more likely to get overheating. I don't see how you can have it both ways.

Let's take an example. A tile or concrete floor. Cold, clear night so the house is at near maximum loss (conducted and radiated). The sun comes up and it clear as a bell so plenty of solar gain. In fact the house goes net positive. Unless the floor is completely shut off the structure is going to overheat.

But if I understand it correctly the purpose of the floor sensor is to make sure the floor doesn't go off to avoid cold floors. Given a room set point of say, 70 that isn't going to be possible. Bare concrete feels cold to stocking feet even at 70.

Forgive me for being dumb as a doornail but I just don't see how it is possible to have a system that is set for a certain comfort level based on either air temperature, radiant surface temperature or some other metric and also demand that the floors be set at a given temperature. Sometimes one of these has to give.

What am I missing? It must be something because you're a smart guy.

BTW I think you're dead on as far as the advantage of predictive thermostats doing a better job of holding the set point better although I have to tell you I'm not a big fan of outdoor reset as a method of determining the load. Solar gain or night time radiant losses often are much more deterministic yet a simple outdoor reset does nothing to detect these. Why not calculate the total load on the building directly? There are a number of ways to do this from simple (the rate of change of the room temperature) to more complex (outdoor temperature, solar cell and a clock to tell you if its night or day) but as far as I can tell nobody does this.

sbe raises some good points. seems to me if the floor is a material that gives up heat easily (e.g. tile), then by setting the floor temp, you're more likely to overheat the room than with a material such as wood.
Another question, Rob: for hardwood floor heated from below using plates, where would the sensor go? seems like the surface is where you'd want it but how would that be accomplished? And what does a sensor look like anyway? How big is it?

I'm not giving a college seminar on floor sensors in a forum post

but suffice to say: room temperature is NOT the minimum temperature of most floors. they can get much colder in between heat demands, especially if you are not using outdoor reset so your demands get shorter, peak temp higher, and low temp lower (same average over time).

floor sensors prevent that the bottom swing, and thus improve comfort.

again: 70 is still much more comfortable than "colder than 70". and you can choose which parameter is more important to your comfort as well: it very well may not be the air temp in all cases.

there are indoor temperature feedback systems sbe: honeywell has one, tekmar's teknet 4 system. they are not cheap, but they are great. outdoor reset is not perfect, but it's simple, easy to implement, independent of your zoning, and a heck of a lot better than bang-bang operation.

I'm using floor temp sensing with a Tekmar 542 tstat as part of a whole TN4 system.

The room is set at 70. The floor min is set at 75. What does that mean? Well, since the room is a master bath adjacent to a bed room, overheating is not a problem. When the 75 deg min floor temp pushes the room above 70, the heat is just flowing into themaster bed room. As a result, the master bedroom's heat's duty cycle is lower than it would otherwise be.

Result, the master bed's floors are cooler than they would otherwise be. Its not a problem since they are wood and feel nice underfoot even at lower temps.

I could imagine a problem with setting a minimum floor temp causing the room temp to go too high if the room was isolated thermally. i.e. the door was regularly closed.

Otherwise, I wouldn't worry.

Ironically, this is how I got hooked on radiant heating. I installed an electric mat system (nuHeat, DEFINITELY the best electric retrofit).

The electric system ONLY had a floor temp sensor. All winter, we set it to go to 80 deg in the morning and then turn off for the rest of the day/night. The room never got too hot. But the floor was great. The door was again usually open.

Another example. Last winter, I had a zone, the main livingroom/kitchen portion of my first floor where the electronics weren't set up for a few weeks. The pump and plumbing was in place, but no tstat or wiring into the TN4 system.

So I turned on my one fully wired 1st floor zone and jumpered the livingroom zone so the pump rancontinuously.

Amazingly, it worked great. The loop temp was adjusted by the other zone and the room stayed nice. As the room heated up, it warmed the other room with a proper zone, turning off the boiler. The mix temp then went down.

What does all this mean? That a system will be very satisfying if its not theoretically optimized, especially if you have an open floor plan.

Don