underfloor heating causes black mold on skirting boards
- Thread starter flooringissues
- Start date
Yes but they're concerned with total losses. We're only talking about incremental losses caused by having a slightly higher insulation value on top of the slab (hence the ΔT's in the calcs). As you said yourself: "The higher the insulation value of the floor covering, the higher the percentage of heat energy will be lost through the sub-floor insulation into the ground." You're correct, but the incremental losses are negligible for wood floors, on the order of 1%.
Last edited:
No! 
We're trying to work out the difference in efficiency between having wooden flooring and something more conductive. We don't care what the total losses are (those will depend on the ground temperature, but we don't care about that). We only care about the additional losses due to raising the temperature for wooden flooring. If you want the whole picture including the soil temperature, then assume that the temperature difference between the UFH and the soil is ΔT₀, and the increased temperature for wood compared to some alternative is ΔT₁, temperatures in Kelvin. The subfloor insulation has a thermal conductivity of x W/m²K. The difference between the total losses for wood and the alternative in W/m² is:
So only the elevation in temperature for wood matters to our calculation (as ΔT₀ cancels out), and the additional losses work out to 1% of the total energy budget.
We're trying to work out the difference in efficiency between having wooden flooring and something more conductive. We don't care what the total losses are (those will depend on the ground temperature, but we don't care about that). We only care about the additional losses due to raising the temperature for wooden flooring. If you want the whole picture including the soil temperature, then assume that the temperature difference between the UFH and the soil is ΔT₀, and the increased temperature for wood compared to some alternative is ΔT₁, temperatures in Kelvin. The subfloor insulation has a thermal conductivity of x W/m²K. The difference between the total losses for wood and the alternative in W/m² is:
So only the elevation in temperature for wood matters to our calculation (as ΔT₀ cancels out), and the additional losses work out to 1% of the total energy budget.
Yes it really is that simple. At the higher slab temperature, heat will continue to be pushed through the floor for longer after switch-off, so the cool down time is longer. All that happens is that you shift the heating period slightly. (And this is even ignoring the fact that UFH is designed for mostly continuous operation, without large changes in temperature).
In the mean time, the only difference is whatever additional energy goes out through the subfloor, which we already calculated at 1%.
We can bring the figures from Junckers and Kingspan to bear again. According to them, for a 30 K difference between slab and soil, about 90% of the energy you put into your floor comes out the top, and 10% goes out the bottom. (These are the absolute values, not the incremental values I was talking about earlier). Let's just suppose you had to run the heating for an incredible 10% longer to get the slab up to a higher temperature. You would have an additional 1% of losses (10% of 10%). So you have put in 10% more energy, and lost 1% through the subfloor. Where does the additional 9% go? Up through the floor surface! When does it happen? Well our entire calculation is based on a constant energy flux through the floor surface, so the only way it can happen is by extending the cool-down period!
In the mean time, the only difference is whatever additional energy goes out through the subfloor, which we already calculated at 1%.
We can bring the figures from Junckers and Kingspan to bear again. According to them, for a 30 K difference between slab and soil, about 90% of the energy you put into your floor comes out the top, and 10% goes out the bottom. (These are the absolute values, not the incremental values I was talking about earlier). Let's just suppose you had to run the heating for an incredible 10% longer to get the slab up to a higher temperature. You would have an additional 1% of losses (10% of 10%). So you have put in 10% more energy, and lost 1% through the subfloor. Where does the additional 9% go? Up through the floor surface! When does it happen? Well our entire calculation is based on a constant energy flux through the floor surface, so the only way it can happen is by extending the cool-down period!
Well, you calculated it! I'm saying I don't agree the calculation takes all factors fully into account.
To get the same heat output through the floor for the length of time you want it, the heat of the slab is going to have to be increased to compensate. I don't see how your calculations factor in the additional losses from the slab to the sub-floor temperature as a result of the a higher differential for a longer time. The calculations need to be over an extended period of time rather than on losses in an instant of time. The thermal resistance of the glue isn't factored in anywhere either, assuming the floor is glued to the slab. Membranes such as Elastilion will have a higher thermal resistance. Elastilions test results show their product just falling within the German guidelines for maximum thermal resistance when testing with a laminated floor panel to a combined thickness of 11.65mm (0.14 (MsqK)/W versus the max of .015).
There's a reason manufactures recommend UFH pipes are laid more densely where wooden floors are specified.
Another point not considered for the OP is whether the rest of the insulation and air-tightness detail is sufficient that the reduced effective 70W/square metre will meet the heat requirements for the room.
To get the same heat output through the floor for the length of time you want it, the heat of the slab is going to have to be increased to compensate. I don't see how your calculations factor in the additional losses from the slab to the sub-floor temperature as a result of the a higher differential for a longer time. The calculations need to be over an extended period of time rather than on losses in an instant of time. The thermal resistance of the glue isn't factored in anywhere either, assuming the floor is glued to the slab. Membranes such as Elastilion will have a higher thermal resistance. Elastilions test results show their product just falling within the German guidelines for maximum thermal resistance when testing with a laminated floor panel to a combined thickness of 11.65mm (0.14 (MsqK)/W versus the max of .015).
There's a reason manufactures recommend UFH pipes are laid more densely where wooden floors are specified.
Another point not considered for the OP is whether the rest of the insulation and air-tightness detail is sufficient that the reduced effective 70W/square metre will meet the heat requirements for the room.
I'm more than happy to consider your calculations.
This is a red herring. In the limit, if you heated up once from cold and left the system running forever, any additional time required would average out to zero. If you want to factor in additional heat up time, then you'll have to say how often you're proposing to switch the system on and off from cold. Then you'll also have to factor in a contribution to additional cool-down time. To be honest that's a waste of time in a broad brushstroke calculation. (If you want to propose your own numbers I can easily factor them into the calculation, but in the mean time I'm sticking to my own experience of running one of these systems for a decade, which is that additional heat-up and cool-down time are roughly equal).
We can take an alternative approach to the calculation, and still get the same results. British standards mandate a maximum total loss through a floor, including all perimeters/cold bridging at 0.25 W/m²K (you can read them [broken link removed]). That's a ten percent loss for a 100W/m²K system running at 40 K above ground temperature. The Kingspan numbers I gave are between 0.1 and 0.2 W/m²K (you'd expect them to be better than the absolute limit specified in the regulations). The increased temperature for wood flooring is between 4 and 9 degrees in all the literature I have looked at (again feel free to provide your own numbers). That is between one tenth and one quarter of the temperature difference between UFH and ground, and since U-values are linear in temperature, means an additional 10% to 25% losses. Since the maximum losses are already 10%, that gives an additional 1% to 2.5% losses for wood, and the better end of the Kingspan range is only 40% of that (0.1 vs. 0.25 W/m²K), equating to 0.4% to 1% losses. It seems to me you would need some extraordinary parameters to hike that up to something non-negligible but, as I said, feel free to provide your own numbers.
The figures I gave for the wooden flooring include a soundproofing and moisture barrier layer for a floating installation as well as the boards themselves, which happens to be the sort of system I'm running. Having seen the sheeting go down during construction and being suprised at its thickness I very much doubt it is less insulating than a layer of glue in a fixed installation. You can check the U-values in the link I provided.
None of the ones I checked do.
Not relevant to the efficiency calculation though.
Last edited:
I don't think anyone suggested the system would be run forever, though in constantly occupied spaces, this is the most efficient means of operation provided the system is tuned to prevent the boiler short cycling. In that case you are correct in that you can really ignore that, given it's a one off event whose significance reduces to zero over time. But that's not what I was talking about, and it's not how must people I know run their heating. So you have to factor in increased losses while running at the higher temp, and during the prolonged warm-up time. The prolonged cool-down time would offset those losses to a degree that is dependent on the rate of cooling and any increase in the residual temperature the next time the heating comes on. Again, the more often the heating is used, the more efficient it will be.
I provided a link to the Junkers specs which calculates the temp difference as 10.5 degrees for a 14mm clip system. 14mm is their thinnest product, all their others will exceed the max thermal resistance advised by the German guidelines.
If you're exceeding PassivHaus standards, then perhaps you could skip the heating altogether
Very true, but could be a significant issue if not factored in to the design specs. I pointed it out as the OP's builder has told them they can't use wooden flooring of any type. So perhaps he has cut corners and not installed a system that can deliver the space heat requirements if wooden flooring is used?
Last edited:
You're totally determined to miss the point, so I'll only say it one more time: that is the incremental losses for wooden floors (in response to your original statement that "it will cost you extra to run"), not the total losses. It does not exceed PassivHaus standards.If you're exceeding PassivHaus standards, then perhaps you could skip the heating altogether
No you didn't. Your own quote is: "One example calculation from Junckers states that the underfloor heating needs to run at 37.5 degrees in order to achieve a surface temperature of 27 at the floor." You are assuming that without the wooden surface a water temperature of 27 degrees would produce a floor surface temperature of 27 degrees. With no temperature gradient that would defy the basic laws of physics. You always have to run the water hotter than the floor temp. Your reference is nothing to do with the incremental temperature for wood.
Yep, I did that.
And you've steadfastly not even tried to put a sensible figure on that. Remember even if you got zero additional cool-down time (which is completely alien to my own experience) the additional losses would be 10% (the fraction of heat lost through the subfloor) of the extra percentage of heat-up time. So if you want to claim significant additional losses, say even 5%, you would need 50% additional warm-up time. For instance, you'd need 12 hours extra on a 24 hour warm-up. You'd be a lot better burning your wooden floor for heat at that rate.
My actual real-world experience with a mixture of wooden floors and porcelain tiled floors is that the rooms with wood floors reach a comfortable temperature slightly after the tiled ones, and after switch-off they return to an uncomfortable temperature by about the same amount of time after the tiled ones... which is exactly what the calculations broadly suggest.
I've only ever been talking about incremental losses.
The PassivHaus standard for floors is 0.15 W/m²K, so the 0.1 you've been referring to from the Kingspan documentation does exceed that.
Perhaps I wasn't clear here, but the temperatures referred to are the surface temperature of the concrete to deliver 70 W/m² of heat into the room.
Because there's just too many variables.
Which is as you'd expect unless the rooms are fully thermally isolated.
We're never going to agree, and we've lost to OP a long time ago I fear. Have a good weekend.