Hi everyone, I am in the process of buying materials for my floor radiant heating system install. I have had the plan and design done by a professional and will be doing to actual labor of installing myself. This is a reto fit. Part of my house is on a concrete slab.

Question I am debating is whether or not I should install a thermal break between the concrete and the subfloor. Install methodology:

- 6mill poly above concrete
- 1 inch OSB w/ sleepers for PEX tubing
- Oak flooring above sleepers

As far as thermal breaks, I was thinking either foil faced double wrap, or 1/2 rigid foam. Thoughts?

foil faced double wrap

Do you mean foil faced "bubble wrap"?

rigid foam will always be a better choice. I would do less on the sleepers (3/4" max) and more on the "thermal break"... presuming the existing concrete is not insulated.

if the existing concrete IS insulated, don't bother with the thermal break.

Any competent designer would specify the insulation. Bubblefoil has few legitimate uses, least of all, under radiant slab...or on top of them for that matter.

Thanks for the reply guys.

Sorry, yes I did mean foil faced bubble wrap.

I was actually thinking 1/2 rigid foam. I would like to keep the floor height down as much as possible. The 1 inch of osb (will include the pex in this layer) + the 3/4 oak flooring and I am already at 2 1/4.

Could anyone comment on how good of a thermal break 1/4 inch foil faced bubble wrap versus 1/2 inch blue rigid foam?

yes, the bubble foil would be a nearly complete waste of money.

"...nearly..." ??? Without air gaps both sides it's TOTALLY worthless!

A half inch of blue board is ~ R2.5- enough to be worthwhile, but an inch would be even better. Going with half-inch OSB and a Crete-heat style foam to support the OSB without the need for sleepers might be a better option. You only need 1" subfloor if you're going to be spanning between joists or sleepers- if fully supported (or on a reasonable grid) the compressive strength of rigid foam is high enough (and the concrete slab below stiff enough) that sleepers are overkill. Just mark the tubing locations on the half-inch OSB as you go to avoid nailing the tubing when you put down the oak.

But even 1" of foam might be a be unacceptably lossy in a cold climate- where are you?

I'd be most concerned with the type of flooring you plan on installing, you need a very low moisture content so your flooring expands and contracts as little as possible with the heat / cool cycles involved. Wood flooring above radiant is usually engineered product or bamboo with very low moisture levels. This is if you are talking real "oak" flooring. Check with your manufacturer for the moisture content and recommedations for radiant heat installations.

I am located in north NJ, zone 5. System designer said the heat loss for the first floor (the concrete slab) was not as bad as the basement (also on concrete). I should mention that the house will have fiberglass r-13 in the studs and Poly-iso 1.5" on the outside below the siding. I guess I am just trying to double check before I start the installation (second opinions).

I had read that in order for bubble foil to work correctly, there needs to be an airspace either below or above. I am assuming now that it will not function as a thermal break between the concrete and the osb floor/radiant now as well. I thought the bubble foil needed the air gap to supply its R value but that had no effect on it functioning as a thermal break. For example, a sill gasket between the foundation and the studs on the bottom, is considered at thermal break correct?

Dana, the advantages of something like crete-heat look good on paper. Considering this is a DYI install, is this something we can get our hands on and what would you guesstimate costs to be?

As per my heating guy I need 5/8 minimum clearing to install the pex. We have a huge amount of 7/16 osb. My plan was to put down a 6mill poly, then 1 layer of 7/16 OSB, then slightly groove the pex channels, then rip the reaming 7/16 sleepers 15 inches wide and screw them to the first layer of OSB. This way my pex is sitting below in channels and then install my 3/4 oak flooring.

As per the manufacturer guidelines, the oak flooring should be fine over the radiant heat. Poly above the concrete and below the subfloor should stop moisture from causing problems from below.

+1 to the worthlessness of double bubble wrap.

Go with rigid foam.

Posted By cloves on 22 Nov 2011 08:49 PM
I am located in north NJ, zone 5. System designer said the heat loss for the first floor (the concrete slab) was not as bad as the basement (also on concrete). I should mention that the house will have fiberglass r-13 in the studs and Poly-iso 1.5" on the outside below the siding. I guess I am just trying to double check before I start the installation (second opinions).

I had read that in order for bubble foil to work correctly, there needs to be an airspace either below or above. I am assuming now that it will not function as a thermal break between the concrete and the osb floor/radiant now as well. I thought the bubble foil needed the air gap to supply its R value but that had no effect on it functioning as a thermal break. For example, a sill gasket between the foundation and the studs on the bottom, is considered at thermal break correct?

Dana, the advantages of something like crete-heat look good on paper. Considering this is a DYI install, is this something we can get our hands on and what would you guesstimate costs to be?

As per my heating guy I need 5/8 minimum clearing to install the pex. We have a huge amount of 7/16 osb. My plan was to put down a 6mill poly, then 1 layer of 7/16 OSB, then slightly groove the pex channels, then rip the reaming 7/16 sleepers 15 inches wide and screw them to the first layer of OSB. This way my pex is sitting below in channels and then install my 3/4 oak flooring.

As per the manufacturer guidelines, the oak flooring should be fine over the radiant heat. Poly above the concrete and below the subfloor should stop moisture from causing problems from below.

R-value is 1 divided by BTUs-per-hour persquare-foot-per-degree-F.  The rate of heat transfer at any given temperature difference falls with increasing R value, and R-value is the determining factor of a thermal break. If it has no appreciable ASTM C 518-testable R-value in the stackup as-installed, it has no appreciable thermal break.  Without the air gaps the emissivity of the foils is irrelevant- there is no radiated heat transfer between materials in direct contact with each other, and the only thermal break you'd get out of it is the R value as measure in the ASTM-defined test plates with full contact on both sides. (And is a damned small R-value with bubble-foils.)

With air gaps the amount of thermal break using radiant barriers is partly a function of the absolute differences in temperature and the emissivity of the surfaces that face one another in the gap, but it's not linear with temp the way a simple R value is.  In order to claim an R value for reflective insulation the test must be performed with a delta-T  of something like 30F degrees (?) not much more, and not much less, and must be tested in the orientation of installation. The effective-R in a wall installation is much lower than that in a horizontal installation with the warm side up, and the effective-R in a horizontal orientation cold-side up is lower still, due to convection currents within the air space.

A sill gasket is indeed not a thermal break, but is rather a capillary break- it blocks moisture from wicking into the wood from the porous concrete.  The soft foams used in sill gaskets has some minimal unspecified R value (less than R0.25), but provides some measure of air-seal for the gap between concrete & wood (which it does to a degree, but it's no substitute for a quality caulk or spray-foamed air seal.)

In your DIY stackup, rather than putting OSB under the tubing & sleepers,  put down 1-2" XPS instead.  Then put the ripped down 7/16" OSB sleepers between the tubing, and glue them, don't screw them to the XPS, using purpose-designed foam-board construction adhesive (available at box stores). The solvents used in standard construction adhesive will degrade the XPS.  The XPS alone is an excellent capillary break and sufficient vapor retarder to keep the wood happy- no poly necessary.  The compressive capacity of even low-density XPS sheathing is pretty good, and the OSB sleepers + flooring distributes the weight over a sufficiently wide area the you could literally park your truck on it without permanent deformation of the XPS, and will be fine for residential furniture & foot traffic loads.

In the zone 5 corner of NJ the deep subsoil temps are in the low-50s- it would be better if you could get at least R10 between the tubing slab, whereas a skinny inch is only R5.  Insulating the slab edge at the exterior may be worthwhile here if you're only going R5. In zone 5 there's a 25 year net-present-value argument for R7.5 even WITHOUT the radiant, and that's with a floor that's only 65-68F.  In a radiant floor situation the floor itself is going to be ~10F or more warmer and since heat transfer increases with increasing temperature differences going to ~R12 (2.5" XPS) would be cost-effective here on energy savings, assuming natural gas fuel in a condensing boiler.  If propane or oil or electricity (at NJ power rates) is the energy source there's a long term financial case for going even higher R.

Dana thanks so much for the very detailed reply! The XPS first with OSB sleepers inbetween the tubing on top on top of the xps sounds like a good idea. I have read a few posts and noticed that some people say that the radiant tubing makes a lot of noise as water moves through it. Would the above installation technique require the sleepers to be tightly placed on 16" (touching the tubing) or loosely placed (leaving a half inch on each side of the tubing). I also saw online that one vendor mentions that they have some outer coating on the pex that prevents the noise from occuring.

I recently asked the designer a few questions based on this discussion and this is what he relayed to me. "Heat loss is based on square footage and insulation and type of radiant. The system is fast response there will be minimal loss to the existing slab unless there is a moisture issue. That is why there must've a moisture barrier."

if the existing slab is not insulated, then it is a part of your heat loss.

fast response or not is irrelevant: the slab will pull the heat out of your heated emitter after it shuts off, requiring another heating cycle on a faster timeframe than otherwise, unless the slab is at equilibrium temperature with the radiant on top and its downward/edge loss.

long story short if existing slab is not insulated, you want insulation on top and radiant over that. Roth panel is the only good way to do that in a short height profile, or alternate foam with sleepers for support.

I recommend PAP tubing with lightweight plates in most cases.

Posted By cloves on 29 Nov 2011 03:33 PM
Dana thanks so much for the very detailed reply! The XPS first with OSB sleepers inbetween the tubing on top on top of the xps sounds like a good idea. I have read a few posts and noticed that some people say that the radiant tubing makes a lot of noise as water moves through it. Would the above installation technique require the sleepers to be tightly placed on 16" (touching the tubing) or loosely placed (leaving a half inch on each side of the tubing). I also saw online that one vendor mentions that they have some outer coating on the pex that prevents the noise from occuring.

I recently asked the designer a few questions based on this discussion and this is what he relayed to me. "Heat loss is based on square footage and insulation and type of radiant. The system is fast response there will be minimal loss to the existing slab unless there is a moisture issue. That is why there must've a moisture barrier."

There's nothing in the physics of heat transfer that would make that statement true- the thermal conductivity of a high moisture content slab isn't very different from that of a bone-dry slab, and you're not putting the moisture barrier below the slab in any event. The moisture content of the subsoil will affect it's thermal conductivity, but unless you're digging down 3' to put the moisture barriers there, 6 mil poly or sill-gasket thickness foam between the slab & tubing won't cut the heat loss by a measurable amount.
I s'pose if you had an inch of standing water submerging the tubing there would be better conductive thermal contact between tubing & slab though.

Get a new "designer". This one can't be educated. Heat loss can be affected by many things but is determined by the difference between outside temperature, inside temperature and the materials between the two.

I had Fred Seton's team design the system for the house. I really was not able to dig up any negative comments on his work after doing research. I am sure everyone on these boards has come across posts of his work.

I dug up part of the concrete flooring, and they actually put down a vapor barrier to my surprise. Its looks like 1mill or 2 mill at most. Better then nothing considering the place was constructed in the mid 60's.

Thanks for explaining why insulation on top makes sense. It makes sense that the slab would suck heat from the bottom and into the earth since its not insulated. The Roth panels looked very solidly constructed but very over budget.

Looking at cost and height consideration, I am leaning toward 1" XPS and ripping the OSB sleepers in-between the tubing as Dana mentioned. Should the OSB sleepers be placed tightly against the pex tubing to prevent noise and movement?

no, leave room. however you will want a snug fit on your aluminum plate channel to pipe. loose fit is why many plates don't perform well. Plates is one area, if memory serves, where Seton's website is wildly off the mark.

We see Seton's flawed designs regularly,
His standard 500 foot loop coupled with a 8, 9 or 10 circuit manifold and a 0011 pump, bubble foil insulation and a takagi water heater is bad math. He seems to design to the lowest cost. We have seen large systems, that are destine to fail
Dan