Insulation and hydronic heat on basement slab
I’ve been researching this as a DIYer for a few weeks and this forum has helped me refine my plans. I am reaching the following installation as a final destination. Existing condition is 4-inch uninsulated slab installed per plans in 1974 over a 6 mil polyethelene barrier
6 mil PE or Epoxy paint as vapor barrier (6 mil much cheaper, so if good enough, I really prefer that)
1/2 EPS (some say XPS?) across the entire floor
1×4 as sleepers spaced 12 inches on center (atop rigid foam), anchored with tap cons (or Ramset) into the concrete below
Pex tubing between 1×4 sleepers
Quickcrete or similar on top of pex tub to bring it flush with sleepers, using sleepers as screed.
Durock or Ditra mat
9×38 ceramic tile
Questions:
1) Will this be solid enough for ceramic tile? That is, are sleepers above rigid foam a good construction technique here?
2) For the sleepers, am I better with 1×4 pine or ripped Advantech plywood?
3) 6 mil PE or epoxy as vapor barrier (noting that building plans report a barrier but not insulation beneath slab)?
4) Any downside to laying 6 mil barrier atop slab even if there is one (albeit 45 years old) beneath the slab?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
>"6 mil PE or Epoxy paint as vapor barrier (6 mil much cheaper, so if good enough, I really prefer that)"
If the description "...4-inch uninsulated slab installed per plans in 1974 over a 6 mil polyethelene barrier..." is accurate there is no point to additional vapor barriers. Even after a half-century it's still quite vapor retardent, and the layer of foam is vapor retardent too.
>"1/2 EPS (some say XPS?) across the entire floor"
R2 isn't much insulation, WOEFULLY little for a radiant floor, even in fairly temperate climates. Even in IECC zone 3 you'd want at least R4. What are your deep subsoil temps? For reference see:
https://www.builditsolar.com/Projects/Cooling/US-ground-temps.gif
The performance of XPS declines to that of EPS of similar density as it's extreme greenhouse gas blowing agents diffuse out over a few decades. At 20 years half-inch XPS is only warranteed to R2.25, and (you might even collect on that if you bothered to pay a lab to test it.)
>"1×4 as sleepers spaced 12 inches on center (atop rigid foam), anchored with tap cons (or Ramset) into the concrete below
Pex tubing between 1×4 sleepers
Quickcrete or similar on top of pex tub to bring it flush with sleepers, using sleepers as screed.
Durock or Ditra mat
9×38 ceramic tile"
The concrete would be nearly guaranteed to crumble in that stackup, even with the cementicious board up top. You'd be better off using Roth panel (glued to the slab with foam board construction adhesive) for getting the heat out of the tubing, with 1/4" tile backer glued to the top side, skip the sleepers. See section 6 on page 7:
http://www.roth-usa.com/PDF_Download_Files/Install_Manual_RadiantPanel.pdf
When the EPS (Roth is EPS) is fully supported by a slab (or another layer of EPS on top of a slab) below it really doesn't flex- not nearly as much as subfloor suspended between joists would.
Every quarter inch of thickness in the stackup you can replace with EPS adds another R1. In colder climates R10 (2.5" of EPS total, including the thickness of the Roth panel) would still make a lot of sense. Using 3/8" PEX instead of half-inch is also going to be significant when the total thickness of the foam is less than R10.
Thanks Dana for the reply.
I'm really confused by "The concrete would be nearly guaranteed to crumble in that stackup, even with the cementicious board up top. You'd be better off using Roth panel (glued to the slab with foam board construction adhesive) for getting the heat out of the tubing, with 1/4" tile backer glued to the top side, skip the sleepers. "
I keep getting different advice. Some people say ESP with sleepers above is totally stable. You and some others say not at all.
Aren't I just recreating something like a Crete-Heat panel? Or are you saying those also wouldn't work under tile?
You don't have enough concrete for it to have any strength on its own, this is made worse by the sleepers breaking it up into several smaller sections. It will flex and crack apart from foot traffic, etc. let alone the force of the sleepers shrinking and swelling with moisture changes.
But it's got Durock above and concrete slab below. Concrete is essentially incompressible so there's nothing to flex.
Please allow me to narrow my questions so that they can be asked more directly.
1) Assuming there is already a 6 mil vapor barrier beneath the slab, as per the blue prints, is there any harm to adding a 6 mil PE vapor barrier above the slab beneath any rigid foam installation? If there is no harm, it seems like a cheap peace-of-mind purchase.
2) Given the price of Roth panels, would a DIY "roth-like" system work adequately? I'm envisioning taking 3/4 inch untreated plywood and ripping it into 12 inch steps, leaving 3/4 inch between the strips to lay the PEX. Would a layer or durock or even luan on top be stable enough to support ceramic tile? I would place 1 inch of rigid foam beneath the plywood strips.
3) The plywood strip method involves a lot of tapcons. Probably 8 tapcons per 8 ft by 12 inch strip. Would this many tapcons damage the integrity of the slab at all?