GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon
Q&A Spotlight

Installing a Concrete Slab the Right Way

Should the slab be poured on a base layer that includes sand, or will that only cause problems?

Right or wrong? An architect has provided this detail for a concrete slab. Is the detail flawed?
Image Credit: Torsten Budesheim

As part of a remodel of his San Francisco area home, Torsten Budesheim is converting the 700-square-foot lower level into living space. An existing slab has been removed, and Budesheim has removed a few inches of material to increase the finished ceiling height. Now, he’s nearly read to place a new 5-inch-thick slab that will include tubing for radiant heat.

Budesheim’s architect has supplied a drawing for the concrete slab and its base layers, and the details are making Budesheim nervous. In particular, he wonders about the call for a 2-inch-thick layer of sand immediately below a 6-mil waterproofing membrane and an inch of rigid insulation.

What has the gears turning are two articles Budesheim has come across, one by GBA senior editor Martin Holladay and another by the Building Science Corporation (BSC). In each, Budesheim finds a warning that a layer of what’s known as “blotter sand” beneath the concrete will lead to problems.

“Is the above the right layer structure or do we need to revise?” Bludesheim asks at GBA’s Q&A forum. “One thing we do not want to change is the concrete slab and radiant heat in it. So I’m basically interested in what should go below the slab at this point. We really don’t want to have to worry about ever touching that slab again!”

Is the architect’s recommendation to include a layer of sand a good idea? And what other issues should Budesheim consider before the new slab goes in? Those are the topics for this Q&A Spotlight.

Get the sand out of the picture

Ask your architect to read the same articles you have, Holladay suggests, and notes, “The sand layer shouldn’t be there.”

Although the sand layer is unnecessary, at least the architect has called for a 6-mil vapor barrier above the sand; the 6-mil poly prevents the moisture in the slab from migrating upwards, avoiding most of the problems associated with the sand layer.

Kevin Zorski adds this: “The other problem with the sand is: what’s to prevent the sand from, over time, filling all the voids in the crushed rock layer, and creating possible cracks in your slab? I agree with Martin. Lose the sand!”

Richard McGrath adds to the chorus. “Should it eventually get wet,” he writes, “it will actually hurt the performance. Water within six feet of a slab in contact with earth must be accounted for in the calculations.”

OK, says Budesheim — no sand.

How much insulation should be used?

But what about some of the other details, including the amount of insulation the architect is recommending?

Holladay says that 1 inch of rigid foam isn’t enough under a heated slab, even in Climate Zone 3, where Budesheim’s house is located. Two inches is the minimum, he says, and more would be better.

Dana Dorsett would rather see 3 inches of rigid foam under the slab, and there’s something else about the architect’s proposal that bugs him.

“There are several issues with the detail I don’t like, not the least of which is the lack of thermal break between the slab and the grade beam foundation,” Dorsett says. “Not only does there need to be more than an inch of foam (2 inches [of] EPS would be the minimum for a heated slab, 3 inches better). Even with an unheated slab you’d want 1 inch at the slab edge for comfort, and maybe 1 inch sub-slab on the perimeter, but for a heated slab you’d want 2 inches (R-8.4) under the slab, and at least 1.5 inch (R-6.3) between the slab and the grade beam.

“That ‘floats’ the slab mechanically as well as thermally,” Dorsett continues, “which probably makes it safer from a seismic/structural point of view, but consulting local codes on that would be prudent. If the beam and slab must be in contact for some reason, 1.5 inch of EPS on the exterior would be called for.”

Well, Budesheim replies, that’s going to be a problem. The architect is calling for a strong connection between the existing stem walls and the new slab. He’s not including a thermal break of any kind between slab and the foundation wall.

“I did read somewhere that in my climate, decoupling the slab from the side foundation walls, while helpful, is not as critical as in colder climates,” Budesheim says. “Is that correct?”

Sort of, Dorsett says. Decoupling the slab from an uninsulated foundation isn’t critical in temperature Climate Zone 3C and is not required by the International Residential Code.

“But when the slab becomes a heating radiator, it sure matters!” he adds. “If it were a cast-iron radiator, it’s the thermal equivalent of exposing the ends of the radiators to the outdoors.”

An inch of extruded polystyrene (XPS) insulation doesn’t provide much of a thermal break, Dorsett says. At R-5, it’s “better than nothing, I suppose, but not exactly great.” If Budesheim were to add an inch of XPS on the outside of the foundation wall, plus 1 inch between the new slab and the foundation, however, he’d be in the R-10 to R-11 range, “which is much more reasonable.”

If the architect insists on eliminating the insulation between the slab and the foundation, as Budesheim suggests he will, Holladay tells him that exterior insulation will be a must.

“You need to install insulation on the exterior side of your concrete stem wall, down to the bottom of the footing, and you need to protect the above-grade portion of the insulation with metal flashing or stucco,” he says.

Where should the tubing be located?

Budesheim is committed to a radiant-floor heating system, and the question now turns to exactly where in his proposed 5-inch-thick slab the tubing should be placed. He’s considering a product called Creatherm, EPS floor panels which include molded channels for PEX radiant floor tubing.

Although the Creatherm panels look as if they’d work, Dorsett says, they place the tubing at the bottom of the slab rather than in the middle of the slab, and that’s going to make the heating system a little less responsive.

“In practice that probably doesn’t matter much,” he says. “I have no idea how it’s priced, but it saves labor on the heating installation end. Sheets of smooth 2-inch-thick Type-II EPS under slabs typically runs 75-85 U.S. cents per square foot, installed; Type X (2-pound density) is more like 90 cents a square foot. Type-X is preferred by some radiant heating contractors for its better staple retention, which allows them to simply staple the tubing to the foam (again, at the bottom of the slab) rather than tie it to the metal reinforcing layer, which is more labor-intensive.”

Tubing does not belong in the bottom of the slab, McGrath says.

“Expansion joints are often the topic when this discussion takes place,” he says. “First, there are two types of slabs, those that are cracked and those that have not cracked yet. Nature of the beast. If it were not so, we would not need expansion joints. Expansion joints are just a place for the slab to crack at instead of wherever it wants due to conditions. These joints should have their location identified and tubing should dip below for 6 inches on either side or if they are to be saw-cut the slice needs not be more than a 1/2 inch to 3/4 inch deep.”

David Meiland would probably not be voting in favor of panel products such as Creatherm, in part because there already is an easy solution to the problem of locating tubing in the slab.

“Rip a 45-degree bevel on rigid insulation and install it against the perimeter walls with the bevel at the top,” Meiland says. “Install crushed rock without fines as your slab base, compacted very flat. Place rigid insulation over the rock and then your [vapor barrier] over the insulation. Install a mat of #3 rebar on 1 1/2-inch dobies. Tie the tubing to the top of the rebar. Pour a 4 1/2-inch slab. When you make your 1-inch-deep control cuts, you will not hit the tubing.”

Our expert’s opinion

Peter Yost, GBA’s technical director, adds this:

Unfortunately, there are a number of problems with the slab detail.

Blotter sand: As others have clearly stated: lose this. This layer of sand originally was added as a moisture sink so the top and bottom of the concrete dried more evenly, reducing edge curl (see the Building Science Corporation article, “Concrete Floor Problems”). But it also was thought necessary because a wetter mix was easier to move around. A water-to-cement ratio of 0.5 or less is generally recommended, but this stiffer mix does not make the folks handling it as happy!

Thickness of sub-slab insulation: In this climate, sub-slab insulation is more about thermal comfort than energy savings. However, for a heated slab, it’s a whole different story. The 2105 International Energy Conservation Code calls for adding R-5 beneath a heated slab in Climate Zone 3, but it’s silly to be redoing this slab and not incorporating at least R-10 worth of insulation.

Slab edge insulation: This actually is more important than the under-slab insulation. Every GBA slab detail shows some sort of slab edge detail. Does this introduce any stability issues in terms of tying the grade beam to the slab, especially in San Francisco? Since this is a structural engineering question, I turned to Bruce King of the Ecological Building Network.

King responded, “Here in the land of the dancing earth, we engineers tend to tie everything together, like slab edges to perimeter footings. But there’s nothing sacred about that. Especially for such small spaces, I doubt that it is all that important, and would weigh the large thermal needs (insulate the slab edge) over the modest seismic benefit of tying slab to footing. But again, there’s lots of context I don’t know [for this particular project] that might make me change my mind.”

Finally, both as a builder (not long ago…) and consultant, I have found the “Type” system for the different grades of polystyrene insulation — expanded polystyrene (EPS) and extruded polystyrene (XPS) — confusing. It’s important that you know the Type of either EPS or XPS that you are specifying or buying. I have found this resource helpful.


  1. User avater
    Armando Cobo | | #1

    Sometimes I wonder.... hummm!
    IMO, some of the advise given here, although in good faith, is given under the "Northern" mentality forgetting that in Southern climate zones things are different. SF is in an earthquake zone, and you should not decouple foundations many can do in other parts of the country without some major structural risks, and it is also in heavily infested termite zone. Those were specs given to me by different engineers for several projects over the years.
    I do believe that a good recommendation is to insulate above the concrete slab, R10 min., with Quicktherm, Creatherm, SmartTrac or any other grooved panel.
    What I do see in the detail is that the 6 mil. VPM is above the sand, however, that sand will sieve through the stone in short time with all the soil movement in that area. I would avoid the sand and use all river rock instead.

  2. Zac Blodget | | #2

    The sand is below the vapor barrier
    It seems significant to me that the membrane is specified between the sand and the concrete - thereby protecting the concrete from any moisture problems the sand might create. I'm just guessing but I'd say the sand is there to protect the membrane from the sharp edges of the gravel layer - seems like a good idea to me.

  3. Edward Odgers | | #3

    Wheres the steel for a connection?
    There seems to be something missing in the drawing. If the architect is requiring a "strong connection between the existing stem wall and the new slab," it isn't shown. The cold joint shown in the drawing won't do it. I suspect that there is a note or additional drawing that calls for dowels drilled and epoxied into the existing stem wall that are later embedded in the slab. If so, the slab will need to have a mat of steel to accept any load from the dowels, also not shown. The point is that such an arrangement could be compatible with perimeter insulation between the slab and stem wall. The dowels could pass through the perimeter foam by "spearing" the precut foam over the embedded dowels. Of course, the dowels will create a thermal bridge but not as egregious as the full-contact slab. Such a design detail will likely require thicker dowels, because of the intervening foam, but it should be able to transfer potential tensile and compression loads from the stem wall to slab while reducing the thermal bridge.

    Control joints should be cut or struck in to a depth of 1/4 the slab thickness. If cutting, make sure make them as soon as possible (12 hours max). You can also use a zip-strip product for the joints. To minimize curing shrinkage and cracking, keep the water/cement (slump) as low as practical to achieve proper placement and consolidation. To further reduce shrinkage cracking, ask the ready-mix plant for a low-shrink mix which will likely include a plasticizing admixture.

  4. John Harris | | #4

    Incorrect statement
    "placing a layer of sand over the top of a plastic ground cover under a concrete slab in California."
    "Yet this is exactly what Budesheim's architect is recommending."

    No it isn't. The Architect has ridge foam and poly on top of the sand, this is NOT blotter sand, there is no sand above the poly. When I built in the Bay Area I did the same thing because to the expansive clay soil. The sand layer was just to get a nice smooth surface over the clay and gravel to support the foam.

  5. User avater GBA Editor
    Martin Holladay | | #5

    Response to John Harris
    You're right. I was one of the commenters on the original thread who failed to notice that the 6-mil poly in the architect's detail was above the sand layer, not below the sand layer. My comments failed to take this into account, and I gave bad advice.

    I'd like to apologize to GBA readers -- especially to Torsten Budesheim -- and to Budesheim's architect.

    John, I appreciate your comment and Zac Blodget's comment (Comment #2). Thanks to both of you.

    I have edited the article to eliminate the confusion and bad advice.

    1. Hugh Weisman | | #14

      This concurs, in a way, with the advice I just received from our structural engineer on Martha's Vineyard. I specifically asked him "What should slab underlayment be? Compacted earth, stone, gravel, sand, ping pong balls? " His answer was "6 inches of clean compact sand." There will be a poly vapor barrier on top of sand. The area where this dwelling is being constructed has soil that is extremely no drainage issues, where gravel might be needed.

  6. Carl Mezoff | | #6

    Radiant tubing location
    The depth of the tubing in the slab will affect the heat output of the system. John Siegenthaler, PE, recently published an article reporting on his modelling various tubing depths for a slab under a wood floor. He reported that the optimal tubing depth is at about 1/4 the slab thickness. As the tubing is lowered, the slab output decreases. With the tubing at the bottom of the slab, the water temperature had to be raised about 7F to achieve the same heat output as tubing at the higher level. Elevated water temperatures will cause greater underslab heat losses, and degrade the operational efficiency of the heat source somewhat. Where the tubing crosses a control joints, the tubing can be depressed to the bottom of the slab and covered with a strip of WWF to keep it out of reach of the saw blade. Where the tube crosses a sawn joint, the tubing can be sleeved for added protection.

  7. Michael Clarke | | #7

    Radiant tubing location
    Re Carl Mezoff

    Tubing needs to be sleeved whenever it crosses a control or crack inducing joint. As you can imagine, a crack inducing joint, once cracked will have fairly sharp and ragged edges at the underside of the slab, hence the need to sleeve all crossings.
    As for poly and other means to aid in securing tubing in place, mostly this was developed in Europe where common practice is to have tubing embedded in a screed of 50mm-75mm thick. The screed and insulation usually sit on the concrete slab, though this process is used for above grade floors as well.

  8. Rob Mckelvey | | #8

    Sand on gravel
    I'm curious of everyone's thoughts on putting the layer of sand directly on the gravel. I understand the benefits of providing puncture protection for the vapor barrier and a more level surface for the insulation, but i have concern that over time the sand will sift into the voids of the gravel - thus leaving unsupported voids under the slab.

    What if the VB (10 mil) was laid directly on the stone, then sand, the rigid insulation. This would keep the sand from sifting into the gravel and the insulation board would keep the sand from having direct contact with the slab.

    or is the argument of the need for sand for 'leveling' for the insulation not as big of a concern. thoughts??

    1. User avater
      Jon R | | #10

      Old, but for others: the vapor barrier should be directly against the concrete with nothing that would retain moisture. Small vapor barrier holes don't matter.

  9. Miggida | | #9

    I'm also doing a new heated slab in the n. calif area. I'm not finding a lot of info on how to pour a new slab and new footing that has 2" EPS between the slab and the footing. how do I do that when I'm pouring both at the same time? drive rebar in the ground to hold the foam in place between the two?

    1. Malcolm Taylor | | #11


      Any chance of a sketch? I'm having trouble envisioning what you are trying to do.

      1. Miggida | | #12

        thanks for following up. I'm trying to pour the slab and footing all at the same time so that it looks like this picture (courtesy radiantec) where I have EPS underneath and in between the footing and the slab. I've got wooden forms up outside of the footing, but how do I keep the vertical facing foam (the one between the slab and footing) in place while I'm pouring the concrete?


        1. Miggida | | #13

          p.s. the easy thing to do would be pour the footing first, then the slab later, but I want to do all at the same time with one visit from the truck.

          1. User avater GBA Editor
            Martin Holladay | | #15

            The illustration you posted shows a building with a frostwall perimeter foundation. After the frostwalls are poured, the slab is poured on compacted stone installed between the frostwalls.

            Perhaps you are installing a different type of foundation -- a thickened-edge slab?

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |