GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Building Science

The Concrete-Free Floating-Plywood Slab Assembly

Watch builder Josh Salinger show how to reduce the environmental impact of a new home by omitting the concrete from the ground-level floors

More than 10 billion tons of concrete are used each year, making it the second-most consumed substance on Earth (after water). Just the production process for cement accounts for 8% of all global greenhouse gas emissions.

In this video, Josh Salinger, owner of Birdsmouth Design-Build, is on-site in Portland, Ore., to show us how to create a more eco-friendly slab assembly by laying two layers of graphite polystyrene (GPS) foam insulation on top of compacted stone, covering that with a polyethylene vapor barrier, and finally topping it all off with two floating layers of tongue-and-groove (T&G) plywood.




  1. Expert Member

    Really enjoyable video by a builder who knows what he is doing.

    I wonder if it would be worth it to reduce the width of the foundation at the top so that the foam and subfloor could extend right to the exterior walls?

    1. Expert Member
      Joshua Salinger | | #18

      Thanks, Malcom!

      (I'm just catching up to the fact that this was published a few days ago)

      The reduced width foundation is a good idea. One of the challenges we have is closing the gap between the plywood and the perimeter edges. We solved it by sloping the foam and staggering the plywood so the top piece overhands the lower piece at the edges. All that said, it is a challenge and I welcome any improvements!

  2. Expert Member
    Michael Maines | | #2

    Nice details, and excellent video, Josh! I've never seen black aggregate like you have for your screed layer--must be the volcanic stone you have out there? Did you have to do anything special with the stem wall to make your engineer or inspector happy, such as using a wider footing and/or extra J-hooks?

    1. Expert Member
      Joshua Salinger | | #19

      Thanks, Michael!

      The 1/4 minus gravel is just that color locally. It's not volcanic, but when it is damp (as it was when it was delivered) it gets that dark color. When it is dry it turns a grey, concrete-looking color. I think the dark, winter day and low light when the video was shot is adding to this effect.

      We didn't have to do anything outside of normal for the stem wall and footing. Nor did it require any extra hold downs. We designed this particular home to have no interior bearing points, so all the load is on the perimeter. The engineer didn't have anything to say about it...

  3. Jon_R | | #3

    It's a nice how-to video, but I'd like to see more about "is it a good idea?". How does cost/CO2e compare to the various alternatives?

    Is differential movement between floor and walls a concern? Over time, I'd expect the area under the floor to dry more than the area under the footings. Meaning the floor might sink a bit in relation to the walls.

    > reducing concrete is a good thing

    But at $50/ton of CO2e, a home concrete slab creates about $500 of CO2. If a non-concrete floor costs more than $+500, then there are many ways to save substantially more CO2e for less $. This would be an even better thing.

    1. charlie_sullivan | | #4

      "there are many ways to save more CO2e for substantially less $."

      Good point, but for the most part, I think we have to do those also, rather than instead.

    2. maine_tyler | | #6

      Is the sector of the economy irrelevant? I.e. if the increased cost of this reduced concrete technique is due to increased labor rather than materials/energy inputs, does that matter from an environmental standpoint?

    3. Expert Member
      Joshua Salinger | | #21


      I would certainly welcome a CO2e analysis on this compared to concrete slabs. If you have the time, I am all ears!

      The plywood is floating. It is not connected to the perimeter walls where the footings are. It is my assumption that any movement be it seasonal, drying of initial construction moisture, live loads, etc. will expand/contract and shift as needed. There is a roughly 1" gap between the stem walls and the floating floor to account for this. As with most things, I didn't come up with this idea myself; rather I first heard about it in a 2005 article by Andy Engel called 'The No-Mold Finished Basement' (FHB 169) and it was elaborated upon by Steve Demetrick in the November 2006 article in JLC titled 'A Basement Floor Without Concrete'. In the latter case, Steve talked about this in a B.S. and Beer show and tells about how he was able to visit this project and see that there was no movement, damage or otherwise roughly 15 years later. At any rate, it is a new idea and it is good to think all of these things through.

      In regards to the $500 of equivalent CO2e savings, I will agree with Charlie below and quote him saying that 'I think we have to do those also, rather than instead'. Also, here in Portland, OR the cost of concrete is really high and at the time we built this we were actually able to save a few thousand dollars by going this route in lieu of concrete. Nowadays with the price of plywood being what it is, that calculation may have changed, but it is still more cost effective for us. This is likely different in different areas of the country.

      1. Jon_R | | #27

        Great to see your responses (many authors don't).

        > 2005 article by Andy Engel called 'The No-Mold Finished Basement' (FHB 169)

        Interesting, it suggests no poly and then vapor permeable flooring to allow upward drying. It also warns about flooding.

    4. user-3403605 | | #43

      I had talked to Marc Rosenbaum the other year in regards to this question of comparing the embodied carbon of the assemblies - concrete vs. wood. Here is the response he sent me:

      "The thing I’ve been learning with concrete is that its EC is so variable depending on the psi (how much Portland cement) and the amount of replacement material (fly ash, blast furnace slag).
      David White did some calcs in a back-and-forth we had that had concrete at 231 - 318 kgCO2e per yd3, I think without replacement material. Range for a 4” slab was 2.8 - 3.9 kgCO2e/sf. 23/32” Advantech is listed at 0.74 kgCO2e/sf (if I did conversion right) so two layers is 1.48 kg CO2e/sf. Plywood is lower - but, one thing I see is that often industry-wide EPDs have lower values than specific manufacturer EPD. Advantech is higher EC than industry-wide OSB EPD. It’s a better product with better adhesive.

      So it’s useful to evaluate assemblies vs. just materials. And understand, the way we do with energy, where you get bang/$ - we do a whole building energy model and look to see which components can be changed to the most effect, etc. We don’t yet have that command of EC IMO."

      1. Jon_R | | #44

        Absolutely, the optimization tools need updating. Barring other considerations (comfort, looks, etc) do everything below some $/ton and don't do things above this.

        I came up with 400 lbs of carbon per cubic yard for typical slab concrete.

  4. charlie_sullivan | | #5

    I like the method and I like the little practical tips on how to do it.

    One question: is 1-3/8 really short enough? It seems like if you accidentally bury the head a bit, combined with tolerances in the plywood, that's actually nominally 23/32 rather than 3/4, you could end up with a puncture.

    Minor note: I'm pretty sure I saw a sticker that said the polyethylene is 10 mil, not the 6 mil stated in the narration. (For anyone outside the US trying to make sense of that, 10 mil is about 0.25 mm.)

    1. Trevor_Lambert | | #14

      I was thinking the same thing. I definitely would have gone with 1.25" screws, even if the plywood is a full 3/4. 1/8" margin for overdriving a screw is too close for comfort. If it was shy 1/32, then that leaves only 1/16 margin of error.

      I didn't notice the marking in the PE, but it didn't seem to behave like I'd expect 6mil to behave, seemed at least as rigid as the Super8 stuff I've handled in the past.

    2. Expert Member
      Joshua Salinger | | #22


      Good eye. I mistakenly said '6 mil' on the video when in fact the vapor barrier is indeed 10 mil. I was coming up with this extemporaneously in front of the camera... I notice all these errors all the time when I see myself recorded. Hopefully the larger content supersedes the mistakes!

      In regards to the screw sizing, I'm not going to die on this hill. I suppose it is a balance between getting enough grip with the screws vs. the risk of puncturing. We tested the plywood we used with some scraps and various screw types and we didn't see any of the 1 3/8" screws come through. Of course if one really buried the head that would matter and I also suppose that over the course of the entire floor this may have happened. If one feels better with a 1 1/4" screw, then by all means go for it!

      1. Expert Member
        Michael Maines | | #25

        Regarding screw length--a few holes into a vapor retarder isn't going to hurt anything, and the two layers of plywood should be airtight on its own, so I don't see an issue if they do punch through. Screws often don't "bite" well with only 1/2" embedment anyway. Probably more of an issue with extra-dense Advantech than with the traditional plywood that Josh is using.

  5. maine_tyler | | #7

    It's interesting that with a new construction slab, the order is: gravel, insulation, poly, concrete. With this alternative, it is similarly: gravel, insulation, poly, (wood) floor.

    But the retrofit version of this (floating a floor over an existing concrete slab) often cites: concrete, poly, insulation, (wood) floor. It makes me think perhaps the poly should go above the insulation in this situation as well. Or maybe it truly just doesn't matter...

    1. Expert Member
      MALCOLM TAYLOR | | #8


      The reason cited for placing the poly over the insulation in new slab construction is to prevent bleed water pooling below the foam where it can take a long time to dry. I've never been particularly comfortable leaving the insulation exposed to sub-grade moisture, partly because I've seen how much water the foam in hot tub lids can take on over time. If the slab bleed water is removed from the equation I'd vote for placing the poly under the foam, although as you say, maybe it just doesn't matter.

      1. maine_tyler | | #9

        Good points.

        I'd be interested to hear the reasoning from those that have done this type of assembly for the order they chose. It's not consistent, which is perhaps anecdotal evidence that it's not an issue either way. One such assembly — — places the poly between layers of foam.

        Perhaps the reasoning revolves more around sequencing, materials workability/protection, or is a flip of a coin.

        The only possible issue I can cook up with the foam over the poly is vapor condensing on the colder poly. I don't have any reason to think this IS an issue, only that it seems remotely plausible. Perhaps a complete non-issue.

        1. Expert Member
          MALCOLM TAYLOR | | #11


          Michael Maines pointed out in a another discussion that the poly above helps keep the foam from floating during the pour.

          Part of me is a bit wary about my ability to level the substrate well enough to get the plywood subfloor completely flat, but then the finished slabs I get done are never that flat either. I had a client who was in a high performance wheelchair who found it would move around on it's own following the dips in their slab. I guess it isn't too important on most projects.

          1. maine_tyler | | #13

            To be clear, I'm not suggesting the order 'might not matter' when a concrete pour is involved, which has the issues you mention. I'm just talking about when there's no concrete or the concrete is already poured and below.

            flooding is a good point.

      2. Jon_R | | #10

        > leaving the insulation exposed to sub-grade moisture,

        While there are lots of opinions about which is better, AFAIK, there is no published data of foam R value in a side-by-side test of foam/poly/floor vs poly/foam/floor.

        > bleed water pooling below the foam

        I suspect that a flooding event would similarly trap water.

        1. Expert Member
          MALCOLM TAYLOR | | #12


          "I suspect that a flooding event would similarly trap water."

          True - and probably quite a bit!

          1. Jon_R | | #15

            Say the flooding event was a spill, pipe leak or toilet overflow (all three have happened to me). With poly below and perhaps a low perm finish flooring above, isn't the plywood going to rot?

            Or say that there are never such problems (unlikely). If the finish flooring is lower perms that 6 mil poly, doesn't the plywood slowly accumulate moisture from below?

            With this, possible differential movement and the ability to run ducts, perhaps a crawlspace is the superior no concrete floor? I know it recovers reasonably from flooding.

    2. Expert Member
      Joshua Salinger | | #23


      Good observation. The layers are exactly the same. The reason being is that there are 5 layers to any successful slab on grade: 1. Stable soils/substrate; 2. Capillary break and soil gas depressurization field (usually 3/4 gravel with no fines); 3. Insulation layer; 4. Vapor control layer (a Class 1 vapor barrier); 5. The finished slab (be it concrete, wood or otherwise) that people and items stand on.

      For retrofits, one is better served by skipping the poly and instead going with a liquid applied vapor barrier such as an epoxy coating. Small air gaps between the poly and the concrete will eventually equal out to 100% RH and this can cause mold or if there are minor holes, evaporation which can transport moisture and damage flooring. The insulation goes over the liquid applied coating (usually 2 coats). Then float the floor. Dr Joe Lstiburek has a great article on exactly this at Building Science Corporation. So does his Daughter Christine Williamson at her IG account, Building Science Fight Club. They are much more articulate about this than I am and I encourage anyone to check these resources out.

      Another reason why the insulation should go below the poly (besides the really good points Michael brings up with the floating of the insulation and Malcom's point about the bleed water) is that foam insulation will get saturated if exposed to liquid phase water. Damp insulation begets critters and they will damage the insulation over time. Better to keep it dry.

      1. maine_tyler | | #26

        Thanks Josh.

        I've read Joe's take on the fluid applied VB vs poly over concrete.

        But what's the fundamental difference between that (retrofit) situation and this situation, where there are likewise interstitial air pockets beneath the poly and within the aggregate layer? Pockets that could also presumably reach 100% RH and then 'communicate' with the interior portion of the floor assembly.

        Is it that the aggregate layer in your assembly is assumed to stay drier than an old existing concrete slab? Or that the air pockets between a poly VB and concrete are much smaller and therefore more easily 'saturated' than the air pockets beneath your insulation and poly? Or that the mechanism of communication is somehow not present?

        I've always felt there was a bit of hand waving used by those trying to defend this logic, including by Joe himself (frankly, I think Joe is infamous for hand waving). Maybe I should read his daughters take and it will clear it up.

        >"Another reason why the insulation should go below the poly (besides the really good points Michael brings up with the floating of the insulation and Malcom's point about the bleed water) is that foam insulation will get saturated if exposed to liquid phase water."

        I assume you are referring to when a fresh concrete pour is involved here? And the liquid water in discussion would be the water in the concrete? So what about Malcolm's concern that putting the poly VB OVER the insulation puts the insulation in contact with ground moisture? Is the assumption again that the free-draining aggregate layer is simply not that moist?

        Thanks for the great video and discussion. Cheers.

        1. maine_tyler | | #35

          "But what's the fundamental difference between that (retrofit) situation and this situation, where there are likewise interstitial air pockets beneath the poly and within the aggregate layer? "

          No-one seems to want to take a crack at this? Is it that we lack the knowledge or is the answer so obvious others don't want to take the time to explain? I fully admit it could be the later.

          1. Jon_R | | #42

            Don't know, but this from Huber supports there being a potential issue with concrete slabs:

            "The polyethylene sheet should be glued to the slab with an adhesive that will bond to both concrete and plastic. ... These laps should also be glued to ensure an adequate moisture seal. "

  6. charlie_sullivan | | #16

    Jon, that's a really good point, that crawlspaces should be a bigger part of the conversation when people discuss minimizing concrete.

    1. Jon_R | | #17

      As should "just pour a slab, donate $500 to the right cause and be proud of the fact that you have a rot proof and net zero carbon floor".

      1. Expert Member
        Michael Maines | | #20

        The primary reason I did my first concrete-free slab was to save money and time, because the owner wanted wood flooring. There are various approaches to installing wood flooring over concrete but when it's real wood the best is to cover the concrete with a full layer of plywood. The carbon savings are a bonus but I start the conversation by asking people if they want a wood floor or a low-cost floor. If they want low-cost, we go with a concrete slab. If they want wood, the concrete is simply not necessary.

        1. Jon_R | | #28

          Makes sense. Any comments on "my poly/plywood/finish-flooring floor just got soaked. What should I do?"

          1. creativedestruction | | #41

            Bust out the squeegee, dry towels and dehumidifier.

          2. Jon_R | | #45

            Stephen's experience suggests that you need to rip up the floor. Often even with AdvanTech. More experience/data would be useful.

    2. Expert Member
      Joshua Salinger | | #24

      I agree that crawlspaces are a good option in lieu of this. They aren't always the best option in all circumstances, though. For example, if one is trying to create universal access for folks with disabilities, one needs stairs or a ramp for a crawlspace. Or if excavation is difficult or expensive a slab may be a better option. Also, the only crawl that really works in all climate zones is a conditioned crawl and these aren't cheap or easy to design and build correctly. For those situations where a crawl isn't the best option, I think this can be a good alternative to consider.

  7. user-3403605 | | #29

    Great video and explanation of a really simple system that passes all the BS sniff tests when applied to an appropriate project! And thank you for citing the JLC article written by Steve Baczek and myself about my design/adaptation of this system based on Andy Engal's FH article. There are a lot of good questions in the responses above that I will try to assist with answering as well. I will address the biggest mental hurdle that people have with this system which is the resiliency in the face of flooding. Rightfully so we are confident in our concrete slabs as the indestructible base of our buildings that protects us from the ground and is still faithfully there when the flood waters are mitigated. The installation we did for the JLC article was flooded one year into the completion of the project. The water main in the house burst while the owner was away and flooded the basement with 8" of water. I"ll cut to the chase. The remediation process was the same as any basement flood in a finished basement and the "slab" (2 layers of 3/4" advantech in our assembly- in your video you use fir plywood which would water would affect differently) held up fine. Everything was dried out and replace in kind. The biggest pain was taking out the bottom 2' of plaster and removing and replacing the dense pack fiberglass insulation. This experience gives me confidence that the system has the ability, when properly installed in an appropriate project, to be a smart option to choose with confidence. When I say appropriate project, I am referring to the fact that all assemblies need to be evaluated as an option based on the existing site conditions. This system survived a freak flood in an otherwise forever dry basement, but I would not install it in a basement that has a super high water table that is prone to flooding (just like I wouldn't advise a customer to build a basement in a location with a super high water table). For the same reason, I will change from EPS under slab insulation to roxul insulation if there is a high water table. I could go on...One comment I do have is in regards to the spacing of the plywood layers. All panel products have instructions on them to space them on the edges, which nobody does because it would throw off our 16" on center framing. With this floor, because it isn't attached to framing, we used shim between all of the panels in both directions to allow for expansion and contraction. -Steve

    1. Jon_R | | #30

      > Everything was dried out and replace in kind

      You replaced the advantech, etc or the floor just dried out without damage, similar to what tile over concrete does?

      1. user-3403605 | | #31

        Good question - I wasn't clear with what I wrote. The remediation/repair to this flood had no more impact on the advantech than it would have on a concrete subfloor. The water was drained, the carpets and hardwood floors had to be removed, the baseboards, the first 2' of blueboard/plaster and fiberglass insulation were removed. Dehumidifiers were set up and run for a week or so to finish drying out the space and then we installed new insulation, sheetrock, installed new wood flooring and carpets and painted. The tile floors in the bathrooms that flooded were not impacted by the water at all.

        1. Expert Member
          MALCOLM TAYLOR | | #33

          I think GBA should run a contest to find the one individual in North America who hasn't had to remediate their finished basement after water damage, and give them a free membership.

          Just a heads-up to everyone: When I become world dictator, finished basements will be outlawed.

          1. user-3403605 | | #34

            I agree that finished basements can be a liability in some places. The installation in my project was a walk-out lower level with concrete foundation wall on only 3 sides of the structure. By no means was it a hole in the ground with finished living space. If this was a slab on grade home with a plywood subfloor, the damage would have been the same due to the fact that the insulation/vapor barrier system, air tight construction and triple gasket European entry doors are what held all the water in. In our flood, the water was lapping at the glass above the bottom rail of the exterior door of the walk-out! Air tight = water tight

          2. Jon_R | | #36

            My experience with minor flooding and remediation (all pretty much successful):

            a) concrete slab with carpet and rugs - removed water with wet dry vacuum and ran dehumidifiers

            b) concrete slab with tile - removed water with wet dry vacuum

            c) bare concrete slab - removed water with wet dry vacuum

            d) floor containing plywood above basement/crawlspace - removed water with wet dry vacuum and ran dehumidifiers below

            Contrast these simple tasks to the higher-carbon-and-$-than-anticipated e) "rip up and replace" that is apparently often necessary with a floor with wood over impermeable poly (whether there is concrete in the assembly or not).

            My conclusion: if there is wood in my floor, I want high perms (and drainage) to airspace below it. If not, AdvanTech *sometimes* avoids post-flood floor replacement (per Stephen).

          3. Expert Member
            MALCOLM TAYLOR | | #40


            Walk-out basements need another name. They share so few of the attributes or problems of their buried cousins.

        2. Jon_R | | #37

          > The tile floors in the bathrooms that flooded were not impacted by the water at all.

          No water made it to the plywood below the tile? I'd expect significant horizontal movement of flood water into this area. And little ability for this plywood to dry in either direction.

          1. Expert Member
            Joshua Salinger | | #38


            >My conclusion: if there is wood in my floor, I want high perms to airspace below it.

            High perms to the airspace below it is great so long as there is plenty of air space and that air space can act as a place for vapor phase water to dry to. This would be the case in something like a conditioned crawl, for instance. With an on grade application, you really want an impermeable surface below the slab, be it wood or concrete. The airspace above it (the conditioned airspace) will act as the space for any vapor phase water to dry to. it will also keep vapor phase water from getting to the plywood in the first place.

            I have a separate but related question with tile floors over the floating plywood slab. We are using a decoupling membrane below the tile as we know the wood slab will have movement. The decoupling membrane is attached with a unmodified thinset to the plywood. The decoupling membranes (Schluter Ditra in this example) is a class 1 vapor barrier. If there is enough space for any water to dry to horizontally to adjacent areas I feel like all is well. The question is this: At what point is too much of the plywood covered with an impermeable membrane so that it can't dry to adjacent areas? I'm not sure there is a calculation to be done here, so I'm guessing its a judgement call and a determination of risk. If one was to cover the whole plywood slab with a decoupling membrane (or any impermeable flooring such as linoleum for that matter) and there was a water event, then it seems clear to me that this would be risky (vapor sandwich, anyone?). Just doing a section of floor the size of a bathroom seems fine to me. I'd love a discussion on this if anyone is up to it...

          2. user-3403605 | | #39

            Therin lies the benefit of advantech vs cdx or for plywood as the choice of material for this system. The advantech has a high enough wax context to resist the absorption of water and the chip sized composition doesn’t allow the water to flow through the sheet horizontally like a traditional veneer plywood material. We did measure the moisture context of the subfloor right next to the tile before moving forward with the remodeling and it was the same as the rest of the floor.

    2. Expert Member
      Joshua Salinger | | #32

      Thanks, Steve! I appreciate you jumping in here. Your point about finding the 'appropriate' project is an important point. These aren't for every situation... Also, thanks for jumping on the 'what if it floods...' question.

  8. truthbox | | #46

    Great video. Thanks for sharing this work. Any thoughts on the comfort of this floor underfoot compared to a conventional concrete slab-on-grade and framed wood floors? I'm guessing it feels more forgiving than the concrete but maybe not as forgiving as a framed floor? Thanks!

    1. Expert Member
      Joshua Salinger | | #47

      Thanks, Peter.

      We have used floating floors, tile and carpet on this assembly. The floating floors and carpet are quite comfortable as the floor does have a nice 'give' to it. I would compare it to a framed floor-- certainly more comfortable than a concrete slab.

  9. davec1808 | | #48

    Thanks for the excellent video Josh. I live in a cold climate where frost depth is 4 feet, many go 5 to be safe. When I brought this video up to a friend, he suggested that he'd rather have the concrete floor as extra support against the lateral forces against the basement wall. He implied the forces pushing against the basement wall could move the foundation without the concrete slab distributing the lateral forces. What do you think? Old homes with stone foundations and dirt floor seem fine...

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |