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

Community and Q&A

About those long screws clamping 2x4s down through layers of ceiling foam

user941025 | Posted in General Questions on

I don’t like it.

When those pictures of Dr. Joe’s barn roof went up in another thread, I thought once again of the foam layers in my yard and wondered, once again, when I was going to start to see a problem.

For ceiling layers of polyiso, you’ve got all those staggered and taped layers of foam, through which you send long screws through to the rafters, to clamp 2x4s down tight. I hated sending those suckers through–what kind of problems are going to crop up? On this one, that’s 45 long metal passages for heat / moisture / air / water to pass, especially when you miss a couple of times….

So, as of last night, I do see a couple of water leaks. When the roof over the top of this ceiling foam is intact, there shouldn’t be any quantity of water hitting the top surface of this foam to leak on through, but in the meantime, that water has shown me exactly where the biggest problem spots are. I intend to open those couple of holes bigger from the inside, drain and dry them, and jam a sprayfoam can straw up in there to fill `em up.

Even so, I really hate this “long screws all the way through the ceiling foam” thing. Surely for this approach there typically ought to be an outer layer glued down over the top of these strips to cover the penetrations. It seems asinine to me.


GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. homedesign | | #1

    Jon, My thoughts.....
    Screw the Ginourmous Screws and the Foam

    You should have built your super-insulated Shed Experiment with....
    * A Riversong-Like "ADA plus Goo as you Go" Pressure Boundary on the warm side...
    (or a Lucas Durand plywood,gasket and tape Pressure Boundary on the warm side)
    * a "buildable" double stud wall with densepack cellulose
    * an exterior plywood wind PRESSURE boundary

  2. wjrobinson | | #2

    John is pretty much right on.

  3. user941025 | | #3

    Those all sound pretty much in line with how I'd plan to do this another time around. Agreed.

    That said, here I am with this one, and am now left needing to solve it.

    We got up on the ladder this morning and looked under the roof. It doesn't look like I'll be able to easily get underneath the low end of the roof to back out those 14" screws / foam the holes with straw inserted from exterior side and resend. That might have been effective. I might try to head up there with a ratchet tomorrow and see what I can do.

    Even so, best case scenario, that may solve these water leaks which should be a temporary problem, ideally redressed when the roof is finished, but *would not* solve the heat loss of these penetrations (as gluing foam strips between and foam boards above the furring strips probably would). Unfortunately, the space looks just too tight to get in there to add those layers on top.

    I am still steamed about sending spikes and holes through my carefully lapped and taped four layers of foam.

  4. GBA Editor
    Martin Holladay | | #4

    I can't visualize your problem yet.

    This is ceiling foam, not roof foam. So that means you installed the rigid foam from the interior onto a ceiling.

    Flat ceiling or sloped ceiling?

    Is there an attic above the ceiling or not?

    You don't have a roof leak, so I assume you have condensation. Is that what you are telling us?

  5. user941025 | | #5

    The roof is sloped 1/12, and is clear span above these lapped/taped/and-then-pierced four layers of foam--comprising 11" of polyiso--that sit on top of the ceiling joists.

    Indeed I do have a roof leak, only for the time being. The roof is currently only plywood, yet to be tarped or taped or anything.

    That roof leak will be solved soon enough, but the fact of the ceiling leaks will remain (by which I mean leaks both inbound and outbound). Even if water isn't headed in once the roof leak is gone, carefully lapping and taping foam ... and then sending long screws through it in several dozen places ... is a stupid design.

  6. GBA Editor
    Martin Holladay | | #6

    OK, you have no roofing on your building, and so you have a roof leak. No surprise there.

    The roof leak made your ceiling wet. No surprise there.

    I'm still not sure I understand your point about the screws.

  7. user941025 | | #7

    You are welcome to not understand it if that's what you prefer to do. However, you certainly seem capable of understanding it, if you prefer to do that instead.

  8. user941025 | | #8

    OK, you have no roofing on your building, and so you have a roof leak. No surprise there.


    The roof leak made your ceiling wet. No surprise there.

    * The roof leak made the top of my ceiling foam wet.
    * The ceiling leaks, produced by sending long screws through the ceiling foam, allowed the interior of the space to become wet.
    * And even when the top of the ceiling foam is no longer prone to becoming wet, those same ceiling leaks will continue to do what ceiling leaks do, unless those leaks are solved effectively.

    That is why I have a strong opinion about the screws.

  9. GBA Editor
    Martin Holladay | | #9

    I'm sorry if I seem slow. Now I get what you are saying.

    However, I can't think of any type of insulation that is designed to be waterproof (except, of course, for specialty roofing foams -- dense closed-cell spray foams installed as roofing). Cellulose is no better at keeping out water than your rigid foam. When your roof leaks, your ceiling gets wet.

    I suppose it's possible to put a secondary roof on top of your insulation layer if you want to.

  10. benalbright | | #10

    I don't think his concern is water leaks after the roofing material is installed, his concern is air/vapor leaks in and out what should otherwise be a well insulating cathedral type ceiling. Everywhere he has screwed through is now a very small breach in the thermal envelope. Honestly though, how much space are you really talking about here that can't be addressed and how much of an impact would that really cause? Seems no different than the same effect of adding sheet insulation on the exterior of the roof deck and screwing down into that.

    I think this is a case where a picture would do wonders for understanding how this is constructed as it took me a few times of reading the posting to understand what was going on.

  11. GBA Editor
    Martin Holladay | | #11

    There are several possible concerns, and I'm not sure whether Minneapolis is worried that the foam layer isn't waterproof, isn't airtight, or includes cold metal fasteners that may become condensation sites. Perhaps all three concerns exist.

    I think it's possible for water to leak through the foam, even if the air leakage rate is very low.

  12. user941025 | | #12

    I'm at work, but here's a sketch. Hopefully the picture will show up properly.

    Sloped roof assembly floats over those four layers of foam that are clamped on the top/exterior side of ceiling. Screw goes through 2x4 sleeper and four layers of foam into joist.

    Yes, all the concerns typical to leaks from a conditioned room's ceiling to an unconditioned space above: air/water/moisture in and out, and indeed, possibly compounded by condensation/heat transfer along that long piece of metal.

  13. GBA Editor
    Martin Holladay | | #13

    Here's my take:

    1. It's unnecessary for your insulation layer to be waterproof.

    2. Airtightness can be verified with a blower door or a couple of window fans (you can feel for air leaks). If you taped the foam seams, I bet you are fine.

    3. If your screws hit the ceiling joists, I don't think you'll get any condensation.

  14. user941025 | | #14

    Man, I hope so.

    My concern is that a leak is a leak--especially a ceiling leak.

    I'll get the roof plywood seams taped in the meantime and keep an eye on it for a couple of months.

  15. GBA Editor
    Martin Holladay | | #15

    I'm a roofer. If someone told me they were going to use 4 layers of taped polyiso as roofing on a low-slope roof, I'd say, "You're nuts. That's not roofing."

    However, 4 layers of taped polyiso make good insulation -- and they probably make a decent air barrier (although I don't know where the air barrier is in your assembly).

    Just because it leaks doesn't mean there's a problem. It just means that the materials were never meant to be roofing.

  16. wjrobinson | | #16

    Strange bunch of posts.... Martin is right on. Put a roof on.


  17. user941025 | | #17

    That's cool, AJ, thanks for summing up. I don't want this thread to keep you from the stack of pancakes that issues you commands from the moons of Jupiter, buddy!

  18. wjrobinson | | #18

    No problem......but love Apple filled pancakes.

  19. user-659915 | | #19

    "Even so, I really hate this "long screws all the way through the ceiling foam" thing. Surely for this approach there typically ought to be an outer layer glued down over the top of these strips to cover the penetrations. It seems asinine to me."

    There are so many ways to insulate a ceiling. If this one seems 'asinine' and you 'hate' it I don't understand why you didn't choose another method. Nor do I understand complaining about your ceiling assembly leaking when you haven't yet put a roof over it - THAT's the outer layer you're missing to deal with water leakage from above. And if you're concerned about air leakage from below, there's easy ways to deal with that as well. If you haven't finished the job it makes no sense to complain about performance failures.

  20. toshiyano | | #20

    could you spray foam for the rafters for air tightness or wouldn't that work/is it to impossible to do that?

  21. user941025 | | #21

    Hi Satoshi--yes, I will be including that as part of my strategy.

    Agreed, James, that obviously a ceiling needs a roof over it (which is why there is roof framing built even if it's not finished, so apparently I realize it needs to be there) and that there are many ways to insulate a ceiling--and so in following one of those ways using what seems to be standard practice for the method, I find part of its detailing counterintuitive and counterproductive. And so I am interested exactly in those ways to solve penetrations, which you refer to but upon which you don't elaborate here, and preferably I'd say before the ADA level. My impression is that in fact perhaps it does indeed make sense to be concerned about weak points in a design, which are helpfully shown to me by the inbound water diagnostic tool.

  22. user-659915 | | #22

    MD - to elaborate:

    1. Put the roof on to prevent water moving through the assembly from above.

    2. Seal the joints between the ceiling joists and the foam board with a can of spray foam to prevent air moving through the assembly from below.

    The second step is probably unnecessary: as Martin points out if the screws have not missed the joists you likely have a pretty damn good air seal there anyway, but if you have any doubts just do a blower door test - that's a far better marker than water dribbling through an unfinished roof. Hint: your major air leaks are going to be at rim joists, not at screw penetrations. Foam board clamped to framing with battens and screws is a common component of many exhaustively field-tested high-performance airtight assemblies.

    And then apparently you're doing ADA as well. Belt, suspenders, and another set of suspenders. Good, if it makes you less grouchy.....

    Anything else?

  23. homedesign | | #23

    Mr Disaster said:
    My impression is that in fact perhaps it does indeed make sense to be concerned about weak points in a design, which are helpfully shown to me by the inbound water diagnostic tool.

    I think the liquid water "leak" is helping you to visualize the 3-D network of voids and flaws in the Pressure Boundary

  24. homedesign | | #24

    Where is your air control layer?
    Were you thinking that the multiple layers of staggered joint foam was the air control layer?

  25. user941025 | | #25

    Thanks, James. I like that answer.

  26. user941025 | | #26

    John, there's also 6mil poly to the outside of the plywood sheathing, lapped and sealed with acoustical sealant. Somewhere there's a thread (and featured post thing) about it

  27. user-788447 | | #27

    It is good to scrutinize assemblies that force a tight tolerance of execution. However based on some construction pics of a project I seen recently I still see some promise for the long screws through the roof in retrofit scenarios.

    Don't be too quick to extrapolate from a single DIY effort the feasibility of a construction detail. While I do think there are a lot of hack contractors you need to watch out for there are also good professionals that can build to much higher levels of tolerance than someone with little experience.

    Retrofitting buildings (especially those that are "wonky" due to poor design decisions and crazy additions) often require creative solutions. Overinsulating roofs with a 4-6" of rigid insulation while not ideal may be sometimes the most feasible and least risky measure.

    The project I am referring to above was a residence in the Twin Cities that experienced frequent ice dams do to roof geometry, insufficient insulation, and air leakage. A local contractor in conjunction with some building scientists at the UofM came up with an overinsulation solution (requiring those long screws) that also included continuous soffit to ridge venting. I wasn't on site and don't have all the details (including costs) but I was impressed by the pictures and since the roof has exhibited no snow melting.

    MD you are a good rabble rouser leading to some good discussions ; )

  28. homedesign | | #28

    MD ,
    Yeah...I forgot about the 6mil poly
    and thanks for sayin what you are thinkin (thinking out loud)

  29. user941025 | | #29

    Thanks, J.
    Like St. Thomas Aquinas, I prefer my "truths" in science and dearly-held-beliefs alike well-examined.
    Unlike Aquinas, I am an uneducated boorish monoglot who curses profligately.

  30. cornercanyon | | #30

    Is this what the finished assembly from outside to inside was planned to be-

    Roofing material, roofing underlayment (WRB), plywood/osb, 2x rafters, air gap, 2x sleepers screwed down through polyios (Thermal Barrier), 6mil poly (Air/Vapor Barrier), plywood/osb, 2x ceiling joists, drywall.

    If that is the assembly than to me the issue is the WRB location not being combined with the Air/Vapor Barrier, not the screws. Though I understand the roof 1:12 isn't on yet, I still must be missing some pieces.

  31. homedesign | | #31

    I think as a retrofit strategy that PERSIST (with a fully adhered membrane) may be a good option.
    I just think there are better ways to build new structures.
    I think Joe Lstiburek admitted in his Insight 056 that polyethylene was not a good choice for an air barrier....
    When he rebuilt .. he used a fully adhered membrane

  32. user-788447 | | #32

    MD, entertain me while I reframe the issue.

    I think some of the discussion in this thread implies the following:
    Liquid water penetrating through a nearly horizontal assembly, primarily along a vertical metallic column, illustrates paths of air (and therefore water vapor) exfiltration.

    I'm not convinced of this.

    Liquid water moving downward and air moving outward are due to different forces as defined by Physics. I believe that due to the bipolar atomic properties of water (or as RR would term 'magic') the forces acting on liquid H2O are significantly stronger than the differential air pressure difference possible in the context of your single room abode.

    H2O is moving under the influence of gravity, capillary action, and the sorptions. I bet the chemophysics between H2O and the metallic screw is significant as well. Air under a pressure difference doesn't follow these forces. I'm speculating that the liquid water doesn't even require air gaps (compromises in the air barrier) to penetrate the assembly.

    Therefore I don't think you need to be concerned with re-establishing an air barrier.

    Because of your sequence of construction and decision to locate a layer of poly within the layers of horizontal polyiso you may have trapped water (somtimes ice?) within this assembly. What is the consequence of this? Water may compromise the R value of polyiso but I don't think polyiso supports mold or breaks down due to water. If I store poly in my damp basement it will start to stink therefore I think poly may support mold. The water could percolate down to your plywood ceiling. Polyiso board sits directly on your plywood, if water made it to this transition could it dry before supporting mold?

    In the end I suspect you have a limited amount of water trapped and have since had proper tarping to stop continued water intrusion. I also suspect that only reoccurring wetting events over several years overcome drying potentials.

  33. user-788447 | | #33

    To add to my previous comment:

    The approach of the Twin Cities reroof was basically a PERSIST (or REMOTE I don't recall the difference) approach. Over the roof decking a self adhering rubberized asphalt membrane (Grace) was installed and over that the polyiso insulation was built up. In theory when the long screws pass through the rubberized asphalt membrane (the air barrier) the material has some capacity to 'self heal' about a screw penetration that may miss the rafters.

  34. user-788447 | | #34

    Like Steve Zissou, despite no-one believing me, I know a yet unseen large glow-in-the-dark polka-dot monster ate my closest friend and I won't cease until I'm reunited with it.
    Unlike Steve Zissou, I'm not fit to be a character living in the wonderful world spawned by Wes Anderson's imagination.

  35. homedesign | | #35

    Great Movie
    fantastic soundtrack too

  36. albertrooks | | #36


    2 things:

    "Unlike Aquinas, I am an uneducated boorish monoglot who curses profligately." ...Okay... Best quote of the week.

    Personally, I think roof outsulation is great and I like "warm roofs" where many prefer "cold roofs". They make great living spaces. Of course, being a "wood head" I prefer the euro wood fibre insulation board over foam. Mineral wool is a good option too.

    Regardless of the insulation type, at the final layer, consider a good membrane, taped with a good quality exterior tape that is recommended for the membrane AND... most importantly... put the battens on compressible nail sealing tape. It "self heals" beyond what a membrane can do. This way the long screws are OK and the roof will have a good long life.

    Pics are below and here is a video:

    MD- I've always wondered at the origin of your login. Obviously from your chosen quotes you have an outsized sense of humor. Good for you.

  37. homedesign | | #37

    Albert, you say you "like warm roofs"
    however, if you space your roof cladding away from the roof structure with battens (as the Euros almost always do) ...aren't you actually creating "cold roof cladding"?

    Something that really baffles me about "Joe Lstiburek's Barn" ... he depressurizes his siding(wall cladding)...but chooses not to depressurize the roof cladding.
    He breifly notes that the shingles failed in less than 15 years ...yet he does nothing to address the shingle issue when he re-builds the whole enchilada

  38. user-659915 | | #38


    As always you keep us in touch with the technologies that enable us to push the envelope in - well - pushing the envelope. I don't know whether I'd feel it was worth the trouble on top of a foam board material which is not going to be troubled by a little water from an occasional minor roof leak but certainly the membrane + nail seal tape would give a lot of comfort with a more vulnerable material like wood fiber as the thermal layer. And there's a lot of good reasons to be looking beyond the seductive petrochemical solutions to our insulation concerns. Let's hear it for the "wood heads"!

    Of course you'd omit the layer of poly BELOW the wood fiber, yes?

  39. albertrooks | | #39


    The "warm roof" term means to me:

    An insulated assembly rather than the good old fashion cold attic (for us in yankee climes). I like the feel of living in an upstairs room that has a high pitched ceiling. Add skylights and that is my house. Although... Since it's an early 90's timber frame with T&G + EPS, it does not have nearly the insulation or air sealing levels that I'd choose today. It needs a retro fit soon! At that point, "Disaster" will be answering the questions that I post.

  40. albertrooks | | #40


    Yes on two points:

    This was probably not the ideal assembally to illustrate use of nail sealing tape. I just brought it up since the penetration was a concern and there are probably others who have a similar application worry.

    The NS tape is a bit extreme- even to me. But, when one starts to worry about what happens around those "long screw" perforations, then why not make use of the thing designed for the application. But yes... I see it as redundant on a foam build-up that has nothing that would be harmed with small leaks around the screws.

    I would also leave out the poly and replace it with something that has a fixed perm rating just below 1. There are some pretty good membranes with a fixed perm raring around 0.7 for this application. I'd choose a fixed perm membrane for this application rather than a variable membrane since I'd be looking to consistently limit vapor travel into a roof assembly in all seasons. However, at 0.7 "a little bit of vapor" can pass in either direction.

    The foam makes the vapor question moot but I'm still going with a membrane becuase I've handled both and it's clear to me that these high quality membranes are much more durable than ploy. Durability is always a good thing.

  41. albertrooks | | #41


    "Something that really baffles me about "Joe Lstiburek's Barn" ... he depressurizes his siding(wall cladding)...but chooses not to depressurize the roof cladding.
    He breifly notes that the shingles failed in less than 15 years ...yet he does nothing to address the shingle issue when he re-builds the whole enchilada"

    I guess it shows that we all have a tolerance for assembally failure. Even those of us that are "wicked smart" on the subject.

    I think we have been slowly moving from traditional structures that were durable since there was limited insulation and air tightness - to well insulated air sealed structures... just in this last generation. There have been plenty of problems along the way. We are looking for the magic bullet of "durable - well insulated - air tight structures". But it's still so early and uncommon that we still take premature failure as something to be expected and tollerated. It's our own conditioned expectation.

    We'll get there...

  42. homedesign | | #42

    Albert: "We'll get there..."

    I think that many Europeans, Riversong, Chlupp and yourself are close to "getting there"

  43. user941025 | | #43

    Thanks, everybody. Wonderful reading.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |