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

Please tell me why I have to put so much insulation / outsulation at the roof/celing?

kloopster | Posted in Energy Efficiency and Durability on

This may be a stupid question………. but, I am building a pretty passiv/aggresive house in Zone 3 and my mandated (code) R-values are as follows:

Ceiling R-value 30
Wood Frame Wall R-value 13
Mass Wall R-value i 5/8
Floor R-value 19
Basement Wall R-value c 5/13f
Slab R-value d, Depth 0
Crawlspace Wall R-value c 5/13
Fenestration U-Factor b 0.50j
Skylight U-Factor b 0.65
Glazed fenestration SHGC b, e 0.30

My wall profile has outsulation (outside in) ~ rain screen, 2 inch foam, Emerald Coat WRB, Zip wall, and insulation (rock wool), 1/2 gyp board, either 2×4 or 2×6 OVE or 2×6 balloon framing. My 2:12 or 3:12 roof/ceiling is mandated to have R-30? If I fully condition the attic space using this profile outside in – white galvalume (rain screen), furring strips, staggered layered foam, Emerald Coat WRB, Zip roof panels, (maybe some interior rock wool, maybe), on 24 on center I joists or dimensional lumber, then why from a physics standpoint does the roof/ceiling need so much more insulation that the wall assembly that it so closely replicates?

Also the same question for the floor assembly? I always use open webs at 24″ OC and actually condition the space, so why is the floor so much different than the wall?

Inquiring minds want to know.

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. GBA Editor
    Martin Holladay | | #1

    Heat flows in all directions, as you know. It's not true that "heat rises" -- only that "hot air rises" due to convection loops and the stack effect.

    It would be great if it was easy to insulate walls as well as we do ceilings, attic floors, or roof assemblies. But it's usually easier (and more economical) to add R-value to ceilings than to walls, so most building codes don't require as high an R-value for walls as for ceilings. The analysis is based on the cost of the construction methods required to implement various insulation strategies -- not due to building science.

    That said, you should follow code minimum requirements, or exceed them. You won't regret it.

  2. kloopster | | #2

    Thanks for the quick response Martin. I understand that warmer air moves in relation to cold air via buoyancy and leaky houses exhibit a phenomenon called Stack Effect, but does a super-tight structure with all internal and interstitial spaces fully conditioned have as much? If one wraps the thermal envelope "around the house" how would it be less easy to insulate roof assemblies than walls? If the only reason that r-30 is specified is to "make up for" normal problems with and unconditioned attic to ceiling intersection then this seems very "rule of thumbish" to me.

    With a air tight, sealed, exterior insulated assembly in which the "attic" is fully conditioned I still don't understand why I should spend extra money (and space) for unnecessary insulation? Why then do we not insulate r-30 between floors if "hot air rises"?

  3. GBA Editor
    Martin Holladay | | #3

    As I wrote in my first response, heat flows in all directions, as you know. So you definitely want your walls and floors to be well insulated. The "hot air rises" phenomenon is a red herring when it comes to answering this question, as you know -- so we are in agreement on that point.

    There are several reasons why it is usually more expensive to create a high-R wall than a high-R ceiling. Most ceilings are insulated by adding cellulose or fiberglass to the attic floor -- and it doesn't cost much more to upgrade from 12 inches to 16 inches.

    With a wall, however, going higher than R-21 (nominal) -- a common way to refer to fluffy insulation products that are 5 1/2 inches thick and designed for 2x6 studs -- gets tricky. You need either exterior rigid insulation, or you need to go to a double-stud wall. That costs more than just blowing a little bit more cellulose on the attic floor.

    Building codes reflect this economic reality. When a wall assembly is expensive, the cost-effectiveness of the building method (in terms of energy savings) is harder to justify.

    For builders following the PERSIST method, or builders who plan to install rigid foam on the exterior side of their roof sheathing, the normal economic comparisons differ from those made by builders who install cellulose on the floor of an attic. You are nevertheless compelled to follow minimum code requirements. If you find the cost of roof insulation too burdensome using rigid foam, you are always free to design your house so that you can install a less expensive insulation like cellulose.

  4. kloopster | | #4

    Martin, thanks, you reinforced what I had supposed. Again the code is a little "behind", I think I will use some exterior foam and just use rock wool batts internally up to r-30, then drywall. It just feels "dirty" however to acquiesce to code once again, when it is modeled on non-advanced, high performance home building, and/or building science.

  5. GBA Editor
    Martin Holladay | | #5

    I disagree with you. I have explained the logic of the code, which considers cost-effectiveness (including the cost of common wall and ceiling assemblies) when determining minimum insulation levels.

    I don't think that your desire to install less than code-minimum levels of roof insulation has anything to do with the code being "behind." It just seems to me that you have chosen an expensive type of insulation, and you want to install less than the code requires.

  6. Expert Member
    Dana Dorsett | | #6

    If you use R30 between rafters or joists in combination with exterior side foam, the exterior foam needs to be at least R5 to meet code without a "smart" interior side vapor retarder:

    In a zone 3 climate R30 isn't exactly an "optimal" R-value for the ceiling from a longer term cost-effectiveness point of view. If your local codes were up to date (IRC 2012) for zone 3 you'd be looking at R38 (between joists, not whole-assembly), rather than R30 for a code-min ceiling, which has a very easy financial rationale at recent years' energy costs.

    There is a reasonable argument for R50 (whole-assembly, with the thermal bridging of the joists factored in) for a vented attic, or R45 (whole-assembly) for a cathedralized ceiling. See Table 2, p.10 of this document (but read the whole first chapter):

    Of course how you get there matters. R50 pencils out a lot easier with open blown cellulose (at an installed cost of less than 5 cents per R per square foot), than open cell foam (at 12-13 cents R-foot), or closed cell foam (at 17-18 cents/R-ft^2.)

  7. kloopster | | #7

    I don't really give a hang about how the code is derived. What I worry about is the enclosure, I was just curious if there was a real reason to change the envelope, just because, I am moving from wall to roof. I suspect that the r-30 is a derivative of a "normal" building enclosure design, with, I suspect a dark shingled roof assembly. I have no choice but to comply (actually they don't enforce energy codes in my county) but that doesn't mean that I shouldn't question it. They used to have a r-value for a cathedral ceiling which was in-between the new wall and ceiling r values but they have omitted it as of late. Wouldn't a ceiling assembly (against the rafters) in a fully-conditioned attic qualify as a cathedral assembly?

  8. kloopster | | #8

    you are right, cellulose is cheap!!!!!!! here is the ORNL calculation: I will comply, grudgingly, thanks for the tip on the cellulose (although I really like the rock wool!).

    Insulation Location R-Value* Notes
    Attic 38 -
    Cathedral ceiling 38 -
    Floor 25 Over unheated, uninsulated space.
    Wall cavity 15 -
    OVE wall cavity 21

  9. kloopster | | #9

    Thanks Dana. I will use dense pack cellulose underneath (minimum r-33) and 2 x 1 xps staggered on the exterior for a r-43 foof assembly. But again, just because code says so, doesn't make it realistic in terms of science. The assembly I am building is as cheap, or cheaper (and more high performance) than a normal shingled roof. I read the build America document, which was well written and the financial metrics are certainly there for a "normal roof assembly", but they aren't using a cool roof assembly in that modeling (read page 11 under "roofs").

    I will investigate the ORNL site and the work they have done on cool roofs as i suspect that "heat rises" is only half of this puzzle, the other half is "the sun shines". here is the calculator:

    after all, shouldn't r-value of any structure tale into account all heat gain and loss from all directions and be at least calculatable ( i know that is not a word) as opposed to one-size-fits-all?

    I think all of this questioning of "just do what code says", has to do with my scientific training and I suspect, a unpleasant personality disorder.

  10. GBA Editor
    Martin Holladay | | #10

    I have no problem with anyone who questions to logic of building code requirements. You are quite right that many building code requirements are illogical or unsupported by building science.

    There is no "one size fits all" answer to the question of the cost-effectiveness of minimum insulation R-values -- even within a single climate zone.

    There are many reasons why these requirements end up being simply "the best we can manage under the circumstances":

    1. Fuel costs vary. A homeowner who heats with electric resistance heaters is more likely to invest in thick insulation than one who heats with natural gas.

    2. Insulation costs vary. A homeowner who insulates with cellulose is more likely to invest in thick insulation than one who insulates with spray foam.

    3. Construction costs vary. Some builders have come up with inexpensive ways to build high-R walls -- while other builders charge an arm and a leg for the same specifications.

  11. charlie_sullivan | | #11

    Here's a try at a concise answer to "why from a physics standpoint does the roof/ceiling need so much more insulation that the wall assembly that it so closely replicates?"

    It doesn't. From a physics standpoint, the "right" answer would be to require R30 (or more) for the walls and the ceiling. But then the builders would all complain that making walls R30 is too expensive. So the code cuts the builders some slack on the walls.

  12. Expert Member
    Dana Dorsett | | #12

    Physics doesn't demand anything, nor does it have a standpoint. It simply is.

    Comfort demands some minimization of heat flow, as does economics. Comfort doesn't typically demand as much as the economics of reducing energy use, but that is highly variable with the current local (and presumed future) energy costs. At 50 cents/kwh R30 would be ridiculously low-R for heating & cooling with heat pumps in zone 3, but at 5 cents/kwh it's going to be too much. Over a presumed 50-75 year lifecycle it may never "pay off" in direct energy use savings.

    Builders don't complain about walls being too expensive, but home buyers have issues with affordability, which is usually dependent on financing. What too few home buyers (and their financers) miss is that the total cash flow of utilities + mortgage can be offset substantially by going higher-R than code min.

    The physics of a roof is not simply that of a tilted wall. From a thermal transfer point of view they are somewhat similar, but from a moisture transport point of view it is not. Walls get but a tiny fraction of the annual direct-wetting from rain/snow, and a much lower temperature cycling from solar gains. Walls convection cool much more quickly than roofs too (especially lower pitched roofs), due to their orientation, yet experience much less radiational cooling on clear nights. Even thermally the are not identical. In a zone 3 climate the cooling season solar gain through an R20 2x6 roof is substantial, whereas the solar gain through an R20 2x6 walls is almost negligible (much much smaller than the gains from windows or roof.)

    So, whatever.

    An R30 roof deals with the worst of the cooling season comfort issues of a high-gain roof, but won't hit the sweet spot from a heating & cooling energy bang/buck point of view (except for those only planning to live there less than a decade), even though R30 walls will probably be overkill on lifecycle financial grounds unless energy prices soar over the next 5 decades.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |