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

Lstiburek’s Rules for Venting Roofs

You need an airtight ceiling, lots of air flow, plenty of soffit vents, and deep insulation at the attic perimeter


Building Science Fundamentals: Roof, Part 1: Ventilation

By Dr. Joseph Lstiburek

Dr. Joseph Lstiburek talks about the not-so-controversial ways to maximize the efficiency and airflow of your roof and attic.

Video Transcript:

There’s been so much stuff said about roofs that you sometimes lose perspective. I’m going to start off by saying what might seem controversial but really shouldn’t be. This is a vented attic, and it’s probably one of the most unappreciated building assemblies we have in the history of building science. It’s beautiful. It’s hard to screw this up. For 20% of the effort, it gets us to 80% of optimal performance, and it works in hot climates, in mixed climates, the Arctic, the Antarctic, the Amazonian rain forest — it works absolutely everywhere. The value proposition of a vented attic, meaning the money that you invest in building one of them — it’s hard to argue with the benefits. But for all kinds of reasons, we manage to screw it up. The single most important thing you have to remember about a vented attic is that the ceiling plane — the gypsum board layer, the drywall layer — needs to be airtight.

1) The ceiling plane MUST be airtight. Absolutely airtight. Above the airtight ceiling plane, the only thing that should be seen is insulation and air, nothing else. Not last year’s Christmas decorations, not your high school prom dress, not the tuxedo you were married in and can no longer fit in. Nothing but lots of insulation and air. Just an airtight ceiling and nothing else.

2) If you’re going to vent the roof, then VENT THE ROOF. If you’re actually going to vent the roof, let’s be serious about venting the roof. Wash the underside of the roof deck with air. That means the entire perimeter of the roof needs to have air inlets, meaning continuous soffit ventilation. It’s dumb to have baffles every third or fourth bay, the entire underside of the roof deck should be washed. Where the air leaves isn’t as important — whether it’s a ridge vent, or mushroom caps, or gables. What’s important is that you have continuous air entry at the perimeter of the roof down low.

3) Put more vents down low than up high. This is where the code tends to have it wrong. You want more entry points at the perimeter than exit points at the top. People say you want to balance the lower down ventilation with the upper ventilation, and a lot of people interpret the codes to say that if you get it unbalanced you want more ventilation up high. That is absolutely wrong; you don’t want more places for the air to get out than to get in. The reason is, if you construct a house with a leaky attic ceiling and you have lots of ridge vents or you have lots of vents up high, the makeup air is going to be pulled from the house rather than being pulled from the outside. That scenario is a disaster. Attics should be ventilated with air from the outside, not the inside. That’s why I hate these whirligig turbine vents — because they depressurize the attic, and if your attic ceiling isn’t perfectly airtight, you suck air conditioned air or heated air out of the house. It’s even crazier when the powered attic fans can actually suck on the roof and they’re controlled by a thermostat. How stupid is that? Of course the attic is going to be hot. You turn them on and they suck all the air conditioned air out. No powered attic ventilation; more vents down low than up high; wash the entire underside of the roof deck. But all of that is secondary to having the ceiling plane airtight. This last tip is more important in cold climates than anywhere else. But where the ceiling insulation hits the perimeter wall, you don’t want the amount of ceiling insulation on the top plate to ever be less than the R-value in the wall itself.

4) Put more insulation on top of the wall than inside it. In other words, if you have an R-20 wall, you want at least R-20 on your top plate. A higher R-value is better, but a lower R-value is not. If you have an R-30 wall you want at least R-30 on top of your top plate. A reasonable rule of thumb is: Thou shalt never, according to Joe’s Rule of Thumb, have less R-value on the top of your top plate than in the wall. It would be nice to have even more, but not less. Notice: nowhere in this discussion did the term “vapor barrier” come up. If you really want to have a vapor barrier in the ceiling, limit it to climate zone 6 or higher, but that’s really not important compared to the airtightness of that ceiling plane. The building code calls for a vapor retarder in climate zone 6 or higher. It’s okay to put one in, but if you don’t, take a Valium and relax. You don’t want to go through a lot of brain damage in a renovated house to try and add a vapor barrier underneath insulation in an attic. What you really want to do is make that ceiling plane airtight, make it airtight, declare victory and be done. Don’t mess around with permeability’s and calculations and whatever. To recap, airtightness on the ceiling; washing the underside of the roof deck; unbalanced ventilation should be in favor of the lower vents because you don’t want to depressurize the attic; no to powered attic ventilation or the whirligigs; and you don’t want to squeeze the insulation at the perimeter so it’s less than the R-value of the wall. That’s it. You can build that everywhere in the world and life is good.


  1. user-659915 | | #1

    Good to be reminded
    of the basics once in a while. And even more so of the underlying principles that go with them.

  2. homedesign | | #2

    Service Core & Airtight Sheathing
    You say that "the drywall layer needs to be airtight" ...why the drywall?
    Why not create a 4" to 12" deep service core for can lights, small ducts & wires and then install continuous airtight sheathing on top?

  3. user-723121 | | #3

    Vented roofs
    Dr. Joe,

    Thank you for a nice piece on venting roofs. I like your point about venting between each rafter and washing under the decking. Yea for energy heel trusses, this, along with air sealing will eliminate ice dams in new housing.

    In MN we get attic frost buildup in homes with poor roof venting and leaky ceiling planes, as soon as the first warm day comes, water drips through ceilings and folks think the roof is leaking. This is due to lack of venting and attic bypasses.

  4. user-659915 | | #4

    Service core & airtight sheathing
    John B's notion of a ceiling service core below the air barrier is a great one - future-proofs the installation by allowing easy electrical remodels below below the air-seal layer. Not sure about the truss though. Here's a close-to-standard stick-frame alternative which meets Joes' requirement for depth of insulation at the attic edge above the wall but which avoids the structural redundancy of the truss above the encapsulated ceiling joist. Just have to figure out how to tell the client they're going to to pay for a fully-sheathed attic floor but they can't store anything up there.

  5. 5C8rvfuWev | | #5

    In a renovation
    Joe L. mentions renovation, but I've wondered --

    if the original envelope lacks an exterior air barrier, does installing one (non-continuous obviously) at the attic (like John or James describe) create any problems? I can't see why it would but wonder if there are consequences I'm not understanding.

    Joe W

  6. homedesign | | #6

    I agree with the stick frame option
    James, I envisioned the concept being used with stick framed roofs as well.
    The cladding structure can be trusses or stick framed
    It is very similar to what Thorsten Chlupp is doing on the SunRise Home

  7. user-659915 | | #7

    Another one:
    A double or triple plate over the attic sheathing may be needed to lift the rafter high enough to clear a thick layer of insulated sheathing on the outside of the wall. Alternatively I suppose you could make the seat cut longer on the rafter notch. A baffle of board insulation above the air barrier will eliminate thermal bridging at the raised plate.

  8. homedesign | | #8

    offsetting the cost of the sheathing
    James, my thinking would be to ask the homeowner to pay for the sheathing instead of paying for sprayfoam.

    I would try to convince the framing crew that they would save time with a method like this by having a SAFE stage from which to build the roof.
    I also think the house could be "dried in" quickly and tested for airtightness.

    The tape on the deck would not be subject to a lot of stress because the deck and the tape would be close to the interior temperature.... and if the tape did fail in 50 years .. it could be redone.

  9. user-659915 | | #9

    Reply to John B:
    All good points.

    I very much like the baseline idea of the service core within the air barrier as I have seen so many attics where an otherwise satisfactory depth of insulation has been kicked out of the way for later electrical work and never replaced, and where can lights have been cut in without any attempt at air-sealing. This may prove to be a more robust setup. Dare I say more idiot-proof? I've long advocated for the attic as a storage-free zone with a tightly-sealed inspection scuttle rather than the standard pull-down stair. This looks to be another good step in that direction.

    By the way, I never specify spray foam anyway, at least if there's any other reasonable option.

  10. josefchalat | | #10

    Service Core
    Why not make the walls 6" higher and fur down with 2x6s under a continuous gypsum board ceiling that provides the air barrier. You could use a truss with a raised heel and you would not need to frame an attic floor. You could omit the furring in places to create a tray ceiling.

  11. Expert Member
    ARMANDO COBO | | #11

    John,A way do it is to
    A way do it is to install furred down hallways to reach rooms and blow the air through the rooms. In larger rooms you need to build a coffer ceiling or in the case of a vaulted ceiling, I’ve designed a duct chase at the bottom of the vaults. You can install ambient lighting on the coffer ceilings and above the chase, that creates an impressive lighting effect washing out the entire ceiling or vaulted ceiling.
    Some of the problems with furred down ceiling are bringing the framer for a second time to close the fur downs, and a close attention to the air sealing detailing in the air barrier and penetrations.

  12. user-659915 | | #12

    The problem with a drywall air barrier above furring:
    two visits by a sheetrock crew. Or am I missing something?

  13. hUd6rCjJMg | | #13

    Roof Thrust Connections
    One problem with some of the built up plate options is the elimination of the ceiling joist as the rafter tie, which would be particularly problematic in high wind or hurricane zones. The toe-nails from the rafters to the plates are insufficient. Metal hurricane ties would be necessary and these would create thermal bridges.

  14. hUd6rCjJMg | | #14

    Roof Washing
    Joe somewhat contradicts himself. He's emphatic about the need to wash every rafter bay with ventilation air but then doesn't care whether the exhaust vents are continuous ridge, "mushrooms" or gable vents. Only continuous soffit and continuous ridge vents will provide complete washing of the underside of the roof deck.

    And most of the ridge vents on the market, especially the most popular ones, are inadequate for their purpose. Unless they have external wind baffles, they will allow entry of wind-driven rain and snow into the attic - undermining one of the primary purposes of roof venting.

    Now Dr. Joe might not like the fact that externally-baffled ridge vents also have the property of increasing negative pressure as the wind is lifted over them by the baffles, since he doesn't want any negative pressure to suck air from the conditioned space. But, if the ceiling air barrier is intact (no can lights, and no unsealed attic hatch, roof flashing on all plumbing vents at the ceiling), then this is not a problem.

    I do not use attic ceiling hatches. Instead, I build a weather-sealed "hay-loft" door in at least one gable for service entry from a ladder. Since no attic storage is possible, there's no reason for an inside hatch, which has to be perfectly both air-sealed and as well insulated as the rest of the ceiling.

  15. user-659915 | | #15

    Hurricane ties
    Hey Robert, welcome back.
    In the zones you mention hurricane ties will probably be necessary anyway for reasons of wind uplift, especially with a wide eave. Thermal bridging will be minimal and easily managed with exterior insulation - yes I know your preferred techniques like to avoid foamboard at all costs but many of us are OK with thoughtful and limited use of this material. And rafter spread can always be handled with collar ties.

    I do like your idea of exterior attic access though. Makes a lot of sense in appropriate situations.

  16. Joe Lstiburek | | #16

    Reply to Riversong
    Continuous soffit vents are significantly more important than continuous ridge vents in washing the underside of roof decks. Are
    continuous ridge vents better than button vents and gable vents?
    Yes. Does that matter much with continuous soffit vents? No. So do I care much? No. That was pretty clear in the presentation. The reasons for all this are pretty obvious.

    Do externally baffled ridge vents and the venturi effect create
    suction at the ridge when the wind is blowing? Yes. Do I like that?
    Yes? Is that a problem with continuous soffit vents where there is more soffit vent area than ridge vent area? No. Is it a problem when there is more ridge venting than soffit venting? Yes, if the ceiling is not airtight. The reasons for all this are also pretty obvious.

  17. Joe Lstiburek | | #17

    Response to JoeW
    Yes there can be some problems with installing an effective air control layer in an attic - mostly with combustion safety and air change. See the links below for more information and how to stay out of trouble.

  18. Joe Lstiburek | | #18

    Reply to John Brooks
    I have no problem with your approach. The air control layer does not need to be drywall. And I certainly would never argue against service cores. Good luck with winning that argument with a production builder or most architects.

  19. Joe Lstiburek | | #19

    Response to James
    Yes, this is a nice detail. See my earlier comment to John. Well done, both of you. But good luck convincing other folks besides the clearly smart and handsome ones like yourselves.

  20. gman571 | | #20

    What's the cost?
    If some is good; then more is better. We’re so focused on saving a roof that we're willing to sacrifice livability, lifestyle and the planet.
    It is wise for builder to ignore that there is a cost associated with not being able to use the dry, secure, unused attic for storage?
    In a cold climate, ventilating a roof more than needed requires more energy to heat the house. That costs the homeowner more money and contributes a little bit more than necessary to global warming.
    Any time the wind is blowing, the house (and attic) are pressurized. Lookup Bernoulli's principle. As such, air will be escaping from inside the house, through any crack or vent (including the soffit vents) to the outdoors. It's wise to not pull air from the living space, but simply increasing ventilation is a costly oversimplification.
    The proper amount of venting is a function of many variables including vent area, interior and exterior temperatures and relative humidities, construction and associated materials, wind speed and direction, etc. This obviously changes with the seasons, climate and lifestyles of the occupants. It's hard to imagine how any passive system can adjust to these dynamic variables. Smart ventilation seems to be the answer, but building codes won’t allow it (sigh).

  21. yonyonsin | | #21

    Why can't the attic be used for storage?
    As long as there is sufficient insulation on top of the ceiling, what is the problem with placing a platform above the the insulation and using it for storage?

    I ask because I am planning to do just that in the attic space of an addition I'm planning for my house.

  22. Joe Lstiburek | | #22

    Reply to Ray Sten
    I argue against storage in the attic because the hatch rarely gets
    sealed tight - especially if you actually use the attic for storage.
    And even more rare is limiting the "stuff" being stored in the attic to the attic platform. Every time I have seen this done it sucks (actually it blows - but you get the idea). In theory it works -
    build the platform above the insulation and have a tight attic hatch.
    But communism works in theory too.

  23. Joe Lstiburek | | #23

    Reply to Gerald
    Listen to the video clip and read the text - again. I think I was pretty clear - the ceiling needs to be airtight so that air is not pulled out of the house. Then all those other variables you mention don't matter. And guess what? The passive approach works just fine - everywhere. No fan. Amazing, eh?

  24. Dayton | | #24

    Insulating top plate
    So what is the reason for making sure that the top plate is better insulated than the wall? Are there moisture issues for the top plate if it is insulated less than the wall? What is suggested for renovations were there is very little room for insulation above the top plate. Believe it or not my house has 2x4 rafters sitting on the plate (in Seatlle, so low snow load), and there is very little room for insulation above the top plate. If you add exterior insulation to the wall, there is no way to get higher R value above the plate, even if you use closed cell spray foam for the attic perimeter.

  25. Joe Lstiburek | | #25

    Reply to David Argilla.
    Yes, the issue is mold on interior surfaces where the ceiling meets the exterior wall due to the cold spot created at the top plate. And yes, renovations are a bear, and you do the best you can. If the moisture levels in the house are low during the coldest month of the
    year you can avoid the mold spot. What happens in some cases after
    the renovation airtightness goes up, air change goes down and interior moisture levels go up and the cold spot can be an issue. If you add a controlled ventilation system and keep the RH below 30 during the coldest period of the winter you can pretty much relax - unless you are in Barrow. In Seattle, the coldest part of the winter means you put on a long sleeve shirt and you get your latte double hot. So I wouldn't worry much.

  26. Dayton | | #26

    No insulation and no central heat = In the winter routinely waking up to a 45 degree F house. Maybe you can tell my wife to put the long sleeve shirt on, I do it one more time and my behind is single again ...

  27. homedesign | | #27

    Vented Attic in a Humid Climate
    Dr Joe,
    I will ease up on trying to sell the Perfect Ceiling ... getting back to a good ole vented attic with airtight drywall and Ample Cellulose above. You mentioned that a vented Attic "works" in Hot climates...
    What about Hot Humid Climates? Any worries or other considerations for a Humid climate?

  28. Joe Lstiburek | | #28

    Response to John Brooks
    In hot humid climates during the day when the sun is shining the attic air temperatures run to 130 degrees F and the sheathing temperatures run to 160 degrees F. The attic becomes a solar kiln. Let us assume that the outside air is at 90 degrees F and 90 percent relative humidity. When this air enters the attic space through the roof vents and is heated to 130 degrees F its relative humidity drops like a rock. The sun heating the roof sheathing drives the moisture out of it into the air in the attic. So even though attics in humid climates are ventilated with humid air, incident solar radiation leads to a huge drying potential.

    Now at night, the situation changes. Venting at night brings moisture into the attic and the wood materials pick up moisture. But this wetting effect is significantly less than the drying effect during the day.

    So, on balance, even in hot and humid climates venting attics make sense. Even if it is cloudy during the day for weeks. But only if the attic ceiling is airtight and there are no ducts in the attic.

  29. dsmcn | | #29

    Ridge vents in snow?
    Joe, I'm a fan. I've bought your "Builder's Guide" for my climate and it helps me to sleep at night.

    But here's my question: when there's 3' of snow on the roof, how much performance can the ridge vent provide?

    And since I'm asking, how does one keep the vents going through the roof from melting the snow and creating ice dams downhill? How do you convince the snow to not notice the hot air from the furnace vent, the water heater vent, the range hood, the bathroom fans...

    And finally, shouldn't they be called "ice: damn!-s"?

  30. vZNXFmZfGh | | #30

    Using the attic as storage.
    Possible on new construction or an extensive remodel this would all make sense. But when you live in a small house that storage area is important. My house is also 100 years old. I bet you anything that my energy bills are smaller than the overly large houses that are being built today. If I take reasonable measures to stop air leakage and to upgrade the limited insulation that my home has now, I can still use the attic for storage, make the hatch as tight as needed and keep my coldest month energy bill near 100 dollars. This makes more sense then to put the attic off limits and build a storage unit or pay for offsite storage.

  31. user-730247 | | #31

    If I may brevify my understanding:
    V1: Do not manage unconditioned attics by bringing the outside into the attic; manage them by putting the attic outside.
    V2: Complete the building enclosure either at the ceiling plane or at the roof roof plane, but for god's sake complete it.

  32. homedesign | | #32

    Durable,Affordable,Buildable & Best use of Material
    Which High Performance Roof Assembly has the highest Rating?....
    The Vented Attic

  33. dburgoyne | | #33

    How do I vent around solid blocking at roof?
    Is seismic zones, we usually install solid blocking between the top plates and the roof sheathing to transfer shear loads between the roof and walls. What are some effective methods to vent rafter bays and still provide this shear transfer?

  34. 7rCT9ChJ7d | | #34

    Venting at solid roof blocking

    I usually plan way ahead of time for this and coordinate with the structural engineer to have oversized blocking, so that I can design venting notches along the top of the blocking, say 1.5" deep by 4" long for example, really depends on my calculations. It gets complex, in that the notches can't land on a truss or joist connection, so they have to be thoughtfully spaced.

    If the blocking is oversized, then the 1.5" +/- depth taken out of the top of the blocking doesn't compromise the actual depth needed for structural integrity. As long as they get the nailing they require, which also gets complicated, the engineers are okay with it.

  35. Joe Lstiburek | | #35

    Response to David McNeely
    Check out:

    Do not vent your vents up through the roof - that is why God invented walls. Yes, I know, I know plumbers are Gods unto themselves and there will be plumbing vent stacks penetrating the roof. Not much heat with that. B-vents up through the roof and chimneys in general not a big deal for reasons that will not make sense unless you read the referenced article above. The melt water up above in the field of a roof does not lead to an ice dam most of the time due to the capillary uptake of the snow melt into the snow pack. Ski Patrollers who deal with avalanche control know about this stuff - speaking as an old and crotchity Ski Patroller - who only wished he skied in a place with enough snow to worry about avalanches.

    In terms of ridge vents covered with snow - snow, like dense pack cellulose is not an air barrier. A couple of inches of snow covering a ridge vent is not a big deal if the ceiling is airtight and the roof is well insulated. Again, see referenced article above. In high snow load areas - like in ski resorts or in Syracuse - ridge vents are a bad idea and you need gable vents or ridge vents that vent into monitors or cupolas or vent at their ends as mini gable vents.

  36. Joe Lstiburek | | #36

    Response to Vincent Alvarez
    I don't have a problem with your logic and approach. But in defense of my views, most folks are not as fastidious as you or as savvy.

  37. Joe Lstiburek | | #37

    Response to Ed Voytovich
    Yes. It takes people with funny last names to make English as clear as you just made it. Joseph Conrad must have been a relative.....

  38. Charles_Shade | | #38

    Habitable Attic
    If I understand my building official correctly and you are required to build under the 2009 IRC if an attic is big enough to be considered a room; i.e. minimum 7' dimension minimum 70 sqft and the ceiling for 35 of those square feet is over 7' high, you will have to now design this as a habitable space AND provide a code compliant stair or access to the space. Therefore the whole idea behind the insulation being free to live as insulation wants to do would be useless.
    Trusses would be more accommodating to this idea.
    If indeed this is the interpretation of the building code.

  39. user-963050 | | #39

    Great article and video
    Good piece. It should be required watching for roofers. I can't tell you how many homes we go into with improperly vented (often time over ventilated) attics. My personal favorite is the homeowner that wants to keep her attic cool so she insisted on the installation or a gratuitously large power ventilator. There is a reason that your attic is now within 10 degrees of conditioned temperatures and it isn't a good one.

    Most recently done roofs we look at have the standard retrofit ridge vent because roofers are afraid of the warranty exclusions from shingle manufacturers. All they have done in most cases is increased the stack effect and pulled more conditioned air from the living space below. Trying to tell a roofer that attic temperatures aren't really the issue usually nets the blank stare.

    Good read as always.

  40. user-982477 | | #40

    Just installed a powered attic fan, like, yesterday. I also have those stupid tiny soffit vents every three or four bays. And stuff in the attic. Drats -- and that's not what I wanted to say.

    Great video / presentation!

  41. user-982477 | | #41

    ... Oh and yes, we got sold ridge vents when we added a new roof about 5 years ago. And, when I added insulation last year, I also put in platforms for storage, of course (So far so good in terms of keeping all the Xmas decorations etc etc on deck.) Gosh, we are a case study for doing almost everything wrong!

  42. PDBZQGHNaV | | #42

    Unvented attic/roof assmbly
    With the above on a stick built or w/ perlins roof (tung n grove ceiling on top of roof rafters, sheathing then 6" phyliso, paper then slate or 3 tab asphalt ) what is a good way to keep the "as much insulation on top of the plate as in the walls"?
    The building has the craftsman wall brackets external to this area. The attic is habited space in that it's part of the lower room (i.e. no attic). We call it cathedral ceiling. I am imagining a shamwiched top plate with more phyliso board, may be extra external phyliso board at the sheathing and for the 2 ft sofits to be ladder design added after the external insulation. Can't see how to get that top plate area up to the R-30 walls.

  43. dsmcn | | #43

    How do you vent valleys?
    Every valley results in an area of unvented roof that is the square of the distance to the ridge (i.e. a 15' distance to the ridge results in 225 sq.ft. of unvented roof). Plus, snow likes to collect in valleys and takes longer to melt there also. Amazingly, ice dams are often densest in valleys!

    And yet, I have yet to read an article on ice dams that mentions this problem, let alone a solution. Joe, in your "Builder's Guide" book you suggest doubling the valley rafters and reducing their dimension from the jack rafters, so there is a space above them (resulting in about a 2"x3" channel). But this doesn't obey your command to have more vents at the bottom than at the top. Wouldn't you be reduced to the soffit vent in the first adjacent bay to feed all the subsequent bays created by the jack rafters (and the area available to that vent is further reduced by the diagonal...)? In my example of 15' to the ridge, with rafters @16"o.c., that means 11 bays on each side!

    And for all the homes already built, is there retrofit that would solve this problem?

    This is not an academic challenge. My MN in-laws live in an expensive house built by a successful builder for himself (not for profit), and there are still major ice dams every winter.

  44. Joe Lstiburek | | #44

    Response to David McNeely
    The valley issue can be handled by dropping the valley rafters. See figures from my Cold Climate Builders guide. Hips are handled the same way. This also extends to framing for dormers and skylights.
    See images below. Notice the horizontal framing on the large dormer below.

  45. user-963050 | | #45

    Unvented attic/roof assmbly for Chad Fuller

    I would first apply a peal and stick membrane (Ice/Water) to air seal the T&G boards. You can then apply your 6" of Poly Iso on top of that sealed T&G and then put down your decking or purlins/battens for the roof system.

    You can run your venting over deck and you should not have any ice damning issues after that.

    6" of Poly Iso is over R-30.

  46. dsmcn | | #46

    But what about the valleys?
    I had already seen these schematics (as I mentioned in an earlier post, I have bought your book and I am a fan). I'm disappointed that you did not respond to my concerns. One soffit vent feeding 3 rafter bays as in your drawing might work, but what happens when you have a real valley that has 11 or 12 or 15 rafter bays all being fed by one little vent? This combined with all the snow melt feeding from a large roof area, like streams becoming a river...

    Valleys seem to collect ice. When you combine this with the very large hurdle of making drywall airtight despite everyone wanting cans everywhere, plus relying on all future homeowners to be cognizant of the need to maintain this barrier, and never store their stuff up there; well, you've helped me make up my mind. Seems to me the airtight barrier should be around the perimeter of the whole house, as in your "perfect wall" description, and the insulation should be up at the roof level. All problems solved.

    I've always preferred the idea of a sealed crawlspace—why would anyone want to invite hot humid air into the environment they are "housing?" I think the same reasoning applies to attics: why would you want to superheat the air immediately adjacent to the space you are trying to cool? And in winter, why do you want to worry about how perfectly you've controlled that same environment exposed to the outside?

    And rewiring a fixture when all the wiring is below 30" of very dusty cellulose is like trying to re-plumb a bathroom on a slab. Except on a slab at least you know where you can walk.

  47. user-963050 | | #47

    Insulated roof decks
    Are certainly and option, just more expensive. If you valleys and ventilation limitations dictate moving your envelope and insulation layer to the roof deck, you can certainly do that.

    Spray foam is just a bit more expensive than cellulose and foam combo for sealing and insulating the attic floor.

  48. homedesign | | #48

    Questions for David McNeely
    McNeely:"This is not an academic challenge. My MN in-laws live in an expensive house built by a successful builder for himself (not for profit), and there are still major ice dams every winter."
    David, does your in-law's house have a simple vented attic?

    What does the price of the house ...or the fact that it was the personal home of a "successful builder" have to do with the quality of the Enclosure?

    What was the Air Control Strategy for the Ceiling? ADA,spray-foam?
    Does the house have Can Lights? How were they Air-sealed?
    Are can lights really necessary?
    Was the house Blower door tested?... what was the ACH50?

    Have you read the Straube/Grin Report #1006 for High-R Roofs?
    What's not to like about a vented attic?

  49. GBA Editor
    Martin Holladay | | #49

    Missing photos
    Three missing photos have just been added to Joe Lstiburek's last response. GBA apologizes for the delay in attaching these photos.

  50. homedesign | | #50

    Dropping Hip & Valley Rafters
    Martin or Joe,
    Is the detail from post #46 mainly intended for vented attics in cold climates?
    What about Climate Zones 2 & 3?..
    Most of the roof designs in Texas are Not-So-Simple
    Lots of Hips and Valleys and most roof rafters are only 2x6

  51. Joe Lstiburek | | #51

    More David McNeely
    If the ceiling is airtight and the assembly has a high thermal resistance the valley rafter problem is not a problem with the detail I recommend. How high an R-value? Depends on the climate zone. Work done by Tobiassion with the Army Corps of Engineers in the 70's pretty much established these R-values. In the ice dam article I referenced earlier I talk about those R-values. About R-40 in climate zone 5, R-50 in climate zone 6. When you do this the snow melt does not happen. The stream river analogy does not apply. Valleys and hips are only a problem with cathedralized assemblies - they are obviously not a problem in the example I used as a basis for my presentation. In complex cathedralized assemblies the unbalanced ventilation approach I recommend for simple attics almost never occurs and is typically not possible. You do the best you can and rely on your airtighness strategy.

    I have no sympathy for the difficulty in establishing airtightness. You either provide it under the roof deck with the gypsum board ceiling with a vented space below the roof deck or you provide it at the roof deck with a vented space above it (an "over-roof") in a high snow load area to control ice damming. The airtightness is necessary regardless.

    When we do complex roofs in high snow load areas like ski resorts we typically have the roof deck act as the air control layer and provide a high level of thermal resistance at the roof deck (either rigid insulation directly above the roof deck or high density spray foam directly under the roof deck - and then we provide an over roof creating a vented space over the top of the entire assembly. In essence we created a vented unvented roof hybrid. An unvented primary roof assembly with a vented over roof.

  52. JohnWalker | | #52

    What about better insulation over top plate?
    All the diagrams I see show a squiggly line suggesting fiberglass or other loose fill type insulation. My first thought is to upgrade the insulation over the top plate to something with a higher R value per inch thickness like XPS. Two or three layers of 2" XPS over the top plate would boost the R value substaintially.

    *Is there a good reason not to do this? Is there a wall dew point/condensation issue?



  53. Joe Lstiburek | | #53

    to John Walker
    There is no good reason beyond convenience or laziness not to do this. There is no hygro thermal reason not to do so (note the $100 dollar hyphenated word replacing "dew point/condensation" - chicks dig this - ok so that is not true - dew point calculations are a horrible way to do analysis and so "hygro thermal" is the "politically correct"
    phrase - and I am all about political correctness.....).

  54. wjrobinson | | #54

    I first went to a Joe seminar
    I first went to a Joe seminar sponsored by NYSERDA I think at West Point decades ago if that is possible.

    Joe, thanks for all your years helping us all out with the common sense of dealing with moisture and high levels of insulation.


  55. D6RFv8MtRu | | #55

    Venting a vintage house
    My 1914 house needs to be re-roofed this summer but I am at a loss re ventilation. Roof is 9/12. Floor area is 1200 sq ft. The attic space contains no mechanicals and is not used for storage. The ridge is about 40 feet long and there is currently a ridge vent but no other roof vents, no soffit vents, and no gable vents. The roofers who have given me bids have all recommended either cutting holes in the soffit to install [ugly] pressed metal vents, adding a bunch of vent boxes/mushrooms both low and high on the slope, and/or installing whirly-bird things to ‘promote airflow’. What is the best solution for the lack of air intake? The soffit is the original painted wood tongue & groove - do I really have to cut into my TG soffits? Are any of the deck-mounted products out there, such as Coravent’s In-Vent beneficial? I cannot install gable vents because the house is in an historic preservation area and such a modification would not get approval from the historical commission. House is located in Minnesota – hot summers / cold winters / lots of snow. The attic floor insulation will eventually be beefed up but [thank you, Mr L.] now I know that I must seal the ceiling plane first.

  56. GBA Editor
    Martin Holladay | | #56

    Response to Jeanne Kruchowski
    Assuming that there are no existing problems, just relax. You are overthinking this.

    Proceed with your plan to air-seal the ceiling and beef up the insulation on the attic floor. I don't think you need to do anything to change the ventilation situation in that attic. Nothing is broken.

  57. D6RFv8MtRu | | #57

    thanks - so I will


    thanks - so I will really be OK without any intake vents? It that is the case, that's wonderful. This is going to be an expensive enough project already, so I while I want to make sure I'm getting everything right up-front, I of course don't want to wrech my TG or spend on adding vents if I don't have to.


  58. tfarwell | | #58

    Workshop Venting
    I built a workshop last year (700 s.f.) with a vaulted ceiling. I decided to spend the time venting every joist bay of the ceiling at the side walls and a ridge vent across the top. I put 3/4" airflow channels integrated into the skylight trim as well, so every bay would have air flow. I installed OSB as the ceiling with battens covering the seams, caulking gaps. I also rain-screened the wall with the soffit vents doing dual duty as the outlet (good idea, bad - not sure - it's an experiment...). So I basically have a full-building skin, then airflow, then insulation, then air-tight (theoretically) interior plane.

    So far, this system is working great. I have no cooling system other than two ceiling fans. On 90+ deg days, the shop remains comfortable to work in (mid-upper 70's) - opening the windows heats it up. I guess I met the ultimate goal of comfort without additional systems / cost. I am also hoping this causes the roof material (30 yr premium comp) to last a real long time. The extra material cost was basically nothing - it was attention to design and detailing.

  59. gchamplin | | #59

    Hybrid above and below deck insulation II
    It looks like the picture didn't get uploaded, I'll try again.

  60. GBA Editor
    Martin Holladay | | #60

    Response to George Champlin
    It looks like you are in DOE climate zone 4C.

    According to the 2009 IRC (Section R806.4), it’s possible to build an unvented roof assembly with a combination of rigid foam insulation above the roof sheathing and air-permeable insulation (like fiberglass) in the rafter bays. If you build this type of roof, the code requires that “rigid board or sheet insulation shall be installed directly above the structural roof sheathing as specified in Table R806.4 for condensation control.” The table calls for a minimum of R-10 for Climate Zone 4C.

    So, your plan to install R-13 polyiso above the roof deck will work -- although the total R-value (19+13=32) is pretty low. The code calls for at least R-38 in your climate zone, so if you can make the polyiso thicker, your plan would be improved.

    Whatever you do, don't install any venting between the roof sheathing and the fiberglass -- venting would introduce exterior air between the two layers of insulation, and greatly reduce the thermal performance of your roof.

  61. gchamplin | | #61

    Hybrid above and below deck insulation
    I am rebuilding a cathedral ceiling in Portland Oregon I'm wondering if a hybrid approach is advisable. I am replacing a rotted plank roof deck with 2x6 rafters and sheathing. I'm planning on putting R-19 fiberglass between the rafters and 2" R-13 foil-faced polyiso or XPS on top of the deck, as illustrated in the recent Fine Homebuilding article on roof venting. I'm a little nervous about an unvented design and wonder if there are any advantages/disadvantages to adding a vent channel to each rafter bay. Does this defeat some of the advantages of the above-deck foam layer? Also does this affect any advice on vapor barriers?

  62. EtQyFdEVpC | | #62

    vent or not with closed cell foam
    Joe, thanks for the great article. I'm building a house which has a steel roof over a frost barrier (i.e. no permeability). The frost barrier is laid directly onto the plywood decking. The roof is nearly flat (1.2/12). I've been strongly considering using a closed cell insulation product (2" thick) combined with a blown in fiberglass batting in the 2x12 rafters(10"). I live in Seattle which, as I understand is zone 4C.
    Question: My house is set up for venting but some of the literature speaks to the fact that a closed cell foam material can be sprayed directly onto the roof decking without venting. Given this situation would it be your recommendation to vent or is that not necessary?

  63. GBA Editor
    Martin Holladay | | #63

    Response to Erik
    Your plan will work, as long as the total R-value of your proposed roof assembly meets the minimum R-value requirements in your local building code.

    The 2009 IRC allows the use a combination of air-impermeable and air-permeable insulation in unvented rafter bays, as long as the minimum R-value of the air-impermeable insulation that is "applied in direct contact with the underside of the structural roof sheathing" meets the requirements for condensation control shown in Table R806.4. Table R806.4 calls for a minimum of R-10 for Climate Zone 4C (that would be for the spray foam only).

    The minimum R-value for ceiling insulation (the entire roof assembly -- in your case, spray foam plus fiberglass) in your climate zone is R-38.

  64. EtQyFdEVpC | | #64

    Martin, thanks for the help.
    Martin, thanks for the help.

    As we're planning on at least 2" of closed cell foam applied directly to the structural roof sheathing which I believe has an R value = 6.5/inch the minimum R value of 10 is met.

    One other question... 2009 IRC R806.4 also states...

    "No interior vapor retarders are installed on the ceiling side (attic floor) of the unvented attic assembly."

    Does this refer to the sheetrock material or the paint primer?

    Again thanks for the comments and helpful insights.

  65. GBA Editor
    Martin Holladay | | #65

    Response to Erik
    The sentence in the code you quoted -- "No interior vapor retarders are installed on the ceiling side (attic floor) of the unvented attic assembly" -- is poorly written and confusing, and I have already had discussions with Joe Lstiburek and code officials about it. Hopefully the sentence will not appear in future code editions.

    The intent appears to be to warn builders about the installation of polyethylene -- yet even that warning seems unnecessary. Clearly, as written, it appears to forbid builders from painting their ceilings.

    Anyway, ignore it. As long as you don't have poly in your ceiling, no code official is going to have a problem with your ceiling.

  66. EtQyFdEVpC | | #66

    can lighting
    One other concern I've got is that I'm going to have a number of can lights (Juno) installed into the ceiling. As I've mentioned, I'm planning on 2 inches of closed cell spray in foam. Do I need to be worried about the heat generated from the cans being close to the foam? I wouldn't want to burn my house down because the can lights could be hot and close to the foam.


  67. GBA Editor
    Martin Holladay | | #67

    Can lights in an insulated ceiling?
    Can lights in an insulated ceiling? What were you thinking?

    Call up your electrician now and pull them all out. Substitute surface-mounted fixtures. Yes, the substitution might be expensive. But it will never be cheaper to do it than now. After you finish your house, you'll be kicking yourself for years if you don't take this advice.

  68. EtQyFdEVpC | | #68

    Sorry, I'm a bit of a novice
    Sorry, I'm a bit of a novice here. Are you suggesting that I remove them because its a fire hazard or because I've penetrated the ceiling? Or some other reason? Not sure why I would regret this for years to come?

    - confused.

  69. GBA Editor
    Martin Holladay | | #69

    Response to Erik Nelson
    I apologize for not explaining better; my fault.

    An insulated ceiling is ideally airtight. That means you need an effective air barrier.

    Such a ceiling should also be well insulated. These days, that means that it should have an R-value of R-30 in Florida, or R-38 to R-60 in a cold climate.

    Installing a recessed can light in an insulated ceiling interrupts the air barrier, allowing escaping interior air to enter the roof cavity. Escaping air often leads to condensation, mold, and rot.

    Recessed can lights also makes it difficult or impossible to achieve R-38 to R-60 above the can light, leading to energy waste and ice damming problems. That's why energy experts advise that no recessed can lights should ever be installed in an insulated ceiling.

  70. 35DPnw5bNq | | #70

    Insulated ducting within the atic
    Thanks for the great video. Is is practical to assume you can seal up an attic when you have insulated HVAC running through the unheated attic in a Zone 3 climate? Am I better off going with a hot roof or is the risk of moisture /condensation from the insulated ducts to high?

  71. GBA Editor
    Martin Holladay | | #71

    Response to Gerald Pluard
    Q. "Is is practical to assume you can seal up an attic when you have insulated HVAC running through the unheated attic in a Zone 3 climate?"

    A. If you have ductwork in your ventilated unconditioned attic, you have several problems. Sealing air leaks in your ceiling is just one problem. That problem can be solved, although the duct penetrations certainly complicate things.

    The more difficult problem is to seal all of the leaks in the duct seams, as well as the leaks between the ducts and the register boots. Then there is the problem of installing enough insulation around all of your ducts to keep your energy bills reasonable.

    The best approach is to keep all of your ducts inside your home's conditioned envelope. If your ducts are already located in your attic, the best solution is (usually) to create an unvented conditioned attic.

    For more information, see Keeping Ducts Indoors and Creating a Conditioned Attic.

  72. [email protected] | | #72

    Attic/Roof Venting
    In the Northeast (Long Island, NY) can i use 1" foil faced foam (Dow Thermax) to create 2" air channels in the attic rafter bays (2x12) then add fiberglass batt insulation in the bays, sheetrock and latex paint? Will the low perm of the thermax cause any problems at this location? Also, if desired can I add 1" rigid foam (dow blue styrofoam) between the rafters and sheetrock to help with thermal bridging?

  73. GBA Editor
    Martin Holladay | | #73

    Response to Gary Steinfeld
    Yes, you can use rigid foam (for example, Thermax) to create a ventilation channel.

    If you have 2x12s with a 2-inch ventilation channel, 1 inch of Thermax will give you an R-value of R-6.5, and the remaining 8 1/4 inches of fiberglass will give you about R-30.5. So you end up with R-37. I don't know if that is enough to meet minimum code requirements; it depends on your climate zone and the local code.

    Yes, you can cut down on thermal bridging through the studs by adding a continuous layer of rigid foam insulation under the rafters. I don't think this type of assembly will have moisture problems.

  74. phil17 | | #74

    Baffles in combo with closed-cell spray foam?
    This thread is a little bit old now, but still very helpful!

    You talk about doing an over roof in areas with high snow loads, Effectively creating a vented unvented roof hybrid. Is is possible to effectively do the same things (although perhaps to a lesser degree) simply by installing baffles from eve to ridge before spraying the foam?

  75. GBA Editor
    Martin Holladay | | #75

    Response to Phil Lanier
    The roof assembly you describe can work well, as long as the following conditions hold:

    1. You provide at least 1.5 inches -- 2 inches is better -- of depth to your vent space, with the vent space extending the full width of each rafter bay. This means that you need to install site-built baffles, not commercially purchased ones.

    2. Your roof should be a simple gable or shed roof, without any valleys, hips, dormers, or skylights. If your roof includes valleys, hips, dormers, or skylights, it's not a good candidate for this approach.

    3. You must be able to install enough R-value to at least meet minimum code requirements.

    For more information on this topic, see How to Build an Insulated Cathedral Ceiling.

  76. rdcrouch1 | | #76

    To Vent or Not to Vent
    Dear Joe,

    Here is a resend with an attachment drawing. pardon previous post.

    On the topic of a Perfect Roof, please picture this unique commercial application. Large existing middle school. ASHRAE Zone 4. Roof retrofit. Steep sloped roof planes to be finished with conventional asphalt shingles. The attic will be ventilated to code in this retrofit. Cold formed steel framing truss “girders” are at 3’-0” OC create the slope. A galvanized steel deck 1.5” thick spans between the truss girders. The air barrier is the gypsum on the bottom of the trusses and insulation is between the truss bottom chords. A vapor barrier is integrated on the warm in winter side of the insulation so as to take the attic ventilation requirements to 1/300 vs. 1/150 using IBC rules.

    There are two camps on how to do address the work above the steel deck:

    Camp 1: Plywood sheathing placed over 2’-0” OC wood furring. The furring is placed over the steel deck perpendicular to the flutes. The gaps in between the furring could promote some air flow from eave to ridge. This is a second zone of ventilation in addition to the entire ventilated attic. Both ventilation zones would co-mingle at the top and bottom and have the same fresh air source low and same vent source high. We do need minimum thickness for this furring -- meaning thick enough only for roofing nails to avoid contact with the steel deck. The existing copper stepped flashing at the abutting walls beyond could be negatively impacted if furring is too thick. Somewhere between 3/4” to 1 ½” we would hope to stay. Under review.

    Camp 2: Two layers of staggered plywood of sufficient thickness so that the roofing nails do not strike the steel deck. No furring - so no ventilation between the steel deck and plywood. The thought here is that the venting of the entire attic space below the steel deck is sufficient.

    One variable to both trains of thought is uncertainty as to how well the existing air barrier at the bottom of the trusses was installed. More moisture vapor may be carried into the attic via air flow than would be desired if the Contractor did not seal it well.

    Any thoughts are appreciated. Discussions with major shingle manufacturers got mixed comments. One initially jumped into Camp 1 then backed off later. The other seeming to not care either way. No real help there.


    Ron Crouch

  77. GBA Editor
    Martin Holladay | | #77

    Response to Ron Crouch
    I've forwarded your question to Joe by e-mail; I don't know if he'll have time to respond.

    Here's my take: I agree with the shingle reps who say, either way will work.

    As a former roofer, I'd prefer the approach using the 2x4s above the steel, because it would allow the use of longer roofing nails.

  78. GBA Editor
    Martin Holladay | | #78

    Second response to Ron Crouch
    I had a chance to talk to Joe Lstiburek on the phone today, and he agreed with me that either approach will work.

    However, he had some advice for you if you choose to install two layers of plywood without the furring strips: since the plywood moves hygrically -- in response to changes in moisture -- while the metal moves thermally -- in response to changes in temperature -- you can get stresses that lead to problems.

    The solution is to install the plywood in 4'x4' squares, not 4'x8' sheets, and to include gaps between adjacent sheets (gaps wide enough to insert a nail). The second layer of plywood should also be installed in 4'x4' squares, and should be similarly gapped. The two layers should be staggered (offset 2 feet in both directions). Following these recommendations, you should be fine.

  79. Paul e. | | #79

    Attic ventilation
    Great article on this topic. This is my situation. Please let me know your thoughts. I have a split level with three separate unattached attics. Each attic has vented soffits, gable vents and power fans. I am having the house resided. Should I keep the gable vents or close them up?

  80. GBA Editor
    Martin Holladay | | #80

    Response to Paul Nigro
    The most important thing you can do is to get rid of the fans. More information here: Fans in the Attic.

    Second, most experts prefer ridge vents to gable vents. Removing the gable vents is probably a good idea -- but if you do so, have some ridge vents installed.

    Finally, attic venting is relatively unimportant compared to making sure that your ceiling is as airtight as possible and that your insulation layer is deep. If you can make sure of these two factors, the presence or absence of attic vents hardly matters.

  81. Paul e. | | #81

    Thanks for the info. For now I am stuck with the attic fans as I can't afford to redo the roof. The good part of my situation is that I have a very tight ceiling. No can lights or cracks and the hatches are sealed. With the gable vents and soffit vents do you really think I will depressurization the attic?

  82. GBA Editor
    Martin Holladay | | #82

    Response to Paul Nigro
    Q. "For now I am stuck with the attic fans as I can't afford to redo the roof."

    A. No, you are not "stuck with the attic fans." Unplug them! Or disconnect the circuit breakers for a few minutes and cut the electrical cords.

    Q. "With the gable vents and soffit vents do you really think I will depressurize the attic?"

    A. No -- as long as you unplug the attic fans, or cut their cords.

  83. dankegel | | #83

    1912 craftsman in Los Angeles has gable vents, but gets hot...
    Our house is lovely, but has a serious heat problem upstairs in the summer. It has big gable vents but no other venting (other than leakage) and no insulation at all. We're reroofing with energy star shingles, air sealing the attic floor, then insulating the attic floor. Our contractor surprised me by also proposing powered attic vents. I'll push back on that, but it brings up the question: since we're going to have scaffolding up anyway, and will be restoring the raftertails, would it pay to also drill soffit vents?

  84. GBA Editor
    Martin Holladay | | #84

    Response to Dan Kegel
    Your plan to push back against your contractor's suggestion to install powered attic ventilators is a good one. If you need more ammunition to resist the suggestion, you can find it here: Fans in the Attic: Do They Help or Do They Hurt?

    Your plan to perform air sealing work and to add insulation to the attic floor -- I hope that you're planning to install at least the minimum code requirements for attic insulation -- should adequately address your "upstairs heat" problem. I don't think that you'll see much benefit from adding soffit vents (especially if the attic has existing gable vents). For more information on this issue, see All About Attic Venting.

  85. brighteyed | | #85

    Leaf Blower to clear ridge vent?
    I live in a small colonial with ridge and soffit vents in Massachusetts and get ice dam damage every few winters whenever we get 36" of snow in a 2 week period like last year 02/2015. So this is when the ridge vents are covered with snow. Can I direct my leaf blower toward the ridge vent from inside the attic to clear the ridge vent without damaging anything? Thanks for your consideration, Joel

  86. GBA Editor
    Martin Holladay | | #86

    Response to Joel Heller
    Give it a try and let us know what happens.

    You should know that operating an internal combustion engine indoors is dangerous (due to the exhaust fumes in your attic) -- so I hope that your leaf blower is electric.

  87. fisher444 | | #87

    I know this is an old post...
    I'm in a 1 1/2 storey post-war home, which means I can only access 2/3rds of the roof. I can access the lower third of the roof through crawl spaces on either side of the house. I can put in baffles to let air in. I can also access the top third of the roof through the attic and make sure air channels exist in there as well. However I cannot access the middle portion of the roof, since it's right above the sloped ceiling of the upstairs bedrooms (basically roof deck on one side of the rafters, drywall on the other side). Will I be doing anything useful by venting the lower and upper portion of the roof, knowing that cold air will go through the insulation in the middle portion before getting sucked up into the attic?

  88. Expert Member
    Dana Dorsett | | #88

    Not really (response to Steve Smith)
    Yours is a common construction issue. Usually the rafter bays are blocked a the kneewalls, and there is no free-flow of air even if there WERE a sufficient air gap or sufficient R-value. The best solution is usually to go with an unvented roof assembly insulating between the rafters, then add rigid insulation above the roof deck when it's time to re-roof. The details of how to do this in a moisture-safe way prior to adding the above-deck R varies with climate zone.

  89. whitenack | | #89

    Gable Vents only?
    Am I reading this correctly that I don't have to have vents in the actual roof plane? It is acceptable to use only the gable end vents with the soffit vents? I'd prefer to make as few holes in my roof as possible from both a water intrusion issue as well as the looks of it.

  90. GBA Editor
    Martin Holladay | | #90

    Response to Clay Whitenack
    Ordinary gable roofs don't have vents in the roof plane. The usual intakes are located in the soffit; these are the soffit vents. The usual exit points are at the ridge; these are the ridge vents.

    Joe Lstiburek notes that you can use gable vents instead of a ridge vent if you want -- but you still need soffit vents as air intake locations.

  91. whitenack | | #91

    Does the type of gable vent matter?
    Right. Still need soffit vents. Does the type of gable vent matter? I assume the triangle vents that sit at the very top of the gable allow for the best removal of hot air collecting in the top of the attic, but what about the rectangle ones that are placed a little further down from the top?

  92. GBA Editor
    Martin Holladay | | #92

    Response to Clay Whitenack
    I'm not sure what Joe Lstiburek thinks. But here is what I think: the obsession with attic ventilation details is misplaced. Most attics stay dry, as long as the ceiling has a good air barrier.

    If the ceiling doesn't have a good air barrier, you can get into trouble quickly. But if you understand air barriers, these questions about attic ventilation details fade into insignificance.

    Here is a link to an article that explains more on the topic: All About Attic Venting.

  93. Jayock | | #93

    I found the video very insightful, but I still have a couple of questions. My wife and I live in Michigan and just put an addition on our home with a true cathedral ceiling (i.e. great room). We put an energy efficient metal roof on, fully ran ridge vent, and baffled every square inch of the roof decking from the soffit to the peak. The plans called for the cavity to be filled with R-49 batts and we purchased tongue and groove wood to finish the ceiling. The question I have is would you recommend that I use a house wrap (i.e. tyvek) between the insulation and the tongue and groove wood to act as a air and moisture barrier?

    1. GBA Editor
      Martin Holladay | | #94

      You absolutely need an air barrier between the tongue-and-groove boards and the insulation. The usual recommendation is to install taped drywall in this location -- a far more robust air barrier than Tyvek (which is more likely to get damaged during installation than drywall).

      For good cathedral ceiling performance, nothing is more important than the tightness of the ceiling air barrier under the insulation. Don't penetrate your air barrier with recessed can lights, or you will live to regret it.

      Two minor points: metal roofing is not particularly energy-efficient, and Tyvek is not a moisture barrier. (Tyvek is resistant to liquid water, but it is permeable to water vapor.)

      For more information on getting the details right on your cathedral ceiling, see this article: "How to Build an Insulated Cathedral Ceiling."

      1. Jayock | | #95

        Ok. The company that we hired to draw up the plans called for us to use a 6mil type plastic sheeting but I have a significant amount of tyvek leftover and was hoping to use it up. We are not doing recessed lighting and they told us to stay away from drywalling it as it adds weight, which is why I'm thinking they recommended the 6mil visqueen type material. Your thoughts? I appreciate the assistance

        1. GBA Editor
          Martin Holladay | | #96

          My advice? Don't use polyethylene as an air barrier -- it's too easy to damage (especially considering all the nails you'll be installing when the boards go up).

          I'll repeat: For good cathedral ceiling performance, nothing is more important than the tightness of the ceiling air barrier under the insulation. If you have ceiling air leaks, you'll have condensation, frost, and a raining ceiling.

          Use drywall.

  94. mmccarthy00 | | #97

    I have a 1956 "single-wall construction" home in Hawaii with a flat-roof. The ceilings are very low (between 7'6" and 7'4") and you can feel heat radiating from the ceilings during the day. The original roof system was pitch and gravel, but was recently replaced by the previous owner with an asphalt roll-roof product. From exterior to ceiling it is roll-roof, 1x6 shiplap sheathing, 2x6 rafters, 1/4" plywood, latex paint. Every rafter cavity is vented at each soffit end with continuous soffit vents. I'm renovating the interior so have removed the plywood ceilings and plan to replace them with 1/2 gypsum. I am hoping there is something I can add to the rafter cavities to help reduce the radiant heat from the roof through the ceilings. Would adding Foamular 250 sheets in each rafter cavity help? If so, should I maintain an air channel between the foam and the sheathing or between the foam and drywall? Thanks for your help!

    1. GBA Editor
      Martin Holladay | | #98

      M. McCarthy,
      Ideally, this roof would be insulated on the exterior (using one or more continuous layers of rigid foam on the exterior side of the roof sheathing). This approach requires new roofing, so you probably won't pursue it. More information here: "Insulating Low-Slope Residential Roofs."

      If you want to install insulation between the rafters (working from the interior), you'll need to seal up the ventilation openings. (A vented approach won't work with a flat roof or a low-slope roof.) You'll need to install closed-cell spray foam against the underside of the roof sheathing.

    2. Jon_R | | #99

      Insulation will certainly help, more is better. Your vent will always have some effect - you can retain it and if you do, it should be between the insulation and the sheathing. Air sealing is always important, even more so since your venting is marginal (but still code complaint).

  95. dwelling_in_pdx | | #100

    I've been reading and following several threads on this subject ever since we bought our home in Portand, Oregon climate zone 4c. We have a early 1900's arts & craft home, with the architecture of the house and being located in a registered historic district it brings some unique challenges. The first part as it relates to this is that it has a complex roof with 8.4:12 (35-deg angle) slope, the primary roof structure runs N-S with 2 gable dormers in the center. The eves have very wide overhangs with no soffits, with artistically cut rafter tails that extend beyond the roof line and 1/2 round gutters that fit into the tails. The house has the original double hung windows, currently with triple-track aluminum storm windows.

    Currently it is best to assume that the exterior walls have no insulation, the extent of insulation is what is offered by cement asbestos board siding that covers cedar shake siding, over 1x6 sheathing, over full dimension 2x4 studs with an interior finish of the original lathe and plaster walls and ceilings. The outer asbestos cement board is likely to be removed in a future project, restoring the house back to the original cedar siding that will be patched as needed. Assuming we can easily trace/remove any remaining knob and tube (currently feeds ceiling light fixtures only) we would likely get cellulose added to the exterior walls at that time.

    One of the previous owners installed some ~R9 craft backed fiberglass insulation poorly and directly against the roof deck (was previously unvented), with the craft facing inward and plastic vapor barrier. The attic is conditioned space and contains one forced air furnace for the attic itself and upper floor of the house. No known moisture issues, other than what is perhaps (assumption) forming between the fiberglass insulation and the roof deck. Considering the state of insulation, age and size of the house the heating/cooling costs are fairly reasonable.

    We put a new asphalt composite roof on a few years ago, which was a big project as it had original cedar shake and skip sheeting with several layers of asphalt over it. New plywood sheeting was installed with GAF Tiger Paw underlayment. The roofing company installed a ridge vent not the primary ridge, and some for of eave vent on the lower roof under the shingles. During the re-roofing project, sections of the fiberglass batt were disrupted by a roofer stepping through the skip sheeting during the tear off...and the fiberglass is no doubt full of dust/debris. Thankfully the roof only has 3 penetrations, the chimney (just offset of the ridge to not break it), one furnace vent, and one soil stack.

    The rafters are full dimension 2x4 24O.C., and being a conditioned space forces that it would be cathedral ceiling with knee walls. Due to the complex roof with valleys/gable dormers it is impossible to vent 100% the roof. Being limited to 4" of total depth forces the options to be high efficiency foam based insulation, as it would likely be a daunting task to try to fur the rafters out to achieve more depth.

    Initially I had assumed the best option would be to just seal the vents and apply 3-4" of close cell spray foam, but after reading this and other newer articles (and checking local code) it appears that I am best to have some ventilation.

    This comes to my questions:

    1) is my best bet to use closed cell spray foam for the unvented bays around the dormers and valleys and then to go with a vented method for the rest of the roof?

    2) I was considering going with cut-and-cobble polyiso, perhaps just going with 3" to reduce the labor involved and allowing for 1" of air gap below the sheeting and covering joints/seams with c.c. foam. The risk here is that it seems to place a great amount of risk on the two mid-seams between the polyiso pieces, as it would take 2.5 8' runs to go from top plate to ridge. Another option would be to add a continuous 1" polyiso face across the rafters below the 3", giving 4" in the bays and 1" over the thermal bridge of the rafters.

    3) It seems another, more costly option, would be use the AccuVent Cathedral Ceiling baffles and to just go with ~2-3" of closed cell spray foam over it.

    The goal is to have minimal ceiling penetrations when finished, we are likely going to have pendant lights and a galvanized spiral duct down the peak of the cathedral for the HVAC. There will be no recessed lights, not just because it wouldn't be feasible in a 4" rafter, but that we just don't like them (and have more than enough installed by P.O. in the basement to make up for the entire house).

    Am I on target for 1? And which of 2 or 3 are the better strategy, or a hybrid or something entirely different? Obviously cost does matter, but more so does doing the right thing to keep this house around for another 100 years while improving livability.

  96. saleen21 | | #101

    Joe or Martin, I have a question, all the advice is great assuming you have a drywall ceiling. But in my 24x24 two car garage all I have is exposed trusses/rafters of 2x4 lumber, and I don’t plan on finishing/installing dry wall on the lower side of rafters, plan on leaving the garage open (I don’t the money or resources to pay to have drywall installed or the manpower for me to attempt to hang them).

    So with that being said and with my garage having no vents anywhere, no ridge, no soffit, and no gable vents. But I am wanting to still try and warm up garage in winter and cooler in summer. I live in Louisville KY.

    So here was my thoughts or wanted to see what options I have.

    Install soffits along under each side of roof/eave. Install gable vents on each end of garage.

    Install R13 in walls, and insulate garage door.

    Install rafter baffles in each rafter cavity.

    Install R13 insulation in the rafters on top rafter baffles.

    I know not ideal but has to help and make some noticeable difference.

    What are your opinions/suggestions/recommendations?

  97. GBA Editor
    Martin Holladay | | #102

    When you write that you want to "install soffits," I assume you meant to write that you intend to install soffit vents.

    For the venting to work, the vent channels that begin near your soffit vents need to connect to either (a) a ridge vent, or (b) an open attic with generous gable vents.

    When you wrote that you intend to "install R13 insulation in the rafter on top of rafter baffles," I assume that you intended to write that the insulation will be under the rafter baffles, not on top of them.

    R-13 insulation for a roof is woefully inadequate. In your climate zone, you should aim for R-38 or R-49 insulation.

    If you install fiberglass batts, you still need an air barrier (that would be drywall) on the interior side of the batts.

    If you install spray foam, you still need a thermal barrier (for fire safety) on the interior side of the spray foam. The thermal barrier would be drywall.

    All of this information can be found in this article: "How to Build an Insulated Cathedral Ceiling."

    1. saleen21 | | #103

      Yes soffit vents around perimeter of garage overhang, with gable vents in open attic, and and rafter baffles on the underside of roof sheathing than insulation, unfortunately my roof construction is 2x4 trusses which only allows R13, not planning on installing a drywall ceiling just not feasible as a one man operation and can’t afford to hire anyone. So everything above top plate would be open/exposed.

      1. GBA Editor
        Martin Holladay | | #104

        Without an air barrier under the fiberglass, you have no hope of controlling the conditions in your garage. You need the drywall. Note as well that kraft facing is highly flammable -- so if you were thinking of using kraft-faced batts, and leaving the kraft paper exposed, you would be creating a fire hazard.

  98. thrifttrust | | #105

    I don't understand the resistance to rocking the ceiling do to manpower issues. Using a drywall lift and 8' rock it can be done by one person. We just finished our garage ceiling. We used 12' rock to limit butt joints. I'm 70 and my wife weighs 110 lbs. We had no trouble. Wherever a butt joint landed I sistered a 2x2 to the truss to provide good nailing for the butts. Our rental place charges $40/day for the lift, but by renting it on Friday and returning it Monday morning they only charged $50.

  99. Mbhoss | | #106

    I have an early post and beam house renovation underway. All exterior walls are the finish interior. I am installing Blums skin then skinning walls with 1” ridged insulation, then layer of ½” zip wall then clapboards. Roof will be layered using ½” ply, taped, vapor / air barrier, 1” ridged insulation taped, roof underlayment, strapping then zip roof, underlayment and slate roof. Obviously not going for high r value. This is a guest house used primarily for 3 seasons. I would appreciate any constructive thoughts or advice.

    1. GBA Editor
      Brian Pontolilo | | #107

      Hi Mark.

      I suggest you hop over to the Q&A section and post this there, since it is not about this article. More people will see it and you'll get more responses, I'm sure. When you do that, include your location (climate zone) in the question.

  100. MichaelAndrew | | #108

    I've read so much of Joe & Martin Holliday's posts on ventilation/insulation. I'm hoping to get some updates & advice on my situation..

    photos can be seen here:

    I am in Climate zone 2 I believe in New Smyrna 2 miles off the coast. I purchased a house that has a 25'x30' Gambrel roof garage with apartment above which is partially complete. I am completing this. It is 60% insulated & sheetrocked with R19 batts, It has AC ducts ran already as well as IC rated can lights. The roof currently has ridge vent & two gable vents on each end. There fortunately is a 8 inch over hang at the bottom creating a soffit but it is currently closed off. I am contemplating putting in soffit venting & baffle vents at the eaves as it seems some insulation batts block this at the break of the roof where it goes from more flat to sloped this area does have a 1-2 inch air gap at the top plate allowing air from side attic to get to the upper attic. I do have a garage/workshop below which is venting up into this space because the sub floor for the 2nd floor apartment runs to about 2' shy of the width of the building.

    I have chosen to cut into the wall structure and install a header to allow for more space in a few spots. Primarily to inset the bed area, toilet, & a closet space. In these area's I'm taking the wall almost against 2x4 rafter for the sloped roof so figuring out venting/insulation for these areas has been a hurdle. Posts on these sites have been helpful but there is old/mixed information

    I am value oriented and feel keeping as much of the work someone else has done makes sense, unless it's going to cause future challenges.. because I am keeping this house long term.

    Questions & confusion I have

    1) Should I work with why I have & keep the vented? I'm leaning this way due to the cost/work to encapsulate .. so I would ensure I air seal as much as possible on side walls and attic of upper apartment add soffit vents and ensure air can flow up to the ridge vent? I've also seen online a recommendation of closing off gable vents IF you have soffit to ridge vent?

    2) Can lights & ceiling protrusions are something I've read both Joe & Martin say are liabilities. But these were older publications/posts.. is this still true? Is newer LED recessed/ flush to ceiling lights with gasket still a major issue in your opinions? I don't really have any other option on the AC

    3) Garage below - The garage has 4 ground level cinderblock intake vents & then the 2' gap in sub floor where the garage is exhausting to.. I know I may have to sheetrock the garage ceiling at some point if I get issues with the town or rent it which I would insulate the underside of the subfloor before I sheetrock. is there any other considerations here? Should I complete the subfloor upstairs? if I complete the subfloor air can still get over to the soffit area and travel up to the ridge.

    2) Baffle vents - I have the two typical kinds of baffle vents box stores sell.. I will use those in the attic best I can to ensure air flows from side attic up to ridge.. for the inset areas where my wall will be closer to the sloped roof line. I am deciding between using those baffle vents or creating baffle vents out of rigid foam & also adding more 2x4 to increase the depth of the rafter which are 24" oc

    thanks for your time

  101. Luca174 | | #109

    "Put more insulation on top of the wall than inside it. In other words, if you have an R-20 wall, you want at least R-20 on your top plate. A higher R-value is better, but a lower R-value is not. If you have an R-30 wall you want at least R-30 on top of your top plate. "

    I'm a novice.
    When Joe say's "Put more insulation on top of the wall than inside it", does the "inside of the wall" refer to inside the roof rafter bays?

    Feel free to talk to me as if I was a 5 year old. I'm still learning. Thank you!

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |