Replies

1. 
   Clewless1 | Jun 25, 2008 06:58pm | #1

   Thinking out loud, here. Spray foam insulation in rafters should not result in ice dams as ice dams are usually a sign of 'leaking heat' that warms the roof enough to melt snow until it reaches the eave ... refreezing and subsequently/potentially causing problems.

   A lot of talk ... unvented OK? What I've heard is ... unvented OK IF you can ensure you have a good seal/vapor barrier ... which should be inherent in spray foam. I've heard unvented w/ fiberglass even if the right details are done.

   Hope someone else jumps in here and adds to this.

2. 
   Dinosaur | Jun 25, 2008 08:41pm | #2

   I just explained this to a client two evenings ago. They'd gone south (to Rhode Island) for the winter and came back to their cottage to find about 250SF of their oak flooring stained black and buckled up off the subfloor. Ice dams had caused enough leaking to do that even though the heat in the cottage had been turned down to just above freezing (to protect the plumbing and canned goods).

   Here's the one fact you have to get clearly in your mind:

   INSULATION DOES NOT STOP HEAT TRANSFER;
    IT SLOWS IT DOWN.

   That said, you can understand that no matter how good a job of insulation you do--whether it's foam, FG, cellulose, straw-bales, whatever--heat will continue to move through the insulation, and some of those calories will eventually escape to the surface of your roof where they will melt the bottom of the snowpack, which liquid water will then drool downslope until it hits the eaves.

   The eaves of your house are cold, because they are beyond the heat envelope. Thus water hitting cold eaves freezes, and that's how ice dams form.*

   The primary defense against ice dams is to keep the outer surface of the roof cold: below the freezing point of water. Since there will always be some heat leakage through to the roof, you need to 'air-condition' your roof by circulating cold, exterior air underneath it. That cold air will carry away the escaped heat from the house before it can warm the roof enough to cause the underside of the snowpack to melt. That's why you have to provide vent space in all rafter bays and both inlet and outlet vents in soffits and gables and/or ridges.

   Some of the worst offenders in leaking heat to a roof are plumbing vent stacks. A good approach to these is to cover them with a decorative cupola which will prevent any snow from lying on that area of roof around them which will always be warm.

   People who insist that it is possible to keep a roof cold without ventilation--including building scientists with lots of degrees, and who ought to know better thereby--are living in a dreamworld where everything works like they think it's supposed to.

   I live and build in the real world, in a climate where the winter temps also hit minus forty (Québec), so I know whereof I speak. Forget the unvented roof, Terry. You need full 1½" venting in each rafter bay, with continuously vented soffits, big gable vents, and possibly 'Charley Noble' vent turbines at the ridge.

   ---

   *Note: Solar heating also plays its part in creating ice dams, and there's not much you can do about that except to orient the house so the eaves face east and west, instead of north and south. Or to keep tall conifers close enough to the house so the roof is constantly shaded.

    

   Dinosaur

   How now, Mighty Sauron, that thou art not brought
   low by this? For thine evil pales before that which
   foolish men call Justice....
   

   

1. 
   DickRussell | Jun 25, 2008 08:58pm | #3

   Where was the 250 sqft of damaged flooring - next to the walls where the ice dam backup would be? Or out in the span?Up in NH, we got a fair amount of rain at times on top of heavy snowpack on the roofs. There were many roof collapses resulting from this. Rain on top of snow pack is going to run down to the shingles, then down to the bottom edge. The wet snowpack, if not subcooled, may continue to weep liquid rain for hours afterward.A thick but light and dry snow layer on top of a roof has some insulation value (sled dogs burrow into the snow to keep warm sleeping outside). That pushes the freezing/melting temperature boundary outward toward the roof deck. Depending on the amount of insulation under the deck, the insulating snow above could result in some melting at the deck.If the connection between wall top plate has unsealed holes for wiring, or there are air leaks that let inside air into the roof insulation, and the insulation is FG batts or something loose, then convective transport of heat could be causing localized melting.Bottom line - details are needed before one can conclude that unvented roofs are a bad idea, and in any case, good ice dam protection at the lower edge and at valleys are essential.That said, you do make a good point about things like vent stacks leaking heat. The OP might consider isolating these to insulated locations if they aren't already. Also, those 2x10 rafters, on 16 or 24" centers, are only R11, 1.5" wide, heat leaks by themselves. He might consider a 2" foam board on the inside, to reduce the thermal bridging thru the rafters.

   Edited 6/25/2008 2:05 pm ET by DickRussell

1. 
   Dinosaur | Jun 25, 2008 11:06pm | #5

   was the 250 sqft of damaged flooring - next to the walls where the ice dam backup would be?

   Yes. Although ice-dam leaks do not always emerge directly under the point where they enter; the slope and configuration of the framing members has a lot to do with it.

   A thick but light and dry snow layer on top of a roof has some insulation value

   Snow has a very high insulation value; northern peoples have known this for thousands of years. In the days before manufactured insulation, the old Québeçois habitants designed that classic curved roofline not just because it was pretty, but primarily to create and catch thick wind-drifts of snow at the roof/wall junction as a way of slowing heat leaks at that point. They also shoveled snow up against their houses in the winter, right up to the eaves where possible.

   You should visit an igloo when the outside temp is -35º or so. Light one small oil lamp, and within a short time you can sit there barechested, sweating.

   That pushes the freezing/melting temperature boundary outward toward the roof deck. Depending on the amount of insulation under the deck, the insulating snow above could result in some melting at the deck.

   Not sure I follow your description about pushing the boundary 'outward', but in effect what happens is that heat which leaks through the attic & roof insulation is slowed in its cold-seeking journey by the insulating snowpack, so it hangs around at that 'boundary' long enough to melt the first snow with which it comes in contact. The heat is then carried away downslope in the liquid water--which is denser than snow, and can thus absorb more calories per cc--until it hits the ice dam and is out from under the snowpack...at which point it can make a rapid transfer to the surrounding air and be carried definitively away from the house.

   details are needed before one can conclude that unvented roofs are a bad idea, and in any case, good ice dam protection at the lower edge and at valleys are essential.

   Attention to details can reduce the rate of heat leaking to the roof deck, but nothing short of an uninterrupted, absolute vaccuum will stop convective caloric transfer completely. And that's a very large contract.

   More to the point, by focusing on blocking every imaginable small leak point, people delude themselves into thinking that if one just tries hard enough, one can keep an unvented roof below freezing.

   But long and extensive experience shows that one can't.

   There are only two ways to keep an unvented roof cold: (a) Remove the heat source (don't heat the house under it); or (b) Refrigerate the roof mechanically, say with staple-up Freon cooling coils (the reverse of in-floor heating tubes).

   IMO, it's a lot easier and much less expensive to just let Mother Nature do the job for you. All you have to do is to give her access to the underside of the roof deck. She'll do the rest.

    

   Dinosaur

   How now, Mighty Sauron, that thou art not broughtlow by this? For thine evil pales before that whichfoolish men call Justice....

1. 
   Terrygoud | Jun 25, 2008 11:38pm | #7

   Thanks (to all) for your quick replies! Starting to sound like it is best to be conservative and provide the under-roof ventilation, no matter what the insulating material. But that foam method sure does sound attractive/simple...
   So, maybe as a last query, has anybody "out there" ever done the under-roof-no-venting-foam-application and been consistently successful? Thanks again to all,
   Terry

1. 
   andy_engel | Jun 25, 2008 11:52pm | #8

   I've seen it done many times. But my climate may be milder than yours.One approach is to nail 1x1 slats to the rafters, tight to the roof deck. Fasten 1/2 in. rigid foam to them to create a vent chute, then foam the cavity full. Best of both worlds, but more work.BTW, lots of reasons not to use fiberglass that's exposed to cold air. Mainly because its R-value decreases as the Delta-T increases in that circumstance. If you've got FG in an open attic, at least cover it with Tyvek.Andy

   "Never try to teach a pig to sing. It wastes your time and annoys the pig." Robert A. Heinlein (or maybe Mark Twain)

   "Get off your dead #### and on your dying feet." Mom

   "Everything not forbidden is compulsory." T.H. White, The Once and Future King

2. 
   Dinosaur | Jun 26, 2008 07:14am | #10

   sound like it is best to be conservative and provide the under-roof ventilation

   In our kind of climate, the conservative, tried-and-true approach is the way to go. That said, I want to be clear that Andy Engel's opinions should never be lightly ignored. However, where he lives, an extreme cold-snap means it went down to -15ºC overnight; a foot of snow will close the schools for a week; and the maple-syrup run happens in early February, LOL. Andy is riding his lawnmower and wearing shorts while I'm still using the snowblower and wearing longjohns.

   The other suggestion someone else made--to use ice-and-snow membrane--is a good one...but the membrane should not be thought of as an alternate to roof-deck ventilation, but as an additional weapon in the war against ice-dam roof-leaks. As I mentioned in my first post, not all ice dams are the result of interior heat escape through the roof. Solar heating can put more than enough ice on the southern eaves of the best ventilated roof to cause trouble--even on the roof of an unheated outbuilding.

   Finally, one should remember that ventilating a roof does more than just help prevent ice dams; it also prevents condensation frost build-up. When spring comes and the attic warms up, having a nice, thick coating of frost melting off the underside of your roof isn't something you really want going on....

   My suggestion for how to do the ventilating: nail 2x2's along the rafter/roof-deck junction on both sides of each rafter, then stick rigid foam panels or Black Joe up against those and foam them in place along the edges. That'll give you 1½" of clear ventway above the panels in each rafter bay. Make sure there is a clear path for outside air from the soffit entry vents through the rafter bays up to unheated part of the attic and out via the ridge turbine(s) or gable-wall vents.

   Dinosaur

   How now, Mighty Sauron, that thou art not broughtlow by this? For thine evil pales before that whichfoolish men call Justice....

3. 
   CapnMac | Jun 26, 2008 09:32am | #11

   has anybody "out there" ever done the under-roof-no-venting-foam-application and been consistently successful?

   Yes, about 20 miles south of my house.  Retrofit job.  4" closed-cell on deck and minimum 1.5" on roof framing.  This will be the second full summer tracking temperatures and stat settings and a/c use.  Have a bad previous year in that the stat settings were not held constant (poor a/c balance during that first year).  Attic peaked about 89º last year.  Only hit 84º this year so far (mine, not insulated at all, has gone to 131º for a peak already.

   Now, bear in mind, this is Central Texas, only ice dam to worry about here is in the ice maker . . . Occupational hazard of my occupation not being around (sorry Bubba)

2. 
   DickRussell | Jun 25, 2008 11:56pm | #9

   "Not sure I follow your description about pushing the boundary 'outward',....."If you have, say, 70 F inside the house and -40 outside some bitter night, for a 110 F temperature difference, then somewhere between inside and outside it will be at 32 F. If that point is within the insulation, then all of the snow on the roof will be colder and won't melt.70-32 is 38 degrees, and the overall difference is 70 -(-40) or 110 degrees. So the freezing point will be at 38/110 or 34.5% of the way across the total "R" of insulation in any form between inside and outside. With no snow on the roof, the freezing point would be around 3 inches into the foam or other insulation.Assuming you have 6 inches cc foam in those 2x10 bays, at R6/inch, the insulation would be R36. For the roof deck to be at 32 F when it's -40 F outside, you'd have to have 36/.345 or a total of R104 up there. Subracting the foam insulation value, 104-36= R68 worth of snow piled up there.On the other hand, if the rafters themselves aren't covered with insulation, that 9.5" of wood, plus the sheathing, gives something like R12. R12/0.345 gives total R35; subracting the 12 gives only R23 worth of snow on the roof for there to be little strips of roof surface where there could be melting. Those numbers aren't that great, because there would be some lateral temperature gradients, not just in the direction perpendicular to the roof surface.But you get the point that the less insulation under the roof you have the easier it is to have the pile of snow on the roof actually melt at the roof deck and not just conduct the heat away to the air without melting.Edit: I looked up the R value of snow. Wikipedia says about R1 per inch. I imagine it is lower for more dense snow than for light fluffy stuff. But that R1/inch would mean that with 6" of CC foam, when you get to around six feet of snow piled up on the roof you would start to get melting at the roof surface.With only two feet of snow up there, I calculate that the roof deck above the insulation would be about +3 F to either side of where a rafter lies, with perhaps somewhat warmer temps right above a rafter. so perhaps even then there wouldn't be melting above the rafters either.My take on the OP's dilemma is that if he uses foam between the rafters, right up against the roof deck, he needn't have any venting under the roof deck. But he needs to be sure to have Ice & Water at the bottom edge and in the valleys. He also needs to think about how much space he has for insulation where the roof meets the walls. If this house was built without "energy heel" trusses, those corners are cold spots, and some melting could occur right there.

   Edited 6/25/2008 5:16 pm ET by DickRussell

2. 
   andy_engel | Jun 25, 2008 11:05pm | #4

   It depends on the climate, the level of insulation, and the interior temperature. There's no one right answer to this question. Both vented and unvented attics can work. I've seen ice dams on unheated barns. I've seen a lot of vented attics with ice dams. I've seen SIPS roofs with no ice dams. You just need to design the house for the climate.Andy

   "Never try to teach a pig to sing. It wastes your time and annoys the pig." Robert A. Heinlein (or maybe Mark Twain)

   "Get off your dead #### and on your dying feet." Mom

   "Everything not forbidden is compulsory." T.H. White, The Once and Future King

1. 
   Dinosaur | Jun 25, 2008 11:07pm | #6

   I knew you'd be along for this ride, bro!

    

   Dinosaur

   How now, Mighty Sauron, that thou art not broughtlow by this? For thine evil pales before that whichfoolish men call Justice....