Stopping radiant roof heat
We are re-roofing a summer community hall (by committee, ugh) and have major radiant heat flow from the open roof during hot/sunny summer days. I am pushing to install insulation to block this heat, but there are concerns being brought up by other members that I need some help with.
Background:
Summer use only. Building is never heated. Located north of Boston on an Island.
Building is 1920 vintage, has open studded walls and ceiling, with unevenly spaced rafters, and no current soffit or roof vents, just one trap door roof vent to let hot air out.
Proposed method to insulate is to strip 4 layers of existing shingles (ugh again), lay tar paper over the existing 1×10 roof deck boards, install 2×4 sleepers horizontally (to provide easier nailing for plywood), then install rigid foam insulation (XPS or ISO) between the sleepers to cut the heat transfer, plywood the entire surface and shingle.
Concerns:
1. Blocking the heat will over-heat the new shingles and shorten their life. (I think I have found articles that say this is not true)
2. The rigid foam will allow moisture to be trapped under the new roof decking. ( the building is not occupied or heated ever Sept – June)
One member says we will have to ventilate the new space costing way more than we can afford.
Will using rigid foam in a sandwich fashion like this create moisture or shingle life problems? Are there other ideas as to how to block this heat transfer so we don’t bake inside on sunny days? We want to keep the open rafters inside, so are hoping to insulate outside the existing roof deck.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
The first thing to comes to mind is to change the sequence of layers you are figuring on anyway. I'd be inclined to put the layer of foam directly over the existing roof deck, then the 2x4 sleepers running vertically, screwed through and into the rafters and covered by the envisioned plywood second deck, with soffit and ridge venting to provide a vent channel from bottom to top above the foam layer. The plywood second deck then would be covered with the asphalt felt and shingles applied to that. This gives you a "hot roof" design, with ventilation between the shingles and foam to carry out a lot of the solar gain above the foam layer. This sequence also provides for the future, when the building might conceivably see a different, non-seasonal use requiring some level of heating for at least the shoulder seasons.
Thanks for the reply Dick. The problem is, your recommended design is the ventilated design that is estimated to cost an additional $10,000, which we can't afford. The vertical 2x4's lined up with the uneven rafters would require each sheet of plywood to be custom cut. So if we skip the air gap and just make a foam sandwich between the sleepers, do you think there would be a moisture problem from condensation if the building is unheated.?
I don't see how the slight change that Dick Russell is proposing would cost an additional $10,000. You are already talking about sleepers, foam, and new plywood, and so is he, just in a different sequence. How big is this roof, and what is the pitch?
Placing the foam directly on the roof deck in continuous pieces butted tightly together is the best approach, IMO. I'm not sure I would run sleepers across the roof, as you describe, since you get relatively few attachment points, but it may be OK. You do not necessarily need the sleepers over the foam--you could attach the new plywood directly to the foam with long screws into the existing rafters. Be prepared to add trim to the edge of the roof to cover the additional thickness.
There is some benefit to the air space you gain with the sleepers over the foam--you get the air movement under your new deck, and you get some benefit from the foil facing on the insulation, assuming you use foil-faced. Over time, dust will build up on the foil, and this benefit will diminish. You could also use foil-faced OSB decking, with the foil facing down into the air space, for some small additional benefit.
If solar gain is the main issue, use shingles with a high solar reflectance index. It may be easier to find these in Florida than it will be in Boston, but they are out there. Google "roofing SRI" and get familiar with the rating system. Simply buying shingles without taking this into account could be a huge missed opportunity.
Dick Abbe,
I also agree that the rigid foam should be in a continuous layer, not cut up into thin strips. The 2x4 strapping should go on top of the rigid foam, as Dick Russell suggested. However, if the rafters are spaced at uneven intervals, I see no reason that you can't install the 2x4s on 16 inch centers or 24 inch centers, parallel to the ridge (like purlins). Of course, you won't get soffit to ridge air flow -- but you'll get most of the benefits of the air space, and the lack of air flow shouldn't cause any problems.
Rigid foam will have only a minor affect on shingle temperature -- not enough to worry about. Roofers install asphalt shingles on SIP roofs every day of the week.
A layer of rigid foam will not trap moisture. The underside of the roof assembly will still consist boards exposed to the interior; these will dry rapidly toward the interior.
If you don't want to build a sandwich on your roof, you can just install Cool-Vent vented nailbase from Hunter Panels: Cool-Vent.
Thanks Dick, David and Martin. This feedback is very helpful. The roof on the building is about 15 squares and I will guess the angle to be about 30 degrees. I will check into the SRI shingles and Cool Vent. One of our problems is the need to be low budget. The building is only used 4 hours per day and some are saying "let them bake" instead of spend money. But now is the time to fix this problem.
The latest discussion (last night) is the possibility of installing powered solar vents to push air through the building continually. This will help to exhaust the hot air that currently moves slowly out of an 18 x 24 roof hatch, but will do nothing for the radiant heat coming from the underside of the roof deck. I have used rigid foam on my cottage roof and it has made a huge temperature difference, so I'm kind of stuck on that solution. Now I'm wondering if high SRI shingles and powered roof vents would be a lower cost "good enough" solution. I'll check out how much heat SRI shingles reflect. Any other suggestions are welcomed for this ongoing discussion. Thanks for your help.
I'm back again with some more information after reading about the SRI and cool roof. The existing roof has 4 layers of shingles (the low budget approach for 80 years), but... the top layer is white. So there is some reflectance, but the shingles are definitely "aged" so the reflectance is lower. Now I'm wondering how much new white high SRI shingles will help versus what we have, and for how long. Do any of you know if the current high SRI shingles are much more reflective than white shingles from 30 years ago?
Plus, I know the thick layers of shingles absorb a lot of heat energy, store it, and then transfer that energy into the building for a long time. So removing the layers will at least enable the roof to cool faster when the sun sets.
Sorry for this scattered discussion, but I do appreciate your input.
In my market, a 15-square 4-layer tearoff and re-shingle would cost about $7,000. One of your posts above refers to ADDING $10,000 by adding the venting. This makes me question the numbers you are getting. You mentioned "island" and I live on one, so perhaps you have some extreme circumstances like using a landing barge, small boats, hand-carrying all materials, etc. I know that stuff can add to the cost. In any case, in most circumstances, a 15-square re-roof is very manageable. I was wondering if you might be in the 50-100 square range based on your initial posts.
On the SRI of the old shingles... I have no idea. Hopefully someone else will chime in. White is certainly better than gray.
There are many colors of high SRI shingles, and the white/off-whites can look pretty shabby in short years from algae growth, or even the streaking of seasonal pollen deposits. Search the CRRC listings for shingles suitable for high-angle roofs- they don't have to be very reflective to do MUCH better than the traditional charcoal spectrum shingles.
The shingle-life over insulated roof decks is a complete non-issue, particularly on roofs with pitches greater than 2:12. Almost all shingle cooling even on an un-insulted vented roof deck is re-radiated heat back at the sky plus convection to the outdoor air, not through the ~R1 underlayment & roof decking to be convection cooled on the interior side. On a 4:12 roof in a New England climate you're taking no more than a year or so off the life of a 25 year shingle, independent of it's SRI. But higher-SRI shingle DO run cooler in any side-by-side comparison to low SRI shingles. Both average and peak temps affect shingle life, but it's far less of an effect than the manufacturers were thinking 25 years ago, particularly on the snow-sloughing roof angles commonly found in New England.
But having a vented nailer deck is more resilient than the sandwich approach, since the nailer deck can dry into the ventilation gap- the inevitable eventual wind-blown moisture or snow-drip getting by the shingles can leave in a reasonable amount of time. Hunter, Atlas an several other vendors have 4x8 panelized versions, but depending on your labor costs and the simplicity of the roof lines it's often cheaper to run DIY with standard iso panels, furring, and an OSB nailers.
Active ventilation brings the peak roof temps down, but the efficiency of the resulting cooling is usually lower than actual air conditioning. (A 3/4 ton mini-split does a lot more for you in mid-day than an 800W fan, and it's quieter too.)
There are several vendors dealing in reclaimed roofing foam (at about 1/4-1/3 the cost of virgin stock), which would allow you to go higher R at low cost. (Check the craigslist materials section for "rigid insulation", or see insulationdepot.com.)
To limit the roof deck from radiating onto the heads of people below a spray-on radiant barrier type paint would provide relief that you can feel, even if the affect on the sensible-temperatures & active air conditioning power use would be quite low. (The matte-silver look of it might not make it through the interior-decor committee though. :-) )
Thanks for your input Dana. All good information to consider. This is an off grid Island with limited solar, so some fans yes, mini splits probably not. One of the latest ideas is to shade the roof with solar panels, and use the power to replace propane refrigerators and freezer with electric. A much more expensive option, but with possible payback in 20 years or so. In any event, SRI shingles seem to be a no brainer. Thanks again. PS: Silver would definitely not make it through the committee!
Dick,
What would happen if you just drywalled the ceiling and added continuous eave and ridge vents, thus relying mostly on the stack effect ventilation to remove the heat gain from the rafter space?
The radiation coming from the bottom of the roof deck would heat the topside of the drywall, but the ventilating air would cool the rafters and top of the drywall by conduction, thus reducing the ability of the drywall to conduct the heat through its thickness and re-radiate it to the interior space.
If you used foil facing on the top side of the drywall, you would further reduce its ability to absorb heat radiated from the bottom of the roof deck while still permitting the ventilating air to conduct heat out of the drywall.
If grid-tied and net-metered to an expensive-power island grid the PV shade thing has LEGS. In MA there is a lot of subsidy for grid-tied PV, including ongoing solar production credits. "Payback" is well under 20 years for most MA grid-tied PV at current ~$4-5/peak-watt installed price.
Payback may be much longer than 20 years if it is totally off-grid.
Ron: You simply can't make an appreciable dent in peak roof-deck temps with convection cooling alone- it has to be CRAZY shingle-melting hot to get sufficient cfm to get even a 3F reduction out of stack effect cooling. (And 3F lower than crazy hot is still crazy hot.) Even with venturis or turbine vents it's still far more about drying rates than it is about peak roof deck temps.
Even 3/4" drywall is less than R1, and radiant heat transfer from the roof deck to the topside of the drywall will keep them at pretty much the same temperature, no matter what the rate of convective flow, even with an aluminum facer. The aluminum may reflect more than 95% of the radiant flux back up to the roof deck which only raises the temp of the roof deck to where it's in equilbribrium, but it's still looking at R1 of deck & shingle to get the heat out. The aluminum facer can't radiation-cool itself, and the cooling air flow will still be within 10F of the roof deck temp. The gypsum will still cool primarily by downward radiation.
Aluminized polyester type radiant barrier in place of the gypsum would have a measurable effect on the peak radiated heat sensed at the ground level though, somewhat independent of the temperature of the radiant barrier itself, and would be a far better solution than a drywall air barrier for a slanted convection stack.
This stuff has been studied to death by academics, DOE/Nat'l Labs, and utilities alike since the energy price shocks of the 1980s. There's a ton of data archived online at Texas A & M, as well as the Florida Solar Energy Center- it's mostly smoke & mirrors, but there is some benefit. It takes more than one low-E surface to get any appreciable benefit out of it, but with a mid-to-high-SRI shingles, low-E paint on the under side of the deck, and double-sided radiant barrier stapled to the under side of the rafters (with or without soffit-to-ridge venting in the resulting rafter channel) there will be a significant reduction in the radiant heat transfer from the roof to the floor level where the humans swelter, and at peak solar-gain periods that effect can be the same as adding more than R10 above the roof deck- it's not nothing. If it's cheaper than adding R10 above the roof deck, it can be the right way to go.