Insulating and air sealing attic in a 1930s colonial
I recently moved into a brick colonial in the DC area built in 1935. It’s heated by a boiler through radiators and has a 25-year-old A/C unit in the attic. It has a slate roof and gable vents, but no soffit. The blower door test showed that it’s quite loose, which is no surprise.
My home performance contractor has suggested two routes: encapsulate the attic or pull up the floor (90% coverage) to air seal and insulate. Based on reading here, it seems that spray foam under the roof deck would need an air gap for the slate roof, which doesn’t sound great to me. Despite the headache of dealing with the flooring, my current thinking is to remove the A/C after the cooling season ends, remove the floor and seal everything up. Then add bath fans, insulate and replace the floor. Install a new A/C. Finally, put in an aluminized fabric radiant barrier to reduce heat loading on the A/C and ducts. The ductwork would get R-8 insulation.
We want to keep the attic floor. What would be the best type of insulation for under the floor, given space constraints? Should I be considering encapsulation more seriously?
Thanks for any insights,
Jeremy
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Ducts above the insulation layer suck, even if the ducts are insulated. It's difficult to make all of the ducts and ceiling penetrations reliably air tight for the duration. The AC is probably oversized, and is nearing the end of a typical lifecycle, and it may be easier & cheaper to just yard it all out and go with a ductless or mini-ducted cooling solution than to spend serious money insulating at the roof deck layer.
I assume the slate is installed on skip-sheathing? If yes, slate on skip-sheathing is inherently ventilated to the exterior, and doesn't always need an additional gap below, but you DON'T want to spray the foam directly onto the underside of the slate itself. Installing a mesh type rainscreen material (eg Obdyke RainSlicker) tightly onto the under side of the skip sheathing and spraying the foam onto the mesh may work, but it might not. More info would be useful here.
How deep are the rafters?
How deep are the floor joists?
Are the rafters & joists on standard 16" or 24" spacing, or something else?
I did exactly what Dana suggested. My attic (which had been converted to living space) was also poorly insulated, old AC unit in the kneewall area, probably 30 years old. No insulation in many parts of the attic because the AC and ducts were blocking access to parts of the rafters. I put in ductless heat pumps/ AC. Removing the AC unit allows me to airseal and insulate properly. To me, it made little sense to put another AC and dutcwork unit into the attic. Instead, I have minisplits on first and second floor and I will just put a window AC in the attic for cooling. The attic gets heated from below in the winter with minisplits. Dana helped with an estimate of the heat load based on prior heating data of natural gas boiler Thank you very much for your help Dana ! The upper end of your estimated range of heating load seems to be where I came out based on actual electricity usage of heat pump on the coldest day and an assumption of the COP of the equipment. Matthias
Thanks, Dana, for the response. The rafters are 6" and 16" OC. The joists are 6" deep. The roof deck is individual boards with minimal spacing, not skip sheathing. It looks like there's felt under the slate.
The auditor said that the ducts were sealed well enough that he didn't recommend additional action. It's a definite drawback to have the A/C in the attic along with the ducting. I've only gotten one HVAC quote so far, though going to ducted mini-splits (one per floor) was really expensive before carpentry and finish work. I will get additional quotes, though a big cost gap won't pay for itself. We used 770 kWh over a hot July and stayed comfortable once we figured out the right set point to manage humidity (via going to a Wi-Fi 'stat). Incidentally, the A/C doesn't seem to be oversized (2.5T), though it's hard to know how close it's operating to its nameplate 12 SEER.
By the way, I stuck a remote temp sensor in the attic with datalogging up there and in July would see temps peaking at 115 F at about 3 p.m. in full sun. The majority of the roof doesn't get shaded in the summer.
Without a careful load calculation or direct measurement of the duty cycle at or near the 1% outside design temperature you have no idea whether the 2.5 AC is perfectly sized or 3x oversized, and both are possible here. All you really know is that it keeps up with the load. If you're the type that likes to measure things (sounds like you might be, if you're data logging the attic temperature) measuring the duty cycle of the AC doesn't take a lot of expensive equipment, and you may already have equipment up to the task of logging duty cycle against outdoor temperature. See:
https://www.greenbuildingadvisor.com/blogs/dept/building-science/how-tell-if-your-air-conditioner-oversized
The 1% outside design temperature in the DC area is about 90-91F (yes, I know it gets hotter than that in DC), so the best time to measure the duty cycle for these purposes is during hours when the outdoor temperature is running from 88-92F.
With ducts and air handler in the attic at least some of the load is duct gains, not directly related to the design load of the conditioned space. With air leaks between conditioned space and the outdoors (including a vented attic) some of the load is air-handler driven air infiltration, since no duct systems are perfectly balanced. When you bring it all inside conditioned space and seal the air leaks in the building envelope that portion of the load goes away.
Nobody is suggesting there is any "payback" to replacing the AC system before it's really necessary. The issue is the relative cost of insulating at the roof deck vs. the cost of ductless cooling + insulating at the attic floor. The other issue is that at age 25 the existing system may still work, but it could crap out tomorrow, or go another 10 years. Guessing how much maintenance/replacement cost to factor into your financial analysis clearly has big error bars.
Ducted mini-splits may or may not be necessary to achieve comfort. With ductless heads in only the high-load rooms it's enough to keep the indoor air dry even with fairly low convective air exchanges with the rooms without ductless heads. A room-by-room load calculation can be useful here. A 3 head 1.5-2.5 ton multi-split can handle quite a number of house configurations. Mini-duct cassettes can often split the output between a few lower-load rooms without much duct work or carpentry, but whether that works is very floor-plan dependent.
Are the rafters & joists milled 2x6 (5.5" deep x 1.5" wide), or are they full dimension 2x6?
All lumber is milled.
Regarding load calcs, I intend to have them done. I was just responding to your earlier comment: "The AC is probably oversized..." Out of curiousity, I pulled a graph from one of the hottest days last month. The unit works pretty hard when there are back-to-back 90+ degree days. Our setpoint is 77, though it looks like I raised it that day briefly, or the thermostat thought no one was home.
https://www.dropbox.com/s/hn2zl2o4bpaikfv/AC%20duty%20cycle%202017-07-19.PNG?dl=0
Regarding payback, I was considering the cost difference between a box swap and going to mini-splits. (We are replacing it in the next year, even though it's still working.) I'm having a hard time seeing the marginal utility in spending many thousands more for mini-splits, even if it's technically the better solution. I'm at the front end of this process, so there's a lot more info to gather and analyze. I do like the idea of mini-splits and of getting stuff out of the attic. (Well, and the basement, too, perhaps.)
PS: Here's the temp chart for the attic on the same day:
https://www.dropbox.com/s/cyglnwbic4ofoqm/attic_temps_2017-07-21.png?dl=0
It's the extremely high comparative cost of insulating properly at the roof deck vs. the floor that you're trading the cost of a mini-split to, not the cost of swapping in a new standard split system.
To hit an IRC 2015 code-min R49 on the attic floor means installing a set of 2x10s perpendicular to and on top of the 2x6 joists, topping it up with 15" of low density cellulose and putting down a new floor (or re-using the planking that's there). The cellulose would be on the order of $1.50 per square foot, and the framing & flooring aren't ridiculously expensive, call it $3-5 per square foot, depending on complexity.
To hit code-min at the roof deck starts ab $7-8 per square foot just for the insulation, and it could easily go over $10/foot, and it's more square feet. It's a BIG difference! There may be cheaper ways to get there with a lot of sweat-equity, provided it's possible to get 4' x 8' sheets of foam board up there.
That is VERY helpful, and helps me understand the high level parameters of the two options. Thanks for taking the time to respond. There is quite a lot more roof deck area, for sure.
In the intervening months I've made some progress and now have more questions(!). A load calc was performed and I replaced the A/C with another 2.5-ton unit (but 16 SEER). I was a little hesitant on the load calc, but ended up learning a lot from it, especially for future heating system changes.
After considering all kinds of scenarios for insulating and air-sealing the attic floor, I've all but decided to put SPF on the roof deck and masonry gable walls. The insulation itself is much more expensive, but it solves a few problems and lowers costs in other areas. The biggest being HVAC. The cost to move the AHU and ducts into living space would be FAR greater. I liked the idea of a ducted mini-split per floor, but it's just not cost-effective for my scenario. The other is that sealing the attic door is nearly impossible because of the design of the sliding, non-folding stairs. It also means that I can stick an ERV in the attic later, along with some other potential improvements.
The current proposal is a full baffle (~1" airspace for slate roof) open-cell SPF and DC315 intumescent coating as the thermal barrier. My two main questions are:
1) Do I need the baffles with open-cell since my slates are installed over tar paper? (i.e., no skip-sheathing)
2) Do I want/need a vapor barrier? I take it from Building Science Corp that I don't need one in CZ4, though my goal is to keep my RH up in the winter. Is that applicable, or should I only be thinking about structural damage? I will add a supply and return to the A/C ductwork in the attic, though there will be no direct heating from the boiler.
I'm leaning toward open-cell and painted drywall (instead of DC315) for the vapor retardation. Closed-cell would still require DC315 as a thermal barrier, so I wouldn't be ahead cost-wise with that route.
Thanks for any insights.
Jeremy,
It sounds like your budget is adequate for above-average details, and that's good.
Q. "Do I need the baffles with open-cell [spray foam] since my slates are installed over tar paper (i.e., no skip-sheathing)?"
A. Considering all of the stories of moisture problems in sheathing when open-cell spray foam is installed on the interior side of roof sheathing, I would say yes, you certainly should include the baffles and the ventilation channel. For more information on the problems I'm talking about, see these two articles:
High Humidity in Unvented Conditioned Attics
Open-Cell Spray Foam and Damp Roof Sheathing
Q. "Do I want/need a vapor barrier? ... I'm leaning toward open-cell and painted drywall (instead of DC315) for the vapor retardation."
A. You don't need an interior vapor barrier, but I would recommend the use of an interior vapor retarder. If you can afford drywall with vapor-retarder paint, that is by far the best approach.
Thanks, Martin. That helps me get oriented. And I see that I'd been using the terms vapor retarder and barrier imprecisely. I'm getting a utility rebate to help with ASI and HVAC upgrades, though it does seem like a lot of money.
The "High Humidity ..." article was illuminating. Thanks for providing that among other high-quality resources on your site. Without this sort of info I could have easily made a decision that would have had nasty, expensive consequences years down the line. Seems like closed-cell might be the better route, except that I should still have the drywall for fire reasons, making the combo really expensive. I'm going to see if the insulation contractor has an encapsulation scheme that doesn't use SPF (hopefully cheaper!).
Any suggestions on alternatives to consider?
Jeremy,
Concerning alternatives: You might want to read the article titled "Creating a Conditioned Attic."
Once excellent way to create a conditioned attic is to install an adequately thick layer of rigid foam on the exterior side of the roof sheathing. Of course, this approach requires new roofing (and usually a second layer of roof sheathing), and is therefore expensive.