GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon
Musings of an Energy Nerd

A New Look at Conditioned Attics

Bringing an attic inside a home's thermal envelope

If your builder insists on putting an air handler and ducts in your attic, the insulation should follow the roof line. The photo shows open-cell spray foam, a type of insulation that can contribute to damp roof sheathing in some climate zones. In most climate zones, closed-cell spray foam is less risky than open-cell spray foam. When installed on the underside of roof sheathing, open-cell spray foam is safest in climate zones that are both warm and dry. [Photo courtesy of PNNL / Addison Homes]

Thirty years ago, most attics were vented and unconditioned, so they were cold in winter and hot in summer. These temperature swings didn’t matter much, though, because the attics had insulation installed on the attic floor. The attics were outside of the home’s thermal envelope.

These days, builders can choose between a traditional unconditioned attic and a conditioned attic. Conditioned attics are sealed rather than vented. Since the insulation in a conditioned attic follows the roof slope, the temperature in a conditioned attic is close to the temperature in the rest of the house.

My last article on conditioned attics was written a decade ago. It’s time for an update.

The problem is attic ductwork

The main reason that anyone talks about conditioned attics is that some builders want to install ducts in the attic. Installing ducts in an attic is a bad idea, but many builders don’t know that. While locating ducts in an attic reduces installation costs for the builder, it increases energy costs for the homeowner.

Here are the main points to remember about attic ducts and attic mechanical equipment:

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial

30 Comments

  1. KurtGranroth | | #1

    Are there any resources/calculators that I could use to determine how much my attic ducts are costing me each year?

    I can say, anecdotally, that I feel a blast of super hot air out of the registers whenever the AC kicks in, but that only lasts for maybe a second or two. AC is blowing on average 10+ hours a day in the summer so that only happens a few times any given day. Given that, it's unlikely that it's that initial hot air blast making much of a difference.

    Instead, it must be the heat injected into the ongoing cold air stream, raising its temperature some amount. That's the bit that I'd like to calculate. When I measure the temperature coming out of the registers, it's notably cold -- 50-something F, typically. I'm assuming that that temperature should be even lower but for the attic heat on the ducts... but how much lower and how much does it matter?

    1. Charlie Sullivan | | #2

      We could estimate it from the length and size of the ducts and the insulation thicknesses on them.

      Allison Bailes has an example calculation in this article:
      https://www.energyvanguard.com/blog/the-invisible-problem-with-duct-insulation/

      But that's only the heat transfer through the insulation--likely there are also duct leaks contributing.

    2. GBA Editor
      Martin Holladay | | #3

      Kurt,
      The short answer is that there is no easy way to make this calculation, because the answer depends on the quality of the duct installation work. Ideally, the duct system is airtight, but the vast majority of duct systems in U.S. homes are not airtight. Some duct systems are a little leaky, and some duct systems are very, very, leaky.

      The next variable is the level and quality of the duct insulation. Some ducts have no insulation; some have bubble-wrap; some have R-2 insulation; some have R-4 insulation; and some have R-6 or R-8 insulation.

      The next variable is the quality of the register boot installation, and whether any attempt was made to insulate the register boots.

      Another variable is climate -- which affects attic temperatures.

      Two researchers who have attempted to quantify duct losses are Dave Roberts and Jon Winkler, engineers at the National Renewable Energy Laboratories. In a technical paper titled "Ducts in the Attic? What Were They Thinking?", Roberts and Winkler concluded that ducts in unconditioned attics waste about 20% of the output of a furnace or air conditioner. These researchers report that during peak conditions, the losses are even greater. Roberts and Winkler wrote that in Houston, Phoenix, and Las Vegas, “The average DSE [distribution system efficiency] for the three locations on the design day, which would be considered the day of the season when cooling demand is highest, is 72%. This means that on the hottest day of the summer, 28% of the air-conditioner output is ultimately lost.”

  2. James Howison | | #4

    I think an article on ways to install ducted mini splits in conditioned space with lots of pictures and dimensions would be great. I explored this but was constantly stymied by misunderstanding the service space requirements, ways filters could or could/not be installed and accessed.

    Sort of an idea book for people thinking of ducted mini splits :)

    Also, you don’t mention the bury the ducts in cellulose option, and for that could anyone outline what happens to the blower unit? Assuming it is horizontal in the attic, is it also buried? How does that work for service? Or is the assumption a vertical unit with a central return?

    1. GBA Editor
      Martin Holladay | | #5

      James,

      Q. "[For ducted minisplit units installed in an unconditioned attic,] could anyone outline what happens to the blower unit? Assuming it is horizontal in the attic, is it also buried [in insulation]?"

      A. For answers to this question, see "Multi/mini-split cassette buried under blown in insulation?"

      You're right that my article didn't mention burying ducts in insulation, but I did provide a link to my article called "Solutions to the Attic Duct Problem." If you click the link, you'll find a section called "Bury the ducts in insulation."

      Your suggestion about an article on ways to install ducted minisplits is a great idea. That said, I would probably not be the best author for such an article. In the meantime, check out these relevant GBA articles:

      "Getting the Right Minisplit"

      "Practical Design Advice for Zero-Net-Energy Homes"

      "Goldman on Minisplits"

      "Do Ductless Minisplits Work With Every Floor Plan?"

      "Can a MiniSplit Live Happily in the Attic?"

      1. James Howison | | #6

        Wonderful, thanks Martin!

  3. Jasiu77 | | #7

    Thanks for the update, Martin! This article is extremely helpful.

    Regarding spray foam under the sheathing, I've asked once before and wanted to see if my idea still makes sense with the latest information available. Our old farmhouse roof that has sheathing over slats, so I was thinking of putting XPS foam board (Foamular) between the joists (using can spray foaming around the perimeter) and then covering the entire thing with a thinner layer of closed-cell spray foam. This would reduce the cost and I think make future roof repairs easier, since the closed-cell spray wouldn't be attached directly to the sheathing. The slats also create an air gap between the sheathing and the foam board.

    Does that still seem like a reasonable approach, or any concerns?

    1. GBA Editor
      Martin Holladay | | #11

      Jasiu,
      Your suggested method is dubbed the "peanut brittle" method, and was described in my article, "Cut-and-Cobble Insulation."

      Whether this method is safe or risky depends on the quality of the spray foam job. You need an impeccable air barrier -- and the air barrier needs to be durable. Some cut-and-cobble jobs between rafters have failed, so the approach is somewhat risky for unvented cathedral ceilings. The method is less risky if the cathedral ceiling is vented.

  4. James Morgan | | #8

    Latest issue of Fine Homebuilding has an article showing a version of a vented roof assembly using fiber insulation which incorporates a continuous air barrier and that accommodates dormers, hips and valleys. Apart from the insulation thickness in the illustrated example looking a little lean, Martin, any thoughts?

    1. GBA Editor
      Martin Holladay | | #10

      James,
      This is a version of an old-school vented roof assembly, but the differences in the method -- for example, establishing the vent channels above the rafters and allowing airflow at hips and valleys -- are significant improvements on the old-school approach. Thumbs up!

    1. Expert Member
      Malcolm Taylor | | #12

      James,

      I keep circling back to that assembly as a really good solution to vented cathedral roofs, but can't get my head around having to tightly install the topside WRB with nothing to stand on.

      1. James Morgan | | #23

        Not a builder so I can't speak to the practicality of this but from the photographs it looks like they are stretching the WRB top to bottom then placing the 2x4 furring fastening as far as they can reach then sheathing their way up with further fastening of the furring as they go.

  5. Miles Smith | | #13

    Greetings,

    I'm in Raleigh, NC, climate zone 4A. I have the cursed gambrel roof, no ventilation at the eaves, and ductwork in the attic (furnace in the crawl). I've read so, so much at GBA, Buildingscience.com, and Energy Vanguard on the subject, much of it multiple times, but haven't found "the" answer. Because, I guess, there isn't a "the" answer. (For a gambrel, it seems the answer is often "good luck with that".)

    The house was built in 1972 and we bought it in 1989. At that time there was a PAV, two passive turbines, and the usual screen over louvers on the gable ends. The second floor exterior is shingled and has three dormers each on the front and back and we face south. My objective is to improve things, most particularly the fact that I have leaky (no doubt) and poorly insulated ducts in an unconditioned attic. The PAV and turbines are gone and when it was last roofed they put in a ridge vent, of course.

    I now have PV panels on the front roof so no interest in rigid foam on the roof deck. I can't abide the GWP of spray foam. I don't consider this house a great candidate for mini-splits because of all the individual rooms (4 bed and two bath upstairs). No vapor diffusion ports in CZ 4A.

    My current thinking: have the blown-in fiberglass and the unfaced rolls I put on top of that 30 years ago removed; also to be removed, the fiberglass wrapped around the branch ducts. Air seal the attic floor. I was considering replacing the straight runs of branch ducts with concentric flex ducts to get above the magic R-13 to bury the ducts. But now I'm thinking I'll find and fix the duct leaks then recover them with at least R-13 (worst case, two layers of something). Last step is to have someone blow in a bunch of cellulose (min. R-49).

    For about a year I've been running an experiment and have sealed up the gable vents and 80% of the ridge vent. I would probably have done it all had I not run out of tape and energy on that one cool night last June. I'm also logging the dew point in the attic and the run time of the HVAC (mostly part of my what-size-should-the-HVAC-be? project).

    I don't have much faith that this duct design was optimized for anything. Sadly, I don't have much faith that a new design and implementation would be done much better but would be disruptive and expensive.

    My only real concern is moisture on the underside of the roof deck (T&G 1x6) but I haven't seen any evidence of that since I started messing with things. (And Martin has said most of that comes from the house!) I have a new moisture meter but haven't gotten up there with it yet.

    Do we think this can work? Any reason not to seal the rest of the ridge vent? When the cellulose is blown in, any reason it shouldn't touch the roof deck (plus all that mess around the dormers)?

    1. GBA Editor
      Martin Holladay | | #15

      Miles,
      A few comments:

      1. I don't know what a "PAV" is. I went online and checked a few abbreviation dictionaries, but I didn't come up with a solution that made any sense.

      2. You wrote, " I can't abide the GWP [global warming potential] of spray foam." Fair enough. But remember that the problem only applies to some brands of spray foam. You should consider installing a brand with a climate-friendly blowing agent -- for example, Demilec’s closed-cell Heatlok HFO spray foam. More information here: "Demilec Spray Foam Named Year’s Best Green Building Product."

      3. It sounds like you intend to bury your ducts in insulation. To do the job safely -- in a way that avoids the potential for condensation problems -- you need to replace your current ductwork with insulated flex duct rated at R-8 or better. For more information, see "Burying Ducts in Attic Insulation."

      1. Expert Member
        Dana Dorsett | | #21

        >"1. I don't know what a "PAV" is. "

        From this statement I'm guessing Passive Air Vent (distinguished from a wind turbine vent or powered vent) but perhaps Miles can enlighten us further:

        "At that time there was a PAV, two passive turbines, and the usual screen over louvers on the gable ends."

      2. Mark Hays | | #25

        Miles could also solve his HVAC duct problem by (1) Carefully seal all of the existing connections to block air leaks. 'Painting' on a generous coat of HVAC mastic is fairly quick and easy. I've also found that the long straight pieces of metal duct leak some air along the seam, which simply snaps together. This is easy to fix with a strip of foil HVAC tape on top of the seam. (2) Wrap the ducts with fiberglass duct wrap, e.g. MasterFlow, which is available in R-8. Seal all of the joints and seams with foil HVAC tape. I wrapped ours with two layers for ~R-16. One additional R-8 layer is also a good idea with pre-insulated flex duct. Also seal and doube-wrap the duct boots above the registers. (3) This will provide the insulation needed to allow the ducts to be buried in cellulose, to bring the entire attic up to an efficient level. (4) To help reduce humidity in an unvented attic, it is a good idea to adjust the runs to add one 'return' placed at the peak of the attic. (Or more than one, depending on how the attic may be divided and how large it is.) Note: It is also a very good idea to wrap your plenum with fiberglass duct wrap - and seal it just like the ducts. These large metal boxes are often poorly insulated (if at all), and should be.

        I did all of the steps noted above with our 1950s era home in Massachusetts. The improvement in HVAC performance was noticeable at the registers and in our energy bills, and the house earned a Pearl Platinum certification prior to sale -- based on these and other significant upgrades. This certification was worth ~$60K, so the upgrades saved money while we lived there -- and paid off again when we sold the house.

        Back on 2006 we had the underside of the roof deck and gable end walls sprayed with ~8" thick open cell foam. This was prior to information about potential moisture problems with open cell. When the roof was replaced, we found no moisture / rot issues. This may have been because the roof deck was 'old fashioned' 1x boards --not modern OSB which is much more sensitive to moisture.

        1. GBA Editor
          Martin Holladay | | #27

          Mark,
          My only comment is to clarify that your recommendation on the installation of a forced-air system return near the peak of the attic only makes sense for unvented attics -- as you correctly noted. All your other recommendations (concerning burying ducts in cellulose) refer to techniques usually used for vented attics, not unvented attics -- so the combination of advice is potentially confusing.

        2. Miles Smith | | #29

          Thanks for your story, Mark. That's essentially what I'm planning to do. I'm been pondering whether I think it'd be easier to replace the straight runs with flex duct or first air seal then insulate the existing galvanized ducts. And in any case the boots have to be covered and it all has to be air tight. I'm getting a little old and a little arthritic so I don't relish the thought of any of it. But hire somebody to do it to my "standards"? Anybody else have his self-imposed problem?

          I need to reread the "burying ducts" article as I had in my mind that I need R-13 to avoid potential condensation problems; but maybe only R-8. I'm one county away from CZ 3A plus I always want to over do things a little . . . but a whole second layer of insulation is like, twice the work! I'm not intent on burying the ducts but to get to R-60 in the attic, most of them will be up to their necks in cellulose.

          Thanks for the product name, too.

          1. Mark Hays | | #30

            Dear Miles: If you only need R-8, then R-8 insulated flex duct would be the easiest solution. A caveat: Flex duct should only be used for straight runs and very gentle curves. All 90's should be metal duct elbows. Ideally it should be pulled relatively tight on straight runs between metal boots and metal branch connections. HVAC installers are notorious for making knots of flex duct because it is simple and easy to install, but this creates back pressure and robs your HVAC system of efficiency. See examples in the photos below.

            When you install flex duct, use a band clamp to securely fasten the inner plastic liner (which has the metal wire embedded) to the metal boot or connector. You can hold the plastic sleeve in position with a couple of pieces of foil HVAC tape, as you tighten the band clamp. Then air seal that connection; foil HVAC tape is the quickest and easiest option. Then pull the outer plastic sleeve and insulation over the inner one and the end of the metal boot/connection, and use foil HVAC tape to fasten it and air seal it on the boot/connection. (Never use silver fabric 'duck' tape.)

    2. Miles Smith | | #24

      Sorry, guys; certainly wasn't trying to stump the band. PAV refers to a powered attic ventilator, a plain old thermostatically-controlled attic exhaust fan. I picked up the acronym somewhere in GBA and/or energyvanguard.com. Obviously, that was a little obtuse.

      Any comments on my last two questions?

      I so appreciate the years of effort Martin, Dana, and others have put into helping us nerd homeowners. Also informative to see below the intent to raise attics to R-60.

      1. GBA Editor
        Martin Holladay | | #28

        Miles,
        Thanks for the explanation. I continue my lonely fight against the tide of confusing abbreviations and acronyms, as I advocate for spelling things out -- in hopes of achieving every writer's goal: communication.

  6. Tim Welch | | #14

    Great article!! In defense of builders who put HVAC equipment/ducts in the attic (I have that!): the builder saves money, but the homeowner also benefits by gaining interior space and avoiding bulkheads where duct work is placed. That said, I’m not a fan of the attic HVAC system, though it introduced me to building science so I guess there’s a silver lining. ;-) Over time we have removed one whole HVAC system, 3.5 tons of AC, and 100k BTU of gas heat.

    In our first winter in the house (2002) we experienced severe ice damming. The builder informed me that it was “normal.” I knew it wasn’t “normal” and it took me several years of study, a failed experiment with attic exhaust fans (since removed), air sealing, and repair to the leaky duct system to solve the problem. The fix included plugging literal holes in the duct work where the HVAC contractor relocated takeoffs, but just left the holes and covered them with insulation! Jaw dropper #1 in my10 year attic odyssey. I was expecting to seal connectors, I was not expecting to find 5-6 intentional 6”x10” holes in the duct work!

    I have attached a picture of a solution I installed to address the duct work in the attic. Our attic includes two finished rooms – the picture shows the unfinished attic area with the roof to the right and one finished room to the left. The barrier foil to the right keeps the radiant heat between the joist runs and out of the attic, this keeps the heat off the HVAC equipment and at worst our vented attic is no more than 10 degrees warmer than the outside temperature. I also put the barrier foil on top of the attic insulation which closes and completes the insulation assembly to prevent convective looping through the insulation. Barrier foil is relatively inexpensive and I worked for free.

    Attic jaw dropper #2: While I was installing the barrier foil I noticed that while the house had ridge vent devices visible on the exterior of the house, the roofer had not cut the roof at the ridge to vent the attic. OMG!! That’s the one that really bothered me – somebody actually installed a ridge vent on top of a ridge which had no air vent.

    1. GBA Editor
      Martin Holladay | | #16

      Tim,
      Thanks for sharing the two jaw-dropper stories. Since I used to inspect homes for a living, your stories don't surprise me. Every imaginable omission or screw-up is possible.

      Your story about intentional holes in your attic ductwork is (unfortunately) fairly typical. I've seen ducts that were totally unconnected to any register boot -- just blowing air into the attic whenever the air handler blower comes on.

  7. Ra Van Dyk | | #17

    Great article, Martin. Thanks!

    My mind isn't always so sharp this time of day, but I'm wondering if you meant to say "conditioned" instead of unconditioned in the following paragraph:
    "To create an unconditioned attic, good air sealing is vital. You don’t want to leave any cracks where outdoor air can enter the attic or leave the attic. The result of this air sealing work will be an unvented attic."
    I didn't see any mention of rigid wood fiber insulation, such as Steico, Gutex, or TimberHP, as a vapor permeable and lower carbon alternative to rigid foam in roof and wall assemblies. I've been pleased with the ease of installation and clean-up, and I don't worry about condensation build-up like I would with foam.

    1. GBA Editor
      Martin Holladay | | #19

      Ra Van Dyk,
      You're right, of course, about my accidental use of "unconditioned" where "conditioned" was intended. Thanks. I have corrected the text.

      Everyone is talking about wood-fiber insulation these days -- several recent GBA articles have discussed it. Eventually, the product may make sense in North America; right now, however, it is hard to source and more expensive than products that are readily available everywhere.

      1. Ra Van Dyk | | #26

        Yes, the rigid wood fiber is more expensive! That is set to change when G-O Labs gets their TimberHP production line running in Maine, where they say they have endless waste from the paper industry to turn into insulation, and we don't have to sail the stuff across from Europe. They have pushed their timeline back, but that's to be expected, given the events of this past year.

        By the way, I'm glad you got bored, and came out of retirement!

  8. user-7700281 | | #18

    Good day Martin,
    Our home construction is well underway 400 yards off the ocean just south of Charleston, SC. I've just read two of your articles and would appreciate if you would advise how best to proceed. The HVAC units along with a dehumidifier are installed in an attic that is to be conditioned. The standing seam roof is installed direct to the decking with no ventilation. The attic has not been insulated. The original plan was open cell foam. If I understand correctly the best solution is closed cell foam applied between the rafters or a layer of closed cell topped with open cell. Please advise which approach is best and the inches of material you recommend.
    Thank you, Chip

    1. GBA Editor
      Martin Holladay | | #20

      User 7700281,
      The answers to your questions can be found in several GBA articles, including the one on this page and "How to Build an Insulated Cathedral Ceiling."

      Q. "The standing seam roof is installed direct to the decking with no ventilation."

      A. I hope your roofer remembered to include roofing underlayment (a code requirement).

      Q. "The original plan was open cell foam. If I understand correctly the best solution is closed cell foam applied between the rafters or a layer of closed cell topped with open cell. Please advise which approach is best and the inches of material you recommend."

      A. Since it's too late to install a continuous layer of exterior rigid insulation above the roof sheathing, the best solution in your case begins with closed-cell spray foam applied directly to the underside of the roof sheathing. (Your use of the word "topped" is confusing. The closed-cell spray foam is installed first, so the closed-cell spray foam is on top of the open-cell spray foam.)

      You are in Climate Zone 3, so your roof needs a minimum of R-38 of insulation. The closed-cell spray foam needs to be at least one inch thick (call that about R-6), so you'll need R-32 of open-cell spray foam -- which means you'll need at least 9 inches of open-cell spray foam.

      Alternatively, you can go with 6 inches of closed-cell spray foam and call it a day. Make sure that your brand of closed-cell spray foam uses a climate-friendly blowing agent.

      When it comes to code requirements for R-value, remember that these are minimum requirements. Installing more insulation than the minimum required by code is often a good idea.

  9. Expert Member
    Dana Dorsett | | #22

    >"You are in Climate Zone 3, so your roof needs a minimum of R-38 of insulation. "
    --------
    >"When it comes to code requirements for R-value, remember that these are minimum requirements. Installing more insulation than the minimum required by code is often a good idea.

    I understand that under IRC 2021 that will become R49 in zones 2 & 3, and from zone 4 on up it's now R60:

    https://thehtrc.com/2020/insulation-requirements-2021-iecc

    Wall specs for zone 3 will not change (still 2x6/R20 or 2x4/R13 + R5 c.i.) and have not been changed for zone 3 since IRC 2012, but it is changing for zones 4 & 5, now 2x6/R20 + R5 c.i. or 2x4/R13 + R10 c.i..

    Code max U-factors for windows in zone 3 are changing from U0.32 to U-0.30 under IRC 2021 too.

    I'm not sure what, if anything they are changing about air tightness & ventilation requirements.

Log in or become a member to post a comment.

Related

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |