The Best Way to Cool Your Attic

35 Comments Read/write comments


The most contentious issue I’ve written about since I started blogging isn’t bad Manual Js. Nor is it endorsing government intervention by raising efficiency standards or improving energy codes. Incredibly, it’s not even whether or not naked people need building science. Nope. The topic that really gets readers hyperventilating is powered attic ventilators (PAVs). Some people swear it’s the best way to keep their attic cool and reduce air conditioning costs. Apparently they haven’t seen the research about what works better than PAVs without the drawbacks.

Do you really need to cool your attic?

First off, let’s limit the discussion in this section to homes with unconditioned attics. Conditioned attics are great, but I want to focus on all the homes out there that have the insulation and air barrier at the ceiling, not the roofline.

What about new homes? If you’re designing and building a new home, keeping an unconditioned attic cooler (in summer) is pretty much irrelevant. Just make sure the ceiling is airtight and fully insulated, and the temperature in the attic doesn’t matter much. Whether it’s 110° F or 130° F, there’s not much difference in the amount of heat flowing from the attic to the living space below.

Oh, you also have to make sure not to do something stupid, like putting your air handler and ducts up in the attic. When you’re designing and building a new home, these are choices you can and should make. If you don’t, worrying about how best to keep the attic cool is a bandaid on a self-inflicted wound.

In existing homes, the question is important. (Again, I’m assuming that you’re not considering a sealed attic.) Lowering the temperature of an attic that contains ducts can reduce your air conditioning costs. If the home doesn’t have ducts, there still may be reasons a homeowner would want to reduce the attic temperature. A lot of people like to store stuff up there, for example. I once had a homeowner who complained of the rubber handles on a wheelchair melting because it was so hot in the attic.

OK, let’s look at one more scenario: non-air-conditioned homes. Sometimes people say they want to install a PAV to cool the house because they don’t have AC. In that case, what they really need is a whole-house fan, not a powered attic ventilator. And they’d still want to have the ceiling air-sealed and fully insulated to minimize heat transfer from the attic to the house. As with new homes, those without air conditioning shouldn’t need to worry about keeping the attic cooler.

So, sometimes people do have good reasons for wanting to keep an attic cool. What are the options? Powered attic ventilators seem to be the one a lot of people want. It’s relatively inexpensive. It’s quick to install. And it does lower attic temperatures. But I know of two other methods that both are better than PAVs. One is a radiant barrier. The other is a reflective roof. And one of those is better than the other. (Of course, there's a fourth option: You could install a ductless mini-split air conditioner in the attic, but that's the stuff of April Fool's Day jokes.)

The research on reflective roofs

There’s been some good work done on how the type of roofing affects the attic temperatures and cooling loads in the house. The Florida Solar Energy Center (FSEC) has done quite a bit of work in this area and even built a test building called the Flexible Roof Facility to study different types of roof and attic assemblies. They’ve also done research on a number of real homes in Florida.

One of the earliest studies they did was to look at the difference between black shingles and white shingles. The first graph below shows the results, and you can see that it’s about what you might expect. White shingles reflect more of the incident solar radiation and stay cooler. They don’t show the attic temperatures here, but if the shingles are cooler, the attic is cooler. (You can read more about this in FSEC’s literature review of attic ventilation (pdf). The graph is from page 30.)


In another study, they changed the color of a home’s tile roofing from dark grey to white. (See photo below.)


Here they did measure the attic temperatures, and you can see them in the next graph. The white roof lowered the attic temperature by about 20° F. (These images came from a presentation Danny Parker gave (pdf) on FSEC's cool roof research.)


The third graph shows attic temperatures from the middle of the attic for several different types of roofing as well as a sealed attic. In this case, the most reflective roofing material (ivory IR selective metal tile) is 23.4° F cooler than attic covered with dark shingles. (Read more about this in Danny Parker’s 2004 Home Energy magazine article, Improving Attic Thermal Performance (pdf).)


But it’s not just FSEC that’s been studying what’s going on with roofs and attics, of course. Jeff Gordon gave a great presentation on attic venting (pdf) at the 2011 Affordable Comfort conference and included some data they’ve taken at the Buildings Research Council lab in Illinois. They found that dark shingles are 27% hotter than white shingles. 

My friend Cameron Taylor has done a less formal study on his own home in Fort Worth, Texas. The design temperature there is 98° F, and they've been experiencing quite a few days where it gets above 100° F in recent years. He replaced his roof with a reflective metal roof (shown below) and wrote this to me in an email:

My experience with my own reflective roof is going on three years, now. Each summer has been consistent with air temperatures in attic tracking below outdoor ambient until mid afternoon, where it then tracks with outdoor ambient.


Why reflective roofs beat PAVs and radiant barriers

As you might expect, it’s not hard to show that reflective materials keep things cooler. When solar radiation hits a surface, three things can happen. It can be reflected; it can be absorbed; or it can be transmitted. The part of the solar spectrum that heats things up doesn’t get transmitted through roofs, so the incident solar radiation is either reflected or absorbed there. The more of it you can reflect, the less gets absorbed. Reflective roofing attacks the problem right there where it starts.

Once the solar radiation is absorbed, some of the heat still gets ejected to the outside but much of it conducts downward through the roofing materials and roof deck. There it radiates down into the attic, heating up the insulation, the framing, the ductwork, and the boxes of Christmas tree ornaments. A radiant barrier can help by reducing the amount of that heat that radiates downward. But still, the heat is already in the roof. Radiant barriers are fine, but why not just stop the heat on the side of the roof where it first enters?

Powered attic ventilators are the worst way to try to keep your attic cool. They’ve been well-discussed here and other places, but briefly they’re not cost-effective, don’t deal with the problem where it starts, are using convection to solve a radiant heat problem, and can even be dangerous by backdrafting combustion appliances and putting carbon monoxide in the home.

Reflective roofs are the winner, first because they deal with the problem where it begins. If we’re talking about existing homes, they beat radiant barriers because it’s not something extra. You gotta have a roof. Changing out a roof before it’s due to be changed would rarely be cost-effective, but if you need a new roof, choosing something reflective is the way to go. Still, don’t forget this doesn’t absolve you from fully air-sealing and insulating the ceiling and getting the ducts into conditioned space or sealing and insulating the heck out of them.

Afterword: A note about climate zones

I didn't explicitly state it above, but I'm addressing mainly those in hot or mixed climates, not cold climates. If you're in climate zone 5 or higher, reflective roofs can cause moisture problems if your ceiling isn't airtight. The roof deck will stay colder in winter, so any moisture that gets into it is likely to stick around and cause problems. You can do a reflective roof in a cold climate, but I wouldn't do it on an existing home unless I were certain I could make the ceiling airtight.


Related Articles

The #1 Reason Power Attic Ventilators Don't Help

How NOT to Retrofit a Furnace in a Spray Foam Insulated Attic

What's That Ice Chest Doing in This Attic Duct System?

Case Closed: Get Those Air Conditioning Ducts out of the Attic


Image credits:  1st & last images from Cameron Taylor, all others from FSEC.


NOTE: Comments are moderated. Your comment will not appear below until approved.


Rachel Kingsley
Aug 26 2015 - 11:48am

"Bandaid on a self inflicted wound" This is why you have such a huge following Allison. Thanks for all you do!

Aug 26 2015 - 12:36pm

It's all about the delta T. You mention the design temperature in Fort Worth, TX being 98F; but what if the attic temps are more like what the previous graphs show (more like 130F)? The ceiling design temp should be the greatest attic temp delta T, not the ambient air temps. The closer you can get the max attic temp swings to ambient, the better - but it is really hard in existing housing stock - especially when the existing roofing surface is nowhere near needing replacement. I'm still experimenting with "directed cooling" of the roof deck underside (which I suspect is what is truly happening with the radiant barrier products), but I'm liking the results so far.

David Butler
Aug 27 2015 - 12:09am

@DCE, that first graph shows shingle temperatures. The air below will be considerably cooler. And within the attic, the temperature will be hotter up near the ridge than at the top of the insulation, which is all that matters when it comes to the ceiling delta-T. And finally, the attic air temperature isn't the whole story. A large portion of what ultimately gets conducted to the room below is driven by radiant gains.

I agree it can be difficult to fix poorly insulated and sealed ceilings in existing homes. But it can be done. The worst offenders are bonus rooms and Cape Cods with clipped ceilings and knee-walls. The key there is to achieve six-sided encapsulation on the knee-walls (e.g., close off the backs), and block the ends of the floor joists under the knee walls.

Johnny nguyen
Sep 25 2015 - 12:42pm

I attended a seminar from Yellow Blue company (that's their real name)
They talked about saving heating and cooling bills by installing solar attic fan (SAF) and multi layer insulation (MLI), both at 4000.

Their website is

Is this a good idea to have them both installed?


Thomas Dugan
Aug 26 2015 - 4:02pm

I am trying very hard to simply accept the science from an expert in this case. However, as a very old builder (70+ years), I have known cases where roof mounted attic fans have helped tremendously where the roof is a full hip and the ratio of ridge vents to soffit vents was way out of whack. We closed off (eliminated) the ridge vents and installed powered vents on thermostats. We were able to drop the attic temperature by more than 20 degrees. This greatly decreased the 11 o'clock PM heat beating down into the top floor bedrooms that was keeping the A/C going full blast. Where is the flaw in my assessment?

Walt Stachowicz
Aug 26 2015 - 7:02pm

Mr. Dugan - I'm one of the young whipper-snappers at 66. I will defer to Allison, of course, but introducing a negative pressure into the attic space will probably draw the cooler air from the leaks in the ceiling plane unless they were scrupulously sealed. That is, your conditioned air is providing part of the cooling effect. And, unless it is solar powered, it is using more power than it saves. I still have contractors who don't believe this until they see the pressure differential when the fan is turned on. There are some good hip ridge vents available, or off ridge vents may be used.

mark brown
Sep 17 2015 - 3:34pm

i am a still younger whippersnapper but i believe "probably" and "definitely" have different meanings. as such it is best not to act on what is "probably" true when one has other information regarding the exact case.

as for the pressure differential...that can be true but sealing the attic/ceiling transition is recommended in all cases so ....that is really a non-issue

as for using more energy than it saves....many years ago i installed a PAV. without AC on it reduced the indoor 8F. so at least in this case it could not have been using conditioned air to cool the attic

this also brings up another issue with allison's comments. he said for people in un-conditioned homes what they need is a whole house fan, not a pav. unfortunately the whf will only bring relief in the evening thru early morning. unfortunately the temperatures n the house begin rising in the day time and the whf only has a residual effect from lowering the starting temp in the home....a pav can keep it cooler during this time. as i said it kept my home 8F cooler.

actually i also had a whf which i would use during the hours its use was appropriate (i installed it a year after the pav)

Allison Bailes
Sep 17 2015 - 3:51pm

mark, You're like a dog who just won't let go of that bone, no matter how dirty and old it is. If you want to put a powered attic ventilator in your house fine. You'd be better off insulating and air-sealing your ceiling. You won't ever pay an energy bill for that, unlike you do with your PAV.

If you want the heat out of your attic, a reflective roof or radiant barrier are better ways to get there.

Enough said. I'm not going to let the comments for this article turn into the Allison-is-nuts-because-PAVs-can-sometimes-work charade that has happened with previous articles. If you've got something to say about reflective roofs, let's hear it. We've already heard what you have to say about powered attic ventilators.

Ted Kidd
Aug 26 2015 - 10:09pm

^^ What Walt said. Suck cool air conditioned air into the attic and of course it won't be so hot. It'll also help reduce up/down imbalance by counteracting reverse stack.

But then you still have a really crappy house. It seems like a lot of trouble to AVOID fixing the problem correctly.

Another nicely crafted article Allison.

Thomas Dugan
Aug 27 2015 - 7:57am

All posted here makes sense. I am a strong advocate for sealing the attic space with closed cell foam under the roof deck and using metal roofs for their radiant benefits and storm resistance (foam under 5/8" minimum decking under steel is extremely strong). My question now is for those high pitched hip roofs (12/12 and more) with about 8' total ridge and 120' plus of soffits. If the owner wants a shingle roof and doesn't want to pay the cost of foaming that tall roof, what is the proper way to vent that attic? It doesn't seem to me that any ridge vent could handle the volume required.

David Butler
Aug 27 2015 - 2:55pm

@Thomas, I think one or more passive turbines is the best solution for a roof with limited ridge.

Sep 19 2015 - 11:42pm

Dear Mr. Dugan,
I'm a homeowner in Summerville, SC (Hot,Humid Summers). I live in a 1983 low-country-style (Cape Cod-style) house with a contiguous wrap-around porch with 6 dormer windows and 12 c. 4'x6' knee walls. Leaked like a sieve. you could see the porch from inside the kneewalls. I had a consultant come at least 5 yrs ago. After the consult, had 24 g metal roof installed with Grace underlayment vs roofing felt. Consultant said seal from top down including rimjoists. Totally sealed attic ceiling, walls, soffits, kneewall ceilings and walls with open cell foam (closed cell not recommended for my home). I removed all fiberglass insulation from attic and dormer ceilings. Removed fiberglass insulation from under house, had cement fiberboard walls installed to separate porch from 3-4' high crawlspace under house. 15 mil Raven vapor barrier installed in crawlspace. All ducting in crawlspace converted to hardpipe. No more sealing. Recently had 21 year old HVAC systems replaced with 16 seer units. ($18,000 so far). Downstairs: Gas htr/single stage HVAC. Upstairs: 2 stage HP. Thermastor 98H high efficiency Whole House dehum downstairs. Santa Fe crawlspace dehum in crawlspace. Upstairs HP airhandler and ducts in attic (NOT possible to move out of attic) Problem: upstairs HP maintains 78°/50% consistently (never touch it). Downstairs maintains 76°/50% but dehum runs 80% of time. Do NOT see any reduction in bills/or increased efficiency. BIGGER PROBLEM is that I don't know how to properly address following. Unconditioned kneewall areas and unconditioned Room Over Garage have consistently high temps/RH. On hot summer days it was c.80°/80% while all conditioned spaces upstairs and downstairs stayed at whatever was set on controller. I THINK I have an air sealing problem. However, I do not know how to properly address it, especially sealing the kneewalls/FROG/attic. They have poorly installed fiberglass insulation with vapor barrier laying against downstairs ceiling and 1x6 subflooring. The ceilings and outer walls have open cell foam which the Electric company said was installed very well. I think the RH would go down if I could properly seal these areas off from conditioned areas. I also though there was a problem at the 20x30 return plenum downstairs. I have opened that wall and there is the plenum with no insulation and only an inch or so between its sides and the wall and and a open shaft straight up to second floor subfloor. I have yet to seal/insulate this plenum but I think I need to insulate it and either pack bags of rockwool up this chase or seal off the chase with foam board just above plenum. Sorry to ramble but I saw your post and thought you might be able to give me a bit of guidance on proper sealing. Thanks in advance for any response, Sir.

Aug 27 2015 - 10:48am

I need a new roof and have been trying to decide whether it is worth the extra cost for a metal roof or even the asphalt shingles that are marketed as being more reflective. I am in the Houston area and my 1957 house has metal ducts in an unconditioned attic. I have been trying to find real world data about the savings, but have not been able to find any. Can you provide any more details about the metal roof your friend Cameron has? Is it installed with an air gap above the roof decking? Is his roof white metal? I can't tell from the photo. Has he seen any savings on air conditioning bills? Also, does anyone have any experience on whether the more reflective asphalt shingles lower attic temperatures any?

Cameron Taylor
Aug 27 2015 - 2:18pm


My house is of similar vintage as yours, in that it was built in 1959 and has metal ducts in the attic. Now, answers for your questions:

* Roof material on my house is standing seam metal with reflective "CoolR" coating of the color "solar white". This coating is specified to have a high reflectivity and emissivity rating. I can get you actual numbers if these types of specifics interest you.

* There is no gap between the metal and the roof deck. The underlayment is a synthetic polymer (instead of roofing felt) from Palisade:

* Summer attic air temperatures with old terra cotta colored asphalt shingles peaked at 140 to 145" at ~ 18" above the blown-in fiberglass insulation layer at the attic floor. Summer attic air temperatures with reflective roof peak at or slightly above outdoor air ambient. It often takes until early to mid-afternoon for this peak to occur, where it then tracks with outdoor air ambient until around midnight

* I don't have hard numbers on utility savings handy at the moment. However, I will be honest and say payback for a roof of this type would be pretty long. I will further be honest and say payback was not our primary motive for selecting this roof type. Primary motives were improvement in interior comfort and HVAC performance, along with a roof that now has considerably higher wind resistance and hail damage resistance. The life expectancy of this roof is stated to be in excess of 50 years, meaning if I stay in this home until my preferred form of demise (long, healthy life), it will outlive me. I also enjoy the architectural enhancement it has brought to the style of home we have (mid-century modern/Asian fusion). To me it makes the house appear more substantial over its former asphalt roof appearance, and in truth, it is more substantial structurally and in other means. In severe storms, the basic reality is that if you lose the roof you lose the house, so having a roof with greater high wind resistance improves the structure's survivability in high winds (downbursts, microbursts, etc.) or lower grade tornadoes.

* Benefits we did not expect with this house were that it noticeably reduced outdoor noise entering the house from above. Before, whenever a propeller driven airplane or jet flew overhead, it sounded like the thing was in our attic. Now, you only hear it through the windows. We also live near a major interstate freeway, and when cold fronts blow through the rush of traffic noise carried by the wind to our house greatly increases...but not in the house! The roof is also NOT noisy when it fact the rain rate must escalate to a certain level for us to hear any rain on the roof at all.

I will follow up with another post about my personal analysis regarding living under this roof for three years, in addition to what I've disclosed above. I'm happy for your questions, as this is what I'd hope would occur in the response to Allison's great article above.

Cameron Taylor
Aug 27 2015 - 9:55pm

As promised, my thoughts and analysis on living with a reflective roof.

First, I realize that we’re conditioned to respond whenever a topic like this is discussed to think in terms of “what will this save me?” As I said in my last post, payback on an investment like a reflective roof, in terms of any hoped for savings on air conditioning cost paying for the roof cost, is long. So why do it, aside from the comfort and other benefits mentioned in the last post?

I have been in HVAC and building related work most of my working life. When I began this career, energy was relatively cheap, and the prevailing attitude for comfort problems in buildings was to merely apply more brute force to it. House hot? Get a bigger air conditioner. Cold? Boost the furnace capacity. Two energy shortage events and price shocks in the 1970s altered this mentality some, and was the beginning of the alternate energy and energy efficiency movements in our nation and elsewhere. However, the brute force animal was not ready to die, yet, or as my friend Robert Bean refers to it, one of three little pigs that must die (see Allison’s blog post on this topic for further reading).

In short, I see the inevitability of shifting the balance of interior thermal comfort management more toward the building enclosure, and less on the HVAC system. Cheap energy allowed the latter, but increasing energy and environmental cost compel the scales to shift. Many are already engaged with this in such programs as Passive House and Building America, but my interest is in existing buildings, as that is what most of us live, work, and play in.

A reflective roof is but one step in this process. In my case it is not practical to get my ducts out of the attic. My next effort with those ducts is to air seal them the best I can and replace the 56 year old insulation with R6 or better. The reflective roof already relieves a considerable heat load off the existing ducts, so air sealing and better insulation will be icing on the cake for them.

In closing, I think we need to move the conversation beyond merely “how will this pay for itself?” or the ever ominous monster of “first cost”. Yes, a roof like mine is about 10% more expensive than an asphalt shingled equivalent, but how often do we balk at buying things we really want knowing full well (or overlooking) that the monetary payback will be slow in coming, if at all? Granite countertops, anyone?

I’m no “tree hugger”; everything we do every day has some form of impact on our environment, and not much of it beneficial. Therefore to me it’s a matter of impact mitigation vs. hoping to someday achieve some Utopian rainbow tooting unicorn fantasy of “oneness” with our environment. I also share a dream with my wife of achieving a net-zero home and solar powered cars. A reflective roof in a hot climate is a big player in this dream.

Aug 27 2015 - 11:31pm

Cameron, thanks for the in depth replies. I hope to find a someone who can install a standing seam metal roof for a reasonable price. The price I was quoted was more than twice the price of asphalt shingles on my approximately 1600 square foot house. Asphalt was quoted around $9k and standing seam metal was around $21k!

Cameron Taylor
Aug 28 2015 - 1:04pm


You're welcome. One thing I did not mention earlier is that many homeowners insurance policies will give discounts when hail and high wind resistant roofing material is placed onto a house. Ours did.

Standing seam is not the only metal roofing option out there. If you prefer the look of shingles there are reflective metal options here. Some schemes use darker colors but place battens beneath the metal to allow airflow underneath the metal. What you want to play close attention to with any material is the percentage of solar rays reflected and the emissivity rating of any material you consider. A website with good info is: which is operated by the Cool Roof Rating Council.

One of the motivational factors for replacing our roof was the old one sustained hail damage. Insurance covered cost up to the equivalent of what was on the roof prior to damage, and we paid out of pocket for the remaining cost. Our house also has some unusual roofing details in that the overhangs are flat and project out from the walls three feet the entire perimeter of the house. These surfaces received a PVC roof coating, and it can have a cooler surface temperature at times than the surrounding metal roof.

Thomas Dugan
Aug 28 2015 - 8:08am

Nice post Cameron. I would like to offer another way to look at payback for any of these types of improvements. What I do with my hurricane/disaster resistant homes is find multiple benefits to be derived from various elements. That is how I got into this whole energy thing in the first place.

Examples: I use precast insulated concrete panels for my exterior walls. The benefits are a home that will not rack in high winds, is immune to small and large wind-driven missiles and is virtually air-tight so that I have no ex-filtration or infiltration issues with the walls themselves. They are also fire-proof.
I use closed cell spray foam under the roof decks that literally glues the deck onto the structure for increased uplift strength. It also insulates my attic space and when combined with metal roofing and 5/8" plywood decking, has tremendous resistance to punctures from falling limbs and trees.
My windows are all impact rated and carry a high energy rating. They also make it almost impossible for a wildfire to get in.
Finally, I use standing seam steel roofing to resist against high winds 170 mph+ and I get the radiant reflection bonus points along with the puncture resistance mentioned above. Also fire-proof.

So you can see that there are benefits beyond just energy efficiency that we can use to help pay for these upgrades. The building is a system and all components can contribute in more than just one way.

Jim Luckett
Sep 1 2015 - 1:56pm

Photovoltaic panels on the roof also do a nice job of cooling the attic, and in many states companies are eager to install them for free.

Sep 12 2015 - 11:12am

Allison - I live in lower central Florida in a 2004 era single story home with deep deep blown in insulation - and in-attic flex insulated ducting (no other choice in Florida?). Like most Floridians in post-hurricane bldg. code homes, my home is concrete block with a couple of blow-out wood walls, 100% vented soffits, large hurricane roof vents and 13" - 16" of blown insulation. We have 14' ceilings and ceiling fans that run 24 x 7. Pretty sure our ceiling is not sealed from the attic. I have read your Blogs and every single comment on PAVs, it was a Saturday morning well spent. Lively intelligent debate is wholesome.

I understand you have experience in this sort of climate and welcome input from you and your readers. After reading your blog and then reading the referenced documents, I see why I have to work so hard past sunset to cool the house. We generally run our inside temps at 84 degrees during day and 79 at night, with a punt at bedtime of 77. I updated our a/c and blowers last year but the electricity costs keep going up and up. The thermostat says the house is cool (78 – 79) but our bodies feel like we are still hot so we are constantly adjusting down the thermostats to 77 or lower to get our perceived body temps comfortable.

My key issue is hot spots in my house and I have attributed that to poor circulation of air in the attic, cold/return ducting being subjected to attic heat and in my amateurish view, warm air inside the house not getting pulled out because of limited return vents. So, I’m confused on the consensus from this blog and how to proceed.

Poor attic circulation –
The hotspots in my house correspond to where the ceiling is closest to the roofline, mere feet in some cases, and the cold spots to where there is 10’ – 15’ of attic truss space. In my mind, circulating air in the attic would accelerate the existing convection currents and help distribute the heat load. I’m not sure if this is a PAV or some other mechanism, and comments in your blogs about irrelevance due to proper insulation give me pause. Can hotspots in the attic contribute to uneven radiant heat inside the house with well insulated ceiling?

Hotspots in certain parts of house –
To me it would make sense to redistribute the air inside the house more evenly, however in Florida you cannot run the blower fan alone because you are heating the indoor air by running it through attic ducting. My idea is to vent the build-up of hot air that has risen to the ceiling in the hotspot rooms into the attic so that air from cooler parts of the house can move in. Low cfm similar to a bathroom vent (300cfm) and only in the hot spots – basically by adding an inline ducting fan to the return ducts for the hotspot rooms. But the positive/negative pressure comments make me think otherwise.

Lastly, I am curious about the recommendation on your blog to reverse the ceiling fans to decrease the radiant heat transmitted from the attic through the ceiling. That is a new one on me. Please explain.

Thanks for some very informative posts by all.

mark brown
Sep 17 2015 - 3:53pm

the reversing of the ceiling fans will not lower the the amount of radiant energy. that sounds totally bewildering. radiant energy is dependent on T^4 and not air movement. theoretically the fan could lower the temperature of the ceiling but it certainly wont be appreciable PLUS you will also be increasing the transfer of energy into the room through convection of the moving air past the ceiling surface.

reversing fans have shown through testing NOT to decrease room temperatures (tho if it did lower ceiling temp there could be a difference in the radiant energy incident upon a body and hence feel cooler but again....)

David Butler
Sep 17 2015 - 4:22pm

@Mark, when you say ceiling fan will not lower radiant energy, I assume you're talking about radiant gains from the roof. But those gains are converted to conducted gains at the ceiling, so they're irrelevant when discussing ceiling fans.

Radiant gains INSIDE a room are a function of surface temperatures (MRT). Reversing the fan disrupts the air film on the ceiling, thus lowering the surface temperature slightly. As you acknowledged, this reduces the radiant energy incident on a body.

In particular, when you're (prone) in a bed, it's not uncommon to feel warm due to the incident radiant energy from the ceiling, which may lead to a lower t'stat setting to compensate. Decreasing the ceiling temperature by even a small amount may allow for a higher t'stat setting.

I agree that "cooling" the ceiling increases the heat flux into the room (disrupted air film slightly lowers ceiling assembly R-value, perhaps by R-0.5. But the impact on conducted heat transfer is way more than offset by the savings if t'stat can be raised by even 1 degree.

David Butler
Sep 13 2015 - 3:43pm

VW wrote: "The thermostat says the house is cool (78 – 79) but our bodies feel like we are still hot so we are constantly adjusting down the thermostats to 77"

It's puzzling why you consider 78-79 to be "cool" when your body says otherwise. Comfort varies from person to person... most folks prefer somewhat cooler temps. If 77 feels comfortable to you, then consider yourself lucky.

"We generally run our inside temps at 84 degrees during day"

Wow. I don't think many people with A/C would tolerate 84, but if that makes you happy, then why not! But there's a caveat....

You didn't mention RH, but humidity affects comfort. When the A/C doesn't run, it can't remove moisture. In humid climates, indoor RH is largely the result of infiltration. In particular, when outside dew points are high, opening up the house during the day and then closing it up in the evening to cool down can lead to problems.

As for your attic... your observations regarding hot spots relative to distance between ceiling and roof are likely just coincidental. Room-to-room temperature variations are far more influenced by duct design (supply imbalance relative to room-by-room loads) and variations in ceiling leakage and insulation uniformity, than by variations in roof distance and attic temperatures. Also, duct runs to rooms furthest from the air handler will experience larger heat gains. But a well designed duct system accounts for that.

And finally, regarding ceiling fans... the only logical reason to reverse a ceiling fan is to improve air mixing in winter (with fan on low) while minimizing evaporative cooling of the skin. In summer, you want both: improved air mixing and evaporative cooling. BTW, one of the benefits of improved air mixing is to bring interior surface temperatures closer to air temperature (i.e., mean radiant temperature). In any case, there's no benefit to running a fan when there's no one is in the room. Turn them off to save some energy.

tom phillips
Sep 17 2015 - 2:55pm

(in case these issues were not already discussed)

1. Cool roofs (hi albedo) are great in cooling climates, but can be a MIXED BLESSING OR NET LOSER in mixed climates. When I ran the LBNL Home E Saver model for my older house in Davis CA, a cool roof slightly reduced my cooling costs (which were minimal to start). But it significantly increased my heating costs, which are moderate for this area. Furthermore, since my main Green House Gas emissions come from heating with natural gas, a cool roof would yield a net increase on my carbon footprint. However, after a few more very hot summers and mild winters, I may apply a cool roof coating to my roof!
2. How do attic temperatures translate to the bottom line for comfort (indoor temperature, humidity, radiant temps)? I would guess it is not linear, and is very house dependent. I suspect Parker, Lstiburek & others have measured & modeled this fairly well for most regions. An older NREL fact sheet (1994) estimates that 30% of a home's heat load comes from the attic, a suggests various cooling strategies,

mark brown
Sep 17 2015 - 3:59pm

i think the way to go in situations like this is to have a double roof...maybe design it as a Tromme roof instead of a Tromme wall. then in the heating months one could channel that heated air between the two roofs into the house.

i havent done this but the next time i do my roof i am condsidering it. i live just a bit north of you in chico

Dac C
Oct 2 2015 - 10:26pm

Thanks for all the information you share here. As a builder, it gives me great food for thought as to how to improve what I do. It's amazing how much crap you get from some of these posts. Keep up the good work.

Rhonda Franklin
Feb 15 2016 - 4:40pm

For a lot of homeowners out there, a metal roof is not an option due to existing homeowner association (HOA) guidelines. In those cases, you can't go wrong with using a quality radiant barrier stapled to the underside of the roof rafters in the unconditioned attic space. It's definitely the next best thing.

David Butler
Feb 16 2016 - 12:34pm

@Rhonda, retrofitting a radiant barrier on an existing home is expensive, especially the labor. With prices exceeding 50 cents per ft2, there's no way a RB will save enough to justify the cost. Even in new construction where RB-faced sheathing only adds about 10 cents per ft2 to the roof, the payback is marginal except in the hottest locations in the US.

You're much better off having your ceilings air sealed, insulation upgraded, and ducts sealed than spending thousands to have someone staple RB to the underside of your roof deck.

Steve Kohn
Apr 21 2016 - 3:25pm

David and Rhonda -- You're talking about what I've been considering.

Some background:
1. 2-story house, built 2008, about 2900 sq ft, San Antonio (hot, humid summers, mild winters), with good insulation as indicated by low electric bills (no higher than $175) in summer or winter months.
2. Carrier heat pump, 14 SEER, works just fine.
3. In late winter/early spring, we can go weeks and weeks without using either the A/C or heater by opening windows at night to cool the house, then closing them in the morning to keep the house cool. But while downstairs is just right, upstairs is, I'm guessing, 5 to 10 degrees warmer.

I figure if I can fix that, I can also reduce the load on the A/C in the summer, when temps can be above 95 from July through September.

So ... radiant barriers aren't worth it, huh? Didn't seem to me either, and your post convinces me now (without even adding the cost of fixing ceilings workers fell through).

Replacing an 8-year-old roof surely isn't worth it.

Until now we've closed, a little, the vents downstairs while leaving the upstairs wide open. Helps some, but not really enough. Ends up making downstairs, where we mostly live, uncomfortable.

It's just occurred to me: temps upstairs are fine in winter. That is, if it's 70 downstairs, it's about the same upstairs. It's only when the sun is bright that the temperature difference is felt. That's telling me the reason the upstairs gets hot is because of heat transfer from the roof/attic. (I need to put my hand on the upstairs ceiling and feel if the ceiling is warm or not.)

Any input would be welcomed.


David Butler
Apr 21 2016 - 6:05pm

@Steve, your problem appears to be with the design of your HVAC system. Some states now prohibit a single system for two floors unless there's an automatic zone control system. And even that's a challenge to get right.

Without knowing a lot more about the design of your house, and where the weak points are, it's hard to say how best to improve your situation. Your electric bills actually sound pretty high to me for a house that size in central TX, but I don't know how much you pay for electricity. The usual suspect is leakage at the ceiling plane and any knee-walls. Reverse stack effect (in hot weather) will create a highway of hot air from your attic into the house, especially the upstairs. The opposite happens in cold weather, thus your improved temperature balance in winter.

Mar 12 2016 - 4:52pm

Just read your article and after a couple of months of research have to say I agree with what you say. Several months ago I re-connected my attic fan after finding that it had been disconnected by the person who repaired the outside wall. After researching about how to cool my attic, I realized he had not "neglected" to re-connect the fan, he had left it intentionally. I went back into the attic a few days ago and unplugged it again. Now I don't know what to do. I have a 2 story barn style house built in the late 60's with not a lot of cellulose left in the attic. It has a wood burning heater but not for real heat. The house is total electric and I have been installing through the wall heat pumps for comfort. With the barn style roof, there are no soffit vents. There are two turbine vents and two small gable vents. Since this is NE Alabama, it is hot and humid in the summer. I want to air seal my ceiling but with the heater, wiring and recessed lighting don't know how successful that will be. I thought I wanted a radiant barrier even though it will be a pain to install. After reading your article I think a reflective metal roof would be a better choice. Since I have through the wall heat pumps, would I be better off to spray foam/air seal my attic and turn it into a conditioned space and put a metal roof on? Should I put the metal roof on with a ridge vent and have the roof edge vents installed and blow in more cellulose. Looks like I have decided I want a reflective metal roof, just don't know whether to make the attic breath better or to close it off altogether. Thank you for any insight you offer. greg

Heather Vanskiver
Jun 22 2016 - 10:53pm

House was built in 1906 and now has a new shingle roof. The attic space was nearly 10' high so we reclaimed the space for a new half bath and two bedrooms, essentially creating a functional upstairs. Probably is the rooms are HOT! The shingle roof is right overhead. Each upstairs room has sufficient air vents but apparently the shape of our stairwell allows heat to head upstairs and swelter anyone who sleeps up there. Any suggestions?? Our HVAC guy offered us a second 3,000$ unit for just the upstairs. Please say there's another way! We've got new windows being installed for those rooms but I was really hoping to not hang units out of those since the windows are so small. Any suggestions would be appreciated!

David Butler
Jun 23 2016 - 5:28pm

@Heather, not nearly enough information to make recommendations but here are some general comments...

Even under the best circumstances, attic bonus rooms are hard to cool, especially in older homes that cannot easily be air sealed.

Bonus rooms typically have sloped ceilings and and attic-facing knee-walls. A critical detail is to build the knee-walls so they can be air sealed. This requires a top plate or blocking as well as a back on the wall so each stud cavity is enclosed on all six sides.

The other critical detail is getting adequate insulation in the slopes, while allowing space for ventilation between insulation and roof. This is almost impossible to achieve without furring down the ceiling in order to increase the rafter bay depth.

These challenges go away if the entire roof is insulated, typically by applying spray foam insulation to the underside of the roof deck. This brings the entire attic into conditioned space, and thus the knee-walls do not have to be air sealed or insulated. Furring down the slopes can often be avoided by using high density closed cell foam.

As for conditioning the space... it may be possible to use your existing central AC, but only if your system has excess capacity and there's a way to route adequately sized supply ducts to the rooms. $3k for a separate system sounds very low when you consider what's involved in installing a new duct system.

A couple of ductless mini-split heat pumps would be the least expensive option if you don't want window units and don't mind if the stair landing and bathroom get hot. But I don't know any competent contractor who would install two mini's for $3k.

Bob MannIng
Jul 3 2016 - 5:20pm

I have a house in the north with an exposed-beam cathedral ceiling and asphalt shingle roof. Needless to say the distance from inside conditioned space to outside is only measured in inches. The house faces east-west. I am wondering if it would be beneficial to install a metal roof over the existing roof? What if I included an air space between the old and new? Additionally, what about a damper at the top to open/close air convection in summer/winter? I would only put it on the back, west, face to maintain the appearance from the street. Any comments?