building-scienceheader image
0 Helpful?

The Perfect Wall, Roof, and Slab — Building Science Podcast

Dr. Joe Lstiburek talks about enclosure design principles for energy-efficient houses. Insulation, air conditioning, heating, and forced-air delivery systems have all changed the physics of how houses work. Houses didn’t used to rot, but too many of them do now.

Posted on May 6 2009 by Joe Lstiburek

You are missing some Flash content that should appear here! Perhaps your browser cannot display it, or maybe it did not initialise correctly.

Download .mp3

This podcast series is excerpted from a two-day class called Building Science Fundamentals with Drs. Joe Lstiburek and John Straube of Building Science Corporation. For information on attending a live class, go to This week Dr. Joe talks about enclosure design principles of energy efficient buildings

Let’s start with smart things

The building enclosure has four functions. In order of importance, they are:
1. Rain control
2. Air control
3. Vapor control
4. Thermal control

Thermal control is the easiest to specify, calculate, and measure, so that’s what codes focus on. Codes typically ignore the most important layers because they’re the most difficult to specify.

The vapor control layer is easier to specify than the air control layer, so codes obsess over specifying the vapor control layer and ignoring the air control layer.


How Heat Moves Through Homes

Efflorescence = Water Damage

Insulation Retrofits on Old Masonry Buildings

How Air Affects a House


Insulation Overview

Insulating Roofs, Walls, and Floors

Installing Fiberglass Right"

Insulation Choices

Can Foam Insulation Be Too Thick?

The Global Warming Impact of Insulation


Building Plans for the Energy Star Thermal Bypass Checklist

Energy Star checklist details

Insulating behind tub with rigid foam

Air sealing behind tub

Foundation/Floor Intersections

Roof/Wall Intersections

Wall/Floor Intersections



Blown Insulation

Batt Insulation

Board Insulation

Foam Insulation

The most important factors are often not considered in design, construction, and regulation; and the unimportant ones tend to have an overly enthusiastic and detailed amount of specs associated with them.

The Perfect Wall has all of the structure to the interior and all of the control functions to the exterior. Let’s start at the outside of the perfect wall with the claddingMaterials used on the roof and walls to enclose a house, providing protection against weather. .

Cladding provides three functions:
1. Aesthetics,
2. Protection from UV light
3. Physical, mechanical protection of the other control layers.

Aesthetics matter because people don’t take care of ugly things. Ugliness is not sustainable. The longer something is around, the more resources it consumes, so the more resource efficient it is, and the fewer resources it uses over its lifetime. We want a beautiful building that lasts a long time and is ultra-efficient.

Claddings should be completely open — we want air circulation behind the cladding system. The more air circulation, the better the system works. Sealants are purely aesthetic, they’re not functional. If the sealants fail, the primary air, thermal, vapor and rain control elements are not affected.

If we take the perfect wall and lean it, we get the perfect roof. From the inside to the outside, the control layers are:

  • Structure
  • Vapor control membrane
  • Insulation
  • Cladding

Some of the old-timers will recognize this type of roof as an IRMA — Inverted Roof Membrane system. If you replace the ballast with dirt, grass, and a goat, you would get a green roofRoof system in which living plants are maintained in a growing medium using a membrane and drainage system. Green roofs can reduce storm-water runoff, moderate temperatures in and around the building (by providing insulation and reducing heat island effect), as well as provide a habitat for wildlife and recreational space for humans. When properly constructed, green roofs can increase roof durability because the roof assembly’s air and water barriers are buffered from temperature fluctuations and UV exposure.. (That was a joke.)

Flip the roof and you get the perfect slab:

  • Dirt and stones
  • Insulation
  • Vapor
  • Concrete (structure)

The physics of a foundation, wall and roof are the same (this is an Ah-Ha! moment). When we look at a section of the perfect roof, wall, and slab, and we get the other Ah-Ha! moment — the important parts are the corners. You have to connect the rain control element of the foundation to the rain control element of the walls, the air control element of the foundation to the air control element of the walls, the vapor control element of the foundation to the vapor control element of the walls, the thermal control element of the foundation to the thermal control element of the walls… Pretty fundamental stuff.

Most failures occur where roofs connect to walls
Tip: Buy multi-colored pens
Whenever we do design reviews in our office, we tell the youngsters to take a colored pen and trace the rain control layer around the building enclosure. If the pen has to leave the paper, they’ve identified a discontinuity that needs to be addressed. Use a different colored pen for each of the control layers. Whenever the pen leaves the paper, you’ve identified a flaw. It’s as simple as that. We find that the flaws are concentrated at the connecting elements.

Windows complicate the perfect wall
Now these are pretty easy, but it gets complicated. In the real world, someone pokes a hole in the building and we call that a window. Windows have to do everything that a wall does, and more. It has to control water, air, heat, and vapor; you want to be able to see through it, and every so often someone is going to want to open it too. Windows can actually do all of that stuff, which is pretty amazing. No wonder they’re so expensive. All we have to do is connect the rain control element of the window to the rain control element of the wall, the air control element of the window to the air control element of the wall, the vapor control element of the window to the vapor control element of the wall, the thermal control element of the window to the thermal control element of the wall.

The reason we’ve been having so much trouble with window-to-wall connections is because we’ve been relying on one person to do all of this: His name is "By-Others." Mr. By-Others shows up on all of these specs and you have to make sure he is not going to be responsible for all of these connections. Someone has to be responsible. The window industry doesn’t do us any favors either — they don’t tell us in their window system which part of these windows systems are responsible for controlling water, air, vapor and heat. In the absence of guidance, we have to assume that the innermost component of the window is where all four of those functions collapse. So we wrap the window openings and make the connection at the back — so that if the window should fail, the water will go to the outside.

That’s how you design a building: water continuity, air continuity, vapor continuity, thermal continuity. It can’t be that simple, right?

Well, the answer is, “Yes it is.”

Tags: , , , , , , , ,

Feb 17, 2011 11:31 AM ET

the perfect wall
by Joe McNally

I am working on a LEED platinum/net-zero project...very challangeing...lotsa fun...about the perfect wall...I get it...I really do...but if the CMU is NOT being used for bearing...and we put all of the insulation outside the sheathing/CMU...enough closed cell foam to get to R-32...then why the CMU?...we plan to put all of the structure inside the thermal project manager tells me we are going to fur out the CMU with steel why not just forgo the CMU...use 6" steel insul in the cavity...CMU seems like a total waste of resources...not green to waste and a waste of $$...the only negative I can see is that the brick ties provide a thermal bridge thru the insul to the steel studs...I think we could eliminate or greatly temper that problem by spraying about 1" of the closed cell foam in the steel stud cavity...thank you in advance for your response

Feb 17, 2011 11:50 AM ET

Response to Joe McNally
by Martin Holladay

Q. "Why the CMU?"

A. I dunno .. why? You're building the wall. If you don't need a CMU wall for bearing, it's obviously silly to include one. Or maybe there is more to your question than it appears at first glance?

Clearly, if your bearing wall will be built using steel studs, the steel will need to be designed to be load-bearing -- something that wouldn't be necessary if you had a load-bearing CMU wall. But I imagine that the cost of the beefier steel studs will still be less than a CMU wall.

Feb 18, 2011 9:08 AM ET

Reply to Joe McNally:
by Joe Lstiburek

I completely agree with you - the CMU is not necessary with the approach you describe.

May 11, 2011 9:08 AM ET

Air barrier on metal roof deck
by Monica Harrington

Hi, I am building a pre-engineered metal building for my home. I plan on building a roof in the same fashion as Joe's New Orleans Superdome roof done the right way. A metal corrugated 1 1/2" roof deck, air barrier, 2 overlapping layers of XPS or Polyiso, 1/2" densdeck and standing seam roof with clips mounted through foam into purlins.

My question is in Joe's diagram he shows the air barrier on the roof deck to be a membrane installed on sheeting as say densdeck. Both XPS and Polyiso says I can put insulation right on top of deck with no thermal barrier. It is my present thought that with a metal deck the air leakage would be at the seams every three foot. I would like to know if there is a way to make the metal deck the air barrier by painting or sealing the seams with liquid applied material, or taping the seams. Or would it be better to spray or roll the entire deck with a membrane? or something else so I don't have to use that first layer of densdeck. Or is the layer of densdeck used for something other than a thermal barrier or a flat surface to apply a membrane on? I will use the densdeck if I have to.

The same question is applied to my walls. Metal deck material is spanning wall girts, then air, vapor, drainage membrane, then 2" XPS then furring then metal wall panels. I have to have the metal decking to span the longer distances between wall girts but do I have to apply OSB or exterior gypsum to the metal deck? for a surface to adhere the barrier or can I apply a liquid membrane to the primed metal deck and use that as my barrier layer? The metal deck corrugations are running vertical by the way.

Another words I wish I could get a recommendation of the Perfect Wall with corrugated metal decking used to span the horizontal girts instead of osb nailed to wall studs. I will of course have conditioned attic so I will run this wall all the way up to roof and spray foam between intersection for continuity. By the way I live in Lake Charles Louisiana in Zone 2. I saw the Lake Charles house on Building Science. I am being led by that material.


May 12, 2011 4:12 PM ET

how to convince my boss?
by Joe McNally


I am working on a net-zero (goal LEED platinum) bldg. I have been instructed to achieve R-30 in the walls using steel stud framing. My past experience (and your guidance) tells me that little or no insulation should go in the steel stud cavity. I read where you thought insulation in such a cavity is a "thermodynamic obscenity" (nice turn of phrase that!). I suggested that we could get by with putting 25% max. of the R-value in the stud cavity...3 1/2" polyiso on the outside...2 1/2" fiberglass batt in the stud cavity. Any more than 25% in the cavity and we would surely get condensation in the stud cavity. My boss insists that we fill that 6" stud cavity with fiberglass batt and only use 1-2" of polyiso. I pointed him to your on-line article about the perfect wall. The irony here is that, early in the design phase, he instructed me to use the "perfect wall"...and then told me to fur out the CMU with steel studs and fill the cavity with fiberglass it got "perfecter"...impossible to make perfect better, right? Any guidance as to how I can convince him that batt in those steel studs is a bad idea. I think he sees that empty cavity and thinks..."why not fill it with batt and improve the R-value and lower the cost construction". help1

May 12, 2011 4:27 PM ET

Edited May 12, 2011 4:28 PM ET.

Response to Joe McNally
by Martin Holladay

It's hard to argue with ignorance unless the person is willing to learn. We're talking physics here -- it's not a matter of opinion. The steel studs are thermal radiators, a framework of conductors, and it hardly matters whether you fill them with fiberglass batts or leave them empty, because the steel-stud assembly won't have much R-value -- full or empty.

All of your R-value belongs outside of your building's structure -- if you want the building to perform well. This is 2011, not 1970. We shouldn't be having these conversations.

How can your boss be so ignorant and still be in charge of a net-zero LEED Platinum design team? It's time for the boss to go back to school.

Jun 4, 2011 3:04 PM ET

Inside or outside can work
by james brown

Hi Joe,

you write " air barrier would be exterior plywood sheathing with a liquid applied mesh reinforced water barrier providing continuity at joints and windows. Over that would go insulating sheathing, furring and a back vented and drained cladding (cedar siding in my case). In the cavities would go high density spray foam."

Wouldn't your insulating sheathing perform better if there were an air barrier on top of it (ie; under the furring)? Also you write "It would have been easy to run a heavy bead of sealant between the wall framing and the exterior foam sheathing to address your concerns." Do you mean caulking all the exterior framing that will contact the sheathing and then attaching the sheathing while the caulk is still wet?

May 16, 2012 2:49 PM ET

Edited May 16, 2012 10:05 PM ET.

Northeastern New York Wall Retrofit
by Allen Bradley


I reside in Northeastern New York, roughly 150 miles northeast of Syracuse, an hour or so south of Montreal, Canada. I believe this would be considered Zone 6A, according to the zoning map I located on this site. I own a 1955 ranch style home and am in the process of upgrading the walls by adding additional insulation and installing vinyl siding. The walls are 2x4 and the stud cavities were filled with R-13 faced fiberglass batts. T-111 siding was installed over the studs. No vapor barrier was installed on either side of the wall.

I have purchased R-21 cotton insulation, which will fit in the stud cavity. Most of the T-111 is in good shape and I would like to reuse it if possible. At some point it was painted with what appears to be an exterior latex paint. Can I reuse the T-111 by installing it directly over the cotton batts? Would it be advisable to install a house-wrap over the T-111? Are there any issues associated with the T-111 being previously painted? Since I am already in the wall, I thought now would also be a great time to add 2" of rigid foam, with taped, staggered joint, seams. Is 2" enough? would it advisable to leave an airspace between the T-111 and rigid foam? I will be installing vinyl siding over the rigid foam.

My home also originally had a flat roof. Not real practical in Northeastern New York. Although soffit vents were installed, their was no air flow into the attic. Essentially, the soffit vents were only for show. I am in the process of remedying this situation. Ice damming has been an issue. With adequate soffit ventilation, will this problem persist? Would I still locate the soffit vents as far out from the wall as possible? Since I now have access to the eaves and walls, should I insulate all the way to the attic floor?

Roughly 80% of the project is as described as above. The remaining 20% is as follows: I have a 46’ of walls in the front of my home, 19’ of which is made of masonry block. It was originally installed with ½” furring strips on both interior and exterior, followed by T-111 (interior and exterior). I understand insulating a masonry wall will be a bit different than a stud wall and would appreciate the best course of action here. As I did not have enough serviceable T-111, I purchased enough Zip Panels to complete this 46' wall section.

I would like to take into account water, vapor, air, heat, and incorporate the materials I have at my disposal (if feasible).

Any feedback regarding how to go about completing this project correctly would be most welcome. I am also very open to suggestions as to which rigid foam to use.

Thank you,


May 17, 2012 5:34 AM ET

Response to Allen Bradley
by Martin Holladay

Q. "Can I reuse the T-111 by installing it directly over the cotton batts?"

A. Most walls need sheathing, for several reasons. The most important reasons: for bracing (to prevent racking), and as an air barrier. So yes, you definititely need sheathing -- either carefully nailed T-111 with sealed seams or a new layer of plywood or OSB.

Q. "Would it be advisable to install a house-wrap over the T-111?"

A. Yes. All walls need a water-resistive barrier (WRB). WRBs are required by code. For more information on WRBs, see All About Water-Resistive Barriers.

Q. "Are there any issues associated with the T-111 being previously painted?"

A. No.

Q. "Since I am already in the wall, I thought now would also be a great time to add 2" of rigid foam, with taped, staggered joint, seams. Is 2" enough?"

A. Yes. For more information on this issue, see Calculating the Minimum Thickness of Rigid Foam Sheathing.

Q. "Would it advisable to leave an airspace between the T-111 and rigid foam?"

A. No.

Q. "Ice damming has been an issue. With adequate soffit ventilation, will this problem persist?"

A. Yes. For more information, see Prevent Ice Dams With Air Sealing and Insulation. Flat roofs should not be vented. You need an unvented roof assembly. For more information on unvented roof assemblies, see How to Build an Insulated Cathedral Ceiling.

Q. "Would I still locate the soffit vents as far out from the wall as possible?"

A. No. You don't want any soffit vents.

Q. "Since I now have access to the eaves and walls, should I insulate all the way to the attic floor?"

A. Your question is unclear. You want a continuous thermal envelope and air barrier, so that insulation and an air barrier totally surrounds your walls and roof, without any gaps or breaks in the insulation or the air barrier.

Q. "I understand insulating a masonry wall will be a bit different than a stud wall and would appreciate the best course of action here. As I did not have enough serviceable T-111, I purchased enough Zip Panels to complete this 46' wall section."

A. Include a layer of rigid foam to insulate the wall. Make sure that the thermal barrier is continuous at the perimeter of the wall.

Jun 20, 2012 4:16 PM ET

Vapor Control
by John Metzger

We live in Rochester, NY and are adding a shower to an existing room. The only place for it is on 2 exterior walls. From the outside in its vinyl siding, Tyvek housewrap, sheathing panel, and I filled the 2”x4” stud bays with foil faced rigid polyiso since it’s a small space. We are about to put on the HardieBacker cement board and then tile over it. I will be putting Redgard waterproofing on the cement board and then the tile over that. I am concerned about trapping moisture in the stud wall.

Do I need to put up 15 pound roofing paper over the entire stud wall before putting up the cement board? Should I do the Redgard waterproofing on the back of the cement board instead of the front of it?


Jun 20, 2012 4:42 PM ET

Response to John Metzger
by Martin Holladay

Q. "Do I need to put up 15 pound roofing paper over the entire stud wall before putting up the cement board?"

A. No.

Q. "Should I do the Redgard waterproofing on the back of the cement board instead of the front of it?"

A. No -- just follow the installation instructions that accompany the waterproofing product you are using.

Feb 7, 2013 11:54 PM ET

The Perfect Wall
by Stuart Murray

In your article from 2008 entitled "BSI-001: The Perfect Wall", I really like the idea of Figure 7 "The Clever Wall." I am planning a residence close to Asheville NC and would like masonry walls. I have 2 questions. If I use 4" of closed-cell spray foam on the exterior of the block, what would you think of replacing the 8" concrete block in Figure 7 with 8" AAC block? I am thinking that it will increase my R value of the wall assembly, but decrease the thermal mass effect. Also, how do you create a framework in which to spray the foam and on which to attach the exterior cladding without creating thermal bridging? I have found lots of diagrams on the Building Science website that detail wood frame wall assemblies, but cannot find any for masonry wall assemblies, and I cannot find any reference to AAC block. Thank you for any assistance you can offer!

Register for a free account and join the conversation

Get a free account and join the conversation!
Become a GBA PRO!