Significance of Thermal Bridging?
I would like to better understand the significance of thermal bridging in light frame / wood construction.
Some information about thermal bridges and thermal breaks:
Thermal Bridge
“Characterized by multi-dimensional heat flow, and therefore they cannot be adequately approximated by the one-dimensional models of calculation typically used in norms and standards for the thermal performance of buildings (U-values).”
Thermal Break
“A component that is a poor conductor of heat and is placed in an assembly containing highly conducting materials in order to reduce or prevent the flow of heat.”
Here’s a BSC / ASHRAE article from J. Lstiburek on commercial construction and thermal bridging: http://bookstore.ashrae.biz/journal/download.php?file=building_sciences_1.pdf
And finally, in this GBA post, Peter Yost states that “wood framing reduces in-cavity R-value by a bit less than 10%”.
https://www.greenbuildingadvisor.com/blogs/dept/building-science/thermal-bridging
I think the concept of a thermal bridge is intuitive, that’s if you’re talking about thermally conductive materials—say metal and concrete (a frying pan or an uninsulated slab edge). And it makes sense when you’re evaluating windows and window frames. To me, the idea becomes less intuitive when looking at stud walls.
As a hypothetical example, consider this simple (okay, very simplistic) wall:
* 10′ x 10′ x 10′ – 100 square feet total surface area
* Single 2″ x 8″ top & bottom plate, 2″ x 8″ studs 24″ OC – 10 square feet surface area
So you have a 10% framing factor. The whole wall R-value works out to 20.3, a 20% loss over the nominal R-value of 25.375 (assumes R-1/inch for the framing, R-3.5/inch for the cavity insulation):
How does thermal bridging come into play in this example? What’s the benefit if you take the same wall, use (2) 2″ x 4″ studs instead of a single 2″ x 8″, and separate the walls by just 1/4″? The wall occupies the same overall volume, has the same framing factor (stud surface area), and almost the same framing volume (96.5% of the base wall’s stud volume). How does the 1/4″ gap between the studs translate into the heat loss calculation? What’s the potential improvement over the base case 2″ x 8″ wall?
Anybody using THERM or PHPP willing to take a stab at this? Or reference previous studies?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Daniel,
You raise many issues. A few comments:
1. Researcher Jan Kosny has written, "The most current study performed for California Energy Commission (Carpenter 2003) demonstrated that Framing Factor (fraction of the opaque wall area represented by solid wood used for framing) for residential walls is close to 27%."
http://www.ornl.gov/sci/roofs+walls/NewRValue.pdf
2. I'm not sure of the relevance of your hypothetical wall, with studs only 1/4 in. apart. The idea is to separate the studs enough to get some insulation between them.
Daniel,
Your example is similar to the proposals that have sometimes come across this forum to use an exterior ½" foam board as a "thermal break".
For a thermal break to be effective, it has to be both high enough in R/inch and thick enough to represent a significant improvement over the wood framing.
In your example, there is virtually no thermal break. But, if you offset the studs in those two 2x4 walls, there would be a significant thermal break.
There are three major advantages to a double stud wall:
1) less wood than deep studs
2) significant thermal break potential
3) easier framing with less waste (of sheathing and drywall) if exterior studs are laid out on center from the exterior corner and interior studs are laid on on center from the interior corner.
With staggered studs, if the wall is 8" thick, the studs will be offset by 8". If the walls are 12" thick, then 24" oc studs will be offset a full 12" or half the oc spacing for maximum thermal break.
Even assuming a 25% framing factor (and R-1.25/inch for softwood), your 2x8 wall will degrade the whole wall R-value by 31% to an R-17.5.
With staggered 2x4 studs, the same wall will have a whole wall degradation of about 10% for R-22.8 (a 30% improvement in whole wall R).
Martin - Thanks for the link to the ORNL document.
I realize that a 10% framing factor is impossible in the real world of residential construction.
I think my confusion is related to how the terms "thermal bridge" and "framing factor" are used when people are discussing wood framed walls.
A thermal bridge is about a solid material creating a conductive path for heat flow (path of least resistance). In the outrageous and ridiculous example I gave, a 1/4" gap between studs does break the conductive path between the inside stud face and the outside stud face, right? That same 1/4" gap between the studs would be very meaningful in a steel framed wall, no?
So when we are discussing a wood framed wall---and the ratio between the thermal conductivity of the frame vs. the insulation is very low (1:3)---does it make any sense to talk about thermal bridges? Should we limit our vocabulary to a discussion of the framing factor and the wall thickness?
Daniel, my quest now for quite some time is for the discussion to move to whole assembly R value. What you are saying is the same thing.
State the assembly and or the factors that determine all and then conclude by stating whole assembly R value. From there one can make further use of these factors to plan and build a building.
GBA should at the advisor and blog level move fully in this direction and do so now. Any partial insights less than this lead no where that's ultimately useable.
I know, who will rip into me for this, please try to be polite, if not at least I will be polite. 2011 let's talk green and do so as a group with green in common and a desire to show our best side.
Robert - Thanks for the responses.
You stated, "For a thermal break to be effective, it has to be both high enough in R/inch and thick enough to represent a significant improvement over the wood framing."
With wood at R-1.25, and the cavity insulation at R-3.5, can there even be a thermal break, or is the heat loss calculation simply a matter of overall wood volume (and R-value) versus the overall insulation volume (and R-value)?
Someone posted a link recently that had some information on double-stud walls and the effect on staggering studs in a wall cavity. See the link here:
http://www.100khouse.com/2010/02/25/moving-from-sips-to-double-stud-walls-in-skinny/
The THERM graph references a study by the Consortium for Advanced Residential Buildings (CARB), however, I could not find the source document on their website, so I don't know if it's information or mis-information. See their website here:
http://www.carb-swa.com/
AJ - I don't think we're asking the same question, or pointing in the same direction.
I've read some of your posts on GBA requesting a whole assembly R--value. Although it's not a bad idea, I find the discusssions of component R-values to be helpful, not hurtful.
I think what you are looking for is a whole-house U-value, which can be translated into a Btu/hour rating (loss/gain), given a specific temperature---say a particular location's winter or summer design temperature. That would be a great metric to add to the list, but it only gives you the big picture. It doesn't allow you to drill down to the details in the wall/roof/foundation assembly, or better understand heat flow through an assembly.
For GBA to give the whole-assembly U-value for all the buildings they feature, they would have to hire somebody with PHPP to run the numbers. Sounds like a lot of $$ to me . . . and financially impossible.
To put some units to the conversation, the Passivhaus Institute uses a Thermal Bridge definition (called psi value) of < 0.01 W/mK (Imperial Units: Ψ ≤ 0.006 Btu/hr-ft-°F). This is a linear value (measured along a foundation edge, for example), not one of area. If you conduct more heat than this amount, they consider that to be a thermal bridge and you need to mitigate it or account for it in your energy calculations.
So, in your example of a 1/4" spacing between studs, to determine if you have a thermal bridge or not you would calculate your psi value with a tool like THERM (http://windows.lbl.gov/software/therm/therm.html).
A rule of thumb they use as well is to strive for an unbroken layer of R-10 minimum to avoid these calculations.
For more on these calculations (and why it's cheaper to avoid them during design than calculate them all...) see David White's tutorial: http://rightenvironments.com/low_energy_buildings_class/101109%20Thermal%20Bridge%20Analysis%20for%20the%20PHPP%20v5.pdf
Oh, and if you want to go deep into the subject, this book (http://www.amazon.com/Passivhaus-Bauteilkatalog-Details-Passive-Houses-Konstruktionen/dp/3211994963) contains a library of passivhaus recommended construction details with calculated thermal bridge values for each, as well as a discussion for significant details of primary energy content, global warming potential, acidification potential, weighted waste volume, number of layers, insulation system workplace hazard health evaluation, sound insulation rating and more!
The hatches don't look right from a US tradition, and many of the materials and construction systems are different and inapplicable to our methods, but it's quite an amazing collection of information in both German and English. I don't know of anything like it from over here.
Take one look at a thermal image of a standard insulated stud wall and your question will be answered.
And I just posted this response:
I question the conclusions of the CARB simulation comparing aligned with staggered studs: “The results are that staggering the studs on our 9" wall would result in an increase in R-value of less that 0.5.”
I also question the statement that “this 9" thick wall...will result in a minimum of an R-38 wall”, even assuming the claimed R-4.2/inch.
While one-dimensional modelling isn’t quite as accurate, a UA analysis will show that your wall system, assuming a 16% framing factor, would result in a whole-wall R-value of 31.7. Staggering the studs by 12″ would yield a whole-wall R-value of 35.7.
Jesse - Thanks for the detailed and technical responses. I'm not so sure David White's statement that "you don't have to be a masochist anymore" is accurate---when discussing thermal bridge calculations. ;)
Looks like I have some serious homework if I want to understand this well.
Robert - Point well taken!
So, using the PH rule of thumb (unbroken layer of R-10 insulation), you would need ~ 3" separation between the studs in a double-stud / cellulose wall, before you could say that the thermal bridge is insignificant to the heat loss calculation.
That PH "rule of thumb" is about as useful as the original meaning of the term: how thick a switch could be legally used on one's wife for punishment before it amounted to abuse.*
* actually, this much-ballyhooed origin of the term has been discredited by historians, though it has been propagated by feminists since 1976.
Leave it to Riversong to know the history of some obscure statement! Where you get all this Robert, I'll never know. In your stock of photos, have you any thermal images of a 12" double stud wall, or a Larsen/Riversong truss? Those would be quite interesting to me. thanks. j
John,
I wish I could afford a thermal imaging camera. Sorry, no UV pictures.
Robert,
I think you mean IR (infrared), not UV (ultraviolet).
From Wikipedia: http://en.wikipedia.org/wiki/Rule_of_thumb
"The term is thought to originate with wood workers who used the width of their thumbs (i.e. inches) rather than rulers for measuring things, cementing its modern use as an imprecise yet reliable and convenient standard.[3] This sense of thumb as a unit of measure also appears in Dutch, in which the word for thumb, duim, also means inch.[4] The use of a single word or cognate for "inch" and "thumb" is common in many other Indo-European languages, for example, French: pouce inch/thumb; Italian: pollice inch/thumb; Spanish: pulgada inch, pulgar thumb; Portuguese: polegada inch, polegar thumb; Swedish: tum inch, tumme thumb; Sanskrit: angulam inch, anguli finger; Slovak: palec, Slovene: palec inch/thumb."
From the same Wikipedia page:
It is often claimed that the term originally referred to a law that limited the maximum thickness of a stick with which it was permissible for a man to beat his wife, but this has been discredited. British common law before the reign of Charles II permitted a man to give his wife "moderate correction", but no "rule of thumb" (whether called by this name or not) has ever been the law in England. Such "moderate correction" specifically excluded beatings, only allowing the husband to confine a wife to the household.
Nonetheless, belief in the existence of a "rule of thumb" law to excuse spousal abuse can be traced as far back as 1782, the year that James Gillray published his satirical cartoon Judge Thumb. The cartoon lambastes Sir Francis Buller, a British judge, for allegedly ruling that a man may legally beat his wife, provided that he used a stick no thicker than his thumb, although it is questionable whether Buller ever made such a pronouncement (poor record-keeping for trial transcripts in that era make it difficult to determine whether such a ruling may have existed). The Body of Liberties adopted in 1641 by the Massachusetts Bay colonists states, “Every married woman shall be free from bodily correction or stripes by her husband, unless it be in his own defense from her assault.” In the United States, legal decisions in Mississippi (1824) and North Carolina (1868 and 1874) make reference to—and reject—an unnamed "old doctrine" or "ancient law" by which a man was allowed to beat his wife with a stick no wider than his thumb.
In 1976, feminist Del Martin used the phrase "rule of thumb" as a metaphorical reference to describe such a doctrine. She was misinterpreted by many as claiming the doctrine as a direct origin of the phrase and the connection gained currency in 1982, when the U.S. Commission on Civil Rights issued a report on wife abuse, titled "Under the Rule of Thumb."
I meant IR but I was drunk on UV.
The info here (not the part about the stick) leads me to believe that the Mooney Wall doesn't do much to combat thermal bridging since the horizontal members only provide an 1 1/2" space. Am I correct?
John
John,
You are referring to the cross-hatched wall, which was around long before Mooney. That's the simplest non-foam method for reducing thermal bridging. It's not so much the 1½" of extra insulation (though that adds to the overall R-value) but the fact that thermal bridging is reduced to the 1½" x 1½" intersections of the horizontal and vertical framing.
Daniel,
Great topic, and I'm so with you! An attempt to restate your questions:
- how accurate is a simple weighted average calculation for determining the overall R-value of a wood-framed assembly? how significant are the nonlinear/2-dimensional effects in this case, and need they be considered?
- how much benefit do you get by staggering studs or incorporating continuous insulation?
My short answers: simple calcs are accurate enough, probably, for most purposes. Don't bother with staggering double studs if you have at least 1" insulation between them. Use continuous exterior insulation if needed to achieve your overall goal assembly R-values (but not purely to provide a "thermal break" for the full-depth wood members). Instead, and no matter how your assembly is designed, get serious about minimizing the framing factor.
In terms of building assemblies, I think of thermal bridging as when a material has a higher conductivity relative to surrounding materials in the same layer. It's a given that you need to do a weighted average to account for the different conductivities; but what about the 2-dimensional heat flow effects near the material with higher conductivity? When the ∆u between materials is “small enough”, I feel the effects aren't significant enough to warrant time-consuming THERM models and that simplified calculations are generally adequate for modeling purposes or general design decisions--at least for the level of precision I tend towards. Personally I'd rather spend 10% to get 95% of the way than 100% to get everything supposedly perfect.
To roughly account for the nonlinear aspects of heat flow, reduce the simply calc'd value of a wall that's designed with full-depth studs (and no continuous insulation) by about 5%; or if you have staggered studs, or aligned studs that are separated by 1" or more, then about 1-2% (as you presumably still have top plates etc. that go full depth and should be considered). I base this on my own experience/calcs as well as the CARB staggered stud study you referenced, really interesting; here's the link:
http://www.carb-swa.com/articles/carbnews%20articles/CNAug09.pdf
For a 7" deep dbl-stud cavity w/ aligned (& touching) studs, their THERM analysis shows 3.8% more heat loss than with the staggered stud approach (R-23.8 vs 24.7). The actual case would be worse due to top plates & other full-depth members that aren't considered in these models… so let's say 5% is the reduction you see in this worst-case situation due to the nonlinear heat flow impacts. Now it gets interesting: change to an 8" wall so that 1" of insulation is now between the aligned studs; there's now only 1.4% more heat loss compared to staggered studs (R-27.8 vs 28.2).
Compare these percentages to a simple 2x6 framed wall I've analyzed: wall with 22% framing factor has 8% more heat loss than a wall with 15% FF. Framing factor should be the priority in the design of our wood-frame assemblies; it has much more impact than mitigating the ~1-5% losses from nonlinear thermal bridging effects.
Here's my suggested approach to wood framing design:
1. determine the overall assembly R-values needed to achieve your energy performance goals
2. design the assemblies accordingly, including advanced framing design/accurate framing factor calcs. maybe continuous exterior insulation will be needed to beef up the total R-values, and maybe the added benefit of minimizing nonlinear heat flows will hugely help your overall picture (or meet Passive House if you're really riding the line!), but maybe not…
3. see attached advanced framing guide for your use
Katy,
I appreciate this philosophy, and it is a good argument against the extremism of the PassivHaus approach.
There are, however, some deficiencies in your double-stud prescription. There is no need for any continuous members except plywood door and window boxes and plywood double top plates to tie the two walls together and transfer shear and wind loads. And, beyond thermal bridging, there is a much more compelling reason to stagger studs. One of the advantages of a double wall envelope is that both outer and inner walls can be framed on center from their respective corners, making sheathing, siding and interior finish easier and less wasteful.
And, once the decision is made to use a double wall for optimizing R-value, it's counterproductive to reduce the drying potential by applying exterior foam board when the wall thickness can be adjusted to meet desired insulation levels. In fact, perhaps the primary advantage of a double wall or truss wall envelope is that it allows high insulation levels without the need to resort to low-perm petrochemical foams.
Another advantage is that so-called "advanced framing" techniques become less important . They were developed to maximize the thermal value of stick-framed walls while saving a bit of lumber, but they do not offer as strong a structure as traditional methods nor as much nailing surface for trim and siding. It's possible to build a 12" thick truss wall with no more lumber than a conventional 2x6 plywood-sheathed wall - and without any "advanced framing" techniques beyond 24" oc spacing.
Katy - You got it, exactly:
"How accurate is a simple weighted average calculation for determining the overall R-value of a wood-framed assembly? how significant are the nonlinear/2-dimensional effects in this case, and need they be considered?"
Thanks for the analysis and comparisons between thermal bridging and framing factor, which is what I was trying to better understand.
Daniel, sure thing. Here's another good resource that explains the Modified Zone Method for calc'ing R-values of steel stud walls (but is helpful for understanding the thermal bridge "zone of influence" from any higher conductivity material):
http://www.ornl.gov/sci/roofs+walls/calculators/modzone/index.html
I wish the actual tool were also offered for wood. Anyway, if you have the PHPP, the R-value calculation tool does account for thermal bridging. I'm not sure of the algorithms, haven't dug into the backend equations, but the results are different than what you get from a straight weighted average, so there's something going on behind the scenes, I think...
Robert, in general I think we agree: one, if it makes sense to design/build with staggered studs for whatever your reasons, that's great; I just don't feel it's crucial to go out of your way to do that for thermal bridging concerns (I haven't ever worked w/ a double stud wall so can't speak for which approach is least expensive, costly etc., but in talking w/ 3 different builders on the subject seem to always get 3 different viewpoints). Two, I didn't mean to imply that you would or should use both a double stud wall with exterior insulation--indeed in that assembly, assuming a very cold climate, you'd need so much ext. insulation to not cause a building science problem that I don't see how you could even build it. I agree that if you're doing a double-stud wall, presumably the whole point is to achieve the whole wall R-values you're seeking, so you then don't need to incorporate continuous insulation.
But I disagree with your statement that in a double-stud assembly, advanced framing "becomes less important". Minimizing wood = maximizes insulation and so whole-wall R-value, simple as that. Unless you have a system that utilizes no cavity insulation, this will hold true. I also disagree that an advanced framed structure is not as strong as one built w/ standard techniques... but that's a topic for another thread.
please visit the following link - it has great information on thermal bridging and building energy efficiency. http://www.engineering.com/Videos/LearningSeriesChannel/VideoId/2735/2-Rs-Wont-Make-Your-U.aspx