AECB Website and Point Thermal Bridges
Some interesting thoughts about thermal bridges in this document…..
http://www.aecb.net/PDFs/carbonlite/1420_AECB_VOL_2_PM_V8.pdf
“Point thermal bridges -…..the nails, screws and other fixings in timber-frame construction.”
My Question is Concerning Outsulation Strategies…..
What is the thermal effect of the 100’s of large “nails,screws and other fixings” that penetrate the outsulation and terminate in the studs and plates( which are also thermal bridges)?
In other words…a thermal bridge connected to a thermal bridge.
Is it significant?
Are there any guidelines or field studies that estimate the degrading of the outsulation?
The Europeans seem to believe it is a concern…
Robert Riversong has also brought up the subject.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
John,
Every construction system has its plus and minuses. We're always involved with trade-offs.
Adopting the PERSIST method brings huge improvements in airtightness and insulation levels compared to conventional construction methods. Of course, fasteners slightly degrade the thermal envelope. I'll leave it to the engineers to quantify the problem, but plenty of monitoring studies provide data showing wonderful results achieved by PERSIST buildings.
If you lie awake at night worrying about thermal conduction through screws, I guess you are a prime candidate for SIPs. But then you'll have to lie in bed at night worrying about the durability of OSB.
We're building better buildings, not perfect buildings.
As I often tell people, insulation doesn't stop heat flow; it just slows it down. Heat will always flow from hot to cold. All we need to do is slow it down enough so that the results achieve the goals we set.
Yes, it is signficant - much more so than advocates of "outsulation" will admit. While we add foam board with an R/inch of 5, we're then piercing it with steel that has an R/inch of .0029. And, though the percentage of surface area in fasteners might be as little as 0.04% (16" x 16" pattern), it seriously skews the weighted U-value of the "outsulation".
For instance, nailing vertical strapping with 10d nails through 1" XPS into 16" oc studs, nailed vertically 16" oc, will reduce the thermal effectiveness of the foam board 46.7% (the same percentage degradation holds for thicker foam board).
Nailing 5" exposure clapboards with 8d box nails through 1" foam board into 16" oc framing, will reduce the thermal effectiveness of the "outsulation" by 71.2%.
Robert,
Whoa, slow down. I don't believe your numbers. I'd like to know:
1. What are you assuming for a fastener diameter?
2. Do you assume the point of the fastener is in contact with interior temperatures (unlikely) or that the point of the fastener is buried in the center of a 2x6 stud?
More info, please.
I used 0.113" for 8d box and 0.128" for 10d box or pneumatic. Since R-value and U-value are per F° delta-T, they are independent of actual exterior or interior temperatures. I ignored the steel penetration into framing members, which would also slightly degrade their thermal performance as well.
Feel free to check my numbers and calculations, but this is consistent with information I've come across that nailing through foam board can degrade its effectiveness by 40% - which is what I've been teaching in my classes. This is the first time I've actually written a spreadsheet to crunch the numbers, and it appears worse than the estimates I've seen.
WOW!! This is shocking...simply hard to believe at first glance. The good folks at BSC seem to be big time advocates of outsulation...maybe Martin could get their comments on this.
Before Microwave ovens...
One of the fastest ways to bake a potato was to stick a nail in it!
At first, I thought that as the depth of the foam would increase, that the effect would be less...but then I realized that as long as there is 1" of steel screw for every 1" of foamboard, the effect remains the same??
Yah, the ratio of U-value and the overall degradation percentage remain the same, regardless of the thickness of foam. In fact, the larger nails required with thicker foam board would increase the relative surface area of steel and the amount of degradation.
It's not so surprising when you consider that steel is approximately 1700 times as thermally conductive as XPS, 85 times as conductive as water, and 30 times as conductive as concrete.
Robert,
What calculation methodology did you use? There are a number of techniques described in the ASHRAE handbooks for this sort of thing; my impression is that they are good approximations for timber frame but not all of them are suitable for dealing with the effects of steel due to its higher U. (This is usually a concern in the context of steel studs, but the principle remains the same for nails.)
Also, you said "I ignored the steel penetration into framing members, which would also slightly degrade their thermal performance as well." Where, then, did you assume that the interior end of the nail terminated? Did you let it conduct directly to the space? Having it embedded in the stud would degrade the performance of the stud, it's true, but the total thermal short-circuiting effect for the wall should be less if the nail is embedded in the stud, than if it is exposed directly to the interior condition. Due to the relatively lower U-value of the wood, the stud acts like a thermal break in this context (even though the stud is itself a short-circuit path relative to the insulation).
I simply used area weighted U-values based on standard conductivity values (350 Btu-in/SF-Hr-F° for steel and 0.2 Btu-in/SF-Hr-F° for XPS) and standard nail diameters. It's a straightforward 2D conductivity calculation. The 3D effects of steel studs and heavy framing do not apply.
Since I was calculating only the degradation of the effective R-value of the exterior foam board, where the nails terminate is irrelevant. That becomes relevant only in calculating the whole wall R-value of a specific wall section. It also has an impact on the potential for wall-cavity condensation.
Robert's assesment still makes sense to me...(no debunking to date)
I am an Architect not a Scientist or an Energy Expert....
Any debunkers out there?
Would you share the spreadsheet you referred to? I'd like to see the numbers. I could redact it myself, but I frankly don't have time and probably wouldn't get around to it.
Brent,
I posted my input values. It takes less than 5 minutes with a calculator to do the math. It's a conventional area-weighted U-value computation.
I don't dispute the math , but I do wish to see more discussion on the matter of termination of the nail, as raised in #9 and #10. If we talk about total R along parallel paths (ignoring lateral heat flow for simplicity), the total R of 1" of XPS foam plus an inch of wood would be 6 (round numbers). The path through nail and the inch of wood would be (0.0029 +1) or about 1. The ratio would be 6/1. If we consider only the foam and nail, both terminating in inside air, then the ratio of R values would be 5/0.0029 = 1724 (and 1724 times 0.04% area ratio is 0.69, as RR said, a sizable fraction).
Since total heat transmission rate does involve total wall R value, I would expect that how the very conductive parts are terminated can't be ignored. This would be much like the matter of thermal bridging of studs, with the space between studs filled with some insulation. Percentage-wise, the wood studs have a greater impact on the insulation with higher R. A layer of foam over the whole thing adds to the R of each of the parallel paths. While the thermal bridging of the stud is only reduced (not "eliminated" as often is claimed), the impact of the stud on the whole wall R is reduced greatly.
Now, nails through foam into a metal stud would indeed be a disaster, thermally.
Dick,
I don't know what 1" of wood you're refering to in the parallel heat paths, so I don't follow your reasoning.
But I just found the error in my calculations (I'm surprised no one else has): I failed to convert the nail diameter to nail cross-sectional area. So here is a revised perspective, not nearly as dire as I had suggested, but still demonstrating the weakness of nailed "outsulation".
1" XPS with clapboards at 5" exposure nailed into 16" oc studs with 8d box nails would result in a degradation of the foam board of 18% overall. But the 1½" x 96" (1 sf) of foam over the stud (thermal bridge) will be degraded by 70.9%. If you also consider the reduction in R-value of the wood stud/sheathing by 3/4" or so of nail penetrating it, then there will be little if any change in heat flow through the thermal bridge. So one of the primary functions of "outsulation" - the elimination of thermal bridges - will be effectively eliminated, at least at studs which can be 10% of wall area.
While the uninterrupted foam board over the cavity between studs will increase the whole-wall R-value, with no thermal break at studs the percentage of degradation of whole-wall R-value will be greater. The more insulation we add to a thermal envelope (I like that term, even if Lstiburek doesn't), the more important it becomes not to undermine it.
Now you could be smarter and, instead of nailing the clapboards on top of the foam board, you could nail up vertical strapping 12" o.c. with 10d pneumatic nails for a rainscreen. But, even if you then attached the clapboards with nails that only penetrated the strapping, there would still be a 55.6% loss of XPS R-value over the studs. If the clapboard nails penetrated even part way into the foam board, then you're probably back to where you'd be without strapping - no thermal break.
Is thermal bridging a concern with the hundreds of fasteners used on outinsulation? Yes, it absolutely is.
Is it more of a concern then thermal bridging through solid framing members since metal has a much higher conductivity then wood? Theoretically in form of a excel calculation maybe - in real world conditions I have not seen this confirmed at all. It is surely absolutely crucial that fasteners need to be driven into the stud members (min 1-1/2"embedment) for two reasons - structurally as it holds the furring/siding/roofing and also of course for thermal conductivity.
Outinsulation wall systems in form of PERSITS or REMOTE have been around for a long time and we have done extensive research and testing on these wall systems. By now we now exactly what is going on in the wall and how well they perform. There are 1200 sensors-data loggers installed at CCHRC's REMOTE walls.
Over the years I have build any imaginable wall system from double wall, larsen, SIP, ICF and hybrids of any system. None of them performed as well as the REMOTE walls we are building with today in our extreme cold climate. Does this make me an outinsulation " advocate" with a biased opinion? Maybe. I build in one of the most challenging climates there is with extended periods of extreme cold down to 50 below. This is a outside temperature were you can take a boiling cup of hot water and throw it in the air to see it evaporate in a puff of cold smoke.
So in a sense I feel that I have a very good idea about thermal bridging factors on this specific question and I think that a claim of a degradation factor of 46-72% through metal fasteners in outinsulation wall systems is simply not true and biased. I do not see this confirmed in any real world experience in form of data logging, thermal imaging and many very successful project we monitor for many years. If it would be the case I could not build 4000 SF homes which consume only about 500 gallon of heating fuel per year at 14,000 heating degree days and are still affordable for the average client.
I can understand that the concept of PERSIST/REMOTE seems strange to someone looking at it the first time - but what I simply don't get is why some folks deliberately try to claim that it doesn't work and perform. Reading through many of the latest discussion and arguments just really makes me wonder why were are trying to go backwards and seem not be able to embrace the latest building science and techniques (especially if it is coming from Europe it seems). What we all need to do is let go of our egos and look at the big picture and the challenges ahead of us...
Thorsten,
Maybe you need to thaw out a bit and calm down. I don't seen any evidence of egos here, nor is there any "bias" in stating the simple mathematical and physical facts. Bias is in how one chooses to interpret facts, and no one in this thread has even mentioned the REMOTE or PERSIST envelope systems.
The question was whether nailing through rigid foam board has a significant impact on its thermal performance. The answer is yes. On that, it seems, we agree.
We also agree that "a claim of a degradation factor of 46-72% through metal fasteners in outinsulation wall systems is simply not true", but nobody here made that claim. In fact, there was no "claim" at all, merely a mathematical proof of the loss of much of the thermal break advantage of rigid foam when penetrated by lots of nails.
The high percentage of thermal degradation of foam by nail penetration is due to its relatively high thermal resistance. Nails into wood would have significantly less relative degrading effect because of the much lower thermal resistance of wood.
This is not opinion - it is basic physics. I found an error in my initial caculations and corrected it. If you can demonstrate an error in either my math or my logic, please present it. But defensiveness doesn't settle an argument.
This comment trail is a real work of art.
Thorsten, thank you for adding some clear brush strokes to an otherwise generally hazy and distorted painting.
"Clear brush strokes"? - only if you come to the issue with the kind of bias that Thorsten expresses. The only "hazy and distorted" contributions have been his, perhaps fogged by that tossed cup of hot water that evaporates into a "puff of cold smoke".
What Thorsten attempted to do was to point out that, while it's useful to analyze each element of a building's thermal envelope, it's also crucial to consider it as a whole system. My modified Larsen Truss wall system also performs far better than most others, but that does not mean there are not specific weaknesses that can be understood (and then mitigated) by careful detailed analysis.
But this thread was about an analytical consideration of a single element of the outsulation approach. And, unless one understands the thermo-mechanical behavior of that element one cannot determine its appropriateness in the whole-building system.
For instance, Thorsten said, "It is surely absolutely crucial that fasteners need to be driven into the stud members (min 1-1/2"embedment) for two reasons - structurally as it holds the furring/siding/roofing and also of course for thermal conductivity."
If we agree that 1-1/2"embedment is necessary for structural reasons, then that's a given for either outsulation systems (though it will require a larger diameter nail with greater conduction) or insulation systems. In both cases, there's the same penetration of steel into wood framing (thermal bridges). The only difference is that, in the outsulation system, such nailing significantly weakens the thermal break over the thermal bridge and increases the relative heat loss percentage through that bridge.
If the nails were placed for "thermal conductivity", by which I assume Thorsten means to reduce thermal conductivity through the envelope, then they would be better placed between studs where there would be more R-value interior to the point of the nail (if cavity insulation were installed), hence a lower temperature and less delta-T to drive heat through the conduction path.
Thorsten's plaint "I can understand that the concept of PERSIST/REMOTE seems strange to someone looking at it the first time - but what I simply don't get is why some folks deliberately try to claim that it doesn't work and perform." assumes that I and others haven't taken a good hard look at those systems and completely misrepresents what has been stated because no such claim has been made. It makes him appear to be an advocate for a particular system rather than an objective seeker of understanding.
That, for Thorsten, this discussion "makes me wonder why were are trying to go backwards and seem not be able to embrace the latest building science and techniques" is full of unexamined assumptions and biases. First, what we are talking about in this thread is pure building science. Second, where I have critiqued the REMOTE/PERSIST systems elsewhere, it's been almost entirely based on building science (including the dangers of encasing a wooden frame in impermeable membranes on all sides). And third, his statement reflects an uncritical assumption that progress is defined by using more and more of what worked when we used a little, or that his approach represents "progress" while other less synthetic and extreme approaches are retrogade.
Robert, I'd like to check your math against my own:
I get a surface area of a 10d fastener as .01286 sq.in., and the area of insulation in a 16x16 pattern as 256 sq.in.
The ratio of these provides an area ratio of fasteners to wall of .005%, not the .04% you mentioned above. I see that you corrected the area ratio somewhat later on.
So, for the 16"x16" pattern you started with, with steel having an R-value 1750x that of the XPS, the heat loss through the fasteners will be 1750x.005%, or 8.8% total wall R-value reduction. If you count only the 1.5"x16" stud portion, the reduction in R-value over the stud is about 35%, with no reduction in the between-stud R-value. For your later analysis of strapping nailed at 12" OC with 10d nails, I get a 46% reduction over the studs, not the 56% you posted. Still, clearly a significant number.
But I think Dick's right about the 3D effects of the nail termination. 2D calculations only work if the wall is heated/cooled to the same temperature on both faces, and if the heat supplied is infinite at that temperature. Meaning, if a higher conductivity material withdraws heat from an area, you don't get a cold spot, you pump more heat at the area to keep it at a constant surface temperature. Engineers model this by assuming that there is a hot plate fastened to the wall on one side and a cold plate on the other. The reality of course is that cold spots happen, and we see them easily with IR cameras and soot deposits on walls. Once a cold spot develops, the surface temperature is lower and total conduction decreases compared to the 2d calculations. Sure, there is more heat per square inch flowing through the fastener, but not as much as the constant-temperature analysis suggests.
More important is the path of heat flow to the fastener at the tip. If the fastener tip is exposed to the inside of the conditioned space, then it has full access to the inside temperature and heat source, and the 2D analysis is closer to correct. But if the fastener terminates in the stud, heat must first flow through the stud before entering the fastener.
For an uninsulated wall cavity, the heat path through the wall between studs is through the sheathing and XPS, and the 2D analysis works pretty well. But at the fasteners, the heat first flows laterally through the stud (3/4", about R .75), and then outward through the fastener (about R-0). So the effective R-value at the fastener is actually R.75. The effective R-value for the rest of the XPS over the studs ranges from R5 to R5.75, because the heat still has to flow laterally through the stud before it flows out through the XPS. Of course, the 3D analysis is more precise than these approximations, but I think you get the idea. So, with an effective fastener R-value of .75, the effective R-value reduction of the fasteners is only .03% - an insignificant number.
Thorsten understands this intimately, and has posted in several areas about his experience that fasteners penetrating the outsulation must terminate in a stud, otherwise condensation occurs at the fastener tips. If your 2D analysis were correct, condensation would occur at the fastener tips regardless of their termination.
Of course, the wall above only has a total of R5 of XPS and uninsulated cavities. If you go up to R20 on the outside, the relative R-value loss from the fasteners is more like .1%. But this is still not a high-performance wall at R20. Once you insulate the cavity, the heat flow becomes a bit more complex, because the heat is flowing through the studs and cavity insulation at different rates (faster through the stud, of course), and the inside of the XPS is not at a constant temperature. I suspect that with an insulated cavity, the effective R-value loss due to fasteners will be greater than with the uninsulated cavity, but still will not approach the 50% or so reduction shown by the 2D analysis.
BSC just released a new paper analyzing a number of different wall systems, and using 3D heat flow analysis for the calculations. http://www.buildingscience.com/documents/reports/RR-0903_High-R_Value_Walls.pdf. Some very interesting results.
It's a long paper, but I didn't see any specific discussion of fastener effects in the various wall systems. Might be interesting to discuss with Straube. I noted particularly that the truss wall did very well, except for condensation potential on the OSB sheathing used for the analysis. This certainly supports your use of sawn plank sheathing rather than OSB for your own truss walls.
Most pertinent to this thread is probably the retrofit walls where they install 2x4 studs on lag bolt standoffs on the outside of the sheathing, then foam the sheathing, filling the space behind the hanging studs and stabilizing them with the stiffness of the SPF. This is one of the highest-performing walls in all aspects measured. The 3D analysis absolutely does not show the much larger fasteners having a significant effect on the total wall R-value.
Bottom line for me? I think the question is still open. I don't have any 3D software or the time to do the analysis, but I suspect that with all heat flow taken into account, the effective total wall R-value reduction of fasteners through the XPS will be on the order of less than 5% for most high-performance walls, and less than 1% for many of them. The fasteners certainly do reduce the effectiveness of the outsulation over the studs more than the total wall numbers would suggest and they decrease the benefit of providing the thermal break, though not nearly to the degree that Robert calculated above.
In the original article posted by the OP, UK designers are encouraged to use a factor that increases the wall U-value by about 27% if "accredited" details are used, and 50% if they are not. This increase in U-value includes all linear (studs, etc) and point thermal bridges, and they are very specific that point thermal bridges includes fasteners. Since the studs in a wood frame wall can easily reduce the total-wall R-value by 30%, it seems the point bridges don't have much more effect. The articles does say that calculating actual effects of point bridges requires 2D and 3D analysis of the specifics.
Thanks, Pete, for an excellent contribution!
I just returned from a four-day trip to Illinois, where I attended the 4th Annual Passive House Conference. Last week I was busy finalizing my preparations for a presentation I gave at the conference, and I didn't have time to respond in depth to Robert Riversong's flawed calculations. In any case, your response is a better one than I probably would have managed.
When Robert wrote, "nailing vertical strapping with 10d nails through 1" XPS into 16" oc studs, nailed vertically 16" oc, will reduce the thermal effectiveness of the foam board 46.7%," I knew instantly that something was wrong with his calculations, which is why I posted my quick answer, "Whoa, slow down. I don't believe your numbers."
Your conclusion — "With an effective fastener R-value of .75, the effective R-value reduction of the fasteners is only .03% - an insignificant number" — sounds right to me, although I'm all ears if other engineers want to chime in with their own calculations. Again, thanks for posting.
Here's someone else who knew there were flaws, but it takes valuable time to properly refute things. I knew there was an order of magnitude error, but had no time to help out, so I had aleady decided to almost ignore the whole comment line here (although Thorsten tried to straighten things out a little). I can do the numbers, but I don't care whether or not the reduction is as low as 0.03%, I just know and care that it's a lot better than the 46% or 76% (or whatever it was) that was being thrown around. Lesson to anyone reading anything on this site or anywhere else while trying to learn something: There are professors everywhere, but you are responsible for what you "learn." You had better get a "second opinion" on just about everything. Not only will you find misinformation, but you will find plenty of professors that will answer questions you never asked, or care about. This ain't college, but it's infinitely cheaper, so it's actually a better rate of return....right?
Thanks, Pete.
Response to Pete Engle, point by point:
"So, for the 16"x16" pattern you started with, with steel having an R-value 1750x that of the XPS, the heat loss through the fasteners will be 1750x.005%, or 8.8% total wall R-value reduction."
The 8.8% is the percentage of total heat loss through the installed XPS that occurs through the fasteners. The reduction in effective R-value of the XPS is 8.1% (using an area-weighted U-value average).
"If you count only the 1.5"x16" stud portion, the reduction in R-value over the stud is about 35%"
If you look at the entire stud (1½" x 8' = 1SF) and add an extra nail at the top of the last 16" spacing (for a total of 7 fasteners), then the R-value degradation is 52.2%.
with no reduction in the between-stud R-value. For your later analysis of strapping nailed at 12" OC with 10d nails, I get a 46% reduction over the studs, not the 56% you posted.
At 12" oc (plus one), the degradation is 58.4%.
"But I think Dick's right about the 3D effects of the nail termination."
Dick was referring to whole-wall R-values, not 3D effects (except by inference when he mentioned the case of steel framing). Yes, whole-wall effects are very different from single-layer effects, which is a different discussion and which I'll get to below.
"2D calculations only work if the wall is heated/cooled to the same temperature on both faces, and if the heat supplied is infinite at that temperature."
You're confusing 2D with steady-state. What we're modeling in residential heat loss analysis is steady-state heat loss, because at an average winter outdoor design temperature the interior temperature is considered to be 65° and the heat source is infinite (until we run out of fossil fuels). 3D analysis models the 3-dimensional heat transfer paths, such as at framed corners, band joists and – most significantly – through metal framing or metal fasteners.
"Once a cold spot develops, the surface temperature is lower and total conduction decreases…"
Steady-state analysis models this. Since there is an interior air film on every surface that (by ASTM convention) offers R-0.68, the drywall surface temperature is always lower than the air temperature, and lower yet at a thermal bridge. For instance, at a 65° indoor and 0° outdoor temperature, the drywall on a 2x6, fiberglass insulated wall cavity will be 62.83°, but over a stud will be 60.4° (cool enough to cause condensation in high humidity areas or stagnant corners).
"More important is the path of heat flow to the fastener at the tip. If the fastener tip is exposed to the inside of the conditioned space, then it has full access to the inside temperature and heat source…But if the fastener terminates in the stud, heat must first flow through the stud before entering the fastener."
The fastener tip would never be exposed to inside temperatures, but at most to inside-of-cavity temperatures or inside of stud temperatures. In a 2x6, fiberglass-insulated wall with 2" of XPS outsulation, a nail penetrating 1½" (same conditions as above) would be exposed to temperatures from 24.9° to 35.2°, while inside of the stud it will be exposed to temperatures from 38.5° to 44.7° (ignoring 3D heat movement through the sides of studs into fiberglass, which will reduce those numbers somewhat).
"For an uninsulated wall cavity, the heat path through the wall between studs is through the sheathing and XPS, and the 2D analysis works pretty well. But at the fasteners, the heat first flows laterally through the stud (3/4", about R .75), and then outward through the fastener (about R-0). So the effective R-value at the fastener is actually R.75. The effective R-value for the rest of the XPS over the studs ranges from R5 to R5.75, because the heat still has to flow laterally through the stud before it flows out through the XPS. Of course, the 3D analysis is more precise than these approximations, but I think you get the idea. So, with an effective fastener R-value of .75, the effective R-value reduction of the fasteners is only .03% - an insignificant number."
Wrong on two counts. First, in an uninsulated cavity, the sheathing temperature will have little variation because of the 3D heat loss paths converging from each side of the studs (effectively eliminating any framing R-value contribution). And, second, the XPS with nails (16" oc) does not offer R-5 over the studs but approximately half that. You're thinking backwards. The wood doesn't "insulate" the nail from conductive losses, the nail effectively brings the outdoor temperature to the interior of the stud (every 16"), dramatically increasing the delta-T between inside of stud and face of stud and dramatically increasing the heat flux and the condensation potential at those points. The effectiveness of the XPS thermal break is reduced by an average of at least 50%, though concentrated at the nail penetration points (where the thermal break R-value is effectively zero).
"Thorsten understands this intimately, and has posted in several areas about his experience that fasteners penetrating the outsulation must terminate in a stud, otherwise condensation occurs at the fastener tips. If your 2D analysis were correct, condensation would occur at the fastener tips regardless of their termination."
Not "would occur" but could occur at sufficient RH levels in the wall cavities, and not just at the buried nail tips but at the stud faces that surround the nails.
"I suspect that with an insulated cavity, the effective R-value loss due to fasteners will be greater than with the uninsulated cavity, but still will not approach the 50% or so reduction shown by the 2D analysis."
The average degradation of the thermal break over each stud is 50% or greater (depending on size and spacing of fasteners – with thicker insulation, the fasteners are thicker). The whole-wall effects are, of course, much smaller but not nearly as small as you suggest.
With a 2x6, fiberglass-insulated wall with exterior XPS (interior air film, drywall, fiberglass or framing, ½" CDX, 1" XPS, ¾" wood siding, exterior air film), the whole-wall R-value degradation by nailing 16" oc into framing compared to glued XPS with no nail penetrations is: 1" XPS – 7.9%, 2" XPS – 10.8%, 3" XPS – 12.5%, 4" XPS – 13.7%. Eliminating the cavity insulation decreases the relative effect of fasteners through foam (because the difference between total wall R-value through cavity and framing is less): 4" XPS – 8.8%, 8" XPS – 12.2% degradation.
"BSC just released a new paper analyzing a number of different wall systems, and using 3D heat flow analysis for the calculations. http://www.buildingscience.com/documents/reports/RR-0903_High-R_Value_Wa.... Some very interesting results. It's a long paper, but I didn't see any specific discussion of fastener effects in the various wall systems."
In fact, that research used 2D modeling, steady-state analysis, but accounted for thermal bridging, which is the same analytical approach that I use. "Two dimensional heat flow analysis was conducted for each test wall using Therm 5.2, a two-dimensional steady-state finite element software package developed by the Lawrence Berkeley National Laboratory at the University of California. Therm was used to calculate the thermal performance of each of the different proposed assemblies including thermal bridging effects."
"I noted particularly that the truss wall did very well, except for condensation potential on the OSB sheathing used for the analysis. This certainly supports your use of sawn plank sheathing rather than OSB for your own truss walls."
The study also noted what I've been expounding: "Case 4, double stud construction, and Case 5, truss wall, experience slightly lower relative humidities because of the moisture buffering effect of the cellulose insulation."
"Most pertinent to this thread is probably the retrofit walls where they install 2x4 studs on lag bolt standoffs on the outside of the sheathing, then foam the sheathing, filling the space behind the hanging studs and stabilizing them with the stiffness of the SPF. This is one of the highest-performing walls in all aspects measured. The 3D analysis absolutely does not show the much larger fasteners having a significant effect on the total wall R-value."
As you noted, the study seems to have entirely ignored fasteners in their analysis, so those results are flawed.
"I suspect that with all heat flow taken into account, the effective total wall R-value reduction of fasteners through the XPS will be on the order of less than 5% for most high-performance walls, and less than 1% for many of them."
--------------------------------------------------------------------------------
"In the original article posted by the OP, UK designers are encouraged to use a factor that increases the wall U-value by about 27% if "accredited" details are used, and 50% if they are not. This increase in U-value includes all linear (studs, etc) and point thermal bridges, and they are very specific that point thermal bridges includes fasteners."
That y-factor increase in clear-wall U-values is based on "typical" building practices, not high-R walls, and they further state: "An alternative to using y-values is for the designer to calculate the heat loss through each linear bridge separately and then add them together." This is exactly what I've done above.
I stand by my numbers.
I neglected to respond to Pete's conclusion:
"I suspect that with all heat flow taken into account, the effective total wall R-value reduction of fasteners through the XPS will be on the order of less than 5% for most high-performance walls, and less than 1% for many of them."
As I've shown, whole-wall R-values are degraded by nailing through foam by 8% to 14% in those specific cases. Suspicions aren't sufficient to settle an argument. Even educated guesses aren't nearly as useful as hard numbers. Show me some contradictory data and I'll reconsider.
It seems that the reactions these numbers are generating is because they puncture some sacred cows of "green building" practice.
There's getting to be far more concepts here than we can easily pick apart piece by piece, but I'll focus on just a bit from the above discussion:
First, I'll grant you the slight differences in our numbers caused by the additional fastener at the top or bottom of the wall. Good point. If nothing else, this discussion helps us focus on how small differences in techniques may have large performance effects.
"Wrong on two counts. First, in an uninsulated cavity, the sheathing temperature will have little variation because of the 3D heat loss paths converging from each side of the studs (effectively eliminating any framing R-value contribution)."
It is exactly those 3D heat flow paths that result in an increase in R-value over the framing. The heat flow in the uninsulated sections is almost purely 2D, or even 1D, through the face of the sheathing towards the exterior. At the framing, the requirement that heat flows laterally through the face of the studs increases the length of the heat flow pathway and the effective R-value of that pathway. The fact that the 2D heat flow converges from both sides does not negate the fact that at the center of the stud, the heat has flowed through a minimum of an additional R.75 of wood framing. In an otherwise uninsulated wall, there is most definitely a whole wall R-factor increase for the framing. Try this excercise: Calculate the R-value of an uninsulated wall (interior air film, gypsum, air in the stud cavity, 1/2" sheathing, siding, exterior air film) with no framing to the same wall with framing. The wall with framing should have a higher R-value because of the R-value of the studs. In this case, the studs are not the thermal bridges, the air spaces are.
"And, second, the XPS with nails (16" oc) does not offer R-5 over the studs but approximately half that. You're thinking backwards. The wood doesn't "insulate" the nail from conductive losses, the nail effectively brings the outdoor temperature to the interior of the stud (every 16"), dramatically increasing the delta-T between inside of stud and face of stud and dramatically increasing the heat flux and the condensation potential at those points. The effectiveness of the XPS thermal break is reduced by an average of at least 50%, though concentrated at the nail penetration points (where the thermal break R-value is effectively zero)."
Nay, it is you who are thinking backwards. Or, we are thinking in opposite directions, but actually looking at the same thing. We might be something like the two blind men describing opposite ends of a mule. Obviously, both effects happen. Yes, certainly, the nail conducts the cold to the heart of the stud very effectively. But, the cold nail tip is not exposed to the same conditions as the interior of the sheathing, because it is insulated by .75" of wood. The important point is that there is no pathway that directly exposes the fastener tip to interior conditions, so the assumptions for the total UA method do not apply. The total UA method is a better estimator of actual wall performance that the installed R-value method because it takes into account the framing factors. But, it doesn't take into account the 2D or 3D heat flow pathways, and these are very important.
If the entire interior face of the sheathing were exposed to the same temperature and same heat source potential, the total UA method would be accurate. But the fact of the studs disrupts the surface boundary conditions so that UA is simply not an accurate model to describe this situation. The insulating value of the studs disrupts the heat flow towards the interior surface of the sheathing so that there is far less heat available to the fastener tip for conduction to the outside. Or, from your perspective, it insulates the fastener tip so that there is far less cold available to further chill the stud cavity. The bottom line is that the assumptions underlying the UA method simply do not work at a point discontinuity like a fastener embedded in the framing. You've got to use far more sophisticated analysis. So, until I see a 3D finite element analysis of the actual heat flow around a fastener tip (and the shank in the insulation), I'll have to disagree with your numbers.
Rather than attempting a response, I posed the question to building scientist John Straube. He e-mailed back, "The effect [that fasteners have on the thermal performance of rigid foam sheathing] is on the order of 1 to 2 percent. Variations of this have been modeled repeatedly. 45 per cent [reduction in thermal effectiveness is] technically preposterous."
3D heat loss analysis always models more, not less, heat loss than 2D or 1D UA models, because it includes multiple heat pathways that are missed in the simpler analyses.
In the case of a REMOTE wall system with only outsulation and empty stud cavities, but large-diameter screws penetrating the XPS, sheathing, and at least 1½" into framing, it is true that the air cavities will be the thermal bridges relative to the framing EXCEPT in the vacinity of a fastener, which creates dramatic heat loss pathways through a section of the stud deeper than fastener penetration (angular paths) as well as angular pathways through sheathing toward the steel and angular pathways through XPS toward the steel.
So the heat loss effect of the steel fastener is significantly greater in 3D than in uni-dimensional UA modeling. This is one of the primary weaknesses of the REMOTE system, both in terms of thermal performance and in terms of hygric performance. The weakness is somewhat mitigated by the "optional" installation of cavity insulation, not in percent of R-value degradation (which increases) but by reducing delta-T at the conductive areas and reducing overall heat loss.
If it were possible to glue, instead of screw, the multiple layers of foam board to the exterior, that would eliminate the dramatic thermal bridge, but then you may as well build with SIPS.
Martin,
Unless you ask the right question, you're not going to get a correct answer. You obviously failed to mention to Straube that the 50% range of degradation I refered to in my corrected calculations was not the entire wall area, but the 1SF of foam directly over each stud (the thermal break).
Robert,
Your calculations are a moving target, so it's hard to have a dialogue.
Your first two postings on this question never mentioned the area directly over the studs; instead, you made very general (and incorrect) assertions:
"Nailing vertical strapping with 10d nails through 1" XPS into 16" oc studs, nailed vertically 16" oc, will reduce the thermal effectiveness of the foam board 46.7% (the same percentage degradation holds for thicker foam board). Nailing 5" exposure clapboards with 8d box nails through 1" foam board into 16" oc framing, will reduce the thermal effectiveness of the "outsulation" by 71.2%."
"This is consistent with information I've come across that nailing through foam board can degrade its effectiveness by 40% - which is what I've been teaching in my classes."
If my posts were a "moving target" they moved only once, after I caught my own calculation error (post #15). I didn't initially look for an error since, as I noted, my initial results were consistent with what I had seen years ago from a reliable source.
Then the discussion went well beyond the simplistic findings to debunk the notion that nailed foam board creates an effective thermal break. And, before your reporting Straube's off-the-cuff and misdirected comment, I had shared a great deal more data.
Perhaps you'd ask Straube about these data from post #26:
If you look at the entire stud (1½" x 8' = 1SF) and add an extra nail at the top of the last 16" spacing (for a total of 7 fasteners), then the R-value degradation [of the thermal break] is 52.2%.
With a 2x6, fiberglass-insulated wall with exterior XPS (interior air film, drywall, fiberglass or framing, ½" CDX, 1" XPS, ¾" wood siding, exterior air film), the whole-wall R-value degradation by nailing 16" oc into framing compared to glued XPS with no nail penetrations is: 1" XPS – 7.9%, 2" XPS – 10.8%, 3" XPS – 12.5%, 4" XPS – 13.7%. Eliminating the cavity insulation decreases the relative effect of fasteners through foam (because the difference between total wall R-value through cavity and framing is less): 4" XPS – 8.8%, 8" XPS – 12.2% degradation.
The two sides in this discussion still seem to be a long ways apart. Hard to argue with Robert's math here...I really wish that Straube could be persuaded to do so. I do have one question however. If the degradation in R-value is as severe as Robert indicates, why does it not easily and clearly show up in infrared photos??
Good question, Garth.
On the outside, the nailheads will be the same temperature as everything else. On the inside, the nail points would show up as cold spots, except they're buried inside the studs with several inches of wood between them and the interior. If they were nailed into steel studs, cold spots would show up on a thermal imager.
Sorry, I am not buying this one. If heat is flowing from warm to cold, the heads of the nails or screws should show up like Xmas lights on a thermal imager.
Thermal imagers detect incident medium and far infrared light, which includes emitted infrared, transmitted infrared (such as through glass), and reflected infrared from other warm sources. Emitted infrared is a function of surface temperature and emissivity.
If the fasteners that penetrate the outsulation are buried behind the cladding, they will be shielded from a thermal imager by the cladding. If the fasteners penetrate the cladding and outsulation, then they will exhibit a temperature gradient from the temperature of the embedding material (outside inch of the framing) to the exterior temperature. The surface temperature of the heads of the fasteners, if exposed to exterior air, will be within a degree or two of the ambient temperature.
Hmmm......imagine that one uses 2" of icynene (R12 total) and connects strapping with 1/4" carbon steel screws 12" oc with studs 16" oc.. We will have 9 screws in the stud, but say 11 to allow for top and bottom strapping.
area of the wall is 16"x8' and the heat flow is 0.26 watt/degF
area of one screw is pi*r^2=3.14*(.125*in)^2=.05in^2 and 11 screws=.5 in^2
U for steel is 87watts/(ft^2*degF), so the additional heat loss is 0.3 watt/degF
But wait, there's more!
A U value of 87 is an R value of 0.01(improves to .02 since it is two inches!). If we add .5R to each screw tip (assuming they are buried in wood), we get an R-value of more like .5 and a heat loss of 0.002watt/degF.
This gives us a percentage loss of 100*.002/.261=0.8%. Not so bad..even though I've grossly over-sized the screws.
Some of the analysis in this topic would be more appropriate to the case of a steel bolt all the way through a steel SIP.
Edit, didn't realize this thread was so old.. sorry for bringing it back from the dead.
It looks to me like there is a major misunderstanding between the two sides...
Riversong is talking about the performance of the exterior foam, not the performance of the entire assembly.
The fastener is a very conductive path for heat to travel...
The math is pretty easy, you have X square inches of nail, and Y square inches of foam.
It doesn't matter at all where the tip of the nail ends up as it is still totally bypassing the foam. you could visualize it by "sawing" the foam off of the sheathing.