? for Martin regarding his report on the BSC whole wall assembly R-value testing results.
ElPotrilloDeGringolandia
| Posted in Energy Efficiency and Durability on
Did the Building Science Corporations report on their whole wall performance r-value tests quantify the decrease in R-value below 50 degrees that they found happening with Poly-Iso insulation? What is the R-value curve for Poly-Iso?
Thanks,
Spencer
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
Spencer,
Samples of 1-inch-thick polyiso have a measured R-value ranging from about R-5.4 to R-6.3 at a mean temperature of 55 degrees F (presumably, this translates to a test performed at 70 degrees F interior and 30 degrees F exterior). At a mean temperature of 25 degrees F (again, I presume this translates to a test performed at 70 degrees F interior and -20 degrees F exterior), the polyiso performs at an apparent R-value ranging from R-4.3 to R-5.6 per inch.
I am attaching (below) the graph shared by Chris Schumacher during his presentation at the August 2012 symposium.
Great thank you for the info. That is exactly what I was looking for. So it looks like it is still very effective insulation...just not as good as everyone thought in cold weather. I am not sure what all the lines on the graph represented...were all the other lines the different types of insulation and the Poly-Iso was the red one?
Thanks a lot,
Spencer
Spencer,
As far as I know, the data points on the curved lines all represent different brands of polyisocyanurate. I'm not sure what the red line at the top represents; I can ask Chris Schumacher if it's important.
All different types of insulation, to varying degrees, change their performance characteristics as the outdoor temperature changes. I've always felt that the best way to deal with this phenomenon is simply to install a little more insulation.
Remember, all of the performance data and energy use data collected at actual houses over the last 40 years are still valid. The amount of energy used by a house over 12 months always reflects real-world performance -- including the performance changes that occur at very low and very high temperatures.
Martin,
Is there a reference with a graph available showing the relative R-value shift with temperature differential for all the common types of insulation?
Ron,
Here are three more graphs: fiberglass batt, EPS, and XPS. All of these graphs came from Chris Schumacher of Building Science Corporation.
As I understand the graph, the outside temperature is shown as a variable horizontally. I assume the temperature on the other side of the insulation, though unstated, is constant and perhaps around 70 degrees F.
So, for a fiberglass batt rated at R-13, it achieves that rated value when the outside temperature is about 58 degrees F.
However, it is not clear to me why the outdoor temperature scale continues upward beyond the typical room temperature. So that raises the question about my assumption that the constant temperature is assumed to be around room temperature.
But, aside from that question about the graph; it appears that the R-value of the fiberglass climbs to around R-15 when the outdoor temperature drops to -30 degrees F. So the R-value of fiberglass increases as the temperature difference between the heated side and the outdoor side increases.
However, I have heard people state that the performance of fiberglass radically falls off when the outdoor temperature falls to the lowest levels of the northern climates. The implication is that a user of fiberglass is not getting the full R-value rating at the extreme low temperatures. The graph seems to refute that premise. What am I missing?
Ron,
You wrote, "It is not clear to me why the outdoor temperature scale continues upward beyond the typical room temperature." Why shouldn't the outdoor temperature be higher than the typical room temperature? That's what happens in July (unless you live in Greenland).
Attics can reach 150 degrees during the summer.
"The R-value of fiberglass increases as the temperature difference between the heated side and the outdoor side increases." Not quite; the R-value (which is defined by law) stays the same, but the thermal performance of the fiberglass batt improves (assuming that there are no convection currents degrading the batt's performance). I mentioned this phenomenon in my August 2011 blog, A Bold Attempt to Slay R-Value.
You wrote, "I have heard people state that the performance of fiberglass radically falls off when the outdoor temperature falls to the lowest levels of the northern climates." That's due to convection currents in attics -- a different phenomenon entirely.
More information on the degradation of fluffy air-permeable attic insulation due to convection currents at low temperatures can be found in this article: Convective Loss in Loose-Fill Attic Insulation.
The sales team at John Manville insulation feel that this research is sometimes misunderstood; here is their take on the data: Convection in Fibrous Attic Insulation.
Martin,
Okay, I see the point of running the temperature values higher than room temperature. And I assume that the fixed temperature to which the varying temperature are compared is 75 degrees F.
It appears that the performance does not simply depend on the temperature difference, but rather, on the two temperature values that define the difference. As I understand, the vertical scale is not R-value, but rather, thermal performance, and that performance rises with fiberglass as the temperature drops. And the actual R-value (in the case of fiberglass) is R-13 at all temperatures.
I will read the links you posted on the fiberglass convection issue. But let me ask this one broad question regarding the issue of fiberglass losing thermal performance due to convection: Does this happen only when the fiberglass is not enclosed on six sides with an air barrier?
Ron,
When very fluffy insulation is installed horizontally on an attic floor, it is subject to convection currents in cold weather. Although the phenomenon is reduced by installing a "cap" above the fluffy insulation -- the usual solution is to install a cap of cellulose insulation, which isn't as fluffy as blown-in fiberglass -- even the cap won't stop the phenomenon entirely.
The best solution is just to install cellulose in the first place.
If you already have the misfortune of having blown-in fiberglass in your attic, you should have as much cellulose as you can afford blown in on top of the fiberglass.
Martin,
Just to clarify my question: Are the convection currents at issue moving from within the insulation to outside of it, and thus carrying heat from within the insulation to outside of it? Or are the convection currents at issue circulating totally within the insulation boundaries?
I assume that it is the former, and if it is:
1) Wouldn’t an air barrier on six sides prevent this type of convection?
2) Isn’t an air barrier on six sides assumed to be required for proper installation?
3) Wouldn’t the omission of an air barrier on six sides be an installation defect?
You introduce the idea of creating an air barrier made of a layer of cellulose, but then you say that won’t cure the problem with fiberglass convection. I don’t see how cellulose can be deemed capable of acting as its own air barrier, and yet cannot lend that function to fiberglass when layered over it. But if it can’t, why suggest that fiberglass should be omitted just because cellulose is incapable of performing as an air barrier for fiberglass? Why not just cover the fiberglass with a woven air barrier that is capable of doing the job?
Ron,
If you install a layer of very fluffy blown-in fiberglass insulation on an attic floor, you can get both kinds of convection: a convection loop within the insulation itself (one that would occur even if you installed a layer of Tyvek on top of the insulation), and also convection through the insulation (as often occurs near soffits -- a phenomenon called wind washing).
In very cold weather, the top of the insulation layer might be at 0 degrees F, and the bottom of the insulation layer might be at 65 degrees F. The delta-T between the top of the insulation and the bottom of the insulation can set up a convection loop that results in a performance degradation. Even if you cap the insulation with Tyvek, this convection loop can occur.
The reason that it's helpful to cap the fluffy fiberglass with 6 inches of cellulose is that the cellulose is not only dense enough to avoid this type of convection loop -- the cellulose also warms up the top layer of fiberglass. That reduces the delta-T from the bottom of the fiberglass to the top of the fiberglass, and reduces the driving force of the convection loop.
You ask, "Wouldn’t the omission of an air barrier on six sides be an installation defect?" The answer is no. Neither the Energy Star Homes program, nor any existing building code, nor insulation manufacturers, nor any "best practices" guide, advocates the installation of an air barrier on top of insulation on an attic floor.
The solution to this problem is to avoid the use of blown-in fiberglass and to install cellulose instead.
Martin,
Okay, I understand your information that an air barrier on six sides is not stipulated as a requirement by the manufacturer or the regulations. You also say it is not advocated by any “best practices” guide. But doesn’t GBA advocate it? Or is GBA saying that the air barrier on six sides is only advocated for walls, but not in attics?
I understand your point that Tivek on top of the fiberglass will not prevent internal convection. But will Tivek on top prevent the convection loop through the fiberglass? If Tyvek would stop convection through the fiberglass, why not use it instead of concluding that the only solution to the problem is to not use fiberglass?
Regarding the issue of internal convection which air barriers are fundamentally incapable of preventing:
You refer to internal convection loop as a “performance degradation.” In relation to what, is it a degradation? The term “performance degradation” implies a failure to deliver what was intended. Yet, a lack of internal convection was never intended or promised.
No insulation is perfect at stopping heat loss. Is this lack of perfection considered to be a performance degradation?
Ron,
The reason that GBA doesn't promote the installation of Tyvek on top of insulation installed on an attic floor is because installing a little bit more cellulose is cheaper than Tyvek and more effective.
Q. "You refer to internal convection loop as a performance degradation. In relation to what, is it a degradation?"
A. The internal convection loop that occurs in a fluffy fiberglass insulation degrades its performance -- that means, makes it less effective at slowing the flow of heat. You ask, "In relation to what?" My answer: In relation to cellulose.
If you install a layer of R-38 cellulose on your attic floor, it will perform better at 0 degrees F than a layer of R-38 fiberglass. The reason: the cellulose is less susceptible to convection loops.
Martin,
Again I need a clarification to distinguish internal convection from external convection. When you say: “If you install a layer of R-38 cellulose on your attic floor, it will perform better at 0 degrees F than a layer of R-38 fiberglass,” would this be the case if you did enclose both types of insulation with air barrier on six sides, thus preventing convection loops from inside the insulation to the space outside of it?
Q. "When you say: 'If you install a layer of R-38 cellulose on your attic floor, it will perform better at 0 degrees F than a layer of R-38 fiberglass,' would this be the case if you did enclose both types of insulation with air barrier on six sides, thus preventing convection loops from inside the insulation to the space outside of it?"
A. Yes. The internal convection loop speeds heat loss. The moving air carries heat from one side of the insulation thickness to the other.
Martin & Ron
you may find Lstiburek's latest insight interesting
http://www.buildingscience.com/documents/insights/bsi-064-bobby-darin-thermal-performance/
I am thinking if there were a "good" air barrier on "all six sides" that the internal "looping convection" would not be a big deal... because the cavity is not-so-tall
There are some more on-topic comments at a recent discussion @ HomeEnergyPros
http://homeenergypros.lbl.gov/forum/topics/another-irreverent-whine-from-dr-joe
The exception would be when high-density R38 batts designed for cathedralized ceilings are used rather than low-density blown fiberglass or low density batts, since the HD38s are about as air-retardent as open-blown cellulose. While neither will match the air retardency of a sheet of OSB or Tyvek, at some density and thickness the air retardency is high enough that from a thermal performance point of view the convection loss issues become negligible- the convection still occurs, but at a tiny fraction of the volume rate.
With cellulose even 3" of 1.2-1.5lbs per cubic foot (a typical open-blown density), it's pretty much there, cutting the convection-communication with the attic air by something like 90%. Taking it to 3lbs density or 6" depth would have only modest improvements in air retardency & convection loss. With fiberglass it takes about 1.8lbs density, but open-blown fiberglass is typically half that density (or less)- most AIR FILTERS are more air-retardent than open-blown fiberglass(!).
But the cellulose option is far cheaper than high-density batts, and since it's blown-in, it has a lower risk of undetected uncorrected gaps & compressions. (Any good installer will rake the top level & even to limit thin spots.) Cellulose will settle quite a lot if blown at too low a density, so it's generally blown at a density and depth that settles asymptotically over a decade or two to the labeled R value for it's labeled coverage (but performs better than label when first installed due to the deeper thickness.)
Low density blown fiberglass has a serious infra-red translucency issue too, cutting measurably into cooling season performance. (Which is good for radiant-barrier sales, I s'pose... :-) )
The performance of batts in walls increases with falling temps when they are as dense as R13s, but will fall when fluffed out to hit R19 or R23 in an ASTM C 518 test fixture. An R19 batt hasexactly the same amount of fiberglass per square foot as an R13, but 57% more air space, making it FAR less air-retardent (and R23s are only marginally denser than R19 when installed in 2x6 cavities), which leads to high internal-convection losses. But with 9' of cavity height to impede the loop, it isn't as severe a performance hit as in a cold-side up attic with only a foot or less of depth (especially when it has no top-side air barrier, which is nearly always in the real world.)
Low density fiberglass does OK in a warm-side-up configuration with a top side air barrier, such as between floor joists over an unvented crawl space in a cold climate (or under a heated radiant floor), not so much in most other configurations.
Still begging to differ with Martin, R-value is not a legal definition, but an engineering/scientific definition. The allowable labeled R-value under US law is defined to be it's performance under ASTM C 518 test conditions, but that's it's label, not it's as-used (and very real) R-value. Other countries, engineers, scientists etc. are not bound by that when discussing the actual R value performance as defined by the physics under a particular set of conditions, but for labeling purposes most countries have similar labeling restrictions. So in answer to Ron, in attic applications most fiberglass without topside air barriers the as-installed R-value defined in units of ft²·°F·h/Btu is in fact generally lower than the legally labeled R value, and in attics it's often a LOT lower especially at temperatures & conditions outside the ASTM C 518 test range.
Martin,
From your post #15:
[My question:]
Q. "When you say: 'If you install a layer of R-38 cellulose on your attic floor, it will perform better at 0 degrees F than a layer of R-38 fiberglass,' would this be the case if you did enclose both types of insulation with air barrier on six sides, thus preventing convection loops from inside the insulation to the space outside of it?"
[Your answer]
A. Yes. The internal convection loop speeds heat loss. The moving air carries heat from one side of the insulation thickness to the other.
So, does the R-38 rating of the fiberglass not include the effect for internal convection? I thought it is tested with plates enclosing a sample in a way that produces the same effect as an air barrier on six sides, thus eliminating external convection; and the internal convection is simply a part of the rated performance.
John,
Thanks for the link to the Lsitburek article. As Joe wrote, "You could get some convection in ceiling insulation or floor insulation if there is a huge temperature difference and the insulation is of real low density." That's the situation we are discussing here: really fluffy insulation under very cold conditions (for example, in the depth of a Minnesota winter).
The denser the insulation, and the lower the delta-T, the less of an issue this becomes.
Ron,
Q. "So, does the R-38 rating of the fiberglass not include the effect for internal convection?"
A. No, the R-38 rating includes the effect for internal convection -- but not at very cold temperatures. The ASTM tests are conducted at a mean temperature of 75 degrees F, and the cold plate used in testing isn't as cold as a Minnesota attic. So in the real world, in a very cold climate, the fiberglass insulation is exposed to conditions that differ from those used to measure R-value in a lab.
I have received further information on ASTM C518 testing from Dave Yarbrough. Dave writes,
"The FTC Rule requires that labeling and advertising use the resistance value at 75 F. ...
"C 518 can be used to deteminine the 'apparent thermal conductivity' from which thermal resistivity and thermal resistance are derived over a range of tempertures from about 20 to 120 F. The temperature of the cold and hot plates can be changed by the operator of the apparatus. A temperature difference is required to make a measurement. The minimum temperature difference is specified in C 518. The reason for the minimum difference is that the expermental uncertainty increases (percentage) as the temperature difference decreases. The result of a test is reported at the average of the hot and cold plate temperatures.
"An average temperature of 75 F can be achieved by the hot plate at 100 F and the cold plate at 50 F, a commonly used set of conditions. An average of 75 F can also be achieved by hot plate at 90 F and cold plate at 60 F. There are an infinite number of combinations that average to 75 F.
"The result doesn't change as delta T changes (within the repeatability of the instrument) unless convection is present. The heat flow direction is up. There is also a requirement that the plates be high thermal emittance (black)."
I contacted Johns Manville yesterday about the issues being discussed here. The person I spoke to did explain that the test is done with a mean temperature of 75 degrees F, meaning that that is the temperature midway through the insulation test sample thickness. The high and low temperature can vary inversely as long as the 75 degree mean temperature is fixed. My inquiry was not intended to pin down an answer to ever technical question, but we did discuss the issue of fiberglass underperforming at low outdoor temperatures.
He told me that this was proven to be occurring with the lightest density, loose fill fiberglass, in attics, without any topside air barrier, around 1994. He told me that the problem has since been alleviated, and that the problem never applied to batts of any density. On this specific issue, we did not discuss whether the underperformance was due to internal convection loops; or convection moving from inside the insulation to outside of it.
However, he did tell me that they presently have zero internal convection within the real world temperature range in any of their fiberglass product today, including all batts of all densities.
So my conclusion is that with batts, external convection (convection moving from inside the insulation to outside of it) is possible, and that it can be totally prevented by the use of an air barrier on six sides. And, there will be no internal convection. Therefore, the insulation will deliver its rated value at all real-world temperatures.
Furthermore, the R-value increases beyond the batt rating as the cold temperature decreases.
Ron,
I'm not quite sure why you are so reluctant to accept the fact that it makes more sense to insulate an attic floor with cellulose than with fiberglass batts. However, it seems like you have done your own research and reached your own conclusions. If fiberglass batts are what you want, by all means, use them.
Martin,
I have explained the basis for my preference for using fiberglass in earlier posts. It is because the product has manufacturing level quality control, and I will install it with enough quality to achieve 100% of its performance potential. I would apply this same rationale to the foam board products. However, I would not rule out blown cellulose because it is not far removed from my basic criteria. It is also interchangeable with fiberglass for my basic superinsulated design. Just to clarify, my criteria are centered on the issue of installation quality, and not on the performance of the insulation product per se.
What I don’t understand is your position on this matter. You say you are not sure why I am so reluctant to accept the fact that it makes more sense to insulate an attic floor with cellulose than with fiberglass batts. On what basis do you conclude that it is a fact that it makes more sense insulate an attic floor with cellulose than with fiberglass?
Ron,
Q. " On what basis do you conclude that it is a fact that it makes more sense insulate an attic floor with cellulose than with fiberglass?"
A. Here's my basis: I have spent hours crawling around attics, balancing on joists and the bottom chords of roof trusses.
Fiberglass batts in attics are never installed to fill the joist bays. There are always light fixtures in the way, and wiring, and bath fans. There are narrow joist bays, and wide ones. If the roof is framed with trusses, the 2x4 bottom chords present a huge impediment to deep insulation with fiberglass batts.
All kinds of gaps occur when you try to insulate an attic floor with fiberglass batts.
With cellulose, the hose blows -- the insulation fills the crannies up -- and the insulation just keeps getting deeper and deeper, until the installer shouts -- "Looks good! Shut her off!"
Martin,
When you asserted that it is a fact that it makes more sense to insulate an attic floor with cellulose than with fiberglass batts, you did so without stipulating any conditions on that declaration. You seemed to apply it as a blanket statement that applies to everyone’s use of fiberglass. You applied it to my use of fiberglass.
When I asked you for the basis of your blanket statement, you listed several problems that can be encountered when using fiberglass. However, I will avoid those problems, so they don’t apply to my use of fiberglass. Therefore, your blanket statement that it is a fact that it makes more sense to insulate an attic floor with cellulose than with fiberglass batts does not apply to my use of batts. I can’t be the only one, so I am sure that your blanket statement does not apply to others as well.
Therefore, I must conclude that your blanket statement is not a fact, as you say it is. It is a fact that the circumstances of fiberglass problems that you cite do exist in some applications, maybe the majority of applications, but they don’t apply universally to the use of fiberglass. Certainly they are not fundamentally inherent to the use of fiberglass, or intrinsic to the fiberglass product itself, as many people repeatedly imply.
Ok we get it you believe you have super human fiberglass batt insulation installing abilities. That's great. Unfortunately very few people in the industry have that ability...Poor or underperforming FB Batt installation is one of the best documented and most cited issues in building performance. It looks like you may be a little naive (sorry it does) and are fooling yourself that it is possible for you to install it 100% correct. Martin isn't saying that it won't work just that cellulose is a lot easier and more redundant. I have installed both and I don't know why you wouldn't install cellulose. It is more resistant to air movement, easier to do a perfect job, covers the truss chords (batts definitely do not do this), allows you to get deeper insulation really easily, etc. My favorite part of this is that you have been arguing with Martin about this and will not believe what he is saying, but then when you consulted a fiberglass manufacturer you believed what he said 100%....Just sayin it looks fishy to me. If you want to use it use it...I'm not sure why you were even discussing it on here because you obviously have an unyielding preference of FB.
Spencer,
I am not trying to talk anybody out of using cellulose. But if people try to talk me out of using fiberglass, I would like to hear their explanation. So far, I have not heard a good reason why I should not use fiberglass. There are two categories of objections to fiberglass. One is the quality control of its installation, and the other is its performance. I do not anticipate any problem with the quality control of the installation, but I would like to get to the bottom of the performance issues.
Regarding the performance, I have been told the following here:
1) The R-value of fiberglass insulation is less than the manufacturer claims.
2) The R-value of fiberglass insulation matches what the manufacturer claims, but claim is based on a test that does not match real world conditions. Therefore, the insulation does not deliver the manufacturer’s claimed R-value in real world conditions that fall outside of the tested standard.
3) The R-value of fiberglass insulation is unknowable unless you test it yourself.
4) The R-value of fiberglass is at least what the manufacturer claims under all real world conditions, and the R-value increases beyond what the manufacturer claims as the outdoor temperature drops.
5) Item #4 is true, except it is overridden by the fact that either or both internal or external convection degrade the R-value as the outdoor temperature drops.
I am not just blindly accepting what the fiberglass manufacturer claims. I have asked plenty of questions here and questioned the manufacturers several times as well. And if I can ever get a specific objection to fiberglass defined, I will immediately take it to the fiberglass industry to get their specific response. At least, at that point, we have the issue framed in a way that science can evaluate it.
However, items 1-5 above are summarized from discussions that are full of relative terms and general statements that need many more qualifiers to make their meaning specific. When I consider the list of items 1-5, my main concern is with the part of item #5 claiming that internal convection degrades the rated R-value of fiberglass. That is a fairly specific charge that stands worthy of a direct answer. The only qualification that might be needed would be the amount of alleged R-value degradation due to convection at various temperatures.
When I asked JM about this yesterday, I was told that there is ZERO internal convection with any of their fiberglass insulation products including all batts of all densities. And just to make sure of our terms, I told them that by “internal convection,” I meant convection loops circulating entirely within the insulation mass. They confirmed that understanding.
My basis for item #5 is embodied in posts #19 and 21. In post #19, notice that my question excludes the effects of EXTERNAL convection just to be clear that we are talking only about INTERNAL convection.
Martin says that the R-value is degraded at a lower temperature due to internal convection.
Then I asked him how the R-value delivered can be less than what is claimed by the manufacturer. In post #21, Martin said this:
“The ASTM tests are conducted at a mean temperature of 75 degrees F, and the cold plate used in testing isn't as cold as a Minnesota attic. So in the real world, in a very cold climate, the fiberglass insulation is exposed to conditions that differ from those used to measure R-value in a lab.”
I interpret this to mean that the manufacturer’s claimed R-value was based on a test that did not include the lowest temperatures such as would be found in Minnesota winters. And therefore, at those lowest temperatures the label R-value would not be delivered in the real world due to the degradation caused by internal convection.
JM says that the 75 degree base line is the mean temperature midway through the insulation test sample, but the cold side is lower. I do not know what the test requires as to the low temperature. However, JM said that they test to temperatures much lower than the lowest real world temperature, and the R-value actually INCREASES as the temperature drops. So, at the coldest real world temperature, the insulation delivers a higher R-value than its label rating. I asked them why the internal convection would not be degrading the R-value, as opposed to the observed increasing R-value. They told me that there is zero internal convection.
So, here we have a clear difference between what Martin says and what JM says. I have listened to Martin’s information, and then taken it to JM for their response. Now I am bringing their response back here for further comment. So, while there is a difference of explanation on this matter, at least we have the issue sharply defined. JM says there is zero internal convection in their fiberglass. Can someone here disprove that? It ought to be scientifically verifiable. Can somebody list the specific JM fiberglass insulation products that fail to deliver their rated R-value at any real world temperatures, and if so, can you show the amount of R-value degradation at specific temperatures? If anyone can offer this specific information, I will take it to JM and ask for their response.
Has this horse been kicked to death yet? :-)
Ron is nothing if not persistent. If he spends as much time detailing his batts as he spends defending his use of them he might just get the performance out of them yet! (Larsen managed to use crummy R19s to good effect on his original retrofit wall trusses after all.) But that detaling becomes tedious around real-world interior framing with plumbing & electrical and uneven stud bay issues to work around. Batts always work great in the abstract, or on clear-wall assemblies in a lab (or in Larsen Trusses on the exterior, I s'pose). But the benefits of blown or sprayed goods over batts is pretty obvious to the real-world practitioners.
FWIW: I'm currently working up the drawing for using (the dreaded) batts to insulate the exterior of an antique masonry chimney, since I can't use rigid foam- the back of the firebox and flue liners are less than 12" from the surface I plan to insulate (per R1001.11, exception 2, IRC 2009) and I don't want to add the necessary air gap to combustibles. it looks like it''ll accommodate 3.5" R15 semi-rigid rock wool batts butted against one another (no studs), strapped to the masonry with furring through-screwed to the masonry 24" o.c. on which to hang siding. The furring also establishes the rainscreen gap. (There's no rigid sheathing, but a housewrap will serve as both an exterior air barrier and WRB.) If small quantities were readily available I'd be tempted to use even higher density panelized rigid rock wool, but for only about 80-100 square feet it's not worth the aggravation. At the moment the 6' wide chimney represents an 80-100 square foot ~R2 hole in an otherwise reasonably insulated (for a ~1915-ish full-dimension stud timber framed antique) house. The only penetration of the insulation will be the through screws, and the sealed-combustion-air vent for the wood burning stove.
As-is the stove is set well-into the oversized fireplace (it can't be moved further out into the room) and heats up the masonry just fine, but a good fraction of that heat ends up in the great outdoors rather than in the house it's heating. But with the exterior insulated the whole 6' stripe of wall will become a massive low-temp radiator, limiting the heat loss to the exterior pretty much to the ~12 square feet of cross section that penetrates the roof/overhang.
OK too much thread drift- I'll stop now. :-)
Dana that was most of my point...this conversation went from great to silly/ridiculous/beatingdeadhorse.
Martin, I am confused about what "WE" are discussing here.....
I get the impression that Ron is talking about fiberglass batts enclosed on 6 sides with an air barrier...
similar to what Chris Schumacher is reporting in the attached snipit.
And You are talking about "fluffy" loose fill fiberglass.
If the Vertical walls {tall stack effect} in Schumacher's report are showing good performance with Fiberglass Batts(plus 6 sided air barrier)......
Why would a Horizontal ceiling application {short stack effect} with Fiberglass Batts (plus 6 sided air barrier) perform poorly?
John,
This has been a long thread. I don't think I could say that this thread has been limited to a single topic.
Ron Keagle asked, "Let me ask this one broad question regarding the issue of fiberglass losing thermal performance due to convection: Does this happen only when the fiberglass is not enclosed on six sides with an air barrier?"
I interpreted Ron's question to be a reference to attic insulation. As far as I can remember, he was the one who introduced the example of insulation that is not enclosed on all six sides by an air barrier.
In a later question, Ron asked, "I understand your information that an air barrier on six sides is not stipulated as a requirement by the manufacturer or the regulations. You also say it is not advocated by any 'best practices' guide. But doesn’t GBA advocate it?"
I interpreted Ron's question to be referring to insulation on an attic floor.
You asked, "Why would a horizontal ceiling application (short stack effect) with fiberglass batts (plus 6 sided air barrier) perform poorly?"
My answer: I don't think it would. I think that such an installation would achieve performance that was close to the R-value on the package -- assuming, of course, that the batts were installed according to the manufacturer's recommendations.
Martin, I didn't think this thread was about a single topic....
My reference to "WE" was from your post #20
John,
I'm a little unsure of your point.
The "we" in that post referred to Ron and I. Ron was asking a question that arose when I discussed the performance of fluffy insulation at 0 degrees F. Ron asked a question about installing "a layer of R-38 cellulose on your attic floor" and whether it would "perform better at 0 degrees F than a layer of R-38 fiberglass" -- so, at least for that portion of this very long thread, that did seem to me to be the topic that we were discussing.
I meant no disrespect to other people, including you, who may have wanted to discuss other topics.
Martin and John,
In post #15, Martin answered one of my questions thus:
Q. "When you say: 'If you install a layer of R-38 cellulose on your attic floor, it will perform better at 0 degrees F than a layer of R-38 fiberglass,' would this be the case if you did enclose both types of insulation with air barrier on six sides, thus preventing convection loops from inside the insulation to the space outside of it?"
A. Yes. The internal convection loop speeds heat loss. The moving air carries heat from one side of the insulation thickness to the other.
Answered by Martin Holladay, GBA Advisor
Posted Tue, 10/23/2012 - 13:08
Notice that I am stipulating air barrier on six sides, which has always been my intent. Martin’s answer to my question is that the fiberglass will still underperform its rated R-value of 38 due to losses from INTERNAL convection. As we all understand, internal convection cannot be prevented by an air barrier on six sides.
However, Johns Manville assures me that none of their fiberglass insulation products (including blown fiberglass and batts of all densities) suffer any degradation due to internal convection at all real world outdoor temperatures. They tell me that their tests show all of their fiberglass batts actually increase R-value performance above their labeled R-value all the way down to their minimum tested real world temperature.
Ron,
I don't doubt that changes in fiber length and other fiber characteristics by fiberglass manufacturers in recent years have reduced the convection-at-cold-temperatures problem considerably. They are striving to increase the installed density of blown-in fiberglass. You may remember that I wrote, "The denser the insulation, and the lower the delta-T, the less of an issue this becomes."
You may also remember that (way back in Post #7) I provided a link the the Johns Manville document that states the argument that was told to you over the phone. That document -- the one I linked to days ago -- states, "The JM tests demonstrated that with proper glass fiber diameter and nodule size design, the air permeability of the installed product could be greatly minimized. The findings from both the ORNL and JM studies were used in subsequent years to establish design specifications for all of Johns Manville’s loose-fill fiber glass attic insulations to improve winter thermal performance. The design specifications primarily focused on maintaining an appropriate nodule or tuft size, which decreased the permeability of the installed insulation."
In other words, there was a problem. Manufacturers changed their manufacturing methods to address the problem.
However, I find it hard to believe that there is NO convection at 0 degrees F or -20 degrees F, even with the newer products. I think there is reason to believe that the modern insulations are better than those of a few decades ago.
That said, I would still choose cellulose over blown-in fiberglass, for two reasons: it is even better at resisting convection than fiberglass, and it is a recycled product with lower embodied energy.
So, let me see if I have it now
Properly installed Fiberglass Batts with an air barrier on 6 sides will likely perform as rated....
no matter if it is a ceiling or wall
Loose fill "fluffy" Fiberglass will likely perform poorly (well below rating) at very cold temperatures even if with an air barrier on all 6 sides?
From a green-liness point of view Martin has it dead-right. No virgin-stock pulp/paper is used in the manufacturing of cellulose insulation- it's all reclaimed/recycled goods (though very little of it is post-consumer, from most vendors- it's mostly printing over-runs and misprint newspaper.) About 15% of it by weight is the fire retardents though, none of which is post industrial reclaimed stock, SFAIK. Call it 85% recycled feedstock with very low processing energy for converting the feedstock into finished product.
Second among fibers would be the cotton goods made from scrap denim, but that's a very small player in the marketplace (and still unfavorably expensive relative to cellulose or rock wool). It has the same fire retardent issues as cellulose.
Third would be rock wool, which in most cases is more than 70% slag from iron & steel making (but it varies among vendor and specific product), the rest virgin basalt. Very few rock wool products have less than 50% slag, and Thermafiber boasts 90% for their Special "Green" fiber:
http://www.thermafiber.com/Portals/0/pdf/EPA%20Recycle%20Content%20Wabash.pdf
Most fiberglass insulation is (of necessity, to get the spun fiber quality) mostly virgin stock, with less than 35% recycled or reclaimed feedstock in any current product that I'm aware of.
Both rock wool and fiberglass use quite a bit of process energy though. It's damned-near impossible to beat cellulose on embodied energy, but we're talking something like two orders of magnitude more processing energy to spin glass or rock wool fiber than it takes to shred & mix newspaper into product. Cellulose is cheap stuff, with better inherent air-retardency than the rock wool or fiberglass.
As spun fibers go, rock wool has a performance edge over fiberglass in open blow attics since unlike fiberglass it's completely opaque to infra-red whereas fiberglass is somewhat translucent in the deep infra-red. Fiberglass manufacturers have added materials to reduce that translucency but it's not possible to eliminate it. Most rock wool batts are higher density and more air-retardent than their fiberglass kin too. It's a lot itchier for the installer, but it's better stuff in general. In most of the world it holds a much larger market share than it does in the US. (How fiberglass came to dominate the US market has both a mystery and a history.)
Martin,
This has been a long thread, and I am not trying to put anybody on trial for what they said. It is just that the discussion gets very confusing with the shifting of terms as different people add information. You were indeed referring to the fluffy fiberglass in your example of internal convection degrading its performance in a labeled R-38 layer. I did not know that you were using the term “fluffy” to refer to blown fiberglass as opposed to batts.
In any case, JM told me that there was a problem with convection causing degradation of the blown product that was found in the early 1990s, and that it has since been corrected. Apparently, this problem of convection was internal convection.
Regarding JM asserting that their products have no internal convection today, that raises one fine point question about that statement. It would be possible to have some internal convection, but not enough to degrade the labeled R-value. If that were the case, it would not be true to say there is no internal convection. But if there were internal convection, but not enough to degrade the labeled R-value, the point would be moot.
But from what I have learned with this discussion and from JM, the blown product and batts of all types will deliver their labeled R-value, or higher, at all real world temperatures, if they are enclosed with air barrier on six sides.
John,
Regarding you post #38, and your question as to whether loose fill "fluffy" Fiberglass will likely perform poorly (well below rating) at very cold temperatures even if with an air barrier on all 6 sides.
According to JM, the fluffy product will deliver the labeled R-value at all real world temperatures if it is enclosed with air barrier on six sides.
I think it might be worth bringing up the idea that six-sided encapsulation of fibrous insulation represents a design ideal.
And not all ideals are practical.
I think this is where the "pretty good" concept displays its value...
For walls, complete encapsulation of insulation is easier than for vented attics.
I don't think that means we should be doing mental contortions to encapsulate attic insulation, but rather take the easier road (as suggested by Martin) and simply plan to add more insulation to compensate for a known problem.
Flexibility and compromise are OK - not every aspect needs to be "perfect".
What is important is to have the knowledge to know where or how you can compromise without adding unnecessary or unacceptable risks to the equation.
John,
You wrote, "Properly installed fiberglass batts with an air barrier on 6 sides will likely perform as rated....no matter if it is a ceiling or wall. Loose fill "fluffy" Fiberglass will likely perform poorly (well below rating) at very cold temperatures even if with an air barrier on all 6 sides?"
I think that is a fair summary. The famous cold-temperature convection problems noted by ORNL in the 1990s were problems with fluffy blown-in fiberglass on attic floors.
Concerning batts installed within 6-sided cavities -- obviously, to get the rated R-value, you need to have perfect installation. (Just call up Ron for that job!)
If an six sided insulation bay leaks air it is NOT A SEALED SIX SIDED INSTALLATION.
Missing from this thread is the fact that air infiltration is huge in standard US 2x construction. Huge. As Martin and any of us can report, we all have found huge movement of air through assemblies and THROUGH fiberglass batts. And as mentioned, no one can install batts in a trussed roof perfectly. I may have done the best job anyone has ever done decades ago by using 2x4 width batts between trusses followed by a perpendicular layer above of R-38 I think. No fun. Perfect? We tried but no way. Ron is the only person that does perfect batt installs. The rest of us are humanoids.
And most except Ron would agree that to add a "sixth side to attic installs the item to use is... drum roll please.... cellulose. Period. Try running Tyvek around a zillion truss struts and perfectly attaching and sealing it to all the edges involved. Ron can do it but no humanoids ever could IMO.
https://www.greenbuildingadvisor.com/blogs/dept/musings/one-air-barrier-or-two
Ron, seriously? And this is the assembly you are going to promote? Put me down as not buying. Good luck with that.
I am not for pieces of rigid foam use. Better to spray foam if foam is desired. Spray foam glues into place. Rigid foam does not. Rigid leaks at edges. Air leaks ruin the insulation ability of insulation. Frames move seasonally.
Ron, you are on your own doing a one off install. Good for you. Not so for most. My warning to others is to not follow your lead. Too complicated, too many failure points, to expensive labor wise. And of course cellulose has less impact on the planet. This is a Green advice site after all.
Content removed.
Okay AJ, I took down the plan for the benefit of others. I would not want to go against the grain.
Ron,
It's not such a bad thing to go against the grain.
AJ,
You said:
"Ron, you are on your own doing a one off install. Good for you. Not so for most. My warning to others is to not follow your lead. Too complicated, too many failure points, to expensive labor wise."
Too complicated? Too expensive? I never said this was for everybody. This high performance, superinsulated construction is hard. If you want cheap and simple, that’s a different road.
I believe the worst thing you can do is go for the high performance option, but make a halfhearted effort because you want to keep it cheap and simple.
Ron, your plan is NOT high performance. It is armchair interesting and highly complicated. It mainly is easily outperformed by simple dense packed "way more green" cellulose. It is easily outperformed by simple but expensive spray foam. But your idea is a no go for an insulation contractor's crew so it is a one of a kind idea. And really and finally you never will get an air barrier seal let alone two and not even close, a six sided seal.
You will learn much, that I will be give you.
Cutting corners is not the opposite of complicated. Simple is. Hey, I am all for experimenting. Press on Ron.
I have already been through this once in practice AJ, and things are working fine. I also leared that there is no end of people who will tell you things won't work. Usually, it is the things they have never done themselves.
Sometimes the armchair is a good place to spend some time thinking about the best way to build something before tearing into it with hammers and saws.