Replies

1. GBA Editor
   Martin Holladay | Oct 29, 2013 01:23pm | #1

   Dirk,
   There is no single answer to your question. Fiberglass batts are loose and fluffy, and air moves through the batts easily. Blown-in fiberglass is somewhat denser, and air moves through it more slowly. Dense-packed cellulose allows even less air through it than the other two types of insulation mentioned -- but dense-packed cellulose is still not an air barrier.

   When fiberglass batts are installed in a vented cathedral ceiling, it is absolutely essential (in my opinion) to have a tight air barrier on the top side of the insulation. With the other two types of insulation, the air barrier is less crucial but still helpful.

   Here's my advice: since you have to install baffles anyway (to create the ventilation channel), it makes sense to pay attention to airtightness when installing the baffles. All it takes is a little caulk or tape -- which isn't too much trouble.

2. Expert Member
   Dana Dorsett | Oct 29, 2013 01:56pm | #2

   At dense-pack densities there isn't a meaningful thermal performance difference without a top-side air barrier, as long as the interior side of the assembly is air-tight. (One-side air tightness is sufficient for delivering the R-value with the amount of air retardency you get with dense pack.)

   But I'm not sure how you can even GET dense-pack densities without risk of compromising the depth of vent channel side without using a fairly rigid material for the baffle though, and it's not a lot of labor to air seal it, as Martin recommends. The more layers in an assembly you can make air tight, the more reliable and resilient the long-term performance becomes.

   The baffle/exterior air barrier needs to be vapor permeable (to water vapor) though, or moisture can & will accumulate in the air-permeable insulation under some conditions. The whole point of the vent space is to allow both the roof deck and the insulation to dry.

   Batts are prone to thermal bypass leakage at the edges. Low/mid-density batts can suffer loss of performance to confection and through-the-batt infiltration at the temperature & wind speed extremes. But if perfectly fitted, rock wool or high-density "cathedral ceiling" fiberglass are sufficiently air retardent to deliver the performance. The key is "perfectly fitted", which is an extremely high standard- in the real world most assemblies will have sufficient fit defects they will benefit from a top side air barrier to mitigate the performance hit from the leaky spots.

3. 
   dirkgently | Oct 29, 2013 03:59pm | #3

   Thank You both.
   Dana, so the Accuvent catherdral baffles probably should not be used because they are made out of Vinyl which I assume is NOT vapor permeable?????
   There will be 2" EPS or reclaimed Polyiso on underside of rafters. (if that makes a difference for drying to interior)
   I have a insulation contractor who wants to use accuvent and dense pack fiberglass (1.8+ lbs I assume) into the rest of the 2x8/2x10 rafter bay. He could see no reason for me to go thru the trouble of building custom EPS baffles (neither can I if R value is the same AND I can live with a 1" air gap instead of 1.5 or 2")
   I am liking the higher R-value for dense pack Spider or Optima.
   I NEVER thought I would say "like" and "fiberglass" in the same sentence ever again.

4. Expert Member
   Dana Dorsett | Oct 29, 2013 05:48pm | #4

   The higher R value is undercut severely by the thermal bridging of the rafters. You get a ~9% improvement in R-value over cellulose at center cavity, but with ~12-15% of the area being rafter or ridge, the effect on the whole-assembly R value is pretty tiny:

   With a 15% framing fraction and 9.25" of insulation in a 2 x 12 rafter bay (with 2" of vent channel) allowing ~R1 for the combined effects of gypsum & roof deck you're talking ~ R27 whole-assembly for dense-packed cellulose, ~R28.5 for dense-packed Optima or Spider about a 4% improvement in whole-assembly R, even though the center-cavity R of the fiber is R34 and R37, respectively (at 9% uptick at center-cavity R.)

   And yes, putting vinyl on the outside of the insulation could be a problem anywhere north of the Mason Dixon, unless it's has a pattern of micro-holes to give it a higher vapor permeance (at the cost of some air tightness).

5. 
   dirkgently | Oct 29, 2013 07:01pm | #5

   I am not sure I follow the penalties of the thermal bridging in the above example. I could be completely misunderstanding but, How can thermal bridging from rafters penalize Spider or Optima more than cellulose.
   another way I can phrase the question is. Does the thermal bridging affect Spider and cellulose equally? I would assume that since Spider has higher R value (properly installed) then it would have a higher R value after thermal bridging penalities.
   BTW someone should tell Accuvent to put vapor holes in those things then....or state on package that joints should not be sealed and caulked to allow for breathing.
   Once again Dana Thank You...Black Diamond on me.

6. Expert Member
   Dana Dorsett | Oct 30, 2013 11:13am | #6

   The R-value of the thermally bridging rafters doesn't change as the center cavity R value changes. Thus the percentage of the total heat transfer the thermal bridging is higher with higher-R insulation than it is with lower R insulation. To understand it better, let's take it to reductio-ad-absurdum:

   Say the center-cavity value was R1,000,000, through the 9.25" of fiber, with some magic material made of pure unobtainium. The heat transfer through the 15% cross section that is R1.2/inch timber would still be the same, so the vast majority (call it all) the heat transfer is through that 9.25 x R1.2/inch = R11.1 wood. But since it's only 15% of the area, averaging that heat transfer over the whole area the average R value is (R11.1 / 15% = ) R74. If we guesstimate the R of the gypsum & sheathing is about R1, that means the whole-assembly R is about R75.

   That means the improvement from going from R34 cellulose to R-one million doesn't even quite triple the whole-assembly performance (less than 3x) , despite improving the center cavity R by (R1,000,000/R34=) 29,400 x.

   To paraphrase a Clinton era campaign focus slogan: "It's the thermal bridging stupid!".

   That is why in wood-framed buildings using high R/inch foam, saving the foam budget for the EXTERIOR is a far better use of the material & money than putting it between studs/joists/rafters. Thermally breaking the much lower R framing fraction is more important than bumping up the higher-R center cavity value. Most 16" o.c. stud walls run about a 25% framing fraction, which means even in a code-min R20 2x6 wall with 25% of the area being R6.6 wood, nearly half the heat transfer is through the wood, and the whole-wall R average runs ~R14 (with gypsum/sheathing/siding R added.) If you put ~ 5" or ~R30+ closed cell in foam in the cavities (at $5/-per square foot) the wood losses are well over half the total, and putting even 2" of exterior polyiso over the sheathing and keeping the R20 cheap stuff in the cavities would be much more dramatic, bumping the whole-assembly R to ~R26 rather than merely R17 for the closed cell cavity fill case.

7. 
   dirkgently | Oct 30, 2013 12:18pm | #7

   Dana You crack me up sometimes. Adding humor into this really helps. I am just a simple carpenter trying to figure all this out.....and despite years of reading info...I am still in the dark.
   I was misunderstanding you originally. That is a great example and I hope it will benefit others who stumble on this thread.

   A real world example here is: I used 1" polyiso/xps attached to underside of 2nd floor rafters and all exterior walls back in 2005 on my 1947 era. small cape remodel. I did not know what I was doing at the time was "air sealing" but I caulked, taped and sealed everything....other trades people (friends) thought I was nuts....WELL that was best money and labor spent because It is cheap to heat and cool and very comfortable.
   Back then it was very hard to find information on these subjects...This is such a great site.