Replies

1. 
   DanH | Mar 22, 2010 08:17pm | #1

   I don't see why you'd need a vapor barrier for a roof in Dallas.

1. 
   kleinchen | Mar 22, 2010 10:24pm | #2

   I don't follow.

   Why wouldn't a vapor barrier be needed? The differential in temp between inside and outside in either the summer or winter can easily be 40+ degrees. Given the difference in relative dewpoints of the inside and outside temperatures, wouldn't this create the possibility for moisture to condense out of hot air reaching a colder air temperature by either penetrating inward or trying to escape outward? The extremes may not be a pronounced as in northern climates in the winter where the differential might reach 70 or more degrees, but they exist nonetheless, no?

   Best,

   Andrew...

1. 
   DanH | Mar 22, 2010 10:49pm | #3

   Right, 40 degrees generally difference ISN'T enough to worry about condensation.

   The real issue is dewpoints.  Is the dewpoint inside above the outside temperature in the winter?  Only then could you have condensation, but only after you to take into account the fact that maybe a third of that temperature differential will be across the shingles and sheathing.

2. 
   Martin_Holladay | Mar 23, 2010 08:45am | #4

   Best of two options

   Andrew,

   As the author of the thermal bridging article in question, I supposed I'm prejudiced. But since Strategy 2 costs much less and provides better R-value as well as a thermal break, it's clear that Strategy 2 is the way do go.

   Vapor diffusion is vastly overrated as an issue to worry about. You won't have any vapor diffusion downward through roll roofing, and either of your two strategies will provide a vapor retarder to prevent any problems from interior moisture duiffusing upward during the winter.

1. 
   kleinchen | Mar 23, 2010 05:58pm | #5

   Thanks for the insights ... and the article.

   #2 does seem the better option but I was unsure given the preference I had read for closed cell foam. If 40 degrees isn't large enough of a temperature differential then that makes my job easier. I hadn't read anything that specified temperatures in any detail, which is why I figured I'd ask. Many FHB articles appear to be written from the experience of a northeastern or midwestern winter, so I'm often not entirely sure if the measures stated really apply to the weather here in Texas. 

   Just curious:

   I'm still a little baffled by the accuracy of the super-insulated specs I've read. In the recent passive-house articlethey seem to work around the diminishing returns of spray foam by mixing and match insulation materials, like spray foam and cellulose. Is such mixing and matching the only way to get R60 performance? What does that say for roof SIPs which can be made pretty thick. Is there a really a meaningful advantage to a 10" thick SIP over a 6" one? 

   Thanks again.

   Andrew...

1. 
   DanH | Mar 23, 2010 06:53pm | #6

   There's no "diminishing

   There's no "diminishing return" for most insulations -- 10" gives twice the R value of 5" in virtually all cases.  But foam is desired in part because it seals against infiltration so well, and is relatively immune to problems with condensation.  However, it's expensive (on a per-R basis) so cheaper stuff is often combined with it once the other benefits of foam have been realized.

1. 
   kleinchen | Mar 23, 2010 07:47pm | #7

   Sorry to belabor the issue but I'm trying to wrap my head around this diminishing return concept.

   In a 2009 article in FHB on spray foam there is a graph charting the diminishing return of spray foam. Basically it says that after about 3-4" of closed cell and 5-6" of open cell efficiency levels off:

   “open-cell foam reaches a point of diminishing returns at around 5 in. That threshold is even lower for closed-cell foam, which experiences diminishing returns at around 3 in. or 4 in. Those thicknesses create assemblies between R-20 and R-24, which by the numbers seem a little weak. Each additional inch of spray foam yields little performance. In fact, while the cost of an R-40 wall is indeed double that of an R-20 wall (not factoring in the construction materials used to create deeper cavities for the extra foam), it reduces the conductive heat flow through a wall by only an additional 2%.

   If I understand this correctly, then In terms of actual performance the R40 is only doing a 2% better job of stopping heat transfer. Doesn't it follow then that 10" of foam does not give 2x the thermal resistance (R) of 5".

   Thanks again.

   Andrew...

1. 
   DanH | Mar 23, 2010 09:52pm | #8

   What they're talking about is the diminishing return for ANY insulation.  You get the best bang for your buck (with foam) with the first half-inch or so -- just enough to provide air sealing -- and it goes downhill from there.  And that's even before you consider the heat loss through windows and doors -- at R28 you'd probably be seeing the bulk of your heat loss through windows, in typical circumstances.

   In theoretical terms, if you double the total R value through a wall you cut the heat loss through that wall in half. But if you have a window in the wall that's accounting for half your heat loss through the wall at R20, doubling the insulation to R40 only cuts heat loss by 25%.

   Re their 2% number, I suspect they left out some details -- such as taking into account those losses through windows and doors, taking into account loss through framing, etc.

2. 
   kleinchen | Mar 24, 2010 05:16pm | #9

   That makes sense. Thanks for clarifying the point.

   best,

   Andrew...