Replies

1. GBA Editor
   Martin Holladay | Dec 07, 2012 04:00pm | #1

   David,
   A few years ago I exchanged e-mails with Bruce Harley about the Xi walls. Because of the thermal bridging inherent in the design, both Bruce and I doubted the manufacturer's R-value claims. In my notes on the topic, I wrote, "Bruce Harley figures that the Xi system has a whole-wall R-value of about R-8.1 -- not the R-12.1 to R-12.5 claimed by the manufacturer."

   Unfortunately I did not save all of our e-mails, but I saved this one from Bruce Harley: "Yes, I see what you’re saying. I missed that entirely when looking quickly just at the calculations. The calculations on 8:2 and 8:3 both suffer from this problem, and both are completely bogus as shown. Although it wouldn’t be exactly right, a quick approximation would be to treat the stud sides and face all as exposed surface area with similar net R-value, so you’d have something like 14” at R-3.8 and 20” at R-12.5, for net of 34/(14/5.2+20/13.3) ~= R-8.1. A more detailed analysis would be a little worse than this because a good deal of that fin’s side has less than the 0.6 R-value of the 7.5” stud but it’s not going to add up to much. My comments about the mis-interpretation of IRC/IECC are still correct as well. There is no legitimate calculation using the mehod on 8:3, even if they correct the problem noted above. And the calcs on 8:4 and 8:5 are also incorrect. You can’t add the “R-value” of the stud across it’s whole width, to that part of the surface area, because the heat will flow through the sides and not in parallel all the way through the stud. Basically, you can ignore the stud and the whole assembly is the R-5.82, except that they don’t take the top and bottom “plates” into account here, which I believe are thermal bridges across the R-5 foam, as well as the metal pins that link the “sheathing” surface to the studs, so the net of the “R-5” wall if you don’t put anything in the cavity has got to be less than 5, plus the air films. Note that they also leave out the (small, 0.17 I think) outside air film which they can include in the assembly R."

2. Expert Member
   Dana Dorsett | Dec 07, 2012 04:01pm | #2

   Using cellulose below grade there is always a risk of high moisture accumulation in the cellulose that goes undetected until you have a real mess on your hands. If not moisture-tolerant rigid foam (EPS or XPS) rock wool or even fiberglass would be preferred. On the above grade stuff cellulose is good.

   In any sub-grade application the assembly should be allowed to dry toward the interior, so don't put up any interior facers or other vapor retarders between the fiber and the conditioned space. To skip interior vapor retarders that without running into condensation issues for the ABOVE grade sections in zone 4 as you need to keep the ratio the exterior foam-R to cavity fiber-R to 1/5 or higher, so it looks like you'd have plenty of margin.

   With 1" of foam wrapping the studs on all 4 sides the studs themselves would have an R-value of R10, which isn't bad at all. If it's 2.5" on the exterior edge of the stud and 1" facing the cavity it's more than R12. Without a clearer cross section and spacing numbers it's a bit hard to calculate the full extent of the thermal bridging losses, but bottom line, boosting it to R25+ with cheap fiber is still going to be "worth it" on both comfort and long term energy savings in zone 4, even if boosting it that high with foam-only might not be. See table 2, p10:

   http://www.buildingscience.com/documents/reports/rr-1005-building-america-high-r-value-high-performance-residential-buildings-all-climate-zones

   Mind you while the R12.5 continuous insulation meets IRC 2009 minimum for zone 4, it would be marginal for IRC 2012. If you went ahead and fiber-filled the cavities you'd be ahead of the game for future code specifications, and more comfortable to boot!

3. Expert Member
   Dana Dorsett | Dec 07, 2012 04:07pm | #3

   I cross -posted with Martin- didn't see his response. I defer to his more informed analysis of the thermal bridging of the Xi system.

   At ~R8 whole-wall it's sub-code for IRC2009 (which needs to be closer to ~R10 whole-wall), but would probably still make it for IRC 2012 with the added fiber insulation. The high thermal conductivity of steel studs is killer.

4. 
   dsmcn | Dec 07, 2012 05:57pm | #4

   Thanks all for responses.
   Dana, I have looked at the BSC stuff quite a bit, and have even bought Lstiburek's books. My intent is to meet or exceed the recommendations in the table you referred to. This is what led to my posting this question.

   Regarding cellulose below grade: it is an upslope lot so the rear wall will be acting as a retaining wall and will be the only one below grade, but none of the interior foam will be buried. One of the promises of Superior Walls—and a big reason for my choice to use them—is that the concrete will remain dry on the inside. I will design all the walls to dry to the interior, so I can't imagine that cellulose would be a problem here. Please correct me if I am wrong.

   The steel studs are filled with foam. Would they still be thermally conductive to any significant degree?

   Martin, I'm not sure I understand the calcs without more context (I don't know what 8:2 etc. refers to). I think what Bruce Harley is saying is that the studs, having only 1" of foam on sides and face, are acting like thermal bridges across both sides and across the face, therefore the whole 14" of exposed surface has to be calculated at R-3.8—that is, 6" on each side and 2" across the face.

   If this understanding is correct, then it becomes very important to fill the cavity with insulation, to eliminate the 6" surfaces as thermal bridges. Would the result then be a reduction of the R-3.8 surfaces to just the 2" face of the stud, and the beam at the top (the bottom beam adjoins the slab)? And then would the remaining wall face be at R-36?

   Thanks again. By the way, I loved John Straube's seminar—what an incredible service to make it available to all of us!