I'm building a cabin in the Colorado mountains at 10300 feet, which means 100 pound snowloads.  I'd like to roof a 12 foot span from wall to ridge beam, with no intermediate purlins.  It's a 4/12 pitch, so that means about a 13 foot long panel.  According to deflection tables on various websites, it looks like I could use 12"EPS SIPS with either dimensional or TJI type splines (on four foot centers).

It seems to me an alternative, not as good from an insulation standpoint, but better structurally and a lot cheaper, would be to use TJIs on 2 foot (or possibly 16") centers as rafters, then insulate in between them with EPS foam blocks (no skins), gluing them onto the rafters.  The only way I can think to do this would be to glue the foam block to the previous rafter prior to installing the next rafter, then sliding the (next) rafter up to it and gluing it to the foam.  This technique has several problems that I can think of, and I'm sure you can too.

The other option -- with more thermal bridging but few problems, would be to use dimensioned lumber instead of TJI's, and gluing the blocks of foam in place after the rafters are installed.

Comments? Thanks.

With 100# loads, and a 12' span it looks like a 12.25" SIP would be borderline. But, it would no doubt be less labor than other methods.

You can also make a very nice ceiling with 16" TJI's. Start with 2" of spray foam(R-7.2 to R-14) at the top(to seal), add 1-5.25"(R-21) and 1-8"(R-30) high density fiberglass batt. Right there you've got a R-58(minimum) ceiling. The thin TJI web won't make much of a thermal break, but if it bothers you add a 1" styrofoam board at the bottom of the TJI's. Then you've eliminated most of the thermal bridging, added another R-5, for an R-63 ceiling.

The bad news is that this nice R-63 ceiling, will probably cost more than a 12.25" SIP panel when you figure in labor.

A 100# snow load is hard on any roof. It looks like the SIPs that I used would require a lumber spline for your application. I would recommend using a 4x rather than (2) 2x’s so that the joint does not open due to contraction. OF course, if you have to destroy the thermal envelope with 4x’s every 4’-0” oc, you might as well use your 2x’s at 24” oc with the block EPS if it is less expensive. Jc’s roof also sounds like a solution.

Yes, the 100 lb. snowload is a little ridiculous. This is the equivalent of 22" of water, acccording to my structural engineer. While it may snow a lot in Colorado, it's typically powder (at this elevation), and you'd probably need 25' of it, or more, to reach 100 lbs. In fact, I'll bet the total annual precipitation in this area is only about 30", much of that in the summer.

Anyway, I'm having my engineer design to the maximum deflection allowed (L/240) because I don't think we'll ever see anything close to this, and if we do, we'll still be within code. Hence my thought that 12" SIPs would be OK.

I may take some of JC's ideas to heart and do a conventional roof with some foaming and fiberglass.

What county are you building in? There is at least 1 county where the load is determined by roof pitch. A 12/12 roof has about 25% of the load of a much flatter roof.

Water weighs 62.43 pounds per cubic foot so 100 psf is equal to 100/62.43 x 12” = 19.22 inches. Like all of us structural engineers, your guy probably added a safety factor to get 22”. Anyway, a 100# snow load in the mountains of Colorado is not uncommon and is based not only on snow fall but on snow drift and other factors. The code allows you to reduce snow load based on the pitch of the roof but building officials do not always allow it. The bottom line is that you must design to the requirements of your building official even if it seems too conservative.

JC- I'm building in Summit County. The roof pitches are 4/12 and 9/12.

Avant, thanks for the exact number. You may be right -- with windblown hardpacked drifts, you might get to 100 psf, but I'd sure like to see that tower of snow on a pitched roof :-).

According to R-control's deflection charts:

http://www.r-control.com/downloads/literature/Load_Design_chart.pdf pages 9-10

I should be able to use either a single engineered joist for a spline or if I want to be really conservative, an insulated double spline. However, your comment about contraction opening the joint concerns me -- can you elaborate please? Thanks.

A double lumber spline consists of (2) 2x’s nailed together to create a 3” deep lumber member. This member is then inserted 1-1/2” into one panel and 1-1/2” into the next panel. This lines up the joint between the 2x’s with the joint in the OSB skins. The skins are attached to the 2x splines with nails or screws. As the walls contract and expand throughout the year, stress is created in the wall plane. With 2x lumber splines, the only connection between the panels are nails which run parallel with the wall plane and thus parallel to the stress. The pull out strength of a nail in a 2x is fairly low so the stress will be relieved by pulling the 2x spline apart. Unfortunately, when the panels expand in the summer they do not typically push the nails back in. Over the years the gaps between the panels get larger and create problems, especially in roof applications.

A better solution is to use SIP splines in walls and roofs wherever possible. If a point load on the wall requires a lumber spline, use a 4x or 6x. If roof spans require more strength use the I-spline option that most SIP companies provide. If the roof panels need more strength, start looking at the design to reduce span lengths. With SIP splines, 4x’s, 6x’s, and I-splines, the spline spans the gap between panels as one piece of lumber or OSB. The panles skins are attached with nails or screws to the spline perpendicular to the wall plane. Now the nails are is shear, which has a much larger failure load. The stress in the panels due to contraction and expansion remains as stress in the material and there is little or no movement in the wall panels. Also, if the OSB skins fail in the very close proximity to the nails and do slip, air movement will still need to pass between the skin and the spline, turn and move between the EPS core and the spline, and then pass between the skin and spline again to pass completely through the wall. The space between the splines and the skin is very tight and is sealed with mastic or expansion foam and there should be mastic or expansion foam between the EPS core and spline. Air will need to pass through a minimum of three seals to get through the wall instead of one that is being pulled further apart every year.

Thanks, Avant. Very informative. If I read your post correctly, using the engineered I splines (either single or double/insulated) should alleviate this concern. Now all I have to worry about is the cost (vs. the stick-built version), and those darn overhangs (eaves).

If you use steel SIPS you do not need splines at all and they have greater span/load capabilities, they can be manufactured up to 40' in one pc. - OSB is usually limited to 24ft.
most jobs on our website are steel link

This may have been Summit county. I'd call the building department and see what they say.

Chris,

Can you post a link for span or load tables the steel sips from southernsips?

Dick Mills

I'd be very interested also, Chris. Something similar to

LINK,

pages 9 or 10.

Dick & Cabinboy,

There is no link for the the span charts however, if you want to drop me an e mail to
[email protected] I would be happy to send back the 44 page PDF engineering file along with construction details in PDF format as well
basically a 6" panel will span 10.5 ft @ L/120 and a 98.7 load or a 6" will span 12ft @ L/120 with a 80.7 load

Chuck;

I do not use Structall anymore, they are to hard to deal with and have some lamination problems in the past ...... I use a mfg. in Georgia that has all of the Florida product approvals, I buy thru their local distributor SIP Supply Inc. in Tarpon Springs , Fl. 727-937-3090 link