SIP STRUCTURAL TRICKS THAT RUN RINGS AROUND STICKS
Bill Chaleff, (bcarchtx@optonline.net)
Chaleff & Rogers
There are two structural characteristics of SIPs that are generally not accounted for in creating the usual "stick translations" that predominate in the residential dwelling market. They are the fact that SIPs are exceptionally high performers when it comes to handling inplane loads, and that SIP buildings actually behave as thin shell structures, dispersing point loads throughout the entire surface area. These properties may be exploited more consciously in doing the structural analysis of the project and eliminated members that one wouldn't think of omitting in a conventional stickframe structure.
 The first "trick" is to look at dealing with the resolution of ridge or purlin bearing points usually at a gable end wall. These point loads are usually in the 2,000 lb.  8.000 lb. Range. For an extreme case, consider the ridge beam hitting at the peak of a gable end wall. My SIP load/span charts tell me that for a 6 1/2" thick SIP the perlinearfoot capacity runs from 2000 lbs (12 foot high30psf wind load) to 8,000 lbs (8 foot high30psf wind load). If the design calls for an 8000 lbs reaction set 20 feet off the floor then it looks like we're in trouble, as the chart only allows for 5,898 lbs at that height ( all numbers will presume 30 psf wind loading for the rest of this paper). If the ridge beam is 5" wide, the sill blocking in the bottom of the pocket will be 8" long (1 1/2" each side ) which is 2/3 of a foot and therefore capable of only handling 5989 x 2/3 or about 3932 lbs  way under our requirement of 8,000 lbs.
If we could distribute the load for 2 feet instead of 8 inches we could actually handle 5,898 x 2 or 11,796 lbs, more than adequate with a good safety factor built in. How can we do this? By putting the sloping blocking installed at the top of the wall on each side of the pocket to work, we can safely assume distribution of the load further along the panel and not have to post down to carry the load. With 5/8" lag screws from each sloping top plate into the ridge beam, or a throughbolt going horizontally through from plate to beam to plate, the top plates are put to work and the load distributed at least a foot on each side for low pitches, but perhaps double that for 9 to 12 pitches.
Variations on this theme are to use double top plates  at least for the top four feet  or to cut off the SIP level with the bottom of the beam pocket and run the sill blocking, doubled, out over the top plates. Small triangular SIPs fill in on each side of the ridge beam. For extreme cases it may be possible to combine the last two strategies so that this short "transfer beam"  the sill blocking runs out over the top of doubled sloping top plates. Careful analysis and calculations should be executed in all these cases, based on proper consultations with the axial load chart, so that a conservative safe solution is arrived at. Please! No guessing!
 After all this monkeying around with beams and purlins, one realizes it would be nice to just eliminate them entirely. Here's where we get into "extreme SIP engineering" and we have to be careful, but a little time with some careful analysis can pay off bigtime if the design allows.
SIP connections have been tested both in the lab and over time in thousands of applications over decades and show that a safe assumption is that 90 % of the loads can be assumed to transfer across the joints. Probably that number is higher, but allowing for field errors, 90% is a prudent number to use. This shows us that the planes of assembled SIPS actually behave as huge diaphragms which may be calculated utilizing the formulas for such as published by the A.P.A. What does this do for us? Let's take a standard "home plate" configuration as an example  say a 24' x 36' footprint with an 6/12 gable roof  and first look at the standard solutions. The options are either some kind of ridge beam or purlins to hold up the ridge, or collar ties  minimum of one every 8 feet  to keep the walls from spreading and the ridge from coming down. If we size a ridge beam it would be for the collected load of half the building width, or 12 feet. If we take live loads, dead loads, and wind loads as a total load of 6o psf, the ridge beam should be sized for 60 x 12 or 720 plf for a span of 36 feet. Well this is the correct analysis for "stickthink" but for SIPs it is different.
We should think of the roof planes as huge diaphragms, or shear planes, that are about 13 1/2 on the run and 36 feet long. With a 6/12 pitch, they can be thought of as box beams laying over on their sides with an effective depth of 6 feet, the height of the ridge over the wall plate. This gives a box beam of 6 foot depth for span of 36 feet with a load of 360 plf (half of 720 plf because there are 2 working roof planes), a small load for a diaphragm with a depth to span ratio of 6/36 or 1/6. In other words, we don't need a ridge beam or collar ties!
I am not an engineer and have not "run the numbers" for the example above, in fact my header box beam charts run out for a 10 �" thick SIP at 4 feet deep, 24 foot span with 569 plf. This would probably work for our example, but it needs verification.
In any case, it is clear to me that this thinking puts us in the right direction and that a lot of supplementary large timber framing members normally called for in SIP structures may not be necessary at all.

It must also be noted here that another popular roof form is the hip roof and that the common structural resolution of this shape also calls for the proverbial ridge beam or collar ties at 8 foot O.C. max. If we take our example from above and modify its length down to its width so the building has a square foot print and thus the roof becomes a pyramid, it is easier to see that the roof may be seen to act as a cone, with a radiating horizontal thrust component all around the perimeter. Some kind of tensile ring at the perimeter is a good structural resolution for the horizontal thrust and may achieved quite inexpensively.
What we have used is sheet metal corner strapping at the roof edges. If we assume the entire roof load (most extreme conservative analysis) to be the horizontal thrust vector, the above example would look like this: total roof area = 24 ft. x 24 ft. x 1 . 13 roof slope factor for 6/12 pitch = 651 sq. ft. Times our 60 psf loading = 39,060 lbs. Each of the four corners is subjected to a quarter of this, or 9,765 lbs. If we use 12 inch wide 20 gage sheet metal, which is .035 inches thick, at 24,000 lbs strength this is good for 10,080 lbs. Keep in mind that this result is extremely conservative in that: A) The horizontal thrust is something less than 100 % of the full vertical loading, B) the four roof sides are also joined at the hips and not only at the perimeter, and C) the rooftowall connection will also take some of the load.
However, we can see that this approach for the resolution of horizontal thrust vectors in hip roofs can work relatively easily. One should also check that the proper number and type of fasteners are used to transfer the load to the subfascia, and that this piece is also of adequate size to handle the load. Just to finish up our example; the load to each side of the joint would be half of 9,765 lbs or 4,883 lbs. At 150 lbs capacity for each no8 screw in shear into douglas fir this calls for 4,883 / 150 or 33 screws. A last check of the 2 x 12 subfascia is its area, 1.5" x 11.25" = 16.875 sq. in. times its allowable load in tension parallel to the grain (800 lbs/sq. in. depending upon species) = 13,500 lb capacity, we're in very good shape.
We have built several hiproofed structures this way, the oldest has been up for over 10 years without showing any sign of any problems.
To sum up, the consideration and analysis of SIP structures with methods that more closely reflect the mechanical properties of this material and their assemblies rather than the performance of stick structural assemblies may enable the designer/engineer/builder to significantly reduce or eliminate large expensive structural components that may now be understood to be redundant. We must remember that structures behave according to the laws of physics, not according to our habits and wishes, specifically that the loads always follow the stiffest and shortest pathways to their resolution, no matter what we may wish or otherwise think.
That SIPs are extraordinarily good at handling inplane loads, and that assemblies act as stiff large diaphragms, or box beams, should eventually find resolution in a paradigm shift of the conventional load analysis and "best practices" of the design and assembly of SIP structures. This will further the raising of the bar in terms of the economy of utilizing SIPs for design and construction and further distance them from stick construction.