I think I see what you are talking about from your excellent graphic. So what I took to be a rigge bopard sandwiched between the ends of rafters, like conventional is really below and is supporting the rafter pairs at the peak. So it is like another bent that runs in the center of the buildin, down the length and it supports the common rafters at the peak. Would I still need collar ties? I like this very much as I agree with your thinking that it will simplfy dormers. The only difference is I planned on 2 dormers per side each 6 feet wide. I had also planned on a chiney centered both length and width which would not work but that can be shifted forwardon one side with dormers to right and left. Thanks, I am starting to seem some possibilities. Joe

No, you don't need the collar ties with a ridge. Depending on frame design you may need to have a properly designed bent with a tie beam that is in tension... but that is another matter.

I'm trying to piece together some of the bits and pieces from this thread with regard to collar ties. Jim cited a study suggesting that a collar tie is only effective in preventing spread at the rafter base if the tie is within a foot of the base. The conclusion being that any collar tie position higher in the triangle only serves to prevent sagging of the rafter under load (with the force on the tie being compression).

Later in the thread, Mike says that collar ties are unnecessary with a supportive ridge.

So, taking both points into account, am I to conclude that the only reason I see so many roof structures with collar ties higher up in the triangle is because the rafters are undersized for the load?

No, I think a lot of people do not understand how a collar beam works. I find a lot of stick builders just do not want to grasp the concept.

The German "Kehlbalkendach", literally Collar Beam Roof, is a roof system that relies on Collar ties to somehow strengthen or stiffen the roof structure. The simple Kehlbalkendach consist just of rafter pair joined with a collar somewhere between the mid point and top third of the rafters. The idea here is not that the ties will resist thrust at the plate, but rather create another triangle that stiffens the roof against live loads (triangles are immutable) essentially allowing them to get away with using smaller rafters.

The spans handled by this simple roof structure are surprising.

Variations of the Kehlbalkendach are perhaps the most prominent roof structure in German speaking countries. For larger spans purlins are used to support the collars.

a collar tie works well against live loading. This is because under wind loads, the tie transfers load to the other rafter in the pair, and so 2 pieces of wood can be used to share a single load in that way.

Kevin - I would work to have properly sized rafters, and not rely on the collar tie for stiffening.

DL - all the triangles make some great bracing.
In stick framing they make a great place to hang the sheetrock. ;-)

I'm sure they are offering some sort of stiffening or bracing. The rule I've always worked with is lower third of overall vertical rise.

I'll try an throw a few examples at my FEA software and post some deflection images. In some cases on large spans I've seen bending requirements go way up at the rafter foot as the really stiff upper triangle does more harm than good (when the tie is placed too high).

Yes, the rafters need to be made of the proper size to support structural loads on their own. In the German system mentioned, the point is to resist wind loads, keeping in mind that the roofs they are used on are very steep, often steeper than 45 degrees. Essentially they are acting as another level of bracing, using the rafter to pairs to brace each other. They are also almost always placed about mid span if they are not supported in any way by purlins, and many times may serve the extra function as joists for a loft floor.