I'm on a property in which there is a garden shed and its roof collapsed yesterday because of heavier than normal snow that accumulated on it over the past few months and it looks like yesterday was simply too much for it.

The roof trusses were made in an (unusual?) A shape like this:

enter image description here

I'm reading online that the rafter tie should have been closing the truss triangle shape at the bottom so that the rafter tie touches both the outer walls and the rafters, like this:

enter image description here

The question is just how much of a difference in load weight support is there between the A-shaped (1st picture) vs. triangle-shaped trusses (2nd picture) ?



Ballpark, factor of 4. In general strength goes up with the square of the depth of the web. But this is only an approximation, and very much depends on the failure mode.

Here's a truss calculator: https://skyciv.com/free-truss-calculator/

Edit: I've been mulling this over, and am no longer convinced that my first order answer is correct. A truss is not a beam. A beam has two main failure modes under a transverse load: Either the top crushes, or the bottom tears. A truss has far more failure modes -- in effect each element has those same two failure modes, plus compression and buckling of the element as a whole, tension failure as a whole, and the connectors that tie the elements together.

The truss with the collar tie at the wall is a 'common truss' If you run a post from the peak down to the collar tie, it's a "king truss"

I do know that A-frame (also called 'raised collar truss') are used less often. This source: https://hubpages.com/education/The-Flexible-Roof-Truss-And-Its-Many-Design-Options mentions that you need to use larger rafters (top beams) when using a raised collar. By implication a raised collar truss is weaker than a common truss when built out of the same dimension wood.

If you are planning on building your own trusses, various extension departments for universities offer truss plans. Note that you MUST follow the nailing diagrams with reasonable prescision. It it says 11 4d nails in such-and-such pattern, do it.

Alternately consult with your local truss builder. They are set up to to make them in a reasonable hurry, and the market is competitive enough that it's often not worth making your own. Indeed: Having a pro design them may save enough material to pay for them to make them and deliver them to your site.

Apologies for my initial answer.

  • So you're saying that the rafter tie truss (image 2) could support up to 4 times the snow weight of the collar tie truss (image 1) ? That much? – that-ben Mar 25 '19 at 21:55
  • Thanks for editing your answer, but I'm still asking for clarifications regarding the difference in weight support when comparing the reduced triangulation of raised collar tie vs. the common king truss whose tie touches both outer walls. Am I still correct to think that the full length common rafter tie touching the walls allows for more weight support throughout the whole truss when compared to the raised collar truss? – that-ben Mar 26 '19 at 16:14
  • Yes. But I don't know how much. – Sherwood Botsford Mar 27 '19 at 1:02

The two principles in effect here are: 1) cantilever (top picture), and 2) simple span (bottom picture). Neither are “trusses”!

1) The roof load in the top picture Is being supported by the roof rafter ONLY. The horizontal tie is merely keeping the roof rafters from spreading the walls apart. The roof rafter is acting as a “cantilever”.

2) The roof rafter in the bottom picture is supporting the roof load, but in a “simple span”.

Yes, the bottom picture can support more, but it is based on where the horizontal tie is placed and how long the cantilever is...

I have no idea where a “factor of 4” came from.

  • Neither are trusses, really? Not even the bottom picture? In hindsight, the pictures may be misrepresenting the angles. In the upgraded "truss" I want to build (bottom picture) the length of the rafter tie is 16 feet, each rafter is 9-1/4 feet long and the cantilever extends 1 foot per side. – that-ben Mar 30 '19 at 15:16
  • @that-ben Yes, technically it’s called “stick built”. That is installing one member on top of another. The sloped members resist all the Live Load (snow load) and they are in “bending” (which puts the member in tension and compression). Truss design has the members entirely in tension or compression. – Lee Sam Mar 30 '19 at 16:18
  • @that-ben Oh, if you cut the bottom member at the vertical member, then you’d have a truss. – Lee Sam Mar 30 '19 at 17:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.