Tension Rod Connection to Central Ring

[MrEngineerUS]

Hi everybody,

Well, I'm stumped. I have an architect that wants to connect two tension rods that form the bottom tie of a roof system to a central hanger using a 4" diameter pipe section that is 2" long. He wants the rods to connect through the wall of the pipe and washers and nuts on the inside of the "ring". The tension rods slope away from this central connection radially at a slope of 1/12.

I thought there would be some type of equation out there to check the pipe section but I haven't been able to find anything. Any help greatly appreciated. From other online research, these type of connections don't typically connect through a pipe/tube but there is a center plate that can be shaped like a ring. This seems better as the tension loads are transmitted by shear and there is less of a concern for stresses due to weak axis bending in the walls of the cylinder. Anyway, take a look. If anyone out there has some solid advice I would love to hear it!

See the attached for the connection I'm talking about.

 

[weab]

The first thing that comes to mind is to use Roark's Formulas for Stress and Strain which has stresses on rings. It works best when the loads are at 90 or 180 degrees to each other. It is easy to use superposition for multiple loads, in this case 3 loads.

Since this is not exactly the case here, this may call for a finite element analysis. If so, and since the shape is very stable, I'd compare the service load stresses to an allowable of 0.6 or .66 Fy.

Keep in mind that if the loads are not symmetrical, the pipe may tend to rotate with the loads not applied through the pipe center line.

 

[weab]

Also, how will you tighten the nuts? I assume that you will need to either install rods in the pipe then tighten the other ends of the rods, or you would use washers with a rounded side and a flat side. Just curving the washers don't provide for a good bearing surface though nuts are very hard. If the loads are high, you might have point loads in the pipe walls.

I'm OK with the concept used here but not the approach in your sketch.

 

[dhengr]

MrEngineerUS:
Take a look at Roark’s Stress and Strain, “Circular Rings.” I’m looking at the 5th Ed., pg. 224, Table 17, #5. Carefully read the verbiage before this table, and the header and any footnotes, or you can shoot a bull pretty easily. I’m sure we could find some of these various conditions developed more fully in any number of advanced Strength of Materials and Theory of Elasticity texts. I too, would be worried about the wall bending and deflection of your 4"o.d. x .25" wall pipe. In the extreme, it is going to turn into a stretched, six sided shape.

I would probably use some mechanical tube rather than some generic galvanized pipe, you will probably have better luck with the mech. properties. Your two bot. rods with that shallow slope will likely develop very high tensions/stresses to carry fairly small gravity loads, and they will need substantial horiz. thrust reactions (compression ring like?) at their outer ends. The curved washers and beveled nuts (?) should be replaced by washers made from a piece of pl., with one flat side, for a washer and nut, and the other machined to match the i.d. of the mech. tube. These could be cut from a 3.5" o.d. bar, lengthwise. You probably won’t get all that stuff; your special washers, regular washers, nuts, rod extensions, and wrench clearance into a 3.5" i.d., although I’ve not laid it out. I think I might sooner make the center connector out of 2" thk. stl pl. or some such, six sided, in side view, with a 2" i.d. center through hole, three edges would be perpendicular to your rod axes, and be about 1.5"-2" long. These three edges would be drilled and tapped for the .5" rod, which would extend into the center hole for a keeper pin or cotter pin through the end of the rod. These three edges would be connected by a horiz. bot. edged, and two edges tangent to your 95̊ dimensioning arcs. Again, I haven’t laid it out, so it may need some adjusting of the above.

You should really describe your problem much better than you have. We can’t see it from here. What are the rod forces, how are the rods loaded, what supports the vert. rod, and what’s its load? What does the entire roof system look like? Are there any lateral loads on the rods? You might make the bot. rods in one piece, threaded at the two ends. This bot. rod would rest in a saddle with a slight radius to match the 1:12 slopes. The saddle would have side plates to pick up the upper vert. rod.