torsion on shear tab connection

[NS4U]

I have a light steel frame, which has beams with slightly eccentric vertical loads. I end up about a 170 ft-lb torque at the end of the beams that needs to get passed through the shear plate connection at the end of the beam. This torsion seems absurdly small to me, but it is killing my shear plate capacity, when I calculate the shear stress in the plate due to torsion. right now, the plate is only 1/4 in. thick.

Any thoughts on how to handle this torsion? It seems crazy to use a super thick shear plate for a frame that has loads on the order of 1-2 kips. Using other connections (angles, tees, end plates, etc.) is not an option. Needs to be a shear tab.

Thanks.

 

[DaveAtkins]

What formula are you using? I would use fv = 3T/bt*2.

DaveAtkins

 

[NS4U]

Dave, could you clarify that equations? The units don't seem to work out. I may be reading it wrong.

 

[jayrod12]

I'm fairly certain the units work on that. ft*lbs/(ft*ft^2)=lbs/ft^2

 

[WARose]

Keep in mind: the weak link in such a system will be the web of the beam in bending. For simple (shear) connections the only way to get that torque out is basically for a couple to form by opposing shears cantilevered from the top and bottom flange. The shears get to the connection by bending in (unfortunately) the weak axis of the web (which tends to be thin). If you are getting overstresses in the tab, more than likely it will blow out the web too.

What I have found is the best approach is to put in subframing to take out the torque. Even if the subframing has simple connection you can still take advantage of the fact it has some degree of fixity (assume maybe 10 %?) to take care of it. I've modeled systems like that before and it will go into the subframing before it goes to the end of your girder. (And usually the moment is small enough for those connections to handle. I know it is counter intuitive to use simple shear connection for such a purpose.....but typically it isn't a problem.)

A good post on this:

http://www.eng-tips.com/viewthread.cfm?qid=307795

Hope this helps.

 

[KootK]

Firstly, you get bonus points for going to all this effort to do the right thing even when the loads are small.

Perhaps you could look at it as two cantilevered horizontal beams instead:

1) A horizontal strip of steel at the top of the shear tab where the beam pushes against it.
2) An opposing horizontal strip of steel centered on the bolt furthest from the the horizontal strip in #1.

I've never attempted this myself but, if that check works, I'd be tempted to call it good. For something of the proportions of a shear tab, I would expect the torsional response to be dominated by warping deformation. I'm not sure that a St. Venant shear check even makes sense. I feel like my double cantilever strip couple would pass as a poor man's warping stress check.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.