Posted by: dalecyr ® 01/30/2011, 17:42:11 Author Profile eMail author Edit

Assuming a horizontal steel tube beam is connected perpendicularly to a vertical steel tube column, as shown in the picture; Assuming all welds are full penetration; In the picture shown below, the connection on the left shows the beam connected to the column by welding all four sides of the beam to the column. The connection in the middle shows a 'slot' cut in the beam, with the remaining sides of the beam proceeding to the left side of the column. The beam has been upsized to accomadate this type of connection. The beam sides are thru-bolted to the column. Welds are made at all junctions of the beam to the column. The connection shown on the right is similiar to the first one, except that a gusset has been added to both sides of the column, with welds at all junctions. Intuitively, the second method provides a stronger connection than the first, and the third method provides a stronger connection than the second. Is there a way to quantify this differance?

Replies to this message

: 90 degree connections of steel tube
: 90 degree connections of steel tube -- dalecyr	Post Reply	Top of thread	Engineering Forum

Posted by: jboggs ® 01/30/2011, 22:48:40 Author Profile eMail author Edit

Your gut is right. Here's why: The first point is that generally welds are stronger in shear than in tension. I always try to design my heavily loaded welded parts so that the welds are loaded in shear (a sliding load rather than a pulling load). Your gusset in the third case is on the side of the beam joint, so it is loaded in shear rather than tension. The second point has to do with moments. The joint is a moment connection. A moment is really just the effect of two forces (or anchor points) and the distance between them. The greater you can make the distance between to two fixed points the greater the moment they will resist. The legs of the gusset are longer than the welds that just go across the width of the beam. That's why they result in greater moment resistance.

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

|   |

: 90 degree connections of steel tube
: 90 degree connections of steel tube -- dalecyr	Post Reply	Top of thread	Engineering Forum

Posted by: Pinkerton ® 01/30/2011, 19:05:13 Author Profile eMail author Edit

Dare I? The bolt in the second one would be superfluous once welded unless the bolt was a tight press fit and even then I doubt it would assist much. The cooling and shrinkage stress would load the bolt unevenly and provide little assistance. Generally a joint is never bolted AND welded. In my experience, mutually exclusive. The third one would depend on the yield strength of the combined gussets and tube-ends versus the yield strength of the tube along the angled weld of the gusset to the tube. I would see #3 failing at that angled weld line to the gusset. You would end up with a perfect, super-strong, gusseted corner piece but attached to nothing. For the gusset to be useful, the tubes would instead have to be solid steel bar. In all three, the end strength of the total joint would still rely on the strength of the tubing in any one lap-joint as any load would be a lateral one. Just to put the mind at ease, check the yield strength of a 1" length of 1/8" fillet weld and then see how many inches it needs for the tube yield strength, compared to how many inches of weld there actually is. In my experience, there is usually around 8 to 10 times more weld length than is required. More is OK though. Don't over-think it, use #1. Dave

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

| |

: 90 degree connections of steel tube
: 90 degree connections of steel tube -- dalecyr	Post Reply	Top of thread	Engineering Forum

Posted by: Kelly Bramble ® 01/30/2011, 18:24:14 Author Profile eMail author Edit

Good question - I agree with your analysis on strength characteristics. I think a quantification could be expressed as a moment or distance/area of greater load distribution. The gusseted design (third) spreads the applied loading over a larger moment and area, therefore the applied loading will be less per unit of cross section.

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

| |

: : 90 degree connections of steel tube
: : 90 degree connections of steel tube -- Kelly Bramble	Post Reply	Top of thread	Engineering Forum

Posted by: dalecyr ® 01/31/2011, 13:56:14 Author Profile eMail author Edit

Thank you. I understand all of your comments and am currently working through Dave's suggestion to check the yield strength of a 1" x 1/8" fillet weld (which I have done) versus the yield strength of the tube itself (which I am in the middle of). Again, thanks. dale

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

| |

: : : 90 degree connections of steel tube
: : : 90 degree connections of steel tube -- dalecyr	Post Reply	Top of thread	Engineering Forum

Posted by: dalecyr ® 02/02/2011, 20:52:39 Author Profile eMail author Edit

I understand that the joint on the left above (first post) is adequate to support the intended load, due to the nature of sheer loads using welds. Especially welds all around. But, to sort of finish the topic, and to check my understanding of your comments: If I understand the above comments correctly, the joint type on the left below is stronger than any of the types in the first post. The gussets have been placed *under* the sidewalls of the beam, and along the sidewalls of the column. This configuration addresses the concept of using a greater moment resistance, *and* the concept of the beam itself failing along the angular weld line. In the picture on the right, the beam has been placed *on top of* the column, eliminating the need for *any* of the beam to column welds to carry any (significant) load. The gusset, again placed under the beam sidewalls, extends the moment connection. Of the five type of connections shown in this post, the one on the right is the strongest of all. Or am I missing something?   Modified by dalecyr at Wed, Feb 02, 2011, 20:59:02

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

| |

: : : : 90 degree connections of steel tube
: : : : 90 degree connections of steel tube -- dalecyr	Post Reply	Top of thread	Engineering Forum

Posted by: jboggs ® 02/03/2011, 07:54:30 Author Profile eMail author Edit

When you talk about strength of joints that discussion is meaningless unless you also define what types of loads are under consideration. For example in your diagram above if you are talking about a pure moment load both arrangements have equal strength. If you are adding a vertical force load, I would prefer the one on the right because it loads the columns directly in tension. However, both arrangements still load the welds in tension for moment loads. If my diagram below comes through you will see what I mean by arranging the welds so they are in shear rather than tension. Also, for what its worth, if you are loading the joint to an amount that it is anywhere near failure you are way too close for the actual arrangement to make much difference.

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

| |

: : : : : 90 degree connections of steel tube
: : : : : 90 degree connections of steel tube -- jboggs	Post Reply	Top of thread	Engineering Forum

Posted by: dalecyr ® 02/03/2011, 12:08:34 Author Profile eMail author Edit

"loads the columns directly in tension" You did mean "compression", right? After studying your diagram, I understand now how "your way" loads the welds in sheer, and "my way" loads them in tension. Thanks. Modified by dalecyr at Thu, Feb 03, 2011, 12:16:36

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

| |

: : : : : : 90 degree connections of steel tube
: : : : : : 90 degree connections of steel tube -- dalecyr	Post Reply	Top of thread	Engineering Forum

Posted by: jboggs ® 02/03/2011, 12:22:36 Author Profile eMail author Edit

You are right. I got fat-fingered again. Compression - not tension.

Post Reply Tell a Friend (must be logged in) Alert Admin About Post

Where am I? Original Top of thread

|   |