View Full Version : Structural/engineering question, I think...

06-02-2012, 09:19 PM
Its this third dimension that is giving me fits...spending a good percentage of time welding instead of machining and I am finally getting some results I am OK with...basic ladder "shape" (a series of "H" stacked so to speak) using a series of angle iron but now flat and quite close to square :D

The problem is (not finished so can make some limited changes) as it sits it is quite prone to twisting.
Part (b) is there is, relatively speaking, little depth to the shape (about 2"), what is the best way (time, cost, weight, hassle factor...) to combat the twisting?

The option I am favoring is a series of braces working on the diagonals so as to make an "X".

From an engineering stand point, would it be effective if one were to make a series of holes in the "down" legs of those angles and run a round strut diagonally from a "top" corner to a "bottom" corner through said holes?
Somewhere in the recesses of my mind, I seem to recall that if you were to put a turnbuckle in the middle, you can push the corners up or pull them down...

06-02-2012, 09:44 PM
Russ, i was going to suggest getting an "Extra Large" woman to sit on it to hold it down flat, but then you mentioned the "Hassle Factor", so that idea is out!!:D

06-02-2012, 09:48 PM
Pictures would help!!!!!!


06-02-2012, 09:48 PM
How about a pic or something? I have no idea what you're talking about...

06-02-2012, 10:50 PM
Diagonal bracing can work wonders. Sounds like you have a number of stacked 'cells'. You may only need to brace a few of them, not all. But yeah, pics would tell a lot more than words. This thing could be totally different than I'm imagining.

06-03-2012, 01:38 AM
If the project is a ladder, then it will not have much resistance to twisting moments, regardless of in-plane bracing. If the project is a lattice mast, as on a tower crane, then the diagonal bracing will indeed resist twisting.

We need to know what the project looks like.

06-03-2012, 01:43 AM
You need to get your welding proceedures, set-up and welding sequences right - for twist, shrinkage and twist control

06-03-2012, 02:17 AM
Sorry, no photos.

Will try to explain a bit better.

Series of angle iron, 2"x 2" legs, all are oriented as if you take the letter "L" and rotate it 180* in either direction (they are not all turned the same way, turning the "L" 90* clockwise amounts to the same thing since the legs of the angle are of equal length).

Four pieces make up the perimeter (28" x 36" rectangle) with 4x more 28" pieces as cross members. Two of those cross members are "L" turned one way, two are turned the other way.
Right now it is flat and square, a grid with one angle iron leg of each piece being the actual top and the other leg pointed down.

If you stand it on end the section is 2" (leg length of angles) by 28" and from just moving it around it is apparent it can twist relatively easily which is what I would like to limit without adding a ton of weight. If an "X" member is added, it will be a series of short pieces (but in total going from corner to corner) because of the down angle legs already in place.

P.S. since it is related, I noticed on the way home, overhead traffic signs, the ones that span multiple lanes (curb to median say)...well, there are two types, one type is more or less a box, the other is triangular in basic shape but with cross bracing on some diagonal going through the center of the triangular section shape...of interest because, I suspect, the triangular is a stronger shape (?), now the mast of a crane...

06-05-2012, 02:15 PM
OK, so it's a welding table / similar?

A pic would make it a lot easier - or even a Crap-o-CAD sketch on a beermat?

Triangulation's the way to go - diagonal brace (or two) on each face (top, two sides and back, assuming you want to get your legs under it from the front) - this'll make the individual faces stay square, at least.
To stop it "hinging" on the corners, you'd need to tie the remaining loose corners together - maybe a piece of flat across the floor at the front would do that? You could add some gussets at the corners, too - they can be triangles with lightening holes in the middle, or just pieces of flat to form a triangle at the top corners of the open side?

Dave H. (the other one)

06-05-2012, 02:46 PM
Tubular cross members, say 2"x 2" thin wall square tubes would add a considerable amount of torsional stiffness. But maybe it's too late to do that.

Dave Cameron

06-05-2012, 03:02 PM
Another way to stiffen the structure would be to box in part of it. For example, weld a thin plate, top and bottom, covering the area bounded by the side members and two adjacent cross members. Two or three of these boxes, evenly spaced along the length should make a big difference. The plates don't have to be very thick, but without knowing the dimensions of the structure, it's not possible to make a useful suggestion.

Dave Cameron

06-05-2012, 03:30 PM
Tubes/boxing is the way to go as the previous posters suggested. Tubes have about 20 times the torsional stiffness of open members of the same weight.

Mtw fdu
06-05-2012, 04:01 PM
If this is being welded on a metal table or something flat, I would recommend tacking it to the bench itself. Only if it is a metal 1. That will help in some respect.

Mtw fdu.

06-05-2012, 06:10 PM
If the parts are connected with a weld you only need to square and brace one area.X bracing will suffice in one corner.If these parts are not connected then x brace each one.

06-06-2012, 08:07 AM
The X-bracing will prevent racking but won't do much for resistance to twisting.

Dave Cameron

06-12-2012, 04:38 PM
As follow up...welding in the end perimeter pieces helped some but putting in "X" bracing in the middle and outer two sections (there were 5 roughly even sections) did quite a bit to help both stiffness in general but also greatly reduced the twisting action possible.

Of course after the fact I went through part of a Blodgett book I had not yet read, "How to Design Flat Tables", "How to Brace or Stiffen a Member" and "How to Design Steel Frames" and realized my "brilliant" design is very basic and not a very good choice given all the other options.

06-12-2012, 10:50 PM
Blodgett is the best. Most practical author in the field.

06-13-2012, 06:15 PM
Cameron and Jep have it. You need to incorporate some tubing in the structure.

Tube is strong in twisting. Angle is not. The tubing can be circular or rectangular, the latter possibly fitting in with your design better. I hope it goes without saying that you need to attach your existing structure to the ends of the tubing all round the end of the tubing. You can't just weld one side of the tubing to your structure.

If your structure is long and narrow, you'll get some increase in torsional rigidity by putting a couple of short pieces of tubing crossways. The tubing doesn't have to be lengthways. Look at what happens when your structure twists. The end parts turn relative to each other, but the sides do too, although to a smaller degree. It's that smaller degree that means crossways tubes contribute less than lengthways reinforcements, but are still useful.