Engineering question related to welding

[txpitdog]

In the grand pursuit of improving my welding skills, or lack thereof, from shameful to simply embarrassing, I am constantly at war with distortion. I have felt the effect of distortion while attempting to freehand a 90. I could feel the metal pull as it cooled. So, as a true engineer I wondered is it possible to calculate the effective force or torque that the cooling metal applies so that I can make sure to provide an equivalent opposing force? Surely there’s a way and surely someone has done this before.

I know that a moment can be calculated around the point (weld) at which the force is applied. I.e. the amount of force required to move X lbs of steel 1.5 degrees at 6ft distance. That’s a number that exists, and once I figure out how to determine it, I can plan for it and offset it.

I tack my opposing corners and then weld them in, but I will still get one corner of a truss that is off, or the truss comes out like a pringle. I am sure there is a practical way to accomplish this, but in reference to the first paragraph, I am an engineer.

[BO-N-ARO]

We always ran sample parts of the material we would be welding with different angles and used the one that pulled the parts into spec.

[BBReezen]

Turnbuckle braces. Tack your pieces square, and weld and adjust as needed. If you have good tacks and a well braces structure and skip around, you should be able to weld with minimal draw. And don’t weld so hot.

I figured out a trick when I used to be a welder full time. See which part of the end of your material has the bad angle, tack one side with the material at a greater angle than you need so your material will draw to close your angle. And if it over draws, you should have sufficient gap to draw back the tother way.

****ty tacks aren’t anyone’s friend. You should be able to tack something together and it have sufficient strength for structure without welding. Everyone likes big boobies, nobody likes big tacks

[Draco]

I build quite a few tables for an interior decorator. The legs are always square tube. I have diagonals clamped to the top and two directions to the leg I'm welding. Four quick tacks and ready to weld out with no pulling. After 40 years of doing this stuff, I find clamping and plenty of fast, light tacks the only way to be sure. If you try leaning it away and tacking then pulling to you, your measurements will be off just a little and a little is a lot if you don't want the table to rock.

[easeup]

if you must churn on this question in the OP, the strength of the pull is limited by the ultimate strength of the metal or the filler, usually 60 or 70k psi. you estimate the size of the tack weld and run the pull strength number when cooled using the appropriate formula in the welding geometry tables.
But the reality of it all is more easily resolved as the men above have provided. turn the heat down and make smaller tacks, let cool as you go around the work piece.

[ramrod]

Don't make a mountain out of a molehill. Just leave it a little out of square and tack it on the side that will pull it back in square (tap with hammer to fine tune). It would be hard to calculate as the tack size/length/pentration will always be different.

[Hoggslayer]

Don't overthink it. Just clamp it up, brace it up and weld it. I always try and weld it in the direction that it will draw the least first. then when I weld the other side that it will draw the most on, it pulls it back square.

[WhiplashTX]

Don't overthink it. Just clamp it up, brace it up and weld it. I always try and weld it in the direction that it will draw the least first. then when I weld the other side that it will draw the most on, it pulls it back square.

This, you can counteract the heat draw by adding a weld opposite of the undesired pull. If your welding a 90, weld on the outside first and then the inside. Make sure the metal to metal fit is good, gap filling makes the pull worse. Rome wasn't built in a day and you don't have to fully weld the joint in 1 shot if you know it's going to warp. Experience helps, I'll let you know when I get there.

[Bullseye07]

Maybe you should call one of those new grads if you can’t figure it out?

[Silvey]

You have to gap the piece a little, then square up.