gonzopancho
SILVER Star
Welp, I didn’t get to take it out yesterday. Mostly because we’re trying to address the heat in the cab, and the wife was going to be in the passenger seat for the 100 plus mile loop from home and back again.
We did take the 250 though. It’s a Toyota 4x4 station wagon, right? The rest of the group was our son (in his 80) and the other current (and two former) guys in Bump It, along with their wives or girlfriends, kids and dogs, except TJ, who had other plans. Kind of a company outing.
Don’t apologize for what didn’t occur.
I did say that the whole subject of MIG and making ‘e’ shapes is new to me. I was taught to control the shape of the pool/puddle, mostly by what seems to now be described as “push pull”, but dad was emphatic about consistency of the pool/puddle shape.
My uninformed opinion is that TIG is a better process than MIG for nearly any welding, but it does cost more, which is why MIG is popular, I guess. (Proper prep costs more too.)
I’ve not done it, but TIG seems to me to be superior because you’re directly in control of the heat, vs. a static setting on the machine and trying to move in and out of the pool/puddle to control the heat. The “stacked dimes” of TIG looks to me (again, I’ve only watched video) to arise from the overlapping dabbing of the filler rod into the puddle. The torch (tungsten) often moves in an ‘e’ pattern to control the heat in the area being welded.
But hey, let’s ask Bob, or anyone who is proficient at TIG, really.
A big difference with MIG is the wire moves at a constant rate, and it has to go somewhere, so there is a lot more motion of the wire (gun) than there is in stick, and (again, haven’t thought about it, haven’t done it), but an attempt to emulate TIG with MIG is likely going to result in a less strong weld, especially with an inexperienced welder who is primarily focused on aesthetics (stacking dimes) not the strength of the weld.
How the cap looks visually (short of showing true weld faults) seems very subjective. It's like someone saying you have to have a preference for partners with blond hair, not red heads. Simply because that's what they prefer.
Seems we agree here. Let me know if that’s not the case.
I didn’t take a class in FEA, it was too new a technique then for a class to exist in the BSME sequence. (There might have been a MS level class. I don’t remember.) I know the linear and non-linear optimization classes taught by Dr. Peterson were 400- and 500-level, but he let me in because it was useful for what I was doing for him.
What I did do was modify the code to NASTRAN to deal with things like non-linear loads for Dr. Chase and wrote and incorporated both a linear (based on the techniques in a program called LINDO out of U Chicago) and non-linear (based on Dr. Peterson’s research) optimizer for Dr. Peterson. I also wrote number of output filters to a different software package (MOVIE.BYU) to visualize the output of NASTRAN. Said output was typically 50-250 pages of fan folded computer printout that someone would then dig through with a pencil and ruler to find the required info.
Turns out many professors have consulting gigs on the side. Their customers loved the (very early days) visualization they could show. We were even working on some animations by running the model several times (took a long time, tied up the Vax for hours) and then stitching together the frames into a short movie.
“here is a visualization of the bending moments due to 20cm and 100cm accumulated snow loads on your new domed stadium, as well as my five figure invoice.” (Yes, red is bad news.)
There were a lot of “Green is good, red is bad” pics. These must be right, it came out the computer! <Insert numerical analysis here.>
(Also how I got into high-tech, but you don’t ask.)
Structures do deform under load, sometimes to yield.
TL;dr: the direct answer to your question: No they didn’t talk about “warp” due to welding in any FEA class. No such class when I was there and it’s easier to model stress and deformation due to point loads on, say, a beam or mast, such as a flagpole. This is the type of thing that is taught.
I did, however, take thermodynamics and heat transfer (these are separate courses), and I did a project for a professor (in Civil Engineering, but he was also my Mormon Bishop, and I was trying to keep him happy, as BYU is a religious school, and by that point, I was already considering walking away from that church and everything associated with it) to predict “warp” in a welded structure (a truss) for a client.
The basic answer is that metal expands when heated (and the material can undergo chemical and structural changes due to this heating, but we won’t worry about that right now.) The surrounding, cooler metal acts as a constraint, preventing the heated area from expanding freely.
As the weld (and metal around it) cools, the metal contracts. Because it's restricted by the cooler metal, it can't shrink back to its original size. This creates stresses which can cause the metal to permanently deform or “warp”.
This is a big reason why pre-heating is a good technique. If you wish, I can go into detail why “driving the water out” is all but meaningless *for mild steel*: Pre-heating to an assumed 250-300F is great for alloyed steels because it drives hydrogen diffusion, which will reduce hydrogen-induced cracking. A high carbon steel (which our 1970s truck frames are definitely not made of) will require a higher temperature than this, and I’m of the opinion that crayons or temperature sticks are preferred to IR thermometers or a wild-assed guess when determining if the correct pre-heat temp has been obtained. At the same time, you’re not doing any harm to mild steel with a pre-heat to 250-300F, and *maybe* you’re reducing warping some. You do you.
Let’s return to “not sitting in class”:
The modeling involved application of Fourier's law (a fundamental principle of heat transfer that governs how heat is conducted through a material) into NASTRAN with a relatively simple model of the thermal expansion of the material, constrained by the modulus of elasticity of the material and the rigidity of the truss as it is being constructed. Within limits, hotter areas expand more, especially if they are not already welded, but welded structures also encounter more stress due to the expansion. FEA is fundamentally about sectioning the model of the structure into very small cubes and applying the stress model to each cube, so each cube expands based on the temperature (and presence) of the surrounding cubes, but the metal is also undergoing stress where the structure resists this movement. Repeat for every weld in whatever sequence you defined.
Fourier’s Law, however, does not describe the entire welding process on its own, which also involves other heat transfer mechanisms like convection and radiation. Note that these are both applied to cooling in the welding process, and cooling is the phase that gives rise to most of the stress in the material being welded. Apply what you learned in thermo and heat transfer. Repeat for every weld in the sequence you defined. For the entire truss. Every time.
Once we had a model that would generate verifiable results, it was repeated for a number of weld sequences and cooling times in a linear optimization framework to minimize the resultant deformation (“warp”). Remember there are a lot of welds in a truss and that most trusses are not linear in the load bearing axis (aka “crown”), and that the construction of same is critical to the resulting work piece.
And then when you’re finished, you load test the deformed truss, first with NASTRAN, (in the computer) and then after verifying the deformation model with real world measurements (again) you load test (with strain gauges) in the real world. Way more satisfying than homework… IJS.
Since you also asked for my thoughts on “supporting a frame during welding”, I’ll attempt an answer, though I touched on same above.
By “supporting”, I assume you mean work holding or clamping. Clamping the workpiece is generally a good idea. But, even though clamping will reduce deformation, it also can, in some situations, increase the residual stresses in the workpiece, which can lead to cracking. This is why, for example, there is a rule of thumb about keeping the bead six bolt diameters away from any bolt hole, even if it’s not bolted. (Remember what I said in the other thread about every hole representing a rise in stress concentration?)
If you're making reference to your "welding while the frame is suspending in a rotisserie" BS, well, you already answered that question, it doesn't matter.
Lots more on the subject if you wish. Decent overview:
Only 72 pages, some of them not interesting, though there is a light discussion of notch stress in fillet welds. Something else I mentioned.
I guess I’ve not forgotten everything. Thanks for the trip back down that time in my life. I’m not sure it was a happier time, though it was less … stressful.
Last edited:
walk 55 
that nobody gives 2 ****s about! Especially from someone who quoting you, does not know how to GTAW or GMAW and has'nt welded anything but Stick in 30 ****ing years! Why are you even responding to this. It was supposed to be a thread for learning and a the first sentence in this thread is A Lighthearted response!! The thread was intented for a response from your Newly aquired employee's It was intended as a learning tool for people with less expierience (apperently you need it more than anyone!) So you can read all your theories and other 
