Welcome to the new and improved Proflex / K2 Tech Forum!
Michael, your 300 square wave has a foot control right?Get some 6061-t6 strip to practice on. About 150 amps AC, 70% straight polarity, medium arc balance. Run about 15 cfh argon, with 3/32" x 2% thoriated tungsten ground to a pencil shape point. Trust me on the tungsten, and don't ball back the tip or I will have to slap you.With your practice strip, if you think you need to take the gas flow up to 20 or so, do it but stop there. Don't forget to use 5356 rod, 3/32 diameter.Try to get full penetration into the joint, but stop every 1/2 inch or so to let the heat dissipate or you risk taking the weld into solution which will allow it to soften too much. Remember 6061 alloys take minutes to detemper instead of hours like 7005. If you are doodling in the weld pool and the part overheats, you have cooked the heat treat out of it. To illustrate, say you strike an arc, tack a piece and the tack cracks because it froze so fast that it over aged. If you take longer, the tack area will retain some heat and the tack will cool slower and be softer, but won't crack. Doodle in the tack, and the whole area will eventually go dead soft. Just don't let your weld go dead soft. If done right, the metal will fall from t6 to t0 during your weld, but the heat rushing away will allow the area to strain harden and bounce back to t2, then age to t4.Oh man, my brain hurts. Now go practice. Terry
The bracket puts an prying load normal to the tube wall at that pointthat causes one of the secondary moments I mentioned in the thread, andthis bending plus the stress concentration at the change in area causedby the weld bead and the bracket causes this failure. The bracket onthe left side might not have a symmetric joint, but it's hard to tellfrom the photos.The bent seat tube aggravates the problem by causing a primary bendingload that is eliminated by using straight tubes.