It's not like steel. Not even a little bit. I can't extrapolate my kinesthetics because it *will* *not* *work*. Stop trying.
- Thin sections *will* melt. If I need a thin section I need to weld in a thick section and then cut it out afterwards.
- All joints need filler, and a LOT of it -- more filler than the running length of the joint. If it's possible to weld without filler, boy, I sure don't know how.
- Just as critical: do not start adding filler, don't even think about it, until there's a decent-sized puddle, because it'll freeze *instantly* when the filler drop touches it, and until there's a puddle the filler will just bead up on the surface and be wasted. Stop trying to add filler too early.
- Corollary to 'thin sections will melt' and 'do not add filler until there's a decent puddle' -- don't start at an edge or corner, because they're too thin and you can't run a puddle cold to keep from melting the corner away because you can't get the puddle started. Start in the middle of a joint and weld to each end.
- Don't forget how much aluminum distorts.
- Don't forget how soft it is and how easy it is to flatten it again by welding just a little on the other side.
- The starting arc *must* be with the tungsten pointed at, *and* nearer to, the thicker-section piece of metal. With that said, you can't make a puddle in the thick-section and push it up to the small piece and expect it to bridge over and finish the weld like steel does -- some heat has to go into the small piece to get an adjacent puddle and then add filler there.
- Despite the AC, the oxide layer is still an arc-starting issue. Hit it once with sandpaper just before starting a weld.
- Hot aluminum welds better, faster, than cold aluminum. Hot-restarts are viable.
