FWIW, we just did a 410 to 316L with 316L filler on 1/2" material with no PWHT, 300F preheat and had no embrittlement problems. It is key that the 410 be in the annealed condition, ~ 65Ksi UTS. I didn't have the time or money to try other HT conditions, and I really doubt very much whether we would have successful with hardened or quenced and tempered conditions.
We weld a lot of 410 too with no heat treat. The issue is this is a 430ti to 430 weld currently. I want to kill any possiblity of a problem since we have cracks in the field. I am even willing to consider using a stabilized austenitic filler, but the joint is not designed for filler. 410 might give us better results, but also could bring the same problems. A lot of variability seems to exist between suppliers too. I am afraid it could become quite a science fair project if we have to generate a custom material spec of our own. There is some possibility of finding a smoking gun tramp element, but there may be more practical and easier fixes to implement like using 410. The effort to switch and requalify is a bit of work too.
That is a pretty good idea, but I don't know if it would be any better than a fillet weld. The current joint is an edge flange weld made by a connector nut that pilots over a tube. The resultant weld resembles a fillet weld. That is what I hate about it. The weld symbol calls a butt for a flange that ends up a fillet. It is a constant source of confusion and debates for design reviews and shop inspections. If we could add filler and make it a fillet weld I would be happier, but not all others. The concern would be getting pen on the edge flange, but is pen really an issue on a fillet if the edges are fused? I have only seen fillet weld specs for root pen in the vertex - ie a throat dimension. The problem is the weld is a stack and not a T though.
That gives me the idea of welding a 430Ti to a 321. 321 is available in bar and could be made into our connector nut.