A stainless exhaust section has to meet a mild steel flange. A stainless bracket has to attach to a mild steel chassis point. A stainless heatshield needs a mounting tab that will be painted rather than polished. In each case the design is sensible and the joint is awkward, because the two metals are close enough to be welded and different enough to cause trouble.
They are both steels, so fusion is easy. That is the trap. Because the arc can join them without difficulty, the joint gets made without anyone asking what the weld metal actually becomes, and the answer is a mixture that behaves quite differently from either parent.
Mild steel carries more carbon than stainless. When the two melt together, the carbon from the mild side migrates into the weld and into the stainless side of the heat-affected zone. The result is a narrow band that is harder and less ductile than either parent, sitting exactly where the joint will see bending and vibration.
That band is the reason dissimilar joints crack where they do. The failure is rarely in the middle of the bead. It starts at the fusion line on the stainless side, in a region that looks perfectly normal from the outside and has almost no capacity to accommodate strain.
The size of the band scales with how much mixing took place. A fast, cool bead with minimal penetration into the mild steel produces a thin band. A hot, slow bead that stirs the two parents together produces a wide one. On a joint that has to survive, that difference is the whole game.
The right filler for a dissimilar steel joint is usually an over-alloyed one, richer in nickel than either parent. The reason is dilution: the filler is designed to be diluted and still land in an acceptable composition. Matching the stainless parent with a matching filler sounds correct and produces a weld that is under-alloyed once the mild steel has contributed its share.
The filler also decides whether the joint can be machined afterwards. A hard, diluted weld fights a cutting tool; a properly over-alloyed one cuts like a tough stainless. On parts that need post-weld machining, that is worth thinking about before the rod is chosen.
On dissimilar steel, heat input is controlled for two reasons at once. It limits dilution, and it limits the width of the hard zone. Both push in the same direction: keep the bead small, keep the travel steady, and do not let the arc sit in one place while the puddle grows.
Cooling rate then decides how hard that zone ends up. A joint that cools quickly in still air at a cold bench is not the same joint as one that cools slowly under insulation. Neither is automatically better; what matters is that the choice is deliberate and repeatable, because an uncontrolled cooling rate makes an unpredictable joint.
Once the part is in service, the joint becomes an electrical couple in an electrolyte. Stainless and mild steel sit at different potentials, and where moisture bridges them, the less noble metal gives up material. The mild steel side corrodes preferentially, and it does so faster than it would on its own because it is coupled to a large stainless area.
This is a design problem more than a welding problem, and paint is the usual answer. Coating the mild steel side, keeping the joint out of standing water, and avoiding a small anode against a large cathode all help. A welded joint with a small mild steel tab on a big stainless panel is the worst arrangement, and it is also a very common one.
Dissimilar joints usually bring a thickness mismatch with them, because the two materials are chosen for different duties. The thin side then becomes the heat problem: it reaches welding temperature long before the thick side does, and the arc has to be directed at the heavy section to get fusion without burning through the light one.
Fit-up has to give the arc a place to work. A square butt between a thin stainless sheet and a thick mild steel plate is difficult; a joint with a small land or a stepped edge is much easier, because the thick side can absorb heat while the thin side is brought in afterwards. Adding that step to the drawing costs nothing and saves a lot at the bench.
The two metals expand at different rates, so a dissimilar joint does not pull evenly. It pulls towards the side with the higher expansion, and it keeps pulling as the part cools. On a long seam, that asymmetry accumulates and the assembly curves.
Countering it means welding the seam in a sequence that balances the pull, and accepting that the part will need to be set slightly out of true before welding. Clamping the part straight and welding it is a reliable way to produce a part that springs when the clamps come off.
Visual inspection on a dissimilar steel joint is limited. The bead can look excellent while the hard zone underneath is wider than the procedure intended, and no amount of looking will reveal it. What a careful eye can catch is the set of symptoms that correlate with too much heat: a wide, flat bead, heavy undercut on the stainless side, colour that runs far back from the toe.
Beyond that, the useful checks are dimensional and destructive. A bend test on a coupon from the same setup will open a brittle fusion line quickly. A section through the joint, etched, will show the band and its width. Both are cheap compared with a failed bracket in service.
Dissimilar steel joints are usually specified by drawing and forgotten, which is why they fail quietly. These are the points worth pinning down.
We make these joints mostly on brackets, hangers and flanges that attach to a stainless system. The stainless work is on our TIG stainless steel welding pages, finished systems under stainless steel exhaust systems, and pipe sections under exhaust pipe work. Related assemblies are listed under exhaust pipe sections and fabricated components, and the titanium equivalent of the same discussion sits under titanium TIG welding parts.
