A stained seam on a stainless exhaust gets reported as a welding fault, and often the weld is not where the problem is. Look at where the brown tracking sits. It usually runs as a band just outside the fusion line, parallel to the bead, on both sides, while the bead itself stays bright. That offset is the clue worth following.
The metal that failed is the metal that got hot but did not melt. It is the heat-affected zone, and the change that ruins it is called sensitisation. The word sounds academic for a workshop, but it describes something a fabricator causes and can therefore avoid.
Stainless is not stainless because of a bulk property. It is stainless because chromium at the surface reacts with oxygen and forms a very thin oxide film that repairs itself. That film is the whole barrier between the steel and the weather.
So the corrosion resistance of a grade depends on chromium staying in solid solution, free to reach the surface when the film is damaged. Anything that locks chromium into a compound inside the metal takes it out of circulation. Sensitisation is exactly that, and it happens in a band you can see if you know what to look for.
Austenitic grades carry carbon in solution. At the temperatures a weld produces in the surrounding metal, carbon can move, and it has a strong affinity for chromium. It migrates to the grain boundaries and forms chromium carbide there, and the carbide holds far more chromium than the neighbouring metal can spare.
The result is a narrow depleted zone along each grain boundary, where chromium has fallen below what the passive film needs. The boundaries become the weak network, so attack follows them. That is why the damage reads as a line or a web rather than a general dulling, and in thin tube the network runs right through the section.
Precipitation is diffusion, and diffusion is governed by time as much as by temperature. A brief excursion may precipitate almost nothing. A longer dwell at a lower temperature inside the same range can precipitate plenty. A fast, controlled weld often outlasts a cautious-looking one with a wide weave and a long pause at the crater.
The damage is also not confined to the visible bead. The band reaches as far as the metal stayed warm long enough, which on thin tube can be wider than expected. Keeping travel speed up, keeping the arc tight and not parking at the end are metallurgical decisions, whether or not anyone in the shop calls them that.
Because the mechanism needs carbon, the obvious lever is carbon. Low-carbon variants of the austenitic grades exist for this reason: with less carbon available there is less to form carbides, and chromium stays in solution. Stabilised grades take another route, alloying with elements that combine with carbon in preference to chromium.
In exhaust work, grade choice usually follows availability, formability and price rather than a sensitisation calculation, and for most systems that is a fair trade. What matters is that the risk is not uniform. A grade that suits a cool rear section may be a poor choice at a hot, restrained joint near the manifold, where dwell times are longer and a depleted band costs more.
Heat input is the main variable within reach. Lower current, faster travel and a smaller weave all shrink the volume of metal held in the susceptible range and shorten the time it spends there. Fit-up matters for the same reason: a tight joint allows one pass where a wide gap demands several, and every extra pass re-heats the band the last one made.
Back purge belongs in the same conversation. A clean internal surface holds its passive film far better than a sugared one, and the inside of an exhaust joint is the side that sits in gas and condensate. The outside can look immaculate while the root quietly fails. The assemblies this applies to are the kind shown under TIG stainless steel welding.
When a seam rusts, the useful question is not whether the weld was good but what the thermal history was. Rust tracking outside the fusion line points to a depleted band. Rust sitting in the bead points somewhere else: porosity, an unpurged root, contamination from handling. Rust as scattered spots across a whole panel points to chlorides from outside the part.
That distinction changes what a conversation about a failed component should be about. A sensitised seam is a process outcome, and the process belongs to the shop. Pitting on an unwelded surface is an attack from the environment, and no amount of welding skill would have prevented it.
Proportion helps here. Sensitisation needs a susceptible grade, a thermal history in the right window and an environment aggressive enough to exploit the depleted band. Many systems serve out their life with nothing visible, because the wall is thick enough to tolerate some boundary attack or the climate never turns hostile.
The risk concentrates where those conditions overlap: thin walls, hot locations, repeated cycling, restrained joints, wet or salty service. A coastal car with a thin hot-side section is the textbook case. The same weld on a dry inland car with a heavier wall may never show a mark, which is why reports of this failure seem so inconsistent.
If corrosion life matters more than the lowest piece price, these points are worth putting in writing rather than assuming.
None of this requires a laboratory on the shop floor. It requires the shop to know which of its habits create a depleted band and to be able to say so plainly. We build to this reasoning across our stainless steel exhaust system range, and the grade differences behind it are set out under shop by material. For hot-side work, our exhaust pipe stock and stainless steel exhaust mufflers show where these choices land in practice, and our titanium fabrication components show the same questions answered in a different alloy.
