Any joint that sees reversing load should be welded. A brazed joint in a thin gap has limited fatigue resistance, and a part that flexes slightly with every engine cycle will eventually open that gap. The failure is progressive and often invisible until it is complete.
Brackets, hangers, stays and anything carrying the weight of the system belong in this category. If the joint sits in the load path between the exhaust and the chassis, brazing is the wrong process regardless of how convenient it is on the bench.
Brazing filler softens as temperature rises, and a joint that is strong at room temperature can creep under sustained heat. On a manifold, a downpipe or any part close to the turbine, the service temperature is high enough that this becomes the governing limitation.
Creep does not announce itself. The joint holds, then holds with a slight movement, then opens. By the time a leak appears the filler has already deformed, and no amount of re-brazing will restore a joint that was never suited to the temperature.
A brazed joint cannot be inspected the way a weld can. There is no cap profile to read and no penetration to check, and the only reliable test is destructive. Where a part must be documented as sound, that limitation is disqualifying.
Radiography and similar methods can reveal voids in a brazed joint, but interpreting them requires a baseline most workshops do not have. A fusion weld at least offers a surface a trained eye can judge, which is why it remains the default for anything needing evidence.
Brazing over an old brazed joint rarely works. The original filler remelts or at least softens, the new filler mixes with it, and the result is a joint of unknown composition with no defined flow behaviour. Each successive repair makes the next one less predictable.
The same applies to a brazed patch over a welded repair. Two processes in one location means two heat-affected zones and two failure modes. Once a joint has been reworked more than once, the honest answer is replacement.
There is a commercial dimension too. A buyer ordering a fabricated part expects a welded assembly, and a brazed joint discovered later reads as a shortcut even when it was technically defensible. That perception is a real cost.
It also affects serviceability. A workshop that receives a brazed part without knowing it may attempt a weld repair and destroy the joint. Clear specification avoids that, which is why TIG stainless steel welding is named explicitly on any part where a weld is expected.
Fusion is not automatically better, but it is better wherever strength, temperature resistance and inspectability all matter at once. That covers most of what an exhaust or intake assembly has to do, so the decision is usually made before anyone reaches for a torch.
The useful rule is to reserve brazing for joints that are lightly loaded, cool, cleanable and non-structural. When a joint fails any of those tests the decision is already made, and our fabricated components are specified process by process for exactly that reason.
