A hanger that has separated at the weld, leaving the rod on the rubber and the bracket on the pipe, was not usually torn off. It cracked. A torn hanger is an overload; a cracked hanger is a fatigue failure, and the two are fixed in opposite directions. Overload wants more strength. Fatigue wants less restraint.
The give-away is the fracture face. A fatigue crack starts at one point and spreads slowly, leaving a smooth, often stained patch where the faces fretted, and finishes with a rough granular zone where the remaining section tore. If the whole face is rough and bright, it went in one event.
A hanger rod welded to a can sees vibration at every frequency the engine and the road produce, and never gets a rest. The weld toe is a geometric discontinuity: the load path turns a corner there, and the section changes from rod to weld to thin shell within a short distance. Stress concentrates there, so that is where a crack starts.
Welding also leaves the zone beside the toe in a different condition from the parent metal, and on thin sheet the weld can pull the shell into a slight dish that puts the hanger under standing load before the car has moved. A standing load plus vibration is the classic recipe for a crack in a place that looks sound.
An exhaust grows when it is hot. If every hanger holds the system rigidly, that growth has nowhere to go and is taken up as stress at the stiffest points, which are the welds. The system needs somewhere to move, and the hangers must allow movement in the direction it grows while resisting the directions that would let it swing.
This is why hanger failures often appear after a system has been added to rather than when it was first built. A new section or a repair that replaced a flexible mount with a solid one can remove the last bit of compliance in a run, and the load finds the next weakest weld. The weld did not weaken; it became the only thing left that could move.
Corrosion at the weld is one, and it looks similar until the part is cleaned. A crevice between rod and shell holds moisture, and stainless in a crevice with chlorides can pit beside the seam. That shows as wastage along the joint rather than a clean fatigue face, with the metal around it discoloured.
Poor fusion is another. A hanger tacked rather than welded, or welded with the arc held on the rod and only washed at the shell, has a joint mostly sitting on the surface. It can hold a long time and then peel. A hanger fitted under preload starts life carrying a load it was not meant to see.
Clean the break and look at it in good light. Note whether the crack began at the toe on the rod side, in the shell beside the weld, or at the end of a partial weld. Compare the smooth zone with the rough one, which tells you how long it was growing, and look for corrosion product inside the crack.
Then check the run. With the car cold, see whether the system moves at all by hand in the direction it grows, or whether it is held solid. Check that each rubber works in shear rather than hanging in tension, and that no hanger carries the weight of a stainless steel muffler with its rubber at the end of its travel.
If the design is sound and the weld was the problem, the fix is a better joint: a full weld around the rod, a saddle or gusset that spreads load into the shell over a longer distance, and a rod that meets the shell squarely instead of standing off it. Fatigue cares about the stress at the toe, not the total load.
If the run is over-restrained, no weld will survive it. Leave one hanger locating the system and let the others carry weight without locking movement, and check that the joints in the run can slide or flex as the pipe grows. Good TIG stainless welding is only half the answer; the other half is the freedom the system gives the joint.
