See other failures under 000.html.

Land Shark brazed quill stem failure near joint

Land Shark brazed steel quill stem

Vertical “quill” tube failed where it joins the forward “extension” tube.

No injuries.

More than 5 years regular use including trails. Rider 70+kgs. Used on two bicycles, both with rigid forks.

Discovered by inspection during other work. There were not noises from the stem that would have suggested a problem.

Some observations:

The rider's weight on the handlebars pushes down at the handlebars. In turn, the top of the extension tube pulls away from the quill, while the bottom of the extension pushes in on the quill. Most “pushes in” load is where the lowest part of the extension tube meets the most-forward part of the quill tube. In other words, the load is concentrated, not evenly spread along the joint.
The “pushes in” load is cyclic: every bump causes the load to increase, then reduce.
Cyclic loads can cause fatigue. “Fatigue” is a way of saying a big one-time load is needed to break it, but a smaller load which is repeated can also cause it to break. Fatigue strength is often affected by different things than for one-time load strength. For example, fatigue often occurs where there are shape changes, with some of it stiff and some of it flexible.
The extension tube's pushing is along its length, which is fairly stiff. It meets the quill tube, where it tries to “dent” the quill. A tube flexes more under a denting (sideways) load than under length (axial) load.
The outside of the brazed joint is radiused, which should help spread the load over a larger area, and should also make the transition from stiff to flexible more gradual. However, the inside of the joint probably is not radiused, so does not further spread the load. Also, heat from brazing somewhat weakens the metal near the brazing.
There is more damage than shows on the outside: internally, the quill tube has a large cracked area behind the extension tube. That area is also significantly offset from the rest of the quill.

Some analysis and speculation:

It seems likely the repeated load cracked quill on the forward face, below the radius. Once a small area was cracked, it was no longer carrying load. That shifted the load towards the sides, which then also cracked.
This quill tube is thin walled, which makes it both weaker and more flexible than a thicker tube. More quill flexibility exagerates the difference in stiffness between extension and quill. It may be that the same design with a slightly thicker quill wall would be much more durable.