Originally Posted by
qcpmsame
I think it was the cage catching in the spokes (what reason this happened I couldn't say,) and it forced the derailleur onto its stop and broke the piece off.
Well, the thing is, is that a spoke catching the cage would have pulled it
backwards on the bottom, clockwise viewed from the DS. The B screw which broke off is there to prevent the parallelogram from moving more
forward, counterclockwise from the DS. So a spoke catching the cage would have pulled it away from the stop, and that wouldn't have broken it. It broke from something pulling it forward. The only thing that would do that is the tension from the cage spring, hardly significant, or chain suck. Or an exiting fracture in the aluminum.
That sheared section seems to have two different surface textures. But that still doesn't mean it broke spontaneously from the tension of the cage spring or from fatigue. It should have been stronger than that.
I think it is pretty clear, especially given the lock-up I felt a few moments before.
1. That lockup was due to chain suck. It pulled the parallelogram forward hard enough to break off the B screw mount. This let the entire parallelogram swing forward more than normal.
2. I was on the small ring. The chain suck let go when I stopped to check what happened. With a slack chain and the parallelogram too far forward the GT cage swung much further back than was normal, so far back that the tension pulley was directly behind the cage pivot. With the pulley so far back the chain would be nearly parallel to the cage. Any shortening of the chain would pull the cage along whatever angle (viewed from the rear) happened to be the greatest. Any lateral out-of-trueness of the cage would encourage the cage to bend sideways instead of winding up the cage spring. This is like what happens when you remove the rear wheel and the RD cage springs backwards. Or when you shift to the little/little combination on a triple with a long chain and a RD that can't quite accommodate that many links. You have to help it out to get it back on a downward arc.
3. As long as I pedaled without shifting it could have continued. But when I shifted to the big ring after cresting the hill the chain bent the cage sideways to the bike's center line and forward instead of pulling it down and forward. This moved the end of the cage toward the spoke until one of them caught.
4 Cue the theme music...
Maybe a stronger B screw wouldn't have broken off, I dunno'. The force of even a tiny 160lb guy like myself mashing a 35" gear into a screw through a lever arm that big is bound to break something.