I had never thought much about the geometry, but now it all makes sense. The bend and suspension points are located where the zero nodes of a freely vibrating beam are. That should also lead to the answer of tight vs. larger bend. A larger bend zone sort of restricts the vibration at both sides of the zero node. I knew about the 0.22 factor before. So, when I have a stack of newly cut wooden boards I place two supports at 0.22 times length from the end. That way the natural deflection is identical at both ends and at the center, and the wood dries fairly straight.
Cal's point of the ring not working well: I think there are two issues. One, he may not have it located at the zero point. The other, if it is off the zero point there is vibratory movement at the ring support point, and the friction dampens the motion. A string has less dampening effect.
Personally, I'd harden the whole thing. Drop a hardened chisel on the floor vs. a soft iron one. The chisel rings.
Dietrich