Here is another counterintuitive example and before anyone writes, yes the endmill is a long L:D. The zero projection ER collet chuck (blue) had the lowest performance of the three toolholders tested. A couple things are going on here. With the toolholder ending at the gage line, the assembly did not allow much damping resulting in slower speed stable lobes. Even though the "beam" created by the second ER collet chuck (red) was longer it had more mass and damping. The longest chuck (purple) had the second best performance. In addition, it was discovered that the connection stiffness of the zero projection holder was less than the other two because the spanner nut and wrench didn't allow as much torque to be applied as with the conventional nut. The collet acts like a spring, the higher the torque the greater the grip. The zero projection holder performed better with smaller diameter and with shorter tools.
So many variables to consider!
Sharing information about high performance milling technologies, the result of 30 years of research.