Off topic, but to answer your question JammedStab, there is flow separation on the upper side of the rudder, due in part to interference with the horizontal plane of the elevator. It is mildly corrected by wortex generators fitted on either side of the fin, close to the elevator.
Yet there was still some rudder aerodynamic fluctuations even after this addition, which may or may not be felt at the pedals, since the whole rudder system is on a spring tab design.
One solution to dampen these unwanted fluctuations was to encase a set of springs in a box, with a linkage from the rudder pressed between these two springs. The fluctuations were stopped, but one problem remains: if the system was locked in one position, the rudder travel would be greatly limited (the rudder cam box is about the size of a pack of cigarettes). If left as is, regular trimming of the spring tab would be acting against the springs in the dampening box. Therefore it needs to be moved (or displaced) when the rudder is being trimmed, and re-locked when the new trim position has been attained.
Hard to visualize the whole works without a picture, but this is the very reason why the rudder cam was designed. Very small cost, rather than redesigning the whole rudder-elevator assembly.
Although I do not have documentation to sustain this, I have worked for ATR as an instructor in Toulouse, and got this information first hand from the assembly line. There could be of course other reasons I am not aware of.
Hope this helps.