I haven't been in an R22 / R44 cabin for quite some years (as an enthusiast), but all this discussion made me curious why the Robinson T-bar cyclic isn't well liked. It seems the main issue is that the cyclic grip of the non-flying pilot is high above their knees when the grip of the flying pilot is at knee level during training flights:
In this video of Maria Langer practicing autos and CFI Trevor Hale monitoring, I see that during the autos, he always raises his hand to be ready to grip the cyclic quickly if anything starts to go wrong (see from 7:00 onwards and then again for each subsequent practice auto). But holding arm up would become tiring if doing so for a whole day while flying with less experienced pilots in control than in this case:
Aside from the STC modification images shared by hargreaves99, wouldn't it have been fairly straightforward to have a T-bar cyclic that is actual a "T" shape rather than the "Y" of the R22/44? That could at least have been done for any machines routinely used for training. Even an adjustable "T-bar" cross-bar wouldn't add much weight or complexity if further height adjustment was needed for long or short legs on either side. Then the grip could be at knee level for both pilots. The grip would still be able to flop up and down (the extra DOF that Wide_Moth_Frog noted), but is that an issue once the arms of both pilots can comfortably rest on their knees?