I finally finished the hull of my spaceship, same as the old one ( but bigger and MUCH heavier)
then I decided to run some math.
I solved the problem for rolling around the main ship axis, with 2 engine blocks, of 2 powerful hydrogen thrusters each, floating free. ( so a 2 arm problem with massless arms)
the answer is that rc= sqrt(Is/K) where K is the combined mass of all the engines. 20*4=80.
with a bit ( quite a bit) of exageration in calculating Is ( I assumed it was a cylinder), the result is 36.
considering that I exagerated Is and that I didn't consider arms nor power sources and engine case, the true result is going to be lower.
still likely to be a relatively large number. not even close to what rotation around the other 2 axis would be however. in this kind of design, Is on any axis that is not that of ship's length is going to be utterly massive. I predict that for every reasonable value of r ( reasonable being any distance at which you can still attach stuff to your ship), turning rate is going to increase the farther you go. rolling rate would however decrease.
this leads me to believe that it would be optimal to have rolling engines and turning engines in separate places along the axis.
first you would build the turning engines, with their arms. then calculate the new Is, considering turning engines and arms too, and add rotational thrusters at the optimal point.
of course, all the stuff above refers to my style of ship. Shooer's recurve assault ship would be handled differently for example.
and in general smaller ships have optimal points closer to the axis.
small note: earleir I considered Ia ( moment of inertia of arms) to be function of r^2. that, however, is not true. since mass increases with length as well, Ia grows with the cube of distance.