I had compared the speed of reaction on two rods balanced on a pivot at their mid-point. One rod was two foot long and the other was six foot long. In each case the same weights were attached, one at the end of the rod, the other, two inches inwards towards the centre of rotation, at the other end. The torque was the same in each case but the shorter rod accelerated into its unbalanced position much quicker. This suggested to me that, with the limited range of movement available from my mechanism, it would be react to imbalance more quickly if I placed the weight as close as possible to the axle.