hi Rob,
thanks for reminding of the nice patches-a-day.
I had come up with a similar solution using force matrix, but this central torque force was something so similar to the central_force that I thought there could be a simpler (and lighter) way to handle it...
A small detail yet (that may actually make the case not so similar) is that computing the orthogonal vector to use it as the force causes divergence if damping is not equal to 1. (see patcher below)
I imagine the damping is acting like a kind of smoothing filter on force matrix, which results to this loose steering.
Is there a way to know the relation to compensate it ?