Argh! All the tweaking I did to get the Manta flight model working? Wasted.
Turns out that the Newton physics has quite a large default damping factor which has been affecting the calculations in the aerodynamics model. I thought it was proving a bit hard to tune - basically the drag calcs were getting swamped by the damping, and now the Manta flies like a squashed banana.
Ah well, should be easier to tweak and more stable this time round, with things actually working as expected :)