We flew a roll control module on this flight and we expected to see the control program send the rocket through a set of programmed rolls while in between maintain an absolute heading. Instead we saw the rocket roll in an uncontrolled manner. From the on board video footage it became clear that when the canards were in the counterclockwise (CCW) control position the rocket actually rotated clockwise (CW) and visa versa. This is an effective control reversal and resulted in a loss of roll control.
Since the canards and flight computer both appeared to function as designed the control reversal must be the result of unexpected aerodynamic forces on the rocket.
The layout of the rocket as it flew:
- Video from the rocket: http://www.youtube.com/watch?v=vKw75YMiLA0
Where the canards are tilted they interfere with the downstream air flow. In particular it causes the air near the rocket to spiral in the opposite direction of the intended control. This is normal and, in fact, how control surfaces work. However this spiral stream of air may be interacting with the (much larger than the canards) fins at the rear of the rocket. If the spiral slipstream creates more lift on the fins (in the opposite direction of the intended control) than the lift on the canards then we would see a control reversal.
The canards are a 60° half delta-wing design. They produce most of their lift when a vortex forms along the leading edge of the canard when it is at a non-0 angle of attack. This vortex "sheds" off the end of the canard and continues to exist in the air stream for some time. These vortexes may be interacting with the fins to create a control reversal. There is a lot of documentation about this problem in missile design books.
This possibility has given rise to the 'Spin-Can' described here