I point out that if I did not turn the wheel just after having blocked the throttle up, the voltage on the 3 wires to 0V DECREASE automatically and I am forced to manipulate the joystick from maximum to minimum to maximum, to continue action (Is this normal?)
Greg, I would say it's probably quite normal for the power to cut when the wheel does not move soon after the throttle applied, this would prevent damage to the motor windings, and also stop the controller from being overloaded if the torque required to turn the wheel is too much.
Your hall sensors do seem to be working fine, so it looks as if the controller itself may have been damaged, probably by too much voltage.
I haven't seen inside the regen controller, but I'm guessing it has six
FETs (but it could be more) which switch the power to the phase leads inside the controller.
Each of the phase leads is either switched high
(battery +) or low
(battery -) dependant upon the signal from the hall sensor.
Your first test (and reverse enabled test) on the phase leads would seem to indicate that the "High" switching
FETs appear to be working as expected.
What I was hoping to have seen with the last test was a similar result to the first test, proving that the "Low" switching
FETs were operating correctly as well.
It may not necessarily mean that the "Low" switching
FETs are faulty, as a different result may have been obtained if a resistive load was placed between battery + and the phase output being tested.
Unfortunately Greg, it would appear that your controller has failed, and I'm guessing that it probably won't be feasible to repair it.
A replacement controller is going to be the most straightforward solution.
Alan 
