This was inspired by @John Hestness's comment in my last blog:
"Can a pentacopter survive the loss of one motor?"
I thought it sounded like a very interesting idea so I tried it during the weekend. Short answer: it depends on which motor fails. In some cases, a pentacopter is still fully controllable after losing one motor.
Long answer: since a) we cannot alter the spinning direction of motors during the flight, and b) a pentacopter has 2 CW and 3 CCW rotors (or vice versa), if one of the CW motors fails, we end up having 1 CW only but 3 CCW rotors, and after doing the math I conclude there is no way such a copter can still balance itself.
However, if one of the CCW motors is not functional, we are left with 2 CW and CCW motors. That sounds like a quadcopter and indeed it can still fly.
I turned one motor off at roughly 00:08 and 00:25 (you will notice the propeller stopped spinning if you look closely), and switched the controller to control the resulting nonsymmetric 'quadcopter'. It still flew reasonably well.
I have to mention that this is still slightly different from what we expect to see in a copter that can be called "motor fault tolerated": ideally, instead of telling the copter to turn off a rotor, we want to auto-detect the motor failure maybe using ESC, or evening infer that from IMU data, but still, this video showed the potential and I hope it could be helpful for someone in this community.
As usual, everything is open-source: