I’ve 15 years experience building and flying R/C helis: I’m no expert – but I know lots about analysing fatal crashes . . . At about the 60th flight with my Phantom 2 I had my first senior moment – a minor argument with a small tree, and lost about 15mm from tip of one prop. Keen to investigate the possible consequences of vibration on flight, I checked the machine over, recalibrated the compass, acquired all satellites, confirmed Home, started motors with CSC, and lifted into stable hover. Within seconds the drone performed an unauthorised high-speed climb to perhaps twice the height I’d ever been happy to fly – possibly 500 ft. I had no control - the wind was gusting at about 15mph. The drone wandered a little at this height for about 30 seconds, then very slowly descended towards Home Point. At about 10 feet it repeated the process – went very high, wandered a bit for 30 seconds, then slowly descended again. I experimented switching between GPS/Att/RTH and seemed to have some minor control forward and back, but no throttle control. It landed automatically a little untidily and safely near the Home Point. Back home I took everything apart and checked all the internals, then connected to my PC and checked all the parameters. They were all normal, but I did a Basic re-calibration of the IMU anyway. I decided to investigate further. I tethered the Phantom to the ground by a 12-foot nylon cord tied to a heavy weight, and it behaved perfectly when fitted with undamaged props. However, when I re-fitted the damaged prop, within seconds of lift-off it climbed rapidly until the tether became taught – after which it’s response to inputs was bizarre (as expected while tethered). I repeated this experiment twice with a good and a damaged prop – same result. In the interests of science, and with the drone anchored firmly to the floor, I then removed all props except the damaged one, and spun it up into idle speed. I was surprised by the magnitude of vibration – the frame shook, and it felt like a small washing machine on spin. So, I seem to have proved that an unbalanced prop causes vibration sufficient to interfere with the IMU. I suppose that the gyro/accelerometers get confused. The event seems to suggest that the Phantom went into a corrupted Return to Home procedure, getting the ascent and descent phases at the wrong speeds, rising to an illegal height, and pausing at this height for much longer than usual. A colleague video’d the experimental tethered flights, and I could post this if anyone needed it, but it doesn’t add much to my account here. So, I wonder if some Flyaways might be caused by prop damage or other imbalances. Often the drone is lost or damaged beyond analysing the cause – I was lucky. I’ve not read any accounts of similar experiences, and would be interested in other’s views. Perhaps someone braver (or richer) than me might like to experiment with tethering and with props of different degrees of damage. (Be warned – tethering is risky if the drone misbehaves – wear gloves) At Inquests into the Death of a Phantom, the cause of Flyaways is mostly conjecture: pilot error, faulty batteries, radio interference, a woman scorned etc. Perhaps DJI should investigate vibration.