I've seen someone lecture another poster for not "doing as the manual states" and unfolding the props before flight. I've read it in the manual, but if I recall correctly, even some of DJI's own tutorial/training videos show the MP being started with folded blades. I've yet to manually unfold them, and I've yet to see any issues.
I realize this is a dated thread and perhaps I'm being a bit redundant but it may help someone. I have never unfolded props, and until this morning, have never had an issue. I've used stock props, as well as carbon fiber props. This morning I fired it up with the carbon fiber props folded as I have on the last 50 or so flights, but this time all hell broke lose.
Apparently the props, when folded, can perhaps be over overlapped to an extent that the centrifugal force isn't enough to "unhook" the blades from each other. What ever the reason, the resulting out-of-balance issue caused the Mavic oscillate violently, skidding off my takeoff pad. It eventually slammed into a nearby chair, ripping two of the blades completely apart. I had the sticks in emergency shutdown position immediately, but the actual shutdown took way too long to avoid catastrophe. .
Fortunately there was no structural damage to the air frame or motors. In fact, I replaced all four props with stock and went for a test flight. No subsequent issues.
I feel the props can and will unfold automatically, providing they aren't overlapped. The carbon fiber blades are very rigid, therefore they can overlapped enough that the opposing force created between the overlapping blades is stronger than the centrifugal force needed to sling them into flight position.
IMHO, unfold
Update:
Lots of discussions regarding FW updating. Perhaps a FW to sense a prop stuck in unfolded position would be nice. It Could auto shutdown should a prop not unfold during startup.causing abdominal vibrations.
To have the system have quick look at the accelerameters for excessive and or erratic output caused by violent vibration shouldn't be all that complex.
Even in the unlikely even two opposing props failed, thus countering the vibration (very unlikely) that would result from just one prop failure, the attitude would still be reflected in the asselerometors During lift off in an unstable attitude
I guessing a few milliseconds would be all the time needed to detect an unfolded prop and initiate a shutdown. Much faster that what I was able to do using emergency shutdown. By then it was way to late to avoid making shrapnel our of my carbon fiber props.