Motor Failure
« on: March 28, 2018, 02:37:15 am »
I am using Revo (w/ Black Rhino 16.09) for flying my hexacopter. I was wondering, would the FC still be able to stabilize the vehicle in event of a rotor failure.

I understand the dynamics of it, and inherently the vehicle is destabilized because of imbalance in moments, but the question is, does the Librepilot FC have the capability to still control it.

Another question : Is there a way to disable motors in the code / firmware, based on flip of a switch from Tx ?

Thank you   

Re: Motor Failure
« Reply #1 on: March 28, 2018, 03:39:42 am »
My thinking of hexacopter theory says these things (I may be wrong):
- The hex needs to be powerful enough to fly on 4 motors.  It will be using 4 motors in hover.  The motor opposite the failed motor will almost be turned off by stabilization during hover.
- The yaw will work in one direction, but not the other direction.  The best it can do is avoid rotating in the bad direction.  Better yet to try to rotate slowly in the good direction while changing roll and pitch to get motion in a line.
- If you want full redundancy without a fancy mode that spins yaw constantly, you need an octocopter.

I don't know of anyone that has tested single motor failure on a hex with LP.

Testing to make a motor fail in flight would require source code changes or a simple on/off switch to switch of the signal to a motor.  It would be better to just remove one prop and try to take off to knee high.

Re: Motor Failure
« Reply #2 on: March 28, 2018, 03:50:21 am »
Thanks TheOtherCliff for prompt response. I tried taking of with 5 rotors , but  the hex didnt lift at all  :P  and that asserted that I dont have enough lift capacity.

Using only 4 rotors would be simplest strategy, however, would result in reduced lift capacity. I had read some papers on using 5 rotors in event of single rotor failure and was curious if it was thought of by LP developers.

Also, if I intend to fail it in air, where would you recommend, I amend in the firmware.

Thanking you.

Re: Motor Failure
« Reply #3 on: March 28, 2018, 05:16:08 pm »
You can't disable two motors (opposite points of the hex) without first disabling the yaw mixer and being willing to fly it with yaw that goes crazy on either pitch or roll.  This is because you will have e.g. CW CW CCW CCW motor rotations where a quad has CW CCW CW CCW.  E.g. pitch will decrease BOTH the CW motors and increase BOTH the CCW motors, making a lot of yaw which it will try to stabliize but can't (depending on yaw direction).

Sometimes people have a quad that looses a motor, with the effect that all motors seem to turn off so the user does not imagine a problem with just one motor.  The reason is that the stabilization converges toward that solution (all motors stopped) as the one that is most stable.  Your hex should effectively disable the opposite motor during a stable hover, but as soon as the wrong yaw is requested (at all) the solution is toward zeroing all the motors.  That may be keeping you from lifting off.  It may need a medium amount of correct yaw direction to hover or lift off.  The correct yaw direction is one where there are three live motors driving yaw (imagine only those three motors spinning, you can have a level hover with yaw spinning like a top).  The wrong yaw direction is the one where one of the 3 motors driving yaw in that direction is non-functional.

You should be able to get a hover somewhere between CW X CW CCW X CCW (level hover with no yaw, 2nd and 5th motors stopped) and CW X CW X CW X (level hover spinning like a top, 2nd, 4th, 6th motors stopped).  Note that the second motor (A CCW) is never used in this continuum, that is the bad motor.  In the middle of this continuum, small corrections to roll and pitch (and yaw rate) should be possible.

Be aware that the gyros are only good up to 2000 degrees per second (about 5.56 revolutions per second) and anything faster than that (or close to that) may have problems.

To be useful, loosing a motor on a hex would have to enter a new mode where:
- the hex spins fairly fast in the correct direction
- it either auto-lands, or creates a virtual fore and aft so that e.g. pitch stick moves in a single direction (e.g. virtual forward) even though the hex is spinning.

That sounds like an interesting project for a developer that has a hex.  It makes me wonder how fast a quad with two motors would spin (yaw) and if some rudders would reduce the rate to below 5 RPS.

Re: Motor Failure
« Reply #4 on: March 28, 2018, 10:14:51 pm »
Here are two research videos you might find it interesting

Re: Motor Failure
« Reply #5 on: March 28, 2018, 10:27:25 pm »
Yea, I've seen TED talks on quads with one prop cut off.  :)  That's the idea with the virtual fore and aft.