Okay so I did a test on one of our test rigs- its a DJI 550 with a Pixhawk 2. We set up a motor so that we could make it fail in flight. The copter flew very nicely until we killed the motor (rear left) and the copter was unable to handle this outage. Now I dont have the specifics of the motor thrust available which I know is critical in keeping the aircraft up when a motor fails however, throttle at hover is approx 35% so there should be more than enough power to handle a motor failure. Any thoughts? Does the copter firmware deal with motor failures as well as dare I say it DJI?
Really weird, I have an octo, I hope it will do better in case of an emergency,
Could you publish your log?
Ok the power to one of the ESC on my octo was cut yesterday, and it flew like nothing happened
It doesn’t invoke a recovery algorithm if that’s what you mean.As far as I know that is.DJI bought the rights to an algorithm a while back and I’ve heard no more of open versions.
I would think that if the failure causes enough disturbance to the sensors there’s not much chance of recovery but a five motored hexacopter should be able to take off and fly.
Did you get a log of the flight ? That may be interesting.
Ok, flew on 7 motors again, but this time, if i had not cut the loiter I would have seen my octo crashed. I had to switch back to Alti to recover and land it, it was pretty hard.
So it’s a bit like russian roulette… 
Hello,
Bad luck or not.
I have both Hexa and Octo (QuadX) configurations.
Hexa is a bit overpowered and one motor failure is no fuss, it maintain altitude and continue auto navigation. I flew this frame with another set of motors (smaller) and one motor loss in loiter evolved in uncontrollable yaw then second motor loss and crash.
With my OctoQuadX, one motor loss is no problem regardless of flying mode (loiter or stabilize). It is an heavy one (10 pounds) with 2 to 1 power ratio (AltHold at around 50% throttle).
Marc
On my 20 pounds octo, the first failure wasn’t a problem in any mode.
But on the second one, it was like the FC wasn’t even listening to the IMU or GPS in loiter mode, going 2 m/s to the left without any input from me. With maximum roll to the right it was barely stationary.
I will find both log and publish them
Ben
Hi Jagger, can you provide a link to where you saw this information? I would like to confirm Ardupilot does not support motor redundancy.
Ardupilot support motor redundancy: Hexa or Octocopter configurations can fly with one motor off.
Marc
Nope.It’s about three years old information.A bunch of German univerity guys worked out an algorithm for gently rotating the aircraft in the event of a motor failure.Even a quadcopter.They posted it on this forum.Then they sold to the highest bidder,which is of course DJI. Arducopter does not have this,or any other ,algorithm specifically for dealing with a failed motor that I know of.Arducopter supports motor redundancy if you designed in enough power headroom (brute force method) and enough motors.
Yea if Ardupilot leaves the seized motor pegged full throttle, it doesn’t support it.
I guess I’m gonna have to add my own support, something I really didn’t wanna have to do.
Is there a dataflash log?
There is an issue here on the issues list to do with hexacopters losing a motor and some where in the middle of the discussion I’ve added some thoughts from Leonard and I about how Copter could do a better job. @Iketh has got the symptom right - the single motor pegged at full throttle while other motors are much lower is the area we need to improve. Basically we need to detect which motor has failed and then remove it from our saturation handling (aka “stability patch”) so that it doesn’t stop other motors from reaching their maximum.
It’s not super hard to do but I requires a developer to put in some decent amount of time especially on the testing.
That’s why I asked some time ago about DSHOT support, this is a way to discover faulty motor.
Yes it should be easy to implement. Would only need to be sure a copter flying at maximum airspeed doesn’t enter motor failure mode.
I think the simplest implementation that would be miles better than the current behavior is cut power to failed motor and it’s opposing motor and increase output to the 4 remaining by a percentage derived from MOT_THST_HOVER. It would probably fly fairly well just doing that, but it could be twitchy when the copter needs rotating along the axis that the 4 running motors govern, and sluggish the opposite. So you can further tweak the outputs based on which axis the copter needs adjustments.
I can do it if code is written in Visual Studio. Otherwise, I will just test.
Hi,
Please consider the motor failure in Autonomous missions.
I losted my drone two days ago and just explained how here
Hopefully this huge issue is in the plan to be sorted out in the next FW.
Cheers,
Kirk
I think politics will prevent this from happening. Big companies can prevent a feature like this to force prebuilt system purchases.
I had a hexa weight 4500gr
In this video you can see the front left prop break and the behaviour of the copter
The copter keep it altitude and behaviour until i yaw
My next build was a coaxial x8
I had again a prop failure in the middle of a auto mission
The x8 keep the auto mission like nothing ever happend
Had my first motor failure finally. Check out how ArduPilot handled it. ESC problem, not motor seizure.