Does 3.6 or any older firmware support motor loss in flight?
From a long time ago,
With 3 motors - 1 : No
With 4 motors -1 : No
With 5, 6, 8 or more: if enough remaining power it is OK, sometimes Yaw control is bad.
There is no active motor loss functionality. The more motors you have the greater the chance it’s survivable manually.
Ok, I read somewhere on this site that one person did an algorithm that was sold to DJI and thats why this system dont support it anymore. When 3.1.5 came it was told to have support for motor loss.
Thanks for your anwsers! =) 8 it should be in the future.
Anyway… I have never lost an ESC, motor or propeller.
Here is a video of a quad flying after loosing two motors. Of course you have to be lucky and the two you loose must be on diagonal corners. But certainly if you loose one motor you always have two that are diagonally opposite.
I don’t know if I like this guy’s solution but it is good enough to support “auto land” or maybe whatever is programmed for failsafe.
I’ve lost a prop on betaflight, over a lake!!!, and still managed to cross the lake and land. It was a shakey sideways flight, but I did it.
There was a research paper published in 2014: “Stability and control of a quadrocopter despite the complete loss of one, two, or three propellers” by Mark W. Mueller and Raffaello D’Andrea from Institute for Dynamic Systems and Control, ETH Zurich. If you Google the paper you can get a copy and also look at videos of a device under test. They were able to control a quadrotor with one or more missing propellers by allowing the entire vechicle to wobble (spin) around the center of the working prop(s). I am not sure if anyone has published this algorithm as usable code but if you’re a math whiz the paper might have sufficient info to duplicate the results.
As @mlebret says, AP will handle motor loss on vehicles with 6 or more motors as long as the vehicle has enough power. It doesn’t use a specific algorithm to do it though, it’s a natural outcome of the PID controllers reacting to the vehicle’s loss of control - they respond by increasing the power to the other motors to compensate.
There has been some work by Leonard Hall (our controls expert) recently to improve support for motor loss but it comes down to a prioritisation decision. With a lack of power, either altitude or yaw control must be sacrificed so Leonard’s patch sacrifices yaw.
The comment you mentioned about selling an algorithm to DJI leading to support being removed from AP must be a misunderstanding or just a made up story… nothing even remotely likely that has happened in AP’s history.
Thanks for the good answers. =) Hexacopter should work as good as it can. That works for us. Next heavylifter will be an octa for sure. We always try to have the setup work with 50% throttle at hover, or a little less.
Yaw is better to sacrifice as it gives the availability for more lift and lets it / us land smoother, in my world. =) Maybe Stabilize SS could work to control it somehow. RTL seems to work. Would love to test this some day on one of the hexas. Servo that switches off two cables to a motor or something like that.
The DJI thing… I read it here on the forum “DJI bought the rights to an algorithm a while back and I’ve heard no more of open versions.” Source: Motor failure- to crash or not? When I now read it again, I see I missed some replies. Like “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.”. Sorry about that! It wasnt ArduPilot.
Were flying a 3DR X-8 today (octacopter). One motorn stuttered during first arming. Touched the cables with magic hands and then it worked. Okay, will do. Not the best, but hey, it should handle a motorloss. Dont care! Old copter. We. Fly.
The aim was to test autotune for the first time. 3DR Pixhawk with APM:Copter 3.5.7 and using Position Hold with Autotune. We took the copter out to the windless clearcut woods and it did the autotune very nice. Roll-twitching, Pitch-twitching and then Yaw-twitching. Suddenly it started acting a little weird, flying around and leaning and doing yaws. Got a little nervous. But then we saw that motor had stopped turning. Nice! It continued in a stable manner with autotune until it was done, with 7 motors. Then I flew it back for landing and it was very flyable. Yaw obviously worked a little bit sluggish to one direction. And the arm was hanging some degrees now and then. Thumbs up!
The motor that had stopped working was hot (maybe only 2 cables was okay). The other motor on the same arm was also hot (it had to lift that arm by itself. On the other side, one motor was cold and the other was hot. The four other motors were normal warm.
Cool! Now I want to test a hexacopter… and build a octacopter!