20kg Hexacopter in flight motor failure

Hi everyone,
on our recent flight with our 20kg H2-powered hexacopter we had a big incident. After 45 minutes of flight in windy conditions one of the motors apparently stopped working and we barely managed to crashland, damaging two propellers. It would be great if someone could help me understand what exactly happend and how to prevent that in the future.
We tested the motors and ESCs. All ESCs work fine, except one of the motors that is just dead.
Maybe someone can find an indication in the logs. Can you see the output power of each motor?
Weirdly we were not able to disarm right after the crash aswell…

As a sidequest, I wonder why the copter was not able to fly much more stable with the remaining 5 motors. Should it not be able to fly if one motor fails?

Many many thanks!
Fynn

Logs:
https://drive.google.com/drive/folders/1waWhL8WlIvSh9if2NrKbplwORDFYG8W8?usp=sharing

Hello Fynn, welcome to the community,

Was this vehicle configured using the ArduPilot Methodic Configurator ?

Hi amilcarlucas,
no it was not. But this was not the first flight of that vehicle. We already flew a couple of times before, including endurance tests with about 2h of flighttime. Maybe I should mention, that we have had some crashes due to pilot error before that could have damaged a ball bearing of a motor. But I dont think, that this was the case in this crash.

Which log file shows the motor failure? If you post a link to more than one identify which one is pertinent.

1 Like

Hi Dave,
all of the three logfiles show the same flight. Just different file types.
I would suggest the .bin file. The crash is at the very end of the file. (Minute 48).
Just look at the motor fail flag param. This is where the drone started to crash. :slight_smile:

It’s usually difficult to determine why a motor failed if that’s what happened. Unless you witnessed something, like a prop flying off from failure, all we know is there was a thrust loss. Do you have a log of a normal or just hovering flight? Because there is oscillation in the outputs that could potentially heat them up. But if it was windy it could just be that.

I see the notch filter hasn’t been configured and there are some default tuning related parameters. The vibration levels are way too high and there are clipping events. You know they are high with clipping events on a Flight Controller with isolated IMU’s.

1 Like

Is there any information about the individual motors and their poweroutputs / motor data?
The whole flight was basically hover and loiter on the spot. But I can check if I can find older logs from previous flights. We were flying on the coast directly behind the dike, so it was very gusty and windy.
Regarding the vibration, how do I know which values are acceptable and which values are not?
What exactly do you mean by clipping events? Data cut-outs?
So as I understand the vibration inside the pixhawk are so high, that the IMUs have delivered bad data from time to time? Should I “cushion” my Pixhawk better with sticky foam?

Many Thanks!
Fynn

This vehicle is not correctly configured. That increases the stress, wear-and-tear on the components. Use the software I posted above to correct this and save you money.

1 Like

No. Only the command signal to the ESC (RCout) and in cases with ESC telemetry Motor RPM. But even this RPM is not a direct measurment.

When the vibration levels exceed the full scale capability of the accelerometers.

This is the better of the 2 IMU’s. It’s terrible.

The Z-axis needs to be lowered to where the X and Y are.

That’s always a hot topic. Foam or a isolation mount perhaps but getting at the root cause is the best place to start. Standard stuff; balanced props, cables not tight to the Flight Controller, Nothing loose and moving around, etc. Or it could just be an overly compliant frame.

Hey amilcarlucas,
will definitely do that! Thanks. I will give you an update after our next flight. I hope that everything will look much better.

Thanks Dave!
The props we used, were a bit worn and not without a scratch. I bought a new set that we will put on for the next flight. Additionally I will put some foamy patch under the Pixhawk. At the moment it is just taped on.
Do you have any suggestion of how we could test the motors for damage in the ball bearings? Maybe through current draw or temperature?
Also I dont really understand, why the autopilot was not able to continiue flying level with just 5 motors. Is that not something a hexacopter should be able to?
In our case the copter just banked almost 60 degrees.

It depends on the thrust/weight. Yours looks like it should be able to handle that with a hover throttle output ~.25. Could be that poor tuning isn’t helping.

Hey Amilcarlucas,
I did went through your incredible configurator. I have to admid though that I am a bit overwhelmed. I used the X11 Plus Template for our 20kg Hexacopter. I am a bit scared to just go flying with the PIDs and settings that I just copied from the X11 Template. Any recommendations or suggestions of how to start with our first flight?
If we have a sort of stable flight, should we do the quick tune and auto tune regardless?
I will share the new param file in GDrive.
https://drive.google.com/drive/folders/1waWhL8WlIvSh9if2NrKbplwORDFYG8W8?usp=sharing
Many Thanks!
Fynn

You should not use the PID parameters from the X11 template directly, you should adapt them to your vehicle.

I recommend you follow all the steps as described, and fill the “reason changed” fields as good as possible. Doing so forces you to think!! The best tunes have good “reason changed” fields that can be used in the certification process later.

Yes, that is the idea.

Use the latest version of the software, more steps got automated now.

1 Like