I’ve just finished building an octodrone. I’ve built a quad before, so I’m not a complete beginner, but I’m still learning. The issue is that during takeoff, the drone starts to oscillate dramatically until it crashes. The flight mode is set to Stabilize.
I’ve already checked the motor spin, the propeller direction, and the balance of the drone, and everything seems to be as it should.
I’ve attached the log and a picture of the drone. Any help would be greatly appreciated 12 12-21-2024 1-11-46 PM.bin (888 KB)
Also, the drone has 4 motors upright and the others inverted so the props dont hit each other and I inverted the props and the spin is already taken into consideration as the following picture
Hey, I’m a beginner aswell, but I thought I might add my 2 cents.
You seem to be using MOT_PWM_TYPE 0, but I don’t believe you’ve done an ESC calibration? Try using DSHOT If your ESC supports it, Since you have a large drone try DSHOT150, If you’re using an AM32, then those ESC’s do not support DSHOT150, so go with DSHOT300 instead.
It seems your setup also has a high thrust to weight ratio which may cause issues perhaps adding artificial weight might be worth a shot? Other than that I’d probably say try reducing the
ATC_RAT_PIT_P/ATC_RAT_PIT_I and ATC_RAT_RLL_P/ATC_RAT_RLL_I parameters.
Thank you for your support! I’ll definitely reduce the PID by 50%. When looking for reasons, I didn’t find that link. really helpful! This is the first time I’ve heard about DSHOT. What is it?
Regarding my ESCs, I noticed they might not be suitable for the drone. They are the following: “V-Good RC 32-Bit 40A 2-6S Brushless ESC for RC Airplane.” Their PWM speed seems slower than that of multirotor ESCs. However, while waiting for the new ones to arrive, I’ll try adjusting the PID values.
As for the motor spin, I can confidently say they are correct. I’ll try to post a video tomorrow to confirm.
Thanks again, guys! Doing this by myself has been a challenging task, but these interactions and bouncing ideas around make me feel less alone. This is such a great community
Well in the photo, with the forward direction matching the flight controller and GPS unit, all your props are set to spin the wrong directions.
See in the diagram how motors A and H have the props so they “spin inwards” towards the front of the FC. At the rear, Motors E and D also “spin inwards” towards the FC.
This is the quick/easy visual check for quads and octos - hex’s are a bit different.
Yours are definitely opposite to that in the picture.
Nope, I’m wrong - you seem to have Octo X set which is the same as this diagram and your photo. So the props would be turning in the correct direction after all !
Its a digital protocol that is less prone to noise, This is especially useful on bigger aircrafts with potentially longer signal/power lines, DSHOT isn’t a requirement, but If your ESC supports it then you should definitely be using it.
Hello everyone!
Quick update: the drone was stable enough to fly! I see this as a small win. I changed the flight mode to AutoTune to get better parameters, but unfortunately, at the end of the flight, one of the ESCs burned while the drone was still flying. I managed to perform a crash landing.
This is the second time the ESC for the same motor has burned out. Could this be caused by communication latency or delay? A bad connection? I have a spare ESC of the same brand that I plan to use as a replacement, but I really want to understand why it’s catching fire, and why it’s happening to this motor and not the others, assuming the issue is with the ESC.
I’ve been checking the connections and haven’t found anything making unintentional contact.
Here’s the video of the flight with the AutoTune and the failure at the end, along with the flight log:
Autotune should only be preformed after some other configurations and tests are completed. Did you do them all like explained in the excellent documentation?
If Im completely honest I only did the " Initial aircraft tune" from the documentation, Im just reading right now the “Test AltHold”. I should have done it before, but I was so happy it was flying and wanted to do the autotune that I rushed it, Im replacing the burnt ESC and doing that test this week.
Additionally, I do not think you can achieve the downward-facing propeller mounted in a way you can get the spin direction correctly without modifying it.
The motors are setup properly otherwise he wouldn’t have been able to takeoff like in the video above, There might be some loss of efficiency since the propellers are on top of each other, The setup is unusual but its properly setup and should fly when tuned properly.
Hello,
Yes, I’m sorry about that. I used a different picture and didn’t notice until it was brought up. Yes, there might be a bit of inefficiency because of the overlapping propellers, but it works! I got good feedback from some users regarding the PID, and now I’m using the ArduPilot configurator software.
