Pixhawk CUBE Orange Disarmed Midflight

Umar_Khan · April 2, 2023, 10:46am

Greetings,
We developed a Mapping Drone using Pixhawk CUBE Orange and Here3 GPS. We are using HERELINK for our telemetry and radio controls. We usually fly the drone in AUTO flight mode. We had 50+ successfull flights. Recently the drone crashed while it was in auto-mission. There were no error msg on screen and suddenly it disarmed mid air and droped from 120m height. We wanted to find the cause of the crash. Can anyone help us with the same ? . Both the flight log and flight path are attached.

https://drive.google.com/drive/folders/19bGk-T5FLdy3qGRJBdPKOv83YqszxXJo?usp=share_link

Thanks in advance

xfacta · April 2, 2023, 12:26pm

I think you’ve got a problem with your Arm/Disarm channel RC11
or someone is accidentally operating it. It changes over 30 times during the flight.
Motor cut-off closely aligns with the final Arm/Disarm operation.

Vibrations are also unacceptably high and this could lead the flight controller into believing it’s falling (maybe even landed??)
Motors are also oscillating very badly.
There’s lots of radio failsafe messages.
There’s signs of poor attitude control.

In summary

vibrations
radio antenna placement
RC11 operation, maybe related to radio signal loss.

If the copter is repairable you would do these after going over the above points
Set these for safety

ARMING_CHECK,1
BATT_ARM_VOLT,22.10
BATT_CRT_VOLT,21.00
BATT_LOW_VOLT,21.60
BATT_FS_CRT_ACT,1

and these to aid the tuning process that should be revisited

INS_ACCEL_FILTER,10
INS_HNTCH_ENABLE,1  //set this then refresh params to see the rest
INS_HNTCH_MODE,1
INS_HNTCH_REF,0.18
INS_HNTCH_FREQ,60
INS_HNTCH_BW,30
INS_HNTCH_FM_RAT,0.7
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4

If you rebuild it, yell out for help with the setup and tuning process - this copter could have flown so much better than it did.

Umar_Khan · April 8, 2023, 7:26am

Hey Shawn,

Thanks for your insignts and sorry for the delayed response. We were really busy in getting the drone in the air. We managed to sucessfully do so. Upon inspection we saw that the RCIN wire coming from Herelink Airunit was damaged and it was sending false signals to the Pixhawk. We replaced the wires and also reduced the clutter around the Pixhawk. We are also seeing reduced levels of vibrations.
I am attaching log file of the recent test flights.

https://drive.google.com/drive/folders/16QiQoMZyyPBU0a0DhsldNaMEVvn3mJMu?usp=share_link

xfacta · April 8, 2023, 8:11am

You didnt set any of the parameters I specified - they will make some difference but mainly will enable better tuning.

I understand if the copter was rebuilt and testing started before you read my post.

Umar_Khan · April 8, 2023, 10:23am

I’ll add those parameters and take a test flight soon

Umar_Khan · April 16, 2023, 5:21am

Hey, We had our second crash yesterday. Same as before ( flying in AUTO, completed 4 grid lines, and then fell from the sky) but this time it didn’t disarmed mid-flight, it just came down. We were looking for the cause of the crash, all the inputs and outputs seemed fine apart from the vibration. We compared the logs of the previous flight with the crashed flight and noticed higher values of vibration. Can we say this is the cause of the failure? I am attaching both the logs in the link below. We are rebuilding it and we will use the parameters you listed.
Thanks in advance.

https://drive.google.com/drive/folders/13OOunVs2HIxFEab_dpKmyTsK3HubTlpc?usp=share_link

Truth_T · April 16, 2023, 5:33am

I didn’t see your log, but z axis vibration can affect the height sensing, as far I know

xfacta · April 16, 2023, 8:13am

Tuning is so bad the whole flight and motors are oscillating wildly - eventually motor or ESC 2 goes bad and stops working entirely. Motor 1 is commanded to minimum to try and compensate. During all that the copter loses attitude control and is yawing backwards and forwards and the battery appears to become disconnected → logging stops midair.
There’s voltage drops and current spikes at the same time as the motor2 failure, but battery capacity is not the cause of the crash.

Basically if it was tuned better the motor and ESC would not have been killed. The advantage of good tuning is you can easily see if some behaviour is non-standard. And you get better flight time and longer life time of components.
A crash is never good and I sympathise - however the signs were there in the previous log you posted.

Umar_Khan · May 16, 2023, 1:12pm

Hello everyone,
After such a long time we were finally able to rebuild the drone, we took initial flight with the old parameters. The drone seemed twitchy. After hover testing it we made a small auto mission and then sent the drone in AUTO. We also added Harmonic notch filtering as suggested by @xfacta . One thing we noticed is that the CW motors have higher temperature that the CCW motors. We are also experiencing EKF compass error during the flight, the compass goes in orange. Attaching log files and videos to this post. How can we make the drone more stable ? Listing the specs of the drone as well

Drone weight - 4.3kgs
Drone diagonal distance - 700mm
Motors used - T-motor Antigravity MN5006 450 KV
Props used - T-motor NS 16x5.4"
ESC used - T-motor Air 40 esc
Payload - Sony Alpha A6100 APS-C camera with 16mm prime lens

Can you help us out in making it a more reliable and rugged system . Thanks in advance

https://drive.google.com/drive/folders/1fiVjudeLf5Ed7zoWzmfGbaqh33gbBuLu?usp=share_link

dkemxr · May 16, 2023, 3:44pm

Not when you post a link with 12 log files. Post One (1) log file where the issue(s) you are inquiring about are present.

Umar_Khan · May 16, 2023, 5:51pm

Sorry for posting so many files as once. I thought it will be helpful to get a better idea of the situation. You can refer to “Flight 11” and “Flight 12” as they are the latest files with the above mentioned issues.

dkemxr · May 16, 2023, 6:01pm

Did you at least create a graph from the log as @xfacta has shown in the previous post to determine if your outputs are oscillating? We are here to help you help yourself to advance the performance of your craft. This was an example for you to use for future tuning.

So, I will simply repeat what Shawn told you a month ago “Tuning is so bad the whole flight and motors are oscillating wildly”

I would suggest resetting all the parameters to default and start over. Make sure EK2 is disabled after you do that. And, It’s foolish to run an Auto Mission before a craft is properly configured and tuned.

xfacta · May 16, 2023, 11:20pm

We can usually understand the situation from just one log file.
These copters are a whole system, so you cant really expect a result from changing 1 or 2 parameters and ignoring vibrations.
So you flew the copter with basically the same issues as before, and no surprises - it still behaves the same way!
If that is a Lipo battery then you are killing it by ignoring the correct battery settings.

If you want a more reliable copter you need to start by setting all the safety related params I specified - and these are just common sense and well documented. I’m not suggesting anything crazy or controversial.

X axis vibrations are bad, indicating an issue with the flight controller mounting, or something is touching it or wiring is pulling it one way. Once vibrations have been attended to (but dont fly yet), set all these:

AHRS_EKF_TYPE,3
ARMING_CHECK,1
ATC_RAT_PIT_D,0.004
ATC_RAT_PIT_I,0.12
ATC_RAT_PIT_P,0.12
ATC_RAT_RLL_D,0.0033
ATC_RAT_RLL_I,0.10
ATC_RAT_RLL_P,0.10
ATC_RAT_YAW_I,0.08
ATC_RAT_YAW_P,0.80
BATT_ARM_VOLT,22.10
BATT_CRT_VOLT,21.00
BATT_FS_CRT_ACT,1
BATT_LOW_VOLT,21.60
EK2_ENABLE,0
FENCE_ENABLE,1
LOG_BITMASK,180222
LOG_REPLAY,0
PSC_ACCZ_I,0.4
PSC_ACCZ_P,0.2
PILOT_THR_BHV,7  // you appear to have a spring-centered throttle

Now just do a small test flight starting in Stabilise mode and if that is stable enough, switch to AltHold and try some gentle movements - nothing radical and do not go to Loiter or Auto missions.
The hover thrust might need some time to learn. The pilot throttle behaviour change will really help in AltHold too.

Let’s see that log file.

EDIT:
For any further tuning you need to take the payload off, particularly if it is a camera gimbal. Anything that can move or vibrate is bad.

Assuming the attitude tuning can be completed OK, I think those motors and even ESCs will give you trouble in the long term. They are OK at hover throttle, but when pushed (ascending or fighting any wind) they will quickly overheat and potentially even fail.

Umar_Khan · May 17, 2023, 11:11am

We used the above mentioned parameters and had the test flight as instructed. Sharing the Flight log. We here having somewhat stable flight in stabilize and Alt Hold flight modes. But the vibration levels were the same. Upon checking the logs we got the same motor oscillations. We removed the canopy, and the vibrations were gone.

We are planning to start AUTOTUNE process, Is this the correct approach ?

https://drive.google.com/drive/folders/1rVASen7lryluCuy4B-xSdMOPCHlFMutK?usp=share_link

dkemxr · May 17, 2023, 12:59pm

The vibration levels are way to high in the log you posted. And as you say the Motor Oscillations are still terrible. The Rate PID’s will have to be reduced until this oscillation is gone. And when you do eventaully run Auto Tune set the aggression back to default (0.10). The .05 you have set will almost certainly produce a bad Auto tune if it runs at all.
The notch filter isn’t set right but at this point it doesn’t matter as you have other priorities.