Quadcopter crash in Autotune

Hi everyone,

I have a hard crash after a few minutes in Autotune using the latest Copter 4.1.2 firmware. This is a new quadcopter setup and first test flight. Could you kindly give some suggestions of what might happen? Thank you very much!

I uploaded the log file here: https://drive.google.com/file/d/1Ndn6Sgl7yjbCr5oCFyXwg060KBsRNAL9/view?usp=sharing

Best regards,
Tony.

Very sorry to hear about the crash. This should be the big guy with 12S battery. Could you please provide more information about the drone config (motor, frame size, props…)

By looking at the log, I have something to point out:

It looks to me you skipped the Manual tuning step in the Wiki. These ATC_RAT_PIT/RLL_X seems to be defautl. This is very important to follow the instruction before going to Autotune (and Loiter)

It is not the right way to invoke the Autotune from Loiter, since the attitude tuning is not done yet, Loiter (positioing mode) should not be invoked before AltHold and Stabilize are good to control. I guess you may have some difficulties in drone control but let’s try harder.

I guess the direct reason for the crash is when AutoTune increased D term (ATUN.RD), it tweaked the drone so hard and it caused oscilation/instability in the control loop so the drone can not keep up with that change because PIDs were wrong (no previous Manual tune step is done)

Motors did not looks good too, it is too sensitive. You can see that they bumped from min PWM and max PWM while flying. These may come from

  • Not calibrated ESCs
  • Not suitable motor, ESC, battery, propellers, weight configuration
  • Your MOT_x parameters seemed fine. MOT_THST_EXPO was set to 0.75, is it a reason for it (22 inch pros - I guess)? If not, please go back to default: 0.65.

I also want to check the battery to make sure it is sufficient

Please have some time to check and get back to discuss further.

My 2 cents!
Thanks,
Vu.

Hi @bigboy061293 , thank you very much for your analysis!

About the drone configuration:

  • Quadcopter H
  • Motor: T-Motor MN605S-KV170
  • ESC: T-Motor FLAME 60A 12S
  • Prop: 21inch
  • Battery: 6S 12000mAh (2 batteries in series)

About manual tuning steps:

  • I followed this instruction to set the filters and battery scaling values: Setting the Aircraft Up for Tuning — Copter documentation

  • About PID values: Yes I used default values for the first flight. Is there a wiki page to set up PID values before flight or you mean I should do manual tuning in the air before Autotune?

  • I calibrated the ESC before the flight.

  • Yes my copter is 21" so I set MOT_THST_EXPO to 0.75 as wiki instruction.

  • I saw the Auto Analysis failed in my Compass. Could it be the magnetic interference that causes failure in the compass that can cause the crash? Could you help have a look at the compass log also?

Many thanks,
Tony.

Hi @tony34,
I do not use FLAME ESCs + MN motors but another combinations from Tmotor (U8lite, P80, Alpha ESC)
Tmotor claimed they have linearizing scale for their ESCs - which I am not sure about this. However, I always set MOT_THST_EXPO 0.65 for all the cases. Somebody found that 0.2 works for their FLAME ESCs - which, again, I am not sure too.

For tunning, you need to follow exactly these steps: 1 → 2 → 3 → 4 → 5 Tuning Process Instructions — Copter documentation
That means, you need to conduct manual tune before invoking AutoTune.

How about the drone’s weight?

If this is first drone you have ever worked with, it seem too big. My recommendation is getting a smaller one to practice before tuning this big guy.

Thanks,
Vu.

They made the opposite claim when I enquired about this. That none of their ESC’s have linearization.

This really surprizes me.
Okay, I am not trying to confuse you guys and it should be off-topic but I still want to share this test. A full test result and conclusion maybe presented in another threads.

This is the data I logged from ESC telemetry of:

  • ESC Alpha 80A
  • P80III and 3218 Propellers
  • Sufficient power

There is always be an “anomaly” point at 79 - 80% of the throttle (approximately 1790 - 1800 RC PWM)
Let’s give it a look and think if it is relevant to previous problems (and Tmotor ESC problems) or not.

Thanks,
Vu.