Autotuning Crash / First Flight Crash

ArduCopter
1

Hi,

i’ve just builded my first quad, powered by a raspberry pi 3 with navio2 shield.

Motors are 4x Emax MT2213 935kv.
ESCs are SHINA 20A with BLHeli.

At first my copter was totally not flyable, so i modified PIDs via MissionPlanner and it worked quit well. After that, i was able to hold it in the air for severals minutes (it was very exhausting and my nerves were strained). But it worked…tested Stabilize and AltHold mode.

Then I tried autotune…it was a short fun…
After a few seconds the copter tilted to the right and crashed.
Done it several times. After more or less seconds the copter always crashes, when in autotune (Sometimes on the right, sometimes on left).

I’ve done some research and read about esc sync failures and crashs, caused by vibrations.

So for now i’m waiting for a new frame (last autotune attempt was followed by a very hard crash), i’m thinking about the problem and how to avoid throw out even more money :smiley: )

Are my ESCs the reason (they cost about 30 or 40 € on amazon), and if so, which are the recommended ESCs?
Were it just the Vibrations (Flightcontroller was mounted on antivibration board, the one with the blue rubbers), and if so…how to get rid of these vibrations?

Or is it a combination of everything (bad vibrations, bad motors, bad ESCs)?

I would appriciate any help or suggestions, what I can do.

I will try it again (when my new frame arrived) and take a look at the logs. But maybe you can give me a hint, before any missleading autotune attempts, whether I choosed some really bad parts or did another mistake.

Thank you

2

I’m sorry to hear about your crash!
It’s really impossible to know what went wrong - posting a log would be very useful.

3

Thank you for your reply.
I know the logs would be very useful, and i will make another try, and post the logs.

My intention for doing this post today was, maybe someone can say something about my ESCs. I read several posts about ESCs failure, so maybe someone can recommend better ESCs.
If my ESCs are completely wrong, i would replace them, before doing any further autotune trys.

4

Do you use the antivibration mount with 8 blue rubbers (two in each corner) or one with 4 rubbers? The one with 4 blue rubbers is too soft for the RPi/Navio and can cause the RPi/Navio to shake from side to side.

And a complete list of your setup would help too.

5

Yes, I’m using the one with 4 blue rubbers. I’ve just ordered black rubbers, I think, these are not as soft, as the blue ones.
But yes…I might have to look for another antivibration board.
Any suggestions, where to find the one with 8 rubbers. Unfortunally I can’t 3d print it for myself, like described in the documents of emlid/navio2 website.

Here is my setup:

3s Lipo (one 3000mAh, one 2600mAh)
Raspberry Pi 3 with Navio2
My new frame, will be a dji f450
ESCs are SHINA 20A with BLHeli
Motors are 4x Emax MT2213 935kv
Props are plastic 10x45
Receiver is a Frysky X4RSB

6

@pUiE best would be that you upload the problematic log here.
The logs are generally saved in /var/APM/logs and you can retrieve them with Mission Planner or other GCS.

7

log_vibrations
log_vibrations.png1600×868 72.8 KB

log_vibrations3_3
log_vibrations3_3.png1600×868 78.3 KB

So, first of all thank you for your replys.

I’ve just downloaded the latest log (from my hard crash flight, crash when autotuning) from the controller.

I don’t realy know, where to start analyzing the log, so I’ve uploaded the Vibration graph. I can see this hard Vibration at the end of the flight. But I don’t know, whether this was the reason for my crash. There are also some markers indicating a crash and low battery. Are these the reasons for the crash?

Let me know, if another parameter from the log would be more helpful, than the vibration graph.

//Edit1: Added Vibrations 3.3 graph
What are the differences between “vibrations” and “vibrations 3.3”?

//Edit2:
I took a look at a previous flight with trying to autotune (also a missleading attempt with a not-so-hard crash at the end). There were no battery errors in the log. See the log below

log_vibrations_other_flight
log_vibrations_other_flight.png1600×868 69.3 KB
log_vibrations3_3_other_flight
log_vibrations3_3_other_flight.png1600×868 76.6 KB

8

You need to post the actual logs, so we can look at them. It sounds very much like you’re getting ESC desync - it’s mostly commonly seen in autotune because autotune commands sudden movements that trigger the desyncs. Do you have active braking/damping light set in the esc firmware? That’s the most common cause, first thing is to ensure that’s turned off. However it’s very uncommon to get desyncs in such a small/highkv motor. What size props are you turning?

9

What do you mean with “actual logs”. Did I posted the wrong logs? The Logfile in plain text?

How can I determine Active Breaking/ damping settings?

My Props are 10x45. So i think the size is 4.5".

10

I’ve done an auto analyze of the log file:

Log File C:\Program Files (x86)\Mission Planner\logs\QUADROTOR\1\2018-03-11 17-53-49.log
Size (kb) 2798.86328125
No of lines 33681
Duration 0:04:48
Vehicletype ArduCopter
Firmware Version V3.5.2
Firmware Hash 62a12357
Hardware Type
Free Mem 0
Skipped Lines 0
Test: Autotune = UNKNOWN - [?] Autotune 5763-24270

Test: Brownout = GOOD -
Test: Compass = FAIL - Large change in mag_field (116.09%)
Max mag field length (662.62) > recommended (550.00)

Test: Dupe Log Data = GOOD -
Test: Empty = GOOD -
Test: Event/Failsafe = FAIL - ERRs found: FLT_MODE FS_BATT CRASH
Test: GPS = GOOD -
Test: IMU Mismatch = NA -
Test: Motor Balance = GOOD - Motor channel averages = [1467, 1504, 1513, 1524]
Average motor output = 1502
Difference between min and max motor averages = 57
Test: NaNs = GOOD -
Test: OpticalFlow = FAIL - FAIL: no optical flow data

Test: Parameters = FAIL - ‘THR_MIN’ not found
Test: PM = GOOD -
Test: Pitch/Roll = UNKNOWN - 'BarAlt’
Test: Thrust = GOOD -
Test: VCC = UNKNOWN - ‘POWR’

What I can say is, that there aren’t any bad Parameters…except the compass. But I don’t think this had an impact on the autotune.

Oh and FLT_MODE FS_BATT CRASH …what does that mean?

11

That means the low battery failsafe activated and then the aircraft crashed.

12

“Actual logs” means either the .log or the .bin file that is time stamped for a particular flight.

The text you posted is a dump from running a log file through Auto Analysis.

The “10” in 1045 means the prop is 10" long.
The “45” in 1045 means the prop has a “pitch” or blade angle of attack that will move the prop 4.5" for every revolution.

The only way to examine/change any BLHeli settings is to connect the ESC to either an Arduino, or connect the ESCs to a CleanFlight/BetaFlight/iNav compatible flight controller and use BLHeli Suite.

13

Do you think, this was the reason for crashing?
I’m not sure about it. Because i had the issues of tilting during autotune in other flights, where the Logs don’t show this Parameter / Error.

So are my props the right ones for my Motors? They came together, so I thought, they might fit.

I have an Arduino at home, so I will take a look at the ESC Settings.
Do you recommend to turn off Active Breaking / Damping, when it is activated? I think I read somethin of active breaking feature, when buying my ESCs.

14

What I recommend is that you go back and read the entire Wiki from start to finish so can get the big picture. When you go through Initial Setup > Mandatory Hardware, DO NOT SKIP ANY STEPS. Then go to Initial Setup > Optional hardware and configure the power module.

Second, for an aircraft as large as your 450, using BLHeli serves no useful purpose. The purpose of BLHeli firmware is to provide RACERS with the kind of throttle response and ESC/motor performance they need to win races. Oneshot support was added to ArduCopter to support the PixRacer.

Ditch the BLHeli ESCs and get some plain old ESCs. DO NOT GET ESCs with SimonK firmware. SimonK is a half step between generic and BLHeli and its outdated. I use Hobbywing parts like the X-Rotor 40A OPTOs I have on my Tarot 680 Pro Hex and my stretched Tarot 650 Sport Quad. For “smaller” aircraft I use either Hobbywing Platinum Pro 30A OPTOs, or Hobbywing Pentium 30A with 2A BEC.

15

Post the log. Nobody here can help you until you post the log.

They’re probably fine, I was just making sure you weren’t using ridiculously large props for the motors that might be causing problems

Yes, turn it off. Get your craft flying 100% with basic settings, then you can experiment with ‘nice’ extras like active braking.

Nothing wrong with BLHeli, it’s not just for racers, you just have to be careful of some of the settings. I don’t think oneshot had anything to do with pixracer - it works on other FCs and pixracer works with normal PWM. Agree with you on hobbywing ESCs, they’re fantastic.

16

Right, I have BLHeli firmware ESC’s (BLHeli and BLHeli_S) running on all my multirotors small to large with Pixracers and Pixhawks. All are set for Oneshot 125 in Arducopter and working well.Agree, the Hobbywing ESC I have in my Rover is bombproof.

17

okay. I will give my BLHeli ESCs another try. Would you recommend to not connect the 5V Cable from ESC to Controller, like denoted in some posts?

Here is my log file with the hard crash -> Hard crash log

Here is my log file with extreme tilting, when autotuning -> Extreme tilting during autotune

Thank you all for your help

18

What you think and what I know are two entirely different things. Adding Oneshot support to ArduCopter was all about PixRacer because NO ONE ELSE NEEDS IT. Go back through the AC3.x.x blogs and you can read all about it.

Here’s another clue for ya: OneShot, was created LONG before PixRacer, and for those FCs that can talk to BlHeli ESCs, if you use a receiver that does not support PPM, S.Bus, iBus, or any other serial protocol, you must configure the ESC input signals to be PWM which TURNS BLHELI OFF. BTDT. Maybe you should read up on BLHeli, CleanFlight, BetaFlight, and iNav, too…

As for Oneshot and larger aircraft, BLHeli ESCs choke, blow chunks and desync:

This is identical to the behavior I got using DYS XM20A ESCs with BlHeli 14.1 firmware connected to 5010-530Kv motors swinging 1555 CF props on 4S.

IMNSHO, BLHeli is useless on larger aircraft.

19

For any one else coming across this with a quizzical eye, this is incorrect information. BLHeli has had a few bugs that cause desync in it’s history, along with pretty much any ESC firmware, but both BLHeli and oneshot work fine with larger aircraft. I have hundreds of hours on lowkv motors with large props on blheli and oneshot, as have lots of others. Just be careful of your settings and test well. Always turn active braking off.

20

Fine, you believe what you want. I have a racer that loves BlHeli, and three other aircraft that don’t. I’m not about to risk several thousand dollars to run something that is of no use to me.