Controller crashed after motor start

When motor spooled up, ArduCopter crashed.
Blades were spinning, no control neither from radio nor GCS.
I had to wait for the esc cut off.
What could be the reason?

2019-06-02 19-31-31.bin (236 KB)

log_udp_2019_06_02_19_30_31_886.tlog (344.5 KB)

Thunder Tiger Raptor E550
Matek F405-Wing
ArduCopter 3.6.9

@twistedwings I’m sorry to hear that. Was there any damage to your heli?
I briefly looked at your log and couldn’t see anything that would explain this. Did this occur just after loading Arducopter 3.6.9? Or have you had successful flights with this version of software. If you had successful flights already with this version of software, what was different with your start up procedures for this time? If this was the first time with this version of software, what was the last version of software you had in this aircraft that successfully ran?

No, fortunately there was no damage. But it’s a scary feeling if you cannot stop the blades from spinning.
The last successful flight was with 3.6.7. Since then I upgraded to 3.6.9, added gps and compass and a wifi telemetry bridge.
Maybe I should undo the changes and take a try.

I would suggest stating with just 3.6.9 and ensure it doesn’t crash. Then add the GPS and then finally the telemetry bridge. In all these cases, I would recommend disconnecting your ESC so it is not powered or if you can’t do that then take off your blades, main and tail rotor blades.

How is your board powered?
maybe voltage was going to low or to noisy.
always make sure you have redundant power supply.

Blockquote
added gps and compass and a wifi telemetry bridge.

especially the wifi bride can use a lot of power extra. 500mw or 1 w version?

It’s the compass.
If I disable the compass in MP, everything works ok.
If I enable the compass, the controller freezes as soon
as I start the motor (blades removed).
It’a GY-271 (HMC 5883L) connected to DA2, CL2.
image

@aamadeuss: the F405-Wing has it’s own BECs. The servos are powered from the ESC. The wifi bridge is this one.
0,2A max according to seller.

@rmackay9 and @tridge Have we seen this issue before? Should this be documented as an issue in github against copter 3.6? Any hardware that crashes the flight controller upon motor start I think is a serious issue!

@twistedwings Please post this issue on the Copter-3.6 discussion thread and provide a hyperlink back to this thread. That will give it maximum visibility as a potential issue with 3.6. You may want to search that thread as well to see if anyone has reported a similar issue.

thank you Bill, I have posted this in Copter 3.6. thread

I have had a very similar issue as you @twistedwings. I had my heli sat on the ground rotor at full speed and no control. I spent some time with the help of @tridge and @peterbarker to try and diagnose the problem. I have not yet worked out what caused the issue. I was using an F405-Wing as well. I also had the exact same behaviour with regards to it crashing when the compass is enabled and fine when the compass was not enabled. The compass I was using was a Mateksys MQ8-5883. That compass uses the QMC5883 driver which, as I understand is very similar to the HMC5883 driver. To stop the risk of the F/C locking up with the rotor running again I flashed master to have access to the watchdog reset. That will prevent the situation of the heli being uncontrollable and will reset the F/C instead of locking up. I have since bought a new F405-Wing and am using a Betian BN-880 gps/compass instead. I have had no lock ups or resets since. The BN-880 uses the HMC5883 compass driver as well.

Out of curiosity, how have you plugged your BEC in? Have you plugged it into the servo voltage rail or isolated the VCC line into a separate voltage rail, not on the FC? The reason, I ask is because I originally plugged the positive voltage straight into the FC servo rail, and my theory is that I may have fried something on the I2C line in doing so. My setup works far better with a separate voltage rail away from the FC.

Is it a bad batch of F405-Wings? Who knows…

Hi Matt, that’s interesting, thank you for reporting this.
How did you enable the watchdog, what firmware are you using?

Regarding the BEC: I have connected the power line of the servos to the BEC not to the FC.

Dear all,

i have the same problems. Compass is a GY-273 (QMC5883), FC is an Omnibus F4 V3, Omnibus F4 V3 Pro or Omnibus F4 V5.
Most times the FCs hang up during compass enable, some times i get just “bad compass health”, minor times it works. The compasses are powered via onboard BEC, tested with and without pullups on the compass breakout and shot cables. Tested also to set the compass type explicitly to QMC5883.

A Beitan BN-880 with HMC5883 work flawless but is too big for the build. Unfortunally most compass boards rely on the QMC5883 as the HMC5883 is out of production.

@twistedwings Apologies for not getting back to you sooner.

To enable the watchdog you need to set BRD_OPTIONS to 1.

I was using the master/latest firmware. In the end I choose to use a different flight controller (A pixhack 2.8.4) and all of the issues with the flight controller hanging went away.

Your experience sounds precisely like mine, except I was using an F405 Wing (which is practically the same board). After changing to the Beitian BN-880 the FC crash issue went away.

We’d really like to try and reproduce this issue with the latest master code. We’re trying to track down a possible lockup related to I2C and we need a reproducible test case. So if anyone can reproduce this lockup I’d love to hear about it.
If the lockup happens with master then we should get a WDOG message in the log which will give us some clues as to the cause. Before the lockup it should also produce a MON message. To get that the vehicle would either need to be armed or have LOG_DISARMED=1.

I would try to reproduce the issue.
What is the latest master code, http://firmware.ardupilot.org/Copter/latest ?

  • Is there a master to explain what happened?