Internal errors 0x100000 I:921 flow_of_ctr Twitching motion in loiter during forward flight

Ah, so the F3 shows its weaknesses.

Would this mean that the Cube black is obsolete now? Maybe not for hobby crafts but for commercial once?

The only extra feature I have is the notch filter running on esc rpm data. Not for all motors but only one harmonic notch running of motor 1 frequency.

There was some updates to the notch filter so maybe got better but a bit to much for my F3

Would I be able to see this issue if I look at the processor usage from the log file?

yep cubes are obsolete!

Hope you only mean the cube black than, orange has an h7 so that would be more than sufficient

what would be my best course of action now?

  1. Go back to 4.2.3 (it worked fine than) and wait for 4.3.2 when this might be fixed?

  2. Disable one imu and stick to the latest (if it works)

  3. Go back to 4.2.3 and wait for a fix but still disabling the third imu?

  4. if you have a better sugestion

Thanks!

Cube Black has an F427. But itā€™s still 2 generations old (F4>F7>H7). A H743 flight controller is available for $110. Well, in the US they are.

This aircraft is already quite old and the cube has been really reliable hardware wise. So for the time being i think Iā€™m goong to stick to it.

My new commercial craft has a cube orange. And when I decide to upgrade this rig I will probably spend tje extra money for a new cube for this one too.

For now I guess Iā€™ll try to lighten the processor load and waot for the ā€œfixā€ to get released

Thanks!

This reminds me of an old car I had with the check engine lamp on all the time. I fixed it by removing the bulb :wink:

Hahaha, didnā€™t think of it like that but thatā€™s basicaly what it isā€¦

I guess 4.2.3 was juat under the limit of what the cube black could handle

Yea, joking of course. It seems to me to be entirely plausible that a loop timing problem could cause a momentary stutter or similar response.

Guess weā€™ll see tomorrow.

Iā€™ll try and dissable the third imu and see if that helps

1 Like

Also make sure youā€™ve got FFT_ENABLE,0
If thereā€™s still any doubts, use throttle based HNOTCH instead of ESC/motor RPM

I wouldnā€™t say it is obsolete, you just have to manage the amount of processing power and memory you have. So running all three EKF and a bunch of harmonic notches is going to push things a bit too hard. Back off to two EKF and have a look at the PM message and see if you are still getting a lot of long loops and seeing this problem.

https://ardupilot.org/copter/docs/logmessages.html#pm

I looked at the logs and the PM tab.

The new firmware uses a bit more processing power (usage %) and the long loops occur in large numbers through out the flight.

Bit hard to see on this picture but blue is processor usage, red is long loops and green on the bottom is internal errors detected.

Before on 4.2.3 I didnā€™t get any internal errors but the long loops still spiked some times

How come this internal error didnā€™t show earlier since the loong loops still spiked? Is it because the checks have been made tighter.

Thanks!

You only plotted the number of long loops, not the maximum time.

Ah, sorry.

So the maximum time is important regarding this error?

Iā€™m on the road untill monday but Iā€™ll get the maximum time posted too!

Thanks!

Hi,

I havenā€™t gotten an opportunity to test fly with the third imu disabled, itā€™s been snowing here and quite high winds. Iā€™m going to try and test fly tomorrow and also test with 4.3.1 and 4.4.0 if the imu disabling doesnā€™t work.

Iā€™ve looked at the logs from a pre error flight on 4.2.3 and a flight with internal error on 4.3.0 and I canā€™t find that much of a difference in the long loop max time and peak number. Not much difference in CPU load neither.

Pre error: https://drive.google.com/file/d/1dMUbzCyqMZaJROAVHfK-0Ts1VGSisw-6/view?usp=sharing

Flight with internal error: https://drive.google.com/file/d/14uZRf9g8-GhHSG_-BvS0AFAEev7r67MT/view?usp=sharing

I thought itā€™s better to send the logs rather than trying to plot the values in pictures, Iā€™ll probably manage to make that pretty messy.

Why would this error occur now? There is a higher number of long loops but not higher that it peaked before the issue. The processor load is also a bit higher, but not hitting like 90%. It goes up to 85 but that should work, right?

Itā€™s probably obvious why there is an issue and my lack in log analysis and system knowledge wonā€™t help me. Maybe, if you got the time, you could take a quick look at the PM plots in the log and see if there is an apparent reason to why my processor canā€™t handle the load.

Thanks!

@rmackay9 Also, if you want a pre error and during error log to compare between 4.2.3 and 4.3.0 you can use these. Just in case it would help!

Disabling the third IMU didnā€™t help, I still get the same internal error as before.

Log file: https://drive.google.com/file/d/1-3OJgXUIA8VCzeD5IBiEPEc_XvKz2KQT/view?usp=sharing

That log still has all three ekf running set:
EK3_IMU_MASK 3
INS_LOG_BAT_MASK 0

Make sure you reboot the aircraft after making these changes.

Thanks for your help and patience!

Ah, my bad,

I tested 4.2.3 again with no issues. Im out testing now. Iā€™ll load 4.3.0 again and set the values you suggested

Iā€™m setting ins_use3 to 0 as well. You wrote that eirlier but I seem to have done something wrong when I set it.

Edit: after a reboot, the ins use resets to 1 again
The other two changed after reboot. Maybe I need to change the ekf one before I change the ins one?